 |
| |||||
Genetic Improvement of Cacao. | |||||
| Contributor:Peter Griffee QA and TEM |
Miscellaneous Published item Free form Table of Contents Display Full eArticle |
ID: 197 |
 |
|---|
| Abstract |
|---|
| This is the first publication which deals exclusively with cacao genetic improvement and is notable for its logical sequence dedicated to the subjects in its 13 chapters: 1. Environmental and socio-economic improvement. C.A.S. Souza & L.A.S. Dias. The principal aspects of cultivation and the strategies of environmental improvement are presented. 2. Diversity in the genus Theobroma. C.R.S. Silva, A.V.O. Figueira & C.A.S. Souza. The diversity in Theobroma is focussed with a view to improvement by incorporation of genes from wild species into the genetic make up of the cultivated one. 3. Origin and distribution of Theobroma cacao L: a new scenario. L.A.S. Dias. A new scenario is presented for the origin and distribution of cacao, with important reflections on the collection and conservation of germplasm. 4. The ecology of natural populations. C.M.V.C. Almeida. The ecology of natural populations in its most diverse aspects is dealt with. 5. Genetic resources. C.M.V.C. Almeida & L.A.S. Dias. How to collect, conserve, evaluate, characterize and use germplasm saved are topics developed. 6. Strategies and methods of selection. L.A.S. Dias & M.D.V. Resende. From Ch. 6 onwards the book focuses on the actual genetic improvement. For the first time, the methodology de mixed mathematical models is introduced to cacao breeding, with a view to making it more precise and efficient. 7. Breeding for disease resistance. R.A. Rios-Ruiz. 8. Biochemical and physiological bases of disease resistance. M.A.G. Aguilar & M.L.V. Resende. Chs. 7 and 8 cover the state of the art of resistance to diseases, in particular witches’ broom, emphasizing the heredity mechanism and the bases of resistance. 9. Clonal breeding. A.B. Pereira. Asexual breeding is highlighted and covered. 10 Molecular markers in breeding. A.V.O. Figueira & J.C.M. Cascardo. The introduction of molecular markers in breeding and the possibilities open for these new tools are reported. 11 Breeding research. L.A.S. Dias & M.D.V. Resende. Another grey area, never really covered in cacao breeding, is highlighted. 12 Contributions of breeding. L.A.S. Dias. Breeding success is illustrated by the comparative results of improved cultivars against traditional ones. 13 New paths in improvement. L.A.S. Dias. This capitalizes on all improvement aspects, harmoniously integrating sexual and asexual improvement and biotechnology to project the future of breeding programmes. |
| Table of Contents |
|---|
| Foreword by FAO. | Return To Table of Contents |
|---|
| In 2003, my colleague, Dr. Elcio Guimaraes of FAO showed me the hard copy of the book ‘Melhoramento Genético do Cacaueiro’ which he had received from the Editor, Luiz Antônio dos Santos Dias. Luiz has 20 years of hands-on experience in working with cacao and people of all its disciplines and is now with the Federal University of Viçosa working in BIOAGRO and is the current Editor-in chief of the international journal CBAB (Crop Breeding and Applied Biotechnology). Elcio was a rice breeder at the EMBRAPA Rice and Beans Station and was the President of the Brazilian Society of Plant Breeding (SBMP) from 2001-2003 before joining FAO as the Senior Officer for cereal/crop breeding. I am the Senior Officer for Industrial Crops at FAO and my home has been in Brazil since 1974, so both Elcio and I were able to see, at once, the extraordinary quality of this book in uniting multidisciplinary aspects of cacao - from ancient ethno-botanical records through to modern biotechnologies; a unique work - never before united into one - and with the support from many renowned scientists. So - with total support from Luiz - I entered the entire Brazilian version of 578 pages into EcoPort. (See Dias L.A.S., 2001 for the Portuguese abstract). We all perceived the urgent need for an English version - “Genetic Improvement of Cacao” - to make the knowledge more universally available. With the help a grant from FAO this work has now become a reality. We would like to cite a quote from Mark Guiltinan who is Professor of Plant Molecular Biology, Department of Horticulture, The Biotechnology Institute, Penn State University. Mark also manages the INGENIC molecular biology working group regarding cocoa genomics related research. “I am very interested to see the chapters, I would love to either make a link from our web site to it or host it as well as a mirror or alternative North American site. This sort of cocoa literature is so hard to find, and as such I am investing some time to try to put the Archives of Cocoa Research online as well”. The translation was done by Ms Cornelia E. Abreu-Reichart of Viçosa, with support from Luiz, myself and Elcio. (See Dias L.A.S., 2001 for the English abstract). Per Diemer, FAO consultant, helped with complex in-line formulae. This translated EcoPort version will be archived (and fully available) as a mirror of the original work and a copy created - the great advantage being that the copy can be up-dated regularly as novel technologies emerge. Finally we would like to congratulate the Editor on his capacity, dedication and courage in producing this work. Peter Griffee Senior Officer, Industrial Crops Crop and Grassland Service Plant Production and Protection Division Food and Agriculture Organization of the United Nations. |
| Preface from the Editor and Supporter and List of Collaborators. | Return To Table of Contents |
|---|
| Editor’s preface: Luiz Antônio dos Santos Dias March 2001. In world literature there has never been a work that deals exclusively with the genetic improvement of cacao. Until now, all incentives in this respect have not gone beyond titles inserted in books on agronomy of the crop or on improvement of cultivated species in general. The present work, on the contrary, proposes to be the first that covers, exclusively and in depth, the genetic improvement of cacao. Therefore it can be called the world’s first. The book also had the privilege of being entirely written in Portuguese and mostly* for Brazilian scientists. *Note from the EcoPort editor: this is why the English translation is being done in close collaboration with the Brazilian editor order to make the book more universally available. Even though it has a universal dimension and is applicable to most tree species, its focus is on problems and solutions of Brazilian cacao. It represents an historical mark of the national scientific community. The objectives which guided this initiative were: i) to make the accumulated knowledge on the genetic improvement of cacao available for the scientific community; ii) stimulate discussion on the subject with all interested parties; scientists, extension agents, students and farmers; and iii) open the scientific discoveries on the theme which, otherwise, would be restricted to the group of national and international scientists that are involved in few research institutions on cacao worldwide. This work, through the editor, had the collaboration of renowned specialists in their particular disciplines; all belonging to research institutions of national and international prestige. The collaborators were stimulated to contribute new - state of the art- technology on cacao breeding in their chapters. Conceived as to be encompassing and, at the same time, as in-depth as possible, the work is highlighted by its logical sequence and clarity of the themes developed in its 13 chapters. In Chapter 1, the principal aspects of cultivation and the strategies of environmental improvement are presented. In order to overcome the crisis which assails the cacao economy, it deals with the socio-economic panorama that predicts changes of attitude of producers, researchers and institutions. Chapters 2, 3, 4 and 5 basically cover the collection, conservation and rational use of genetic resources of Theobroma, the genus to which cacao (Theobroma cacao L.) belongs. The diversity in Theobroma is focussed on in Chapter 2 with a view to improvement by incorporation of genes from wild species into the genetic make up of the cultivated one. Chapter 3 presents a new scenario for the origin and distribution of cacao, with important reflections on the collection and conservation of germplasm. The ecology of natural populations in its most diverse aspects is dealt with in Chapter 4. How to collect, conserve, evaluate, characterize and use germplasm saved are topics developed in Chapter 5. From Chapter 6 onwards the book focuses on the actual genetic improvement. For the first time, the methodology of mixed mathematical models is introduced to cacao breeding, with a view to making it more precise and efficient. Chapters 7 and 8 cover the state of the art of resistance to diseases, in particular witches’ broom, emphasizing the heredity mechanism and the biochemical and physiological bases of this resistance. Asexual breeding is highlighted and covered in Chapter 9. The introduction of molecular markers in breeding and the possibilities open for these new tools are reported in Chapter 10. Another grey area, never really covered in cacao breeding, (Chapter 11) is experimentation. In Chapter 12, breeding success is illustrated by the comparative results of improved cultivars against traditional ones. Finally, Chapter 13 capitalizes on all improvement aspects, harmoniously integrating sexual and asexual improvement and biotechnology to project the future of breeding programmes. The Editor is thankful for the criticisms, suggestions, personal communications and help received from Doctors, Professors and Colleagues; in particular to Professors Acelino Couto Alfenas and Cosme Damião Cruz (Federal University of Viçosa) for the invitation to work together at the University. He also acknowledges the contributions given by Professor Magno Antônio Patto Ramalho (Federal University of Lavras), Dr. João Batista Teixeira (EMBRAPA, CENARGEN), Dra. Frances Bekele (University of the West Indies, Trinidad), Dr. Albertus Bertus Eskes (CIRAD-CP, France). The Editor thanks Professor José Baldin Pinheiro (Federal University of Goiás, School of Agronomy) for the stimulation and dedication for publication of this work through FUNAPE. Finally - thanks to Dr. Jairo Cunha (CEPLAC, CEPEC), who guided him to the path of cacao, the subject closest to the Editor. Preface of the Supporter: Prof. José Baldin Pinheiro, President of the Board of FUNAPE. It is with great satisfaction that the Foundation for Support to Research - FUNAPE, views the launch of the book “Genetic Improvement of Cacao”, as it is the first text uniquely devoted to the subject. The inexistence of a publication which deals with the various particulars of the crop is a gap which has now been filled. This book, written by important Brazilian specialists, provides information that scientists, professors, students and producers have awaited for a long time. Plant breeding and the development of new varieties has contributed, in a significant way, to the national economy. Thus, Brazilian agricultural progress, in which the production of cacao has its highlights, was only possible because of the work of quiet people, such as these here, who leave their mark. This work is thus not about a simple gathering of ideas, but an important contribution to scientific information and for technological advance. It would be onerous, and is probably unnecessary, to list the numerous impacts and the resolution of agricultural production problems that have been made possible through the discovery and use of new genetic material. When reading this work, we note that the contributions made to it were well chosen and we congratulate, in the person of its editor, all those who contributed in order that this book could give open access to information that was, until now, never united, and made it possible to enrich it in the context Brazil and overseas. Collaborators: André Barretto Pereira. Genetics Section, Centre of Cacao Research, CEPLAC, CP 37 - 45.600-000 Itabuna, BA, Brazil. E-mail: abpereira@ceplac.gov.br Antonio Vargas de Oliveira Figueira. Plant Breeding Laboratory, Centre of Nuclear Energy in Agriculture, University of São Paulo, Av. Centenário, 303, CP 96 - 13.400-970 Piracicaba, SP, Brazil. E-mail: figueira@cena.usp.br Caio Márcio Vasconcellos Cordeiro de Almeida. SERPE/SUPOC/CEPLAC, Av. Governador Jorge Teixeira, 86. Bairro Nova Porto Velho, 78.906-100 Porto Velho, RO, Brazil. E-mail: ceplac@enter-net.com.br Carlos Alberto Spaggiari Souza. Research Station “Filogônio Peixoto”, Centre of Cacao Research, CEPLAC, CP 102 - 29.901-970 Linhares, ES, Brazil. E-mail: carlosspaggiari@bol.com.br Carlos Rogério de Sousa Silva. Centre of Nuclear Energy in Agriculture, University of São Paulo, Av. Centenário, 303, CP 96 - 13.400-970 Piracicaba, SP, Brazil. E-mail: roger@cena.usp.br Júlio Cézar de Mattos Cascardo. Genetics Department, State University of Santa Cruz, Rodovia Ilhéus-Itabuna, km 16, 45.660-000 Ilhéus, BA, Brazil. E-mail: cascardo@uesc.br Luiz Antônio dos Santos Dias. Grant recipient CNPq, Centre of Cacao Research, CEPLAC. Actual address: Department of General Biology/BIOAGRO, Federal University of Viçosa, 36.571-000 Viçosa, MG, Brazil. E-mail: lasdias@mail.ufv.br Marco Antônio Galeas Aguilar. Vegetable Physiology Sector, Centre of Cacao Research, CEPLAC, 45.571-000 Itabuna, BA, Brazil. E-mail: maga@cepec.gov.br Marcos Deon Vilela de Resende. Nacional Centre of Forest Research, EMBRAPA, CP 319 - 83.411-000 Colombo, PR, Brazil. E-mail: deon@cnpf.embrapa.br Mário Lúcio Vilela de Resende. Plant Pathology Department, Federal University of Lavras 37.571-000 Lavras, MG, Brazil. E-mail: mlucio@ufla.br Rolando Alfredo Rios-Ruiz. Nacional Agrarian University de la Selva, Apartado 156 - Tingo Maria, Peru. Actual address: Plant Pathology Department, Federal University of Viçosa, 36.571-000 Viçosa, MG, Brazil. E-mail: ds20628@correio.ufv.br |
| Chapter 1. Environmental improvement and socio-economy. C.A.S. Souza and L.A.S. Dias. | Return To Table of Contents |
|---|
| Contents: Abstract; Introduction; Cacao; Environmental improvement; Shade; Windbreaks; Liming and fertilizing; Soil water availability; Weed management; Pruning; Disease and pest control; Processing; Social economic panorama; The crisis; How to overcome the crisis and Final considerations. Summary: Cacao is a typical tropical climate tree, indigenous to the humid forest region of America and is the principal source of raw material for making chocolate. Actually, the principal cocoa producing regions are concentrated in America, Africa e Asia. In all three, environmental improvement is practised, to a greater or lesser extent, with a view to obtaining a superior quality, more competitive product that attracts higher prices in the international market. Environmental improvement as discussed here involves various plantation management techniques such as irrigation, pest and disease control, pruning, liming and fertilizing, and processing. The social-economic importance of cocoa cultivation is critical for producing countries, particularly Brazil, which is the fourth world producer with the fifth largest global chocolate industry. In spite of having been a major exporter until recently, in reality the country now imports cocoa in order to keep the industry running. Since 1987 the cocoa sector has been going through the most profound crisis in its history. The principal factors which provoked it were; low prices for the product, monoculture, the disease witches' broom (Crinipellis perniciosa) in south Bahia - the main Brazilian producing region, lack of financing and debts of producers who were discouraged by indolent government policies. Other factors include low competivity of the sector which operates at relatively high cost and the use of old and decaying plantations, subjected to inadequate management. The means of overcoming this crisis, presented in detail in this chapter, will, obligatorily, pass along the path of collaboration and partnership. |
| Introduction, Cacao, Environmental improvement; Shading and windbreaks. | Return To Table of Contents |
|---|
| Introduction: Long before Christopher Columbus discovered America, Native American peoples such as the Mayas and Aztecs knew and appreciated the cacao tree. In some American countries, cacao beans were used as currency. A good slave could be bought for around 100 cacao beans in the 16th century (Bondar, 1938). Peter Martyr of Algeria wrote about the use as currency in 1530: “Blessed money; not only does it provide sweet beverage, beneficial to mankind, but also protect its owners against the hellish pest of greed, since it cannot be stored for years on end or hidden away in cellars.” Indians used to prepare chocolate by roasting the beans and grinding them between two stones. This mass was boiled in water and spiced with vanilla, cinnamon, and pepper from Jamaica. (Bondar, 1938). Later, the Spaniards added sugar and milk to the chocolate. Spain was the first European country to use cacao, and kept its monopoly throughout many years. At the beginning of the 17th century, chocolate became popular in Italy and France, and, a little later, in the Netherlands, Germany and England. However, consumption was a privilege of the richer classes, due to the high costs of the beverage. In London, chocolate-selling houses became famous hot points, such as the White House of Chocolate, and the Piccadilly Cacao Tree Club (Hardy, 1961). From Mexico, the Spanish took cacao along to other conquered countries, among them some Central American: Colombia, Venezuela, and the Caribbean islands. In South America, Venezuela was one of the first countries to cultivate cacao. Mexico continued as the largest cacao producer up to the 17th century. In the same way that demand increased, cultivation expanded quickly to other regions, including Brazil, where it first found its way to the State of Bahia, in 1746. After about one century, in 1822, cacao was taken from Brazil to West Africa by the Portuguese. Seeds of traditional Bahian cultivars were exported to the island of Principe from which cacao plantations were formed and called ‘Amelonados’ of West Africa. After a few years, cacao reached the island Bioko (formerly Fernando Pó), then to Ghana, Nigeria, Ivory Coast, and Togo. Cacao was introduced to the Republic of Cameroon by the Germans, and seeds which left from there formed the basis of plantations in South Asia and Oceania (Wood, 1985 e 1991). Together with the developments of production methods as well as transport and communication means, factors like the pleasant taste and nutritional value of chocolate helped spread it across all continents, being consumed good by nearly all inhabitants of the earth. Actually, cocoa is looked upon as a stimulating product. Cocoa butter and powder, liqueur, pastries, and chocolate itself are derived from the fermented and dry cocoa beans. Its principal use is in making chocolate, a highly nutritious food due to its high sugar, butter, and protein contents. A 100g milk chocolate bar is equivalent to six eggs, or three cups of milk, or 220g of bread, or 750g fish, or 450g of beef. Cocoa seeds contain alkaloids such as caffeine, theobromine, and over 300 chemically active substances. Recent studies (Holden, 2000) reveal that a moderate consumption of bitter chocolate can prevent heart attacks. According to these studies, flavonoids - a kind of polyphenol found in chocolate - counteract the free radicals which are to blame for obstructed arteries. They function as blood filters by reducing fat plates and the ‘bad’ cholesterol, known as LDC. A 40g milk-chocolate bar contains over 300mg of polyphenols. The studies were based on the assumption that plants, including cocoa, produce polyphenols that act as antioxidants in the elimination of the free radicals which damage their DNA. Principal cocoa-producing regions of the world lie between 15o N and 20o S latitude, in spite of some smaller production areas in subtropical regions (23o S), as in the case of the State of São Paulo, Brazil. The densest plantation concentration, however, is found within a range of 10º above and below the equator (Alvim, 1977). Largest plantations lie in regions where the mean temperature oscillates between 22 and 25oC and rainfall precipitation is high (1200 to 2000 mm), well distributed throughout the year, with a monthly minimum of 100 to 130 mm. Geographically, the three great cocoa-producing areas of the world are well demarked; one in each continent: (i) In America, particularly Brazil, and more precisely the State of Bahia; (ii) The west coast of Africa, especially the Ivory Coast, Nigeria, Ghana, and Cameroon; and Last (iii) south-east Asia - a new, but probably the most competitive region, headed by Malaysia and Indonesia. Cacao: Among the 22 species of the genus, Theobroma cacao L. is the only one used at a large commercial scale, although “cupuaçu” Theobroma grandiflorum (Willd. Ex Spreng) Schum. is also being commercially cultivated in some areas. Cacao is a perennial, typically tropical tree with its origin in the rainforests of America, where it grows in the understorey. From its probable centre of origin, the high Amazon region (Cheesman, 1944), it spread in two main directions, which resulted in two principal race groups: the ‘Criollo’, grown in Venezuela, Colombia, Ecuador, northern Central America, and in Mexico; secondly, the ‘Forastero’, cultivated in northern Brazil and Guyana (see chapter 3 on origin and distribution of cacao). A third group, designated ‘Trinitario’, is also presented by some authors as the progeny of a natural cross between the Criollo and the Forastero. The majority (85%) of world cocoa production is covered by the Forastero group, which also prevails on Brazilian plantations. The root system of cocoa normally consists of one main taproot of variable length (up to 2 meters), depending on the soil structure, effective depth, and fertility. Lateral roots, most densely concentrated 20 to 30 cm from the soil surface, extend from this root, subdivide, and form a dense network. These roots are mainly responsible for the plant’s water and nutrient uptake, while the taproot maintains the stability of the tree. The stem is upright, with a smooth, green bark in the first two years, which becomes dark grey and of irregular surface in adult trees. At a variable height (between 1m and 1.5m), the stem emits lateral branches that form the jorquette, and, later, from these branches, others grow out to form the crown of the plant. These branches develop in vertical (orthotropic) and horizontal (plagiotropic) directions. Cacao trees grown from seed reach a height of between 5 and 15 meters, though trees with heights of 50 to 75 meters have been reported in the jungle of Belize (Mooleedhar & Maharaj, 1995). The tree can also be propagated in asexual form, usually by budding and cutting processes (see Chapter 9 on clonal improvement). Cocoa clones are much shorter. Cacao is a monoical species and a typical cauliflower. The flowers are formed on the trunk, in tiny inflorescences called flower cushions, from where fruits develop and form. These flower cushions are able to produce fruits for several years, as long as they are not harmed by cultural treatments or by diseases. The flowers are hermaphrodite and pentamerous, with sepals, stamens and staminoides or false stamens. The ovary on the pistil contains between 30 and 70 ovules. The reproductive organs (stamen and pistil) are isolated by two physical barriers in the flower: the staminoide crown and the petals themselves, which surround the anthers. These barriers favour cross-pollination, even in self-compatible cocoa, although the rate of self-pollination in these trees is relatively high. In young trees, flowers are produced mainly on the trunk. In adult trees, they emerge all over the plant, with a higher frequency on branches of a diameter above 1 cm. An adult tree can produce more than 50 000 flowers per year, less than 5% of which are pollinated, and only 0.5 to 2% bear fruits. If a flower is not pollinated within the first 8 to 10 hours after emergence, it will drop; this abortion normally occurs after 24 to 48 hours (Alvim, 1984). This enormous quantity of flowers that cacao produces calls for more research since, intriguingly, the tree produces only few fruits in its habitat. Why should this plant produce such a bulk of flowers, compared to the number of harvested fruits? This strategy seems to have been developed by the species in its centre of origin for survival and perpetuation. The dim light that penetrates the understorey and the great number of species vegetating and flowering in the various forest strata kindles the competition for light among plants and for nectar among (principally) pollinators. This brings on the great variability of flowering periods observed among the species, and the abundant flowering of some of them within a short period, as in the case of cacao. Thus, the specific pollinator (microflies of the genus Forcipomyia, Diptera, Ceratopogonidae) of cacao, which seems to have co-evolved with the tree, creating a mutual dependence between the two species, and the dispersion of cocoa in the forest increased the selection pressure in favour of trees with abundant flowering (see also Chapter 4). Cocoa produces indehiscent fruits of the drupisarcidic bacoidal type (Figueiredo, 1986), pentalocular, with a large variation in size, format, pigmentation, roughness, furrow depth on the fruit surface, wall thickness, and wax content. The time from flower fertilization to fruit maturation varies from five to six months. Economically, the seeds are the most interesting parts of the crop, diverse in colour, format, weight, and size, according to the race and cultivar group. Environmental improvement: Crop productivity depends on genetic, edaphic, climatic, and management factors. Considerable differences in production among and within years and properties are common. In relation to genetic factors, the chances of a crop’s success are high where a cultivar with advantageous characteristics of earliness, yield, dry bean weight, fat content, and adaptability to local conditions is selected. Regarding edaphic factors, extremely varied soil types can be cultivated, provided they are not a priori inadequate, thanks to the available technology. Obviously, the more distant soil fertility is from the ideal range, production costs increase. Climate factors are limiting, since even the most sophisticated technology can seldom provide changes that would be able to revert their impacts. An exception to the rule is water stress, which can be controlled by irrigation. The factor management, the main subject of this chapter, is the responsibility and duty an engineer should adopt. If the applied management is not efficient, the success of a selected cultivar is doubtful, even in a region of suitable edaphoclimatic conditions. This may bring economic unfeasibility to the crop yield. Shade: In the formation process of a Cacao plantation, two types of shades are used: the temporary, and the permanent shade. In most cases, banana plants are preferred for the former. They can be planted in the same row as the cacao, where the area allows mechanization, or in the midst of the square formed by four cacao trees, where the terrain is hilly. Cassava, pigeon pea, castor, and other plants of regional interest are also used for this purpose. This temporary shade protects the tree for a limited time from excesses of light and wind during the initial years of cultivation. The permanent or top shade consists of larger trees, planted together with the temporary. These trees establish the crop’s micro-environmental conditions, a protection the Cacao tree needs during its entire productive life. In the South of Bahia State, the most commonly used shade tree is the coral tree (Erythrina glauca), since it grows fast and changes leaves only in wintertime. The recommended spacing is 24 x 24 m or 24 x 24 m with an additional tree on the diagonal. Other trees, such as the ‘ingazeira’ (Inga alba), mother of cacao (Gliricidia sepium), avocado (Persea americana), peach palm (Bactris gasipaes), ‘cobi’ (Cassia sp.), ‘guapuruvu’ (Schizolobium parahybum), ‘bandarra’ (Schizolobium amazonicum) and Ecuador laurel (Cordia alliodora) can also be used. Care should be taken so as to use the most appropriate spacing for each species, and to let 50 to 60% of light penetrate the shade, which is a requirement of adult cacao (Gramacho et al., 1992). In the Cacao region of northern Espírito Santo State, along the banks of the rivers Rio Doce and São Mateus the coral tree does not develop satisfactorily due to the lengthy dry period and the devastating attack of the borer Terastia meticulosalis (Agostini, 1971). Forest essences such as ‘boleira’ (Joannesia princeps), yellow mombin (Spondias lutea) and Gmelina (Gmelina arborea) are most promising for top-shading of Cacao plantations in the State (Souza et al., 1990; Souza and Augusto, 1991 and Souza et al., 1995). Spacing varies, in this case, from 12 to 15m between shade trees. Nevertheless, cacao shading is a controversial topic. The practice is based on agricultural rather than on physiological reasons. Cacao is a shade-tolerant tree, but not specifically a shade tree (P.T. Alvim, personal communication). However, in the present chapter Cacao is treated as a shade tree, based on Cuatrecasas, 1964, Okali & Owusu, 1975, Raja Harun & Hardwick, 1988, Serrano & Biehl, 1999 and Augusto, 1997. In his comprehensive and thorough studies on the genus Theobroma, the first author relates the discovery of glabrous leaves (or nearly glabrous, but adapted to tropical forest ecology) in the section Theobroma, which includes cacao. This leaf constitution provides clear evidence that Cacao is in fact a typical shade tree. Okali & Owusu, 1975, studied the effect of light on the photosynthesis of young Cacao, and found saturation at a light intensity of 400 µmol photons.m-2s-1. Later, Raja Harun & Hardwick, 1988, conducted experiments, which corroborated these results. Augusto (1997), who studied gas exchange in cacao trees under supplementary irrigation in the State of Espírito Santo, found that irrigated cacao presents increasing values of net photosynthesis rates, at a light intensity of up to 600 µmol photons.m-2s-1. The photosynthesis tends to stabilize at values close to 5.6 µmol CO2m-2s-1. On the other hand, not irrigated cacao presented increasing photosynthesis values, up to a light intensity of 250 µmol photons.m-2s-1, and became steady at approximately 3.4 µmol CO2m-2s-1. In the tropical regions where Cacao is cultivated, a typical sunny summer day may have a light intensity of approximately 2000 µmol photons.m-2s-1. Since cacao is a C3 plant, it seems to reach its light saturation at approximately 600 µmol photons.m-2s-1. In the light of these facts the need to shade this crop becomes evident. Although cacao is a typical shade tree, the advantages of shading are actually not related to an optimized light intensity level for plant growth and yield (Alvim, 1977). Main benefits are the conservation of soil and water supply, a longer lifetime of plantations, higher production stability, lower incidence of pests and diseases, less wind and weeds, and a diversified agricultural use through the utilization of economically appealing plants. In coffee plantations with correct fertilization as well as efficient weed and disease control, top yields have been achieved under full sunlight in Costa Rica (Fournier, 1988). The production of this crop is on average 10 to 20% higher than that obtained under equal conditions, yet with adequate shade. The same author, however, warns that to date there is no experimental evidence that would allow a general recommendation for the cultivation of coffee under full sunlight as the best agronomic, ecological, and economic long-term alternative. His statement is based on the following facts: i) No information on chemical, physical, and biological soil characteristics of coffee crops cultivated under full sunlight for at least 25 years is available yet; ii) So far, there is no adequate information available that would allow the calculation of a coffee plantation’s lifetime under full sunlight, or the economic assessment whether a profit gain of 10 to 20% of this production type over the crop under shade would cover the cost of renovation; iii) It is necessary to quantify the nutrients, mainly nitrogen, incorporated into the system by the shade trees, when leguminous trees are used for shade. Therefore, the safest alternative for coffee production under Costa Rican conditions is the cultivation under uniform and regular shade, particularly recommended for regions with high temperatures and light intensity, prolonged periods of water stress, and low soil fertility. It is important to emphasize that a shaded coffee plantation will form an agroforestry system with additional benefits, such as: wood, firewood, forrage, besides other, not easily quantified or valued advantages, such as soil protection against erosion and easier infiltration of water into the soil profile. Specifically in highland regions of greater soil fertility with a moderate climate, coffee cultivation under full sunlight may be recommendable. In this case, there are no drawbacks for soil characteristics, due to the lower decomposition rate of the organic matter. Fournier’s statements (1988) on coffee can be extrapolated to cacao cultivation, mainly true for southern Bahia, where soils are shallow, terrains hilly, and there is a dense net of watercourses. In this region, the elimination of cacao shading would affect the entire local physiography. Much has been said about the interaction between shade and witches’ broom disease (Crinipellis perniciosa). It is argued that shade is beneficial because it reduces disease incidence. L.A.S. Dias (Chapter 13) points out that there is still little experimental information on this interrelation, which deserves more attention. He claims that shade reduces the fungal spore production, as well as the peaks of leaf flushing, flowering, and fructification of cacao, due to the stable temperature and humidity in the canopy. Since all these tree parts are susceptible to infection by the pathogen, a more shaded tree will suffer less attacks, and will therefore acquire a lower number of brooms, compared to another, less shaded trees. However, the level of shade that would benefit cacao and and reduce the pathogen incidence is yet unknown (Dias, Chapter 13). The cacao leaf size varies according to the light quantity. Leaves are bigger where trees are grown in dense shadow. The leaf size partially compensates for the lower photosynthesis rate in the shade. Adult cacao develops a self-shading that significantly modifies its relation to light. If shade trees are suddenly removed from a cacao plantation, when environmental factors vary greatly, or when diseases and insects attack the trees, their external as well as internal leaves, un-adapted to the full light, eventually drop. This phenomenon is known as gradual death or die-back; the tree tends to become lanky and senescent. Cacao leaves of trees cultivated under adequate shade have a lifetime of over one year (A.A. Almeida, personal communication). In cultivation under full sunshine, though, their lifetime does not exceed six months. Müller & Biehl, 1993, found that the mean lifetime of cacao leaves (evaluated after leaf emergence) in trees developed under high and low light intensity was 450 and 250 days, respectively. In cacao regions where permanent shade is recommended, these shade trees must be selected according to their economic usefulness, besides serving as shade. Nitrogen-fixing leguminous trees or trees which provide fruit, wood, rubber, palm heart, etc are suitable. Windbreaks: Wind speed can be a serious problem in cacao-producing regions, for example in Linhares, State of Espírito Santo, (Alvim, 1977) or the Recôncavo region of the State of Bahia (Pinho & Müller, 1987), thus protective barriers become necessary. Wind action hinders the establishment and maintenance of cacao, young plants being most sensitive (Sena Gomes & Kozlowki, 1989). There are indications in literature about protective barriers in areas of wind-exposed cacao plantations (Alvim, 1977; Wessel, 1985). Besides causing mechanical damage, the wind can pave the way for the dissemination of diseases in the plantation, as in the case of pink disease, Phanerochaete salmonicolor, (Almeida & Luz, 1986) and witches’ broom disease ([[Luz et al., 1997]r550937). Alvim (1977) comments that the reduced cacao yield in African countries between February and June is not only a consequence of water deficit, but also of desert winds coming from the Sahara in the months of December to March. Instruction manuals on the cultivation of cacao unanimously agree on the point that wind is extremely harmful, mainly for young plants, and windbreaks are recommended in such situations. For adult cacao, however, most information on the detrimental influence of wind is based on visual observations only, without experimental proof. Summing up; shading not only reduces the light incidence, but also reduces harmful wind effects. |
| Environmental improvement; liming and fertilization, soil water availability, weed management and pruning. | Return To Table of Contents |
|---|
| Liming and fertilization: Inadequate or no fertilization at all has been a main cause for the low productivity of Brazilian cacao plantations. The problem is not only the lack of fertilization, but also unbalanced fertilization with careless use of lime, calcium, magnesium, and sulphur, as well as the omission or sloppy use of micronutrients. Fertilization of any crop begins with the soil analysis, proceeds to acidity adjustment, and ends with the fertilizer application (Malavolta, 1993). Before discussing fertilization we must therefore discuss liming. In spite of the relative aluminium tolerance of cacao, liming is an indispensable practice, not only to correct the soil acidity, but also for the supply of calcium and magnesium, to bind toxic manganese, raise the cation exchange capacity, increase the utilization of applied macronutrients, and to accelerate the decay rate of organic matter. Fertilization of cacao plantations on soils with a very acid pH leads to a low utilization rate of the fertilizers, boosting production costs without providing a reasonable yield gain. Research and experience have demonstrated that the effects of light and nutrition in cacao plantations are linked to each other. In practice, cacao shading and the response to fertilizers cannot be dealt with separately (Wessel, 1985 and Matos, 1991). The effect of shading is very complex, since it includes, besides the reduction of the light intensity, temperature, and of air movement (wind), it impacts on the relative air humidity and soil moisture. The reduction of radiation is crucial, since it is one of the main controlling factors of photosynthesis. The higher the radiation, the higher the temperatures, the higher the turnover (metabolism rate), and, consequently, the higher the requirements for water, nutrients, and cultural treatments. In Bahia, Brasil, effects of fertilization were first investigated in 1964. Together with the elimination of shade on the plantations, NPK fertilization gave rise to 40 to 80% higher yields (Cabala et al., 1970). Cabala et al., 1971, evaluated cacao yields on 21 plantations under different soil moistures, during seven years. These authors found that a fertilization of 105 kg of N, 146 kg of P2O5, 64 kg of K2O, 45.7 kg of CaO, and 34 kg of MgO (/ha/year) gave significant crop yield increases, especially in low fertility soils and plantations where the shade had been eliminated. A considerable residual effect of this three-year consecutive fertilization was observed for three more years after the application. The influence of shade on nitrogen nutrition in cacao has also been studied (Santana and Cabala, 1983). These authors evaluated the reuse of nitrogen through cycling in a 17-year-old cacao plantation shaded by coral trees (Erythrina glauca). Nitrogen inputs by vegetation residues and husks of harvested fruits, as well as losses by leaching and bean exportation were considered. Within one year, over eight tonnes/ha of residues dropped onto the soil surface, summing up to approximately 140 kg/ha of N. This quantity is equal to a six to eight-fold export of this nutrient in a harvest of 1 ton of dry beans. It is notable that within six to nine months the stage of dry matter decomposition of almost all residues had reached 50%. An adequate soil for cacao cultivation requires: pH values near 6.2; sum of bases about 12 cmol/100g; a resin extracted P content of 20 to 25 mg/dm-3; sum of bases around 60%; organic matter above 3.5%; relation calcium: magnesium in the 0-15 cm soil layer around 4; a Ca +Mg/K quotient above 25, and absence of physical limitations (Cabala et al., 1989 and Malavolta, 1997). Pebbles or stones in the soil profile, as well as the colour of the horizons are not significantly correlated to the tree yield (Souza Júnior et al., 1999a e 1999b). In decreasing order, the extracted and exported macronutrients by cacao are: N > K > Mg > Ca > P, and > S, and the micronutrients: Fe > Mn > Zn > B > Cu, and > Mo. Fruit husks (broken shells of harvested fruits) contain high nutrient quantities with up to nearly 90% of the total extracted K and Ca (Malavolta, 1997). Fruit husks should therefore remain on the plantation, properly treated with lime and/or copper, and covered with leaves to avoid dissemination of the diseases (pod rot, Phytophthora palmivora, and witches’ broom, Crinipellis perniciosa), and return their nutrients to the crop. Another possibility would be to break up the cacao fruits outside the plantation, compost the husks, and apply them as fertilizer after fermentation. The relation between fertilizer application and the cacao yield with a view to define deficiency and/or toxicity, determine critical levels, and recommend nutrient doses for the crop has not yet been scientifically established. In many trials, no significant answers to the relation production-fertilizer application were found (Morais et al., 1978, Cabala et al., 1982 and Nicolella et al., 1983). The great environmental heterogeneity in small areas, mainly regarding the degree of shade and physical-chemical soil characteristics, is responsible for such unsatisfactory or inconsistent results; this effect is intensive in plantations under “cabruca”, a selective logging, where the native forest is thinned out for cacao implantation, and shade is very heterogeneous. Other factors responsible are climatic variations among years and sites, sometimes inconsistant spacing, and the long cycle that requires several years of production to obtain conclusive results. Other limiting factors are the water availability at the moment of fertilizer application (since fertilizers do not provide positive responses under water deficit or excess) and different methods of conducting and evaluating the experiments. CEPLAC has not recommended the use of fertilizers for cacao cultivation in the State of Espírito Santo, owing to a lack of response to their application - a fact which has caused controversy among technicians. Soil analyses showed low levels of P and medium levels of K, Ca, and Mg. Nowadays, there are no doubts as to the fact that the response to fertilizer application in the referred region had been compromised by the yearly water deficit or occasional flooding of the plantations. For recommendations on cacao fertilization factors such as soil analysis, nutrient export by production, leaf analysis, the level of shade, and the desired yield must be taken into account. To evaluate the the nutritional state of trees it is necessary to monitor each stand through soil and leaf analyses. Incacao plantations attacked by witches’ broom, the virulence of the causal agent of the disease, the fungus Crinipellis perniciosa (Stahel) Singer, was altered by a lack of nutrients. In susceptible cacao plantations of the ‘Catongo’ cultivar, under continuous balance of macro and micro nutrients, the severity of the disease was alleviated, while cacao plantations of the tolerant hybrid ICS 6 x SCA 6 were disease-free. This hybrid is different to others due to its high Mn, N, Ca, Fe, and B contents. It is believed that, of these nutrients, Mn is the nutrient with the strongest influence on witches’ broom tolerance (Nakayama, 1995). This statement is based on the fact that manganese is essential in the formation process of phenol compounds, which are important for the manifestation of tolerance towards diseases in plants. In spite of the variety of research results that confirm how beneficial fertilization is for cacao plantations, there are still technicians and farmers in many countries and regions who are suspicious about the economic usefulness of this practice. Soil water availability: Since cacao is a typically a plant of the humid tropics, its water consumption is high. It is sensitive to lack of moisture in the soil and water deficiency causes dire problems. Quantity and distribution of the cacao yield are more determined by rain than by any other ecological factor. Trees grown on soils with a low buffer capacity and low organic matter content are the most affected by water stress in drier years. Fertilization trials in the Ivory Coast showed that fertilization, with or without irrigation, doubled the number of flowers. With fertilization and irrigation, the number of harvested fruits was 63% higher than in control plants (Jardin & Paulin, 1988). Although cacao irrigation has already become a practice in countries such as Venezuela, Colômbia and Peru (Alvim, 1967 e Wood, 1985), little research into the topic has been done and guidelines for an optimized irrigation management system are not yet available. The scarcity of studies on how soil moisture influences the production of cacao plantations might be due to the fact that cacao is mainly grown in regions where, characteristically, the total annual precipitation outstrips water losses by evapotranspiration. Where precipitation is below 1 200 mm, cacao can only develop successfully under irrigation. This is the case in Venezuela, where the precipitation is 700 to 800 mm/year (Alvim, 1977) and in the North of Espírito Santo, where rainfalls are accumulated in few months of the year, in spite of an annual precipitation of 1 200 mm/year (Augusto, 1997). Cacao cultivars equipped with an efficient stomatic regulation mechanism lose less water by transpiration under water stress, which indicates an important adaptation strategy (Balasimha, 1988, Balasimha & Daniel, 1988 e Balasimha et al., 1988). The tolerance to drought in cacao can be explained mainly by an efficient stomatic regulation, which maintains the leaves’ turgor. Weed management: Weeds compete with the cacao trees for water, light, nutrients, and CO2 and can be controlled either by manual, mechanical or chemical means. A very common practice on cacao plantations is slashing which involves the use of several region-specific tools (scythes, machetes, etc). The traditional management system entails high operational costs during the first four crop years, mainly due to the high frequency which is required to reduce the competition between the invading plants and the cacao. These high costs are due to the lack of and high cost of rural manpower, which compels cacao farmers to search for alternatives to minimize the problem. The use of herbicides (chemical control) has increased significantly in Brazilian cacao regions, not only due to the efficiency of this method, but also because of the cost reduction when correctly applied (Silva Neto, 1994). Aspects which have led to the increased use of chemical weed control are also the high costs of manual weeding, the appearance of weeds that are difficult to control, as well as the advantage in the practice of heaping up mulch which protects the soil against erosion. To avoid failures in the use of herbicides, a locally adequate product must be chosen; errors in the calibration of the application equipment must be avoided; the correct moment of application must be respected and one single active principle must not be used for years on end. However, when a technically well-established crop cultivation is mature (under adequate stand and shade), designated by the familiar name ‘leaf touching stage’, weeds no longer threaten the crop. The dead cover or mulch, which is formed by fallen leaves, and the shade hinder the germination and growth of undesired plants. In such plantations most weeds grow on clearings where a cacao or shade plant died, or in the border lines between stands. Pruning: To prune sensibly, some basic principles of the vegetal physiology must taken into consideration. The plant’s production capacity is a direct function of the leaf surface area, and each crop has its proper leaf area index (LAI). In the case of cacao, which grows in association with shade trees, it is not so easy to determine this index. High yields cannot be expected from small plants with a reduced leaf area. The use of the sunlight should as efficient as possible, which requires a good leaf cover of the soil. On the other hand, the photosynthesis of plants with an excess of leaves inside the canopy is reduced. The small amount of light they receive makes them function more like a drain than as a source of photosynthesis assimilates. There are two kinds of pruning; the formative and the maintenance pruning. The first can cause irreversible damages to the cacao plants, and is therefore not advisable. Presently, only maintenance pruning is recommended. Orthotropic branches that grow from the base of the stem or unproductive branches (locally called ‘chupons’) are removed, as well as dry and ill-placed branches that normally grow towards the plant centre at any time of the year. It is worth remembering that the quantity of ‘chupon’ branches that grow on the stem and branches is large, calling for frequent pruning to avoid reduction of fruit production. Where the ‘chupons’ are not cut away, they grow and emit lateral branches, which form new crowns, giving the plant the appearance of canopy stories. Experimental results indicate that the practice of pruning is imperative to obtain high crop yields. In correctly pruned plantations, the quantity of flowers and fruits produced increase significantly in relation to non-pruned ones. In plantations infected by witches’ broom, the pruning practice is indespensible as the disease causes hypertrophy of the branches, shoots, and flower cushions. Pruning of these plantations by cutting away green or dry brooms is one of the forms of co-existence with the disease, besides being effective in the reduction of the plant height and separation of the crowns from each other. The efficiency of phytosanitary pruning as a control of witches’ broom has already been demonstrated (Almeida & Andebrhan, 1989 and Albuquerque et al., 1995). Still, the recovery of a pruned cacao plant is slow and gradual; the production and level of the losses are only stabilized from the third year after pruning. Until the production becomes stable and the disease incidence drops, investments of the cacao farmer are required, since manpower is expensive and the return flow is only re-established from the third or fourth year on. In areas less than 3 to 5 ha, this practice might best be carried out with help of the farmer’s own family. The practice of pruning is important for clones in order to correct and lift the crown. In clonal plantations, cacao plants are less tall and tend to present an increased growth of side shoots as well as some branches being very close above the soil. Pruning must take place from the inside of the plant outwards. Crossed branches have to be eliminated, but no spaces may be left which would allow light to fall on the plant stem or the ground to avoid necrosis of the stem and a higher incidence of weeds caused by the direct sunlight. In most Brazilian cacao plantations, however, pruning is excessive; even productive branches being eliminated. Such excessive pruning increases the frequency of leaf flushing and reduces the fruit yield. |
| Environmental improvement; pest and disease control. Processing. | Return To Table of Contents |
|---|
| Pest and disease control: There is an economically important disease in each producer region, for example the presence of witches' broom (Crinipellis perniciosa) in the State of Bahia, Brazil, and also in the Brazilian Amazon. Pod rot (Phytophthora palmivora), despite being less harmful than witches' broom, appears in all cacao-producing regions of the world, while witches' broom is restricted to the Americas. Monilia pod rot (Monilia roreri, another serious disease, which attacks the fruits, is spread all over the Americas, except for Brazil. VSD (vascular streak dieback), caused by the fungus Oncobasidium theobromae, is the principal disease in Malaysia (Silva, 1985). Besides these, there are other diseases of minor importance, such as pink disease (Phanerochaete salmonicolor), Verticillium wilt (Verticillium dahliae), anthracnose (Glomerella cingulata) canker and root rot etc. The last two are caused by a pathogen complex; however, in this context, the main ones are Phytophthora sp. and Rosellinia sp. respectively. The incidence of diseases is lower in dry regions, but pests like thrips, 'vaquinhas', lizards, beetles, borers, stink bugs, mites, among others, occur more frequently and reduce cacao yields. Moths, woodworms, and other insects threaten the product quality during the storage of the dry beans. In pest and disease control it is increasingly necessary to apply integrated pest management (Abreu et al., 1989 and Gramacho et al., 1992), with the objective of diminishing the use of agricultural pesticides, which cause harm to man and nature. It is worth repeating that the simple presence of a disease does not justify its chemical control. The level of economic damage done by each one must be evaluated, and, below a certain level, a chemical control could be deemed economically unfeasible. In many cases this level has already been defined, and cacao farmers, assisted by a technical advisor, must know about this level to apply rational control cacao-attacking insects/mites and phytopathogens. Biological control must also be taken into consideration - the cultural control that helps so much to live with these pests and diseases - since it minimizes the quantity of applied fungicides and insecticides/acaricides. To date, there is still no efficient chemical control for the witches' broom disease. CEPLAC has been concentrating on substituting a great part of the cacao plantations in Bahia, susceptible to Crinipellis perniciosa, by tolerant genotypes. This gave rise to the project 'Biofábrica' (Bio-factory), developed by the federal government in collaboration with the Bahia State government and the cacao farmers. It has the 5 year goal of substituting approximately 300 000 ha of susceptible cacao by tolerant genotypes, obtained by means of rooting of cuttings and grafting (see chapters 9 and 13 on the same topic). An integrated management system for witches' broom proposes: i) phytosanitary pruning to remove infected material; ii) application of chemical products to protect developing fruits; iii) use of fungus tolerant genotypes; iv) application of the biological fungicide "Tricovab" (Trichoderma stromaticum) on the material which has been removed and left on the ground. These measures are expected to revert the devastating picture created by the disease in that region. In the State of Espírito Santo technicians of CEPLAC together with the cacao producers are producing witches' broom tolerant clonal seedlings. This is a preventive measure in view of the predicted arrival of the devastating disease, which, so far, has not been reported in the State. It is important that breeding programs, which aim at increased crop yields, must also take into account resistance to pests and diseases. (Chapters 7, 8, and 9). Processing: Processing of the cacao beans involves all the practices that together guarantee a superior quality raw product that conforms to the required industrial standards. A sequence of five interdependent stages makes up the chocolate taste and flavour: fruit harvest, bean extraction (breaking), fermentation, drying, and storage. Fruits harvest and bean extraction: Fruits are harvested when they are fully ripened. Diseased or damaged fruits should always be removed from the tree on this occasion. Overripe fruits sometimes present weight loss of the seeds or internal germination. Thus, during the pre-treatment, the rootlet may break and leave punctures in the seed, which become an open door for insects or pathogens. On the other hand, the harvest of immature or partially ripe fruits also leads to a drop in output, and causes insufficient fermentation as well, due to the low sugar contents in the pulp of immature seeds. After harvest, the seeds must be extracted as quickly as possible, within a maximum of five days. The fruits can be broken up on the plantation or elsewhere. During this process, diseased fruits must be discarded, so as not to affect the fermentation of the healthy fruits. After fermentation, the cacao husks should be returned to the plantation, treated against diseases, and, if possible, ground, since this material is synonym to savings in fertilizers, particularly of K2O. The number of harvests per year varies according to the producer region. In Bahia, due to the almost uninterrupted regime of cacao flowering, at least 10 harvests a year are possible. On the west coast of Africa, where production is concentrated in only three months per year, the number of harvests is considerably lower. Fermentation: This is the initial stage of the formation of substances that contribute to the development of the peculiar characteristics of taste, scent, and colour of chocolate, formed during the industrialization process. Superior chocolate can only be produced with well-fermented cacao. Unfermented cacao is tasteless (Lopez, 1984), and only used for fat extraction. Fermentation causes chemical, biochemical, and physical alterations in the seed, normally manifested in changes of colour, smell, and temperature of the cacao pulp. During the process, the beans must be regularly turned to aerate and homogenize the pulp. The turning must begin on the second day, and continue daily until the end of the fermentation process, which lasts between five and seven days. Its duration depends specifically on the pulp content in the fruits, and the ambient temperature. Seeds with less fruit pulp require less fermentation time, while days with lower temperatures extend the fermentation for one or two days. The seeds, just after the extraction from the fruits, are of a milky white to purple colour on the outside, and gradually turn light to dark brown during fermentation. At the beginning of the process, the cacao pulp has a smell of alcohol and vinegar from the third or fourth day of fermentation. The temperature of the pulp normally rises to between 47-50oC. When the temperature becomes stable, the fermentation process is complete. An un-completely fermented product is violet-coloured on the inside, considerably acid, and produces a chocolate of weak scent and bitter and sour taste. On the other hand, excessive fermentation results in beans of a deep dark brown colour, with an unpleasant odour of ammonia or putrefying material. Chocolate produced with this kind of beans will normally smell of fish and other strange odours that affect the product's quality. In Brazil, fermentation is normally realized in wooden troughs called "cochos", with holes in the bottom to drain the excess pulp. In most African and Asian countries, however, fermentation is carried out either on banana leaves or plastic sheets, with drainage in so-called 'African batteries' or mangers/troughs (Cunha & Serôdio, 1991). Drying: At this step, the excess of water content in the recently fermented beans is reduced from its initial moisture content of 50-55% to 7-8%. This can happen naturally, under the sun, or in artificial dryers. When drying occurs under the sun, the final product quality is satisfactory, provided the beans are turned regularly and frequently and protected from rain or dew. When artificial dryers, of various types and sizes, are used, however, several aspects must be taken into consideration. These include everything from the temperature of the drying air to the level of the oven maintenance, to avoid contamination of the beans with the smell of smoke. The smell of smoke has been, and is still, one of the major problems of the quality of Brazilian cacao. In Brazil, a so-called 'barcaça' is used for sun-drying. The main characteristic of this structure is a mobile roof, which rolls on tracks, while the ground is firm, made of wooden boards. On most cacao farms in Bahia, the 'barcaça' is constructed as a first floor, while the ground floor is used as a simple residence or sometimes a storage room. In both cases, an open cooking fire of the residence or the insecticides/pesticides in the storage room may add the smell of smoke or other substances to the beans, impairing the final product price. In other producer countries, drying is carried out on cement or asphalt areas, or on plastic sheets. There are several models for dryers, from the most simple, suitable for smallholders, up to the most sophisticated ones, which turn the cacao pulp automatically. This larger and more complex equipment is recommended for large estates. It is well worthwhile to remember that a good dryer is not the one that dries cacao very quickly, but the one that supplies an excellent quality product. Storage: Once adequate humidity is attained, the beans must be stored in linen sacks to maintain the quality until they are sold. Standards define that each sack must contain 60.5 kg of dry cacao beans. Well-fermented and dried cacao must be kept in a clean and well ventilated place, free of insects and rodents, which attack and degrade the product. No chemicals, fuel, paint, or smoke emission are allowed close to the storage room, since the beans absorb any smell very easily, due to their high fat content, and thus would lose quality. As the product is highly perishable, cacao storage should not exceed 90 days. The storage phase should not be too long, since the beans can be attacked by insects. Where the relative humidity of the air is high, there is the danger of mould infection which impairs the product quality. To preserve the dry product a little longer, it is recommend to cover it with a plastic, to avoid a re-absorption of humidity. For more details on the technology available for cacao preparation read Cunha & Serôdio, 1991, where this topic is broadly discussed, including references on each stage. |
| Socio-economic panorama; The crisis. | Return To Table of Contents |
|---|
| Socio-economic panorama: Cacao is one of the agricultural commodities with higher price variations. Its history has been marked by a succession of cycles of notable price fluctuations. A historical analysis of the world’s cacao production, with a frequently shifting centre of highest production, reveals a peculiar phenomenon: commonly, a country develops its production, reaches a peak, and then declines. Up to 1990, the only countries that appeared to be an exception to this rule were Brazil and the Ivory Coast in spite of the notable production drop in Brazil during that decade (Jarrige & Ruf, 1990). Cacao is a typically labour-intensive crop of tropical underdeveloped or transitional countries, cultivated by low-income producers under intensive exploitation of unskilled workers. By contrast the main product chocolate, is consumed by people in developed countries with a high purchasing power. Over 80% of the world’s production stem from only seven countries (Figure 1.1). Among these, the Ivory Coast accounts for approximately 50% of the total production. The elevated productivity attained by Indonesia and Malaysia, owing to these countries’ investment in research and technology, deserves mentioning. On the other hand, Brazil and Malaysia are the two countries that have, for different reasons, gone through the most significant production drop. In the late 90s Brazil’s productive structure was shattered by the arrival of the witches’ broom disease in 1989. Malaysia, in turn, reduced its cacao sector by reducing the planting areas as a result of low international prices. Altogether, the cacao production chain of the developed countries has a turnover of over 20 billion dollars per year (Nascimento et al., 1994 and the National Agricultural Forum, 1997) (FNA- topic group cacao). Values are expressed in dollars, at an exchange rate of R$ 1.81. The Brazilian cacao production chain involves, at present, investments of U$ 1.3 billion, 0.93 billion of which flow into the primary sector (soil, trees, and infrastructure). It provides about 300,000 direct jobs, but over 3 million people depend on cacao. Cacao is the economical backbone of practically 100 municipalities in Bahia, where it is cultivated on 29,000 farms, covering an area of over 700,000 ha. Brazil is the fourth world cacao producer, although this position is threatened by Nigeria and Cameroon, and has the fifth largest chocolate industry. Data on Brazilian production, given as 285,000 tonnes of dry beans by FAO (Figure 1.1) arouses controversy, despite the repetition of this same value by AGRIANUAL, 2000, a reputable Brazilian journal which compiles data on national agricultural production. According to data of the International Cacao Organization ICCO, (ICCO, 1999), this production was only 133,000 tonnes (Figure 1.2), 47% of the value published by FAO. In view of such a discrepancy, and with the lack of official statistical data provided by Brazil, information on the cacao volume produced in Brazil is unquestionably being manipulated, and it is up to the parties involved as to which data are presented. As a commodity, the price of cacao is defined at the stock-market. Therefore, the expectation of an elevated production causes reduction of the product’s market price, and lower prices are paid to producers. Since Brazil cannot refute this data, it depends on production statistics of the crop established by other, more organized, members of the production chain, which take advantage of the country’s vulnerability. Brazil was the world’s second largest producer until 1993, when production began to decline gradually, due to several factors previously mentioned. Data on the Brazilian production (Figure 1.2) point out that for the crop of 1999/2000, Brazilian cacao farmers received only 20% of the value paid in dollars/‘arroba’ for the crop of 1980/81. An ‘arroba’ (approximately 15 kg) is the customary weight unit for cacao trade in southern Bahia. In 1979, cacao had brought in U$ 0.94 billion in foreign exchange, thanks to the high price attained by the product on the international market. During the eighties, the annual mean revenue receipts dropped to U$ 0.61 billion, and will probably not even achieve U$ 300 million per year in the 90s. In the last three years, Brazil has begun to import instead of export cacao since the Brazilian production no longer satisfies the demand of its own industry. In 1999, at least 18,000 tonnes were imported from Indonesia and the Ivory Coast (AGRIANUAL, 2000). The present production (99/00) is, approximately, 2.3 million sacks/year, or about 138,000 tonnes. Between 3.0 and 3.5 million sacks, or about 180,000 to 210,000 tonnes are processed by the country’s industry of chocolate and derivatives. There is, therefore, a deficit of 700 thousand to 1.2 million sacks/year, that is, 42 to 72 thousand tonnes, which must be compensated by imports to keep the national industry afloat (data of the Cacao Trade Department of Cargill, Ilhéus, Bahia). For countries such as the Ivory Coast and Ghana, cacao and its by-products are economically relevant, since they account for 40 and 60% of their export earnings, respectively (Alger & Caldas, 1996). The same is true for Indonesia and Nigeria, although the economic importance of cacao has declined in the latter, when oil fields were discovered within its borders. Nevertheless, actual Nigerian government politics aim to intensify cacao cultivation once again. Previsions foresee that Indonesia will soon outstrip Ghana, and consequently become the world’s second greatest producer. The economic importance for Malaysia and India is less essential (Silva, 1985). According to the ICCO (ICCO, 1999) the world crop 98/99 is forecasted at 2,747,000 tonnes of dry beans, 2.1% above the last harvest, for a roasting estimated at 2,795,000 tonnes, that is, 0.4% above last year’s. The production is still slightly lower than the grinding, with a deficit of about 100,000 tonnes. International stocks, however, practically remained at the 1998 levels (AGRIANUAL, 2000). European stocks attained 817,000 tonnes of beans in May 1999, with a slight drop (0.8%) in relation to the beginning of the same year. The Crisis: In the 80s, the world’s cacao production greatly increased due to the remarkable expansion of cultivated areas in the producer countries, kindled by the high quotations of the product in the biennium 1977/78. The Ivory Coast nearly quadrupled its cultivated areas, while Malaysia and India multiplied their areas 16 and 40-fold, respectively. Last but not least, Brazil also promoted the expansion of the crop by means of PROCACAU, founded in 1976, and - in spite of never attaining its initial ambitious aim of producing 700 thousand tonnes/year within 15 years - it managed to increase the then cacao bulk yield by 50% (Santos Filho, 1995). Before analyzing the reasons for cacao’s dire straits in detail, it is worthwhile recalling a seminar of Adriano Romariz, a Professor for economy at the University of São Paulo, held at CEPEC, Ilheus, Bahia, in 1984. On this occasion, he drafted a very daunting prospect for the crop (Santos Filhos, 1995). Cacao was at that time obtaining elevated prices on the international market, but Romariz already predicted the failure of the expansion plan of cultivated areas to triple Brazil’s production until 1993; he mentioned the volatile market, the perishable product, and the risks of extending a perennial crop without analyzing the consequences carefully beforehand. Moreover, he pointed out that the high product prices would stimulate other producer countries to enlarge their cultivated areas as well, which in the end would affect the global market negatively. It was later verified that exactly the amplification of the areas, and, in consequence, the excess of product supply, had been the main factor to spark off the current crisis. At that moment, PROCACAU revealed its much more political than technical character. Actually, the staggering increase of production in practically all producer countries led to a considerable worldwide swelling of stocks. As production surpassed consumption, reserves piled up 1.25 million tonnes (equivalent to an eight-month global consumption) (Santos Filhos, 1995). In view of the increased offer, chocolate industries all over the world altered their purchase strategy, and began to adopt the “just in time” method, so cacao stocks were drastically reduced. The increased production, increased stocks, and, later, the reduction of the buyer’s stocks explain the low quotation of the product during the last decade. Furthermore, the fusion of the world’s chocolate industry down to four or five mega industries can be cited as another factor. These multinationals agree on similar purchase politics, which did not allow any price rises on the market. Summing up, it can be duly said that chocolate industry has had as much and as cheap cacao as it could want in this period, without running any risk of supply problems. In protest against the inexorable price drop, the Ivory Coast retired from the market for four months in August 1988, in the attempt to block the price drop and impose a new price high. This try missed the mark however as it was no more than an isolated attitude of only one producer country. Chocolate industries simply bought the cacao retained by the Ivory Coast from other countries, assisted by the high production and large stocks. It was clear for industry representatives that the country, alone, would not be able to modify the product supply, and besides, that it would not be able to resist for much time, since cacao is a perishable product and mainstay of the country’s export products. This policy of the Ivory Coast might well have born fruit if there had been an agreement among the largest cacao-producing countries. As an isolated position, it failed. In the Chinese idiom, the word crisis has a double meaning: problem and opportunity. In the global cacao culture however, and specifically the Brazilian, it seems as if crisis means only problem. Crisis has always been part of the southern Bahia cacao culture, in contrast to other agricultural export activities of Brazil (Menezes & Carmo-Neto, 1993). In this century, there were three especially critical stages. The first in the late 20’s, when the ICB was founded in 1931; the second in the late 50s, when CEPLAC was created in 1957, and, finally, the third, which began in 1987 and lasts to date, and has been devastating for the region and the crop, due to its intensity and duration. The principal components of this last crisis were: i) accentuated capital loss of the producers, due to the low international market quotations for cacao; ii) the high degree of producer debt with exporters and the industry; iii) low competitiveness of the sector, with high costs on the one hand, and old and decadent plantations on the other hand, together with inadequate management and inert government politics; iv) the accentuated decline of the participation share of the product in the return receipt of the State of Bahia (from 50% to 5%); v) the outspoken indifference of the federal government to search for permanent solutions in a battle against causes instead of consequences; vi) the arrival of the witches’ broom which triggered a socio-economic, agricultural, and ecological disaster in the region; vii) the incompetence of research to spur the hunt for resistant genotypes, and thus anticipate the arrival of witches’ broom; and viii) the water stress in over six successive years in the region which had previously received adequate rain distribution. The extended drought caused death of cacao trees and impoverished plantations, undermining the plantations’ productivity for years to come. Since the end of the 80s, the national cacao culture plunged into crisis and stagnation because of the decline in production and productivity, increased production costs, excessive tax rates, idleness of the cacao processing equipment, dwindling of the national cacao export and industrialization companies, regional unemployment and impoverishment, and the loss of confidence of the producers in the government’s actions. Another aggravating factor at this stage was that the cacao sectors’ economies are isolated and not convergent. As a rule, producers live in the country, while the exporters live in the major cities. The industries form multinational companies, the Brazilian chocolate industry lies in southern Bahia, while the consumers mostly live in other States or countries. The producers are not organized; on the contrary, they have a totally individual and unarticulated profile. This fact caused a great crisis of identity and of political representation. The decisive power of the cacao economy lies in the hands of more organized and a good deal more sophisticated groups. Interests of industry and exporters lie far outside the cacao regions, mostly included in the circuit Salvador - New York - London. In an analogy to coffee, one observes how coffee producers got themselves organized in over a hundred cooperatives, renewed the coffee plantations with highly productive lines, and applied high-tech in the plantations. Moreover, in spite of coming from different sectors, the public, private industry, exporters, as well as the producer group itself spoke the same language, with the only objective of increasing the competitiveness of Brazilian coffee, obtaining top productivity and superior quality. Cacao farmers failed to get organized into cooperatives and did not come to terms with the capitalist structure, which is now the basis of the international cacao market, with its multinational companies, industries, stockbrokers, exporters, and stock market. One proposal for the consolidation of a cooperative network was rejected in 1968 by exporter representatives, which forced CEPLAC to beat a retreat and practically deactivate its Department of Cooperation (Nascimento et al., 1994). Even worse, the cacao farmers themselves did not recognize this weak point of theirs, nor did they have a perspective for survival in the long term. Their cultural formation was based on a peculiar individualism, which makes them behave as if their problems were the most relevant and had to be solved their way, in an office of CEPLAC. They were not aware that they are the people who should control, own, and keep the command of the economy and wealth they generate with their own hands. This is a cultural component that calls for changes. Confirming the information above, Willumsen & Dutt, 1991, have compared the Brazilian coffee and cacao economies, and concluded that cacao farmers do not have an entrepreneur’s vision, opposed to the entrepreneurial spirit of coffee farmers, which might explain the lack of development of the cacao sector. All crises faced by the cacao sector had the same fundamentals, but the battle was constantly fought against the consequences: low prices, market oscillations, low productivity, and indebtedness. An average production of 401 kg/ha of dry cacao beans, which is the national productivity in an average stand of 600 trees, is simply unacceptable. The core question has always been the incapacity of the producer to deal objectively with these crises, and today one acknowledges very clearly that the problems of the sector could have been levelled out, if solutions had been got under way by the mid-eighties. This was not the case, because producers awaited, in vain, for some intervention from the State. The absence of measures entailed a lack of financial sources, with extremely negative impacts on society. Everything indicates that the situation for a revival of the crop requires, even more, the self-management of the producer, although it is necessary that the government promotes this process. Without this help, the goal will not be reached, in view of the insufficient levels of know-how, education, information, and culture that cacao farmers usually have. Cacao cultivation is in crisis as a consequence of inefficient management. Factors like witches’ broom and the policy of reduced export subsidies have merely helped accelerate its decline. Symptoms were evident, and the earlier crises had already demonstrated this by unveiling the all too vulnerable agrarian exportation model, predominantly in the cacao-producing region (Mesquita, 1998). CEPLAC is upheld as the largest and most successful intervention of the federal government in favour of cacao. Founded in 1957, during the crisis spawned by the plunge of international prices and the drop of cacao production in Bahia, the company, unparalleled, managed to raise production and productivity of cacao plantations to unrivalled levels among all producer countries. Thus, in the 80s, Brazil regained the position of second greatest cacao-producer worldwide. However, the institution was not able to rise to the challenges, missing the opportunities that this cacao region offered (Menezes & Carmo-Neto, 1993). Its proposals did not envisage more than stimulation and support of infrastructure (roads, schools, communication, health, sanitary installations, harbour), and worse, without order of priorities. Prosperity, resulting from the modernization of cacao plantations or by the technological effort induced by CEPLAC, benefited, to a higher degree, the cacao producers, export companies, large multinationals and the banks. As the local consumption market was small, due to the remuneration of the workers, most of the cacao-generated profit was invested into other regions or even abroad. The absence of surplus re-application into the region generated an economic imbalance, interrupting an adequate development of the second and third sectors. A market of consumer goods and services that would attract investment chances and stimulate the regional market failed to be consolidated (Mesquita, 1998). CEPLAC was on the wrong track in making the region a “domain reserve”; no other state and national development organizations were tolerated, nor the entrance of funds of international organizations, such as the World Bank. The cacao region was one of the few Brazilian regions that had never received any investments in dollar until the 90s. CEPLAC did not integrate the cacao region into the national and international capital flow (Menezes & Carmo-Neto, 1993). In this sense, CEPLAC can be considered a chief culprit for the isolation of the cacao economy in Bahia. The cacao region of southern Bahia comprises an area of 100 km2. Of this entire area, over 2 million hectares lie fallow, in spite of their suitability for diversification with oil palms, tropical fruit, food, and animal husbandry, and to generate income and employment (Silva & Leite, 1988). The region consists of a coastal strip, broader than that of other Brazilian states, and is dotted with mangroves, where crustacean and molluscs could be raised, and bays for the production of shrimp and fishing. The soil is rich and the climate also suited for the establishment of forests. With such favourable characteristics, southern Bahia could undertake diversification, following the example of Londrina, in the State of Paraná, which broke away from coffee to be transformed into one of the richest regions of the country; or as Ribeirão Preto, State of São Paulo, known as “The Brazilian California”. In its 89 townships (similar areas to the South of Bahia, but land-locked), Ribeirão Preto triggered a development based on industries of orange, milk, sugarcane, soybean and maize, providing a per capita income of U$ 5.47 thousand/year. The cacao region, with its enormous potential, attains a per capita income of under 0.9 thousand/year. Its income is unevenly distributed, unjust, and doomed, in the long run, to stagnation, since there is no perspective of increased social improvement, not to mention the educational level of the region, which lies far under that of Ribeirão Preto. |
| How to overcome the crisis and final considerations. | Return To Table of Contents |
|---|
| How to overcome the crisis: Proposals for the solution of the Brazilian cacao crisis have, historically, all displayed the same shallowness. Is there anything new now? No - nothing. The same reasons, the same consequences and the suggestions how to supersede the crisis are still the same. Interventions were efficient in the short, but not in the mid and long-term, as they did not deal with the crisis causes, however did impede, for some time, the manifestation of the phenomena which characterize the causes (Nascimento et al., 1994). It is useless to insist on the theory that the release of state funds would solve the problems of the cacao crop. Easy solutions, which promote an illusionary popularity, like cheap long-term loans and State release from debts, are no longer available and can’t be considered. Cacao farmers in Bahia have got along with the crisis by prolongations of their debts by the agricultural pawnbroker, without adjustment for inflation. By this mechanism, the debt was cleared at the end of the year, and the next production period was started under liquidation of all debts. This means that the profit was individualized, while the loss was socialized. With this background, the cacao farmer, ill-advised by CEPLAC, has neglected the function of forming more modern and efficient ways to further the plantation for far too long. For example; the renewal of the cacao stands, crop diversification, agro-industrialization, and the use of innovative technologies and management methods. Moreover, the development of protection mechanisms and price stabilization in cooperatives and future markets were inhibited. Summing up, the cacao culture needs a high standard of efficiency and competitiveness, at a management level that strives for profit. Every cacao farmer will have to work on the agro-industrialization of the product, diversify production and transform the farm into an enterprise. It is imperative that he leave the “cacao plot” phase behind, and enter the stage “cacao enterprise”, since cacao production no longer tolerates amateur management. It is fundamental to be competitive, and whoever is not, will not survive in the sector. There is no other way to attain these goals, than by organization. A cooperative network needs to be developed to empower producers the right to receive the income they generate, which is almost constantly transferred to other exportation sectors of the bean and its by-products. Only 5% of the proceeds of the cacao agribusiness return to the producer countries; the other 95% are in the hand of multinationals’ chocolate sales. With a market share of only 5%, political and economic chances to obtain important structural changes in favour of the producers are tiny, unless they act in groups. The absence of organization means that the workers and regional economy will be indefinitely coupled to each other in dependency and instability (Nascimento et al., 1994). Common interests of cacao farmers must be fostered, to build up cohesion, mutual trust, and solidarity to a level that ensures the success of the cooperative network. The second stage of modernization can be initiated by CEPLAC. In the 80s, the company lost its financial autonomy and credibility, with negative effects on the efficiency of its actions and functionality. Until the early 80s, the institution managed to handle the crises in Brazil’s cacao production, since it generated its own resources, resulting from exchange confiscation or incidental export taxes on the export value of the beans and cacao by-products. When the institution’s budget became part of the fiscal budget of the nation, CEPLAC had to compete for the release of financial sources to realize its programs together with the other institutions linked to the Ministry of Agriculture. That was the end of its financial and political independence. The institution came under political manipulation, so that its focus gradually shifted away from the technical aspects of the crop (Vieira, 1999). However, CEPLAC should be remodelled and resized to face the new reality. A modernization of cacao production must be driven by changes at the agrarian regional basis, in its economic, social, and power relationships. The modernization the CEPLAC induced was conservative, and did not promote social changes. It is imperative that a new movement of modernization in the cacao economy, which could be directed by this organization, be a lot more advanced in its conception and objectives. This modernization will require a form of capitalism in the field which guarantees profitability for the cacao enterprise, associated with, above all, well being in the rural environment. After the Second World War, there was a tendency to form agro-industrial complexes. With this scenario, the cacao region of Bahia should have adopted another model to structure its activities. The civil society and the State are mainly to blame for this failing to happen. The civil society, in this case represented by the cacao farmers, organized or not in advisory boards and associations, was responsible for not paying attention to the management problems of absenteeism, excessive indebtedness, and lack of reinvestment of the returns into the proper sector, thus increasing the patrimony of cacao farmers, but not furthering the cacao production. The State, principally represented by CEPLAC, was responsible for not clearly recognizing the problems on the international market, and stimulating the over-expansion of the cultivated area by means of the PROCACAU project (Mesquita, 1998). Instead of promoting new plantations with subsidies, the organization should have channelled funds to renew old and decadent plantations. In this way the traditional cultivation areas would have recovered competitiveness and, simultaneously, elevated the national output, which had been one of the main goals set by the PROCACAU. As a demand of the buying market for grinding, cacao in the form of dry beans is classified in three types: type I or superior, type II or good, and residues that are not recommended for chocolate production. With the increasing prices from the seventies onwards, the intermediate trade market expanded in a disorganized manner, speculation took over, and the classification criteria were no longer respected. No matter what kind of cacao was produced, a good, satisfactory, or inferior quality; the basic price was always the same, with only slight variations among buyers. This fact made it impossible for the technicians to demand superior quality from producers, since such a production quality would have implied higher costs but no return guarantee. But the poor quality and irregular standard meant the decline of Brazilian cacao on the international grinding market. African or Asian cacao may be identical to a Brazilian one in quality, but will obtain a higher quotation, mainly in the European Community. This is a consequence of the altogether inferior quality that has been thrown on the market, caused by inadequate fermentation, smell of smoke, wrong drying, and pests. As if the lower price paid by the European Union, making the product less competitive on the international market, were not sufficient, production costs for Brazilian cacao are also higher than for the African or Asian product. To obtain by-products like liquor, butter, pastry, or cocoa powder, or the proper chocolate of good quality, the processing of top quality beans is basically the most determining factor. If the remuneration of the basic product is inadequate, this standard will most likely not be met. If, however, the remuneration were quality dependent, the results would be beneficial for the whole production chain: the producer would receive a worthwhile value for his work, while a high quality product would be processed without cost increase for the industry. Regarding infrastructure and tenure, there are marked differences between the Brazilian and the other cacao-producing regions, mainly African and Asian. In the African countries most farmers are smallholders. Besides cacao cultivation, they raise other crops which warrant their subsistence. The work force is normally the family itself, and even the price of a worker from outside the family is very cheap, compared to the Brazilian worker, which diminishes the production costs remarkably. Among the Asian countries, Malaysia developed furthest in using tropical soils rationally and continuously (Alvim, 1989). This ex-colony of England, with less than 4% of the Brazilian surface area, has climatic and edaphic conditions comparable to the Brazilian Amazon region. This country uses a system called FELDA, successfully applied with oil palm and rubber and also tested with cacao, sugarcane, and coffee, which establishes settlers on so-called FELDAS. These FELDAS belong to the Malaysian government, which chooses families and also organizes the rational use of the human, financial, and soil resources, provides credit facilities, plans and helps with services, enables the creation of the physical structure by modern processes, provides superior genotypes for planting, and directs the development of the communities. The Malaysian government gives settlers the possibility to become owners of the area, in a hire purchase plan of up to fifteen years (Silva, 1985). This description illustrates the enormous difference to the funding structures of cacao in Bahia, where large estates prevail, with mostly degenerated over-aged plantations, thin stand densities, and inadequate crop management, in spite of the highly fertile soils in an extraordinary climate. In the Amazon cacao region, however, the properties are also small, as fruits of settlements organized by the government. However, settlers do not receive any support at all to establish adequate production and commercialization conditions, or financial and social stability for their families. Other Asian countries, such as Indonesia and Papua New Guinea also implanted vast areas of cacao plantations. Obviously, the Brazilian cacao production strategy must urgently be reviewed, once having learnt the lesson that competence is a must. At least 1,500 kg/ha of dry cacao must be produced, for which the technology is available. The current stand density on traditional Brazilian plantations is 600 trees/ha, but should reach 1,100 to 1,200 trees in technically enhanced stands. Among the three production factors, land is still the most expensive and its use must be maximized. In the context of the cacao ecosystem, the reserve of the Atlantic forest biosphere is one of the greatest in the world but is threatened by extinction. It is one of the five natural areas most threatened by extinction on the planet and comprises around 30 million hectares in the surroundings of the largest Brazilian cities. In spite of 92.5% of its original area having been destroyed, the Atlantic forest is still exuberant, with its compact forest of frondescent trees covered by lianas, orchids, and bromeliads. It is of great importance, be it because of its extensive area, or due to the characteristics of its individuals. It is the greatest reserve of plant biodiversity, with the highest index of species per hectare, and an elevated percentage of endemic plants. In Bahia, the fragments of the Atlantic forest are not only restricted to areas of accentuated slopes. On the contrary, they cover most of the coast line in southern Bahia. This region holds the forest’s most substantial reminiscences, which provides shade for most of the 700 thousand ha of cacao and preserves soil and water. Here, cacao cultivation was mostly implanted under thinned forest, in the “cabruca” system, and its conservation must be persued at any cost (Silva & Mendonça, 1998). Returning to the socio-economic issue, the actual low prices seem to be associated to the existing stocks and on new plantations coming into production in Africa and Asia. Stocks are relatively high, approximately 800,000 tonnes (AGRIANUAL, 2000) and provide a certain tranquillity to industry. In the late 70s, when international prices soared, warehouses had stockpiles equivalent to 2 to 3 months of consumption. Cacao is one of the agricultural commodities which suffers most under price variation, calling for urgent measures to create instruments that reduce these intense ups and downs on a short, mid, and long-term basis. Prices are established at the stock market, and alterations in the physical offer only will therefore hardly stabilize them. Main agricultural commodities have reacted to the increase of the global consumption in the last years, to the stabilization and/or reduction of the global production, and to the consequential dwindling of global stocks, as in the case of coffee. However, this rule does not appear valid for cacao, since the prices did not reflect the higher consumption, stabilized production, and reduced stocks adequately. At present, stocks are relatively high, while production remains stable, resulting in low prices. Only the incidence of a drought in some of the larger producer countries or an agreed market intervention by the most weighty producer countries could destabalize the market. With an offer below consumption, price increases were to be expected, but the market has not reacted since the global chocolate industry cartel controls the price prediction. Industry representatives get hold of very precise information on the crop prospects in the principal producer countries, sometimes clearer than the producer country itself. Industry, armed with the statistical data of the forthcoming crop, make plans on how and where to buy, and how much will be paid for the cacao of the following year. This mechanism has a freezing effect on prices. The proposal considered by the European Union to use 5% of hydrogenated plant fat in chocolate, thus substituting cocoa butter, has kept the market on the alert. Therefore: either the principal cacao-producing countries get organized, as did the oil producers in the Organization for Petroleum Exporter Countries, or they will remain at the mercy of the small number of cartelized chocolate companies which have the upper hand on the world market and are extremely well organized. The defence mechanism of Brazilian cacao farmers to outlive the crisis with its low product prices has been the extraction of wood from their farms. However, with the appearance of witches’ broom, this practice has become unsustainable, since it compromises the survival of the cacao crop. The only options to enable the national cacao production to survive the installed crisis are either the control of or co-existence with witches’ broom. There are the possibilities of genetic, chemical, cultural, and biological control or, in the ideal case, a combination of the methods or the cultivation proposals suggested in Chapter 13. Further possibilities include renewal of decadent cacao plantations, agro-industrialization, the alteration of the profile of the Brazilian cacao farmer, who had always been absent from everyday life on his property, or even an administrative re-structuring of the cacao estates, in form of a rural partnership, giving birth to profound changes in the worker-owner relationship. A worker who is integrated in the profit distribution has commitment to the performance of the property, and it is well-known that this is a success factor for any enterprise. To date, with rare exceptions, only the bean is used for industry, which represents barely 10% of the fruit’s potential. But fruit pulp can be used for juice, ice cream, liquor, jam, vinegar, oil, amongst other end-uses, while the fruit husks can be applied as fertilizer on the plantation or be used as animal food - uses that must not be overlooked. (Freire et al., 1990). In this context, Nascimento et al., 1994, suggest the following measures to ensure the sustainability of cacao production: i) commission institutions such as CEPLAC, EMBRAPA, UESC, and Cooperatives to initiate integrated research with the private sector; ii) train new human resources in all cultivation segments; iii) create information systems that actually reach the cacao farmer; iv) enhance the extension services of the public sector - linking them to the cooperatives, introducing a co-responsibility of the technician for the success or failure of the technique recommended; v) stimulate and develop community cooperatives, integrating them horizontally and vertically. This last measure aims at raising the income of the producer member and at enabling savings through technology and costs at buying input material and selling the product. It also envisages obtaining systematic market information, diffusion of technique knowledge among producers, through economical and not only agronomical efficiency. Mesquita, 1998, claims that even if the agro-industries were not implemented, the cooperative networks would benefit the stage of primary production as well as trade and consumption. The sale of dry cacao in great quantities, classified by a trustworthy institution, and the acquisition of inputs and equipment in a larger scale and with better contacts to the supplier net would provide a new base for the community at bargaining over purchases and sales. One can conclude that the cacao sector must seek greater efficiency in the link of the primary production, intensify the line of agribusiness, renew its leadership, and establish less client-orientated State relationships. The present-day requirement is to implant an innovative management in all senses, and for this to become reality there is only one path to follow, that of cooperation. Final observations: It is imperative that the cacao-producing countries assume a better organization in a global context, as did the oil-producing countries. The way they are, they will continue to be at the mercy of the chocolate company cartel which controls the world market for cacao, and which is an extremely flexible and organized sector. In Brazil, in particular the Southern Bahia region, the history of this crop in the past two and a half centuries has seen moments of glory and of crisis. In the first, forests were felled, cities founded, roads constructed - summing up: the ‘tree of the golden fruit’ was responsible for a regional boom. At a second stage, when product prices dropped, the impact was sweeping, and a new crisis was installed. The reasons were simply that the region had never diversified its production, and the owners of cacao farms, with rare exceptions, were absent from the administration of their estates. They did not worry about a correct management, and always found a way to restructure their debts by cheap or free federal agricultural credits. The ICB (Cacao Institute of Bahia, now extinct) and CEPLAC were founded in an effort to find ways out of the crisis. In spite of all pains taken by these two institutions, especially CEPLAC, they imposed a disastrous dominance in the region, assumed the monopoly for research, dictated rules, did not promote social changes or even a decentralization of the regional income, and maintained the existing (failed) agrarian export model. The cacao region in southern Bahia is one continuous non-stop plantation with an extension of over 300 km. This situation threatens the crop constantly with pest or pathogen attacks of destructive consequences, spelling regional disaster - which was bound to happen one day. The appearance of witches’ broom (Crinipellis perniciosa) in 1989 in a phase of low product prices was the onset of a socio-economic and ecologically regional tragedy. Today, the disease has affected about 95% of the region, and shows that a co-existence with it entails radical changes in the cacao farming practices used so far. Cacao farmers have to look upon the estate as an enterprise, which must be managed competently and rationally, otherwise, the activity is doomed to failure. There is a need to diversify crops in the region. Cacao monoculture cannot and must not persist in the presence of witches’ broom. To achieve changes in the regional agricultural paradigms, the region itself has model examples to be followed, such as in Ituberá, Bahia. This municipality has the greatest crop diversity in the country. Besides this diversity being more economically advantageous, it makes the incidence of witches’ broom less drastic than elsewhere, where cacao is cultivated as a monoculture. Well-structured and united producer associations were formed, which should serve as a model for the rest of the cacao region. When coffee rust (Hemileia vastatrix) established itself in Brazil in the early 70s, many thought that it would be the end of coffee crop. However, the disease turned out to be the crop’s salvation, becoming more technical and competitive. Certainly, witches’ broom will be responsible for the modernization of cacao farming in Southern Bahia as well. The cacao production in Bahia, a sorry sight, seemed to be doomed to extinction in the region. This proved wrong though – cacao is still alive. However, to overcome the crisis, the existing vision and practices cannot continue. Cultivation areas must be reduced, become more efficient in the link of primary production, intensify the agribusiness line, renew leaderships, and have less client type relationships with the State. Summing up, cacao production needs management innovation in every sense, without missing out the dimension of sustainability, be it the co-existence with the Atlantic Forest or be it the reduction of social segregation. There is only one way to bring these objectives into effect: that of cooperation. One hopes that witches’ broom will soon be no more than yet another chapter in the history of cacao, of which the people living in the region have merely a dim recollection. |
| Chapter 2. Diversity in the genus Theobroma. C.R.S. Silva, A.V.O. Figueira & C.A.S. Souza. | Return To Table of Contents |
|---|
| Content: Introduction; Taxonomy of Brazilian species; Species identification; Origin and distribution, Economic importance; Evolutionary hypothesis; Biosystematics; Interspecific hybridization; Identification of hybrids and Hybrid description; Storage proteins and phylogeny; Molecular markers and diversity; Seed fatty acids and proteins and Final considerations. Summary: The diversity of the genus Theobroma, the most important genus of the family Sterculiaceae, is described. The genus contains 22 species, classified into six sections by the classic method of comparative morphology, mainly based on flower and fruit structure and relevant vegetative characters. The classification is also based on the evolutionary hypothesis of the genus. Studies with molecular and biochemical markers confirm the morphological division of the genus into sections. The section Theobroma is the most evolved, while {{Glossopetalum}e} is the most antique. All sections except for {{Andropetalum}e} are represented in Brazil, restricted to the Amazon region. A theoretical key adapted for the identification of Brazilian species of the genera Theobroma and Herrania is presented. All Theobroma species produce fruits of commercial value, but only T. grandiflorum, T. bicolor, T. angustifolium and T. cacao are cultivated. Theobroma cacao is the only species planted on a large scale. Interspecific hybridization in Theobroma is possible. A theoretical key for the identification of hybrids is presented, with photos and a description of interspecific hybrids. The total seed proteins of the Theobroma species show polymorphism - a qualification of these markers for phylogenetic studies. The fatty acid composition of Theobroma and Herrania seeds suggests the existence of differential metabolic biosynthetic pathways of these fatty acids in the various species. Therefore, there is considerable scope for research into the application of fats from these Theobroma species as potential substitutes for cocoa butter extracted from T. cacao. |
| Introduction; Taxonomy of Brazilian species, Species identification, Origin and distribution, Economic importance e Evolutionary hypothesis. | Return To Table of Contents |
|---|
| Introduction: The family Sterculiaceae contains around 50 genera, comprising 750 species and consists mainly of small tropical trees and bushes (Purseglove, 1968). Theobroma is the most important genus of the family, because of the economic importance of cacao (T. cacao). Theobroma comprises 22 species, divided into 6 sections: {{Andropetalum}e} (T. mammosum); {{Glossopetalum}e} (T. angustifolium, T. canumanense, T. chocoense, T. cirmolinae, T. grandiflorum, T. hylaeum, T. nemorale., T. obovatum, T. simiarum, T. sinuosum, T. stipulatum and T. subincanum); Oreanthes (T. bernoullii, T. glaucum, T. speciosum, T. sylvestre and T. velutinum); {{Rhytidocarpus}e} (T. bicolor); {{Telmatocarpus}e} (T. gileri and T. microcarpum) and Theobroma (T. cacao) (Cuatrecasas, 1964). Representatives of all sections, except for Andropetalum, occur in Brazil. The Brazilian species (T. grandiflorum, T. obovatum, T. subincanun, T. speciosum, T. sylvestre, T. microcarpum, T. bicolor, T. cacao, T. glaucum and T. canumanense)* are limited to the Amazon region (Cuatrecasas, 1964). *In Brazil, T. glaucum occurs sporadically in the western Amazon, near the borders to Peru and Colombia, and the existence of the species T. canumanense is based on the description of two exsiccates collected near the Canumão river, near Borba, State of Amazonas, Brazil. Taxonomy of Brazilian species: Species identification: A theoretical key for the identification of Brazilian species of the genera Theobroma and Herrania was adapted from Ducke, 1953, with an update of species names, inclusion of observations on fruit abscission and flower position in T. speciosum, and the genus change proposed by Cuatrecasas (1964)*. *Original species names used in Ducke's classification (1953) are set in square brackets [ ] and the observations in the original key are in italics. Figure 2.1 shows pictures of fruits of species of the genera Theobroma and [Herrania]. Synopsis of the Brazilian Theobroma and [Herrania] species: A1. Multi-stemmed tree; simple leaves. Petal ligules are smaller than three times the length of the calyx, which is upright or reflexed during the pre-flowering stage. B1. Tree: a single erect trunk or several inclined trunks, emerging from a unique base; 5-toma branching. Flowers: on the trunk and on old branches. Ligules spatulate, genuflected, reflexed, with a long unguis. Stamens geminate with 4 anthers. Staminodes subulate. Fruits: in spontaneous Amazonian as well as cultivated plants they are ovoid, acuminate, with 10 longitudinal ridges, glabrous, yellow when mature; bitter seeds, violet-coloured at cutting (forma leiocarpum Bern). Fruit dimensions of the spontaneous trees up to 15 cm x 7 cm. There is no fruit abscission at maturity. T. cacao L. B2. Trees: one single upright trunk; 3-toma branching. Ligules not spatulate. C1. Leaves: broad, frequently cordiform at the base. Flowers: on the smaller branches. Ligules orbiculate, sessile or subsessile. Stamens geminate with 4 anthers. Oblong, narrow staminodes. Fruit: subglobular ellipsoid, slightly pentagonal, with thick reticulate green veins, in relief on the slightly tomentose yellow undercoat; dimensions of the examined fruits up to 19 cm x 11 cm. There is abscission of ripe fruits at maturity. T. bicolour H.B.K. C2. Leaves: never cordiform. Fruit plain coloured. D1. Staminodes subulate. Fruit with more or less distinct front or back and with tender tomentum. Leaf underside glaucous or pale rusty, glabrous. E1. Ligules largely elliptic, sessile or subsessile. Long leaf petioles (1.5 to 6 cm), lamina up to 30 cm long and 15 cm wide. Young twigs and petioles only slightly tomentose. Mature fruit ellipsoid or subglobose, slightly pentagonal, with 5 shallow furrows, not reticulate; dimensions of up to 10 cm x 8 cm. F1 Flowers in dense and multi-floral cushions on the trunk, with a strong lemon scent. In rare cases there are flowers on branches young trees; Petals are dark red, about 1 cm long. Stamens trigeminous with 6 anthers. Fruit yellowish when ripe. There is no mature fruit abscission, so bunches of many dry fruits on the trunk are a common sight. T. speciosum Spreng. F2. Flowers mainly on small branches, some on the trunk, in small fascicles, scentless or with a faint odour of vanilla; light pink brownish petals, with a length of about 5 mm. Stamens geminate with 4 anthers. Fruit: glaucous, unchanged colour when mature. There is no fruit abscission, so after harvest, large quantities of dry fruit, mainly on the branches, are a common sight. T. sylvestre [T. spruceanum Bern]. E2. Rudimentary ligules, reduced to a triangular inflexed tip of the calyx. Stamens trigeminous with 6 anthers. Short leaf petiole (up to 1 cm), lamina with 10-18 cm x 5 cm. New twigs and petioles with tender tomentum, although somewhat conspicuous. Very small flowers, isolated, or some with special short twigs brought forth by smaller branches. Fruit nearly round with a diameter up to 6.5 cm, with 10 accentuated lengthwise ridges, and in-between a reticulate of smaller veins, green-yellowish when mature. There is no fruit abscission when mature and they are commonly seen dry on the branches after harvest. It is also common to see a large quantity of small dry fruits, which did not develop. T. microcarpum Mart. D2. Petaloid-lanceolate staminoides. Stamens trigeminous with 6 anthers. Petal-lamina triangular in keel form, with a tender unguis at base, erect in pre-flowering. Fruits are neither ridged nor reticulate. G1. New branches and petiole moderately tomentose, underside of new leaves faintly tomentose. Staminodes are lanceolate with an acute and reflexed apex. Obovoid fruits, yellow-brownish when ripe, sweet scentless pulp. H1. Fruit with granular, fragile husk, 7 x 4 cm. Leaves rarely over 20 cm x 7 cm. Some fruits fall after maturity but it is common to see others, already dry, still attached to the branches of the plant. T. obovatum Bern. H2. Fruit with thick and resistant husk, not granular, but mildly tomentose, 10 x 6 cm. Leaves frequently over 30 cm x 15 cm. There is no fruit abscission when mature, so after harvest large quantities of dry fruits are commonly seen attached to the branches of the plant. T. subincanum Mart. G2. Small branchlets densely tomentose; leaf underside somewhat tomentose. Flowers and fruits are the largest of the genus. Staminodes with an abruptly acuminate apex and very tender. Fruits measuring 24 x 12 cm, ellipsoid, subglobular or subcylindrical, with a thick woody husk covered with chestnut brown tomentum; acid pulp, strong and pleasant scent. There is abscission of fruits upon maturation. T. grandiflorum (Willd. ex Spreng.) Schum. A2. Single stemmed shrubs only a few metres tall, slender trunk crowned by the leaves; only in the case of stress divides into 2 or 3 branches, always erect. Compound palmate - digitate leaves. Flowers on trunk with long petal lamina exceeding many times the length of the 'cógula' - petal hood; rolled up during pre-flowering. Large petaloid, curved staminodes. I1. Free stamens, but juxtaposed, forming pairs; one of the stamens with 2 anthers and the other with 4, or both with 2 anthers. Oblong-ovoid fruit of up to 10 cm x 5 cm, acuminate in the apex, strongly pentagonal, with 10 lengthwise elevated ridges (5 accentuated ones) and with few transversal furrows, green, yellowish when mature, sweet pulp. There is no abscission of mature fruits. Herrania mariae [T. mariae (Mart.) Schum]. I2. Stamens: joined in pairs, each stamen with 2 anthers. Fruit: red, subglobose, 'lamelado' dentate with acid pulp. Herrania camargoana (Schultes) [T. camargoanum (Schultes) n. comb]. Origin and distribution: The present knowledge on the origin and distribution of the Theobroma species is based on herbaria surveys and reports of botanical expeditions (Cuatrecasas, 1964 and Reksodihardjo, 1964). Theobroma is considered to be of exclusive neo-tropical origin, with a natural distribution in rainforests, extending from the Amazon basin to southern Mexico, between 18o N to 15o S (Cuatrecasas, 1964). However, since this genus contains species of pre-Columbian use, its distribution may have been influenced by humans and most likely does not represent the natural dispersal. The southern and eastern distribution limits stretch out from the foot of the Bolivian Andes, across the State of Rondônia and northern Mato Grosso, to the north of the State of Tocantins in Brazil, turning away to the northeast to the north western part of the Maranhão State. The distribution of Theobroma spreads over the Cordilleras and to the region of Chocó north of the Equator to the Pacific coast of Colombia. In Peru and Bolivia, the occurrence of Theobroma is restricted to the Amazon side (Ducke, 1953). With exception of the cultivated species (T. cacao, T. grandiflorum and T. bicolour), the species found on the eastern region of the Andes are not found on the western side (Baker et al., 1954 and Cuatrecasas, 1964). The geographic distribution of genus Herrania is similar to that of Theobroma, with only one species found in Central America and 16 others only in South America (Schultes, 1958). Herrania species are morphologically similar to Theobroma, distinguished by a few morphological traits, such as the single stemmed slender trunk without branches; the compound palmate leaves; and flowers with long petal laminae, exceeding the length of the petal hood many times (Baker et al., 1954 and Ducke, 1953). Theobroma cacao, the most studied species of the genus Theobroma, extends from the rainforest of the Amazon basin to southern Mexico. Cuatrecasas (1964) proposed that plants found in Central America and Mexico were classified as T. cacao subsp. Cacao and its cultivated forms would represent the Criollo group. Plants from South America were classified as T. cacao subsp. sphaerocarpum and its cultivars would represent the Forasteiro group (Cuatrecasas, 1964). Theobroma subincanum is the species with the most frequent and largest distribution, besides cacao. It is considered one of the most ancestral species of the genus (Cuatrecasas, 1964). Its natural distribution area covers the whole Amazon basin (Figure 2.2), from the State of Pará to the western tributaries of the Amazon river, the upper Orinocco, Venezuela and French Guyana. The species occurs under forest shade, in non-flooded areas, in rich soils with humus, and on small sandy elevations along small rivers and streams. The recognized natural distribution of T. grandiflorum (cupuassu) extends from the southern half of the State of Pará to the adjacent Amazon area to the state of Maranhão (Figure 2.2). Wild cupuassu is only found on higher ground of the middle region of the rivers Tapajós (waterfalls of Mangabal and Itapacurá), Tocantins, Guamá river (between Ourém and Bragança), Xingu river (between Vitória and Altamira) and the river Anapú. The tree has a low population density in its natural habitat. Theobroma obovatum occurs throughout all of the western region of the Amazon basin, on elevated and wet ground, in fertile forest soils (Figure 2.2). In Brazil, it is frequently found in the western half of the Amazon region from Tefe, Amazonas, and from the river Jaú, tributary of the river Negro. The Anglo-Columbian expedition of 1953 found T. obovatum, although not abundantly, along the rivers Caguán, the upper and lower Caquetá, and the Putumayo river. In Peru, the species can be found in the lowest part of the rivers Huallaga, Ucayali, Itaya, as well as the river Putumayo. Theobroma sylvestre is frequently found along the Amazon River and the lower part of its tributaries, from Santarém to Tocantins in an easterly direction to the mouth of the river Içá (Figure 2.2). Theobroma speciosum extends across all the Amazon jungle, except for the northwestern part, from Cururupu in the State of Maranhão to Acre, Madre de Dios in eastern Bolivia and the Ucayali river (Loreto) in Peru (Figure 2.2). Theobroma bicolor is frequently found under non-extensive cultivation (few trees) across tropical America (Figure 2.2). The centre of origin of T. bicolor is uncertain; it possibly lies in the eastern region of Peru and Ecuador, or in Central America, where the tree is found in primary forests (Cuatrecasas, 1964). Theobroma microcarpum occurs in the Brazilian southern and western region of the upper Amazon, along the rivers Solimões, Yapurá, Purús, Madeira, and Tapajós, and the Caquetá river, in western Colombia (Figure 2.2). Economic importance: All Amazon Theobroma species produce fruits of commercial value (Ducke, 1953; Le Cointe, 1934; Calzavara et al., 1984; Venturieri & Aguiar, 1988 and Nazaré et al., 1990). Of these, only T. grandiflorum, T. bicolor, T. angustifolium and T. cacao are cultivated, while only T.cacao is widely cultivated (Baker et al., 1954 and Reksodihardjo, 1964). The economic importance of T. cacao is expressed through chocolate and confectionary consumption in the most varied forms all over the world, also through utilization of cocoa butter in the pharmaceutical and cosmetic industries. The fruit also contains a highly appreciated seed pulp, which is conquering national and international markets. The juice extracted from the seed pulp is used for the production of wine, vinegar, liqueurs, and jams of excellent quality (Menezes & Carmo-Neto, 1993). The demand for cupuassu (Theobroma grandiflorum) has increased its economic importance in the Amazon region and in other Brazilian states and is beginning to attract the attention of the world's market of exotic tropical fruits. Cupuassu seed pulp is used in the manufacture of ice cream, juices, and jams, besides a multitude of homemade sweets and desserts. The use of the seeds however is still incipient, but a product similar to cocoa powder can be prepared from them (Venturieri & Aguiar, 1988), and the extracted fat can be used to manufacture skin cream (Berbert, 1981). Cupuassu ice cream and juice are already on the market in the United States, and skin creams can be found in the United Kingdom. Cupuassu is considered one of the most profitable crops of the Amazon region, remarkably suitable for agroforestry systems and intercropping (Venturieri, 1993a and 1993b). It is estimated that at least 16,000 hectares (without considering the wild harvested area) in the Amazon region are under cupuassu cultivation. The States Pará and Rondônia alone cultivate over 6,000 ha, followed by the Amazon region with 3,000 ha (Ribeiro, 1997 and Rodrigues & Santana, 1997). Theobroma bicolor is rarely grown in the State of Pará, but is frequently found cultivated along the river Negro and mainly in the villages of the western part of Solimões for the production of pulp for juice substituting cupuassu, which is not found in the region. In Iquitos, Peru, and surroundings, T. bicolor is one of the most common fruit trees in urban gardens, and on smallholdings on the outskirts of the city (Ducke, 1953). Evolutionary hypothesis: The phylogenetic relationships of Theobroma were defined by Cuatrecasas (1964) by the classical method of comparative morphology, based mainly on flower and fruit traits and relevant vegetative characters. The sections Rhytidocarpus, Oreanthes and Theobroma are considered the most evolved, presenting epigeal germination; sub-terminal growth type (orthotropic shoots arising below the jorquette - see diagram) and young glabrous leaves. The sections Telmatocarpus, Glossopetalum and Andropetalum have hypogeal germination, pseudo-terminal growth (orthotropic shoots appear near the centre of the jorquette) and young pubescent leaves. Section Theobroma is the most evolved and presents a jorquette with 5 branches, while the other sections display 3 branches. The evolution of the Theobroma and Telmatocarpus sections was analogous, as inferred from their glabrous or nearly glabrous leaves, more adapted to tropical forest ecology. The section Telmatocarpus is considered more advanced in another direction due to the reduction or absence of ligules and the discontinuity of the vascular woody system in the pericarp, which is partially woody and more vulnerable. The species of the Glossopetalum section are considered as morphologically the most ancestral of the genus. There are no reports about Theobroma fossils, and the geographic distribution of the sections does not provide data that could help elucidate the evolution of the genus, since all sections have representatives in regions on either side of the Andes, with the exception of Andropetalum (T. mammosum) which occurs only on the western side of the Andes. The appearance of the Andes at the onset of the Tertiary period might have separated the previously dispersed Theobroma populations, thus favouring speciation by isolation. Vicariant species separated by the formation of the Andes include T. subincanum (western side of the Andes) and T. hylaeum (eastern side); T. microcarpum (eastern side) and T. gileri (western side) (Cuatrecasas, 1964). The subdivision of T. cacao in T. cacao subsp. cacao (Central America and Mexico) and T. cacao subsp. sphaerocarpum (South America) may have been derived from a sub-speciation of the original wild population after the emergence of the Andes (Cuatrecasas, 1964). The independent domestication of these subspecies brought forth the two main morpho-geographic groups, Criollos and Forasteiros. However, the largest genetic diversity of plants in the upper Amazon region, estimated by molecular, isoenzyme, agronomical, and morphological markers, have supported the hypothesis that the upper Amazon is the probable centre of diversity of T. cacao (Cheesman, 1944 and Figueira et al., 1994). The possible predecessor of the Criollo group could have been transported by humans to Mesoamerica, where it would have been domesticated by the Mayas, about 2000 years ago (Cheesman, 1944 and Whitkus et al., 1998; see also Chapter 3 on origin and spread of T. cacao). The complexity of the mountains in northern Colombia and Central America was an isolating factor which favoured the speciation of Theobroma in this part of the hemisphere, where areas of regional or local endemic distribution of Theobroma are found (Cuatrecasas, 1964). Compared to Theobroma, Herrania is a more evolved genus in relation to flower characteristics, but not in view of the unbranched single stemmed growth, with compound palmate leaves and hypogeal germination (Cuatrecasas, 1964). Hybridization experiments between the species of Theobroma and Herrania detected genetic incompatibility at varied levels among the species studied. However, in crosses with T. cacao, Herrania mariae produced a high percentage of fruits in relation to the number of pollinations (Addison & Tavares, 1951), and the pollen of Herrania balaoense proved to be compatible with flowers of T. cacao of the Amazon and Trinitarian types (Posnette, 1945). These findings demonstrated the high affinity between the two genera, and refuelled the discussion on the very nature of the genus Herrania; even though its existence had been assumed by Schultes (1958) and corroborated by Cuatrecasas (1964). |
| Biosystematics; Interspecific hybridization; Identification of hybrids and Hybrids description. | Return To Table of Contents |
|---|
| Biosystematics: All species of the genus Theobroma, with exception of T. cacao, have white cotyledons. The number of chromosomes in all species is 2n = 20 (Muñoz Ortega, 1948). The intra- and interspecific genetic diversity of Theobroma species other than T. cacao have never been evaluated, and there are only few collections with a limited number of species in the world, which represent a restricted sample of the populations. A partial assessment of purine alkaloids in 11 species of Theobroma and 9 of Herrania indicated that only T. cacao contains detectable levels of caffeine and theobromine in the seeds. However, tetramethyluric acid (theacrine) was found in all Theobroma and Herrania species, with the exception of T. obovatum, which has no alkaloid at all (Hammerstone et al., 1994). The seed fat composition of Theobroma and Herrania was analyzed for the profile of fatty acids and for levels of sterols, tocopherols and tocotrienols, used as a chemotaxonomical criterion to separate both genera, and to differentiate the subgroups in conformity with the existing section classification (Carpenter et al., 1994) and according to the evolutionary hypothesis of the sections (Cuatrecasas, 1964). The subdivision of Theobroma in sections based on morphology was confirmed by using the ribosomal (rDNA) gene (Figueira, 1992 and Figueira et al., 1994). Although the data corroborated the taxonomy classification and the evolutionary hypothesis of the genus, there are restrictions regarding the data. This had been gathered from collections with a limited number of species, representing a restricted sample of populations, and without evaluating the intra and interspecific genetic diversity. Recently, the phylogenetic relationships of Theobroma and Herrania have been investigated using the sequence of the vicilin gene, which encodes a seed storage protein (Whitlock & Baum, 1999). Herrania and Theobroma were monophyletic. The monophyletic nature of all sections was supported, with the exception of Andropetalum, which was grouped together with the section Glossopetalum. Interspecific hybridization: Natural hybrids within Theobroma species are extremely rare. According to Reksodihardjo, 1964, there is only one report that describes the discovery of a tree with intermediate characteristics between T. obovatum and T. subincanum on the banks of the Caquetá river in the forest of Remolino (southwest of Colombia), thus suggesting its possible hybrid origin (Baker et al., 1954). Reksodihardjo (1964) also mentioned the existence of other T. obovatum and T. subincanum hybrid exsiccates deposited in the herbaria of ICTA, Trinidad, and the Instituto de Ciencias Naturales of Bogotá, Colombia. Two others are conserved in Brazil, in the herbarium of the former IAN (now 'EMBRAPA Amazonia Oriental'), besides the two other exsiccates of a probable natural T. glaucum and T. sylvestre hybrid, also deposited in the herbarium of 'EMBRAPA Amazônia Oriental'. The artificial hybridization of cacao with wild species was proposed as a strategy to find new disease resistance sources (Addison & Tavares, 1951 and Martinson, 1972). Besides, there are Theobroma species that present natural abscission of mature fruits, such cupuassu, while others do not drop fruits spontaneously, such as cacao, as well as plants with a reduced canopy size like T. microcarpum and nearly no lateral branching, like T. speciosum. These are promising attributes to be used in breeding programs of cacao to obtain plants with spontaneous fruit abscission and reduced canopy for elevated yield in high density plantings. In the Amazon region, there are two Theobroma collections of extreme importance. They contain interspecific hybrids obtained artificially or naturally, involving species that occur naturally in the Amazon. These collections are the "Basil Bartley" and "Addison O'Neill", which belong, respectively to CEPLAC (Marituba; 1' 12" S, 49' 30" W) and 'EMBRAPA Amazônia Oriental' (Belém; 1' 20" S, 48' 30" W), both in the State of Pará. Hybridization experiments between the Brazilian species of Theobroma were carried out between 1945 and 1951 by Addison and Tavares at IAN and represent the most significant work on interspecific hybridizations of Theobroma and Brazilian Herrania species (Addison & Tavares, 1951 and 1952). These hybridizations included eight Brazilian Theobroma species and demonstrated the relationships between them. The affinity among the species can be summarized as follows (Addison & Tavares, 1951): - Close affinity between T. grandiflorum, T. subincanum, and T. obovatum; - Less evident affinity between T. grandiflorum and T. cacao; - Close affinity between T. sylvestre and T. speciosum (both less related to T. bicolor); The other species presented less pronounced affinity among each other. Hybrids were obtained from T. grandiflorum x T. obovatum; T. subincanum x T. obovatum; T. speciosum x T. sylvestre; and between T. grandiflorum x T. subincanum. Theobroma grandiflorum presented a certain affinity with T. cacao. The plantlets obtained from T. cacao x T. grandiflorum crosses were very fragile and only a few plants reached about 15 cm. Adult plants of this hybrid were obtained later in Ghana (Martinson, 1966). Plantlets of T. grandiflorum x T. obovatum developed, flowered, and gave fertile pollen. These hybrids produced fruits after backcrossing with T. grandiflorum. The abscission of fruits at maturation segregates in T. grandiflorum x T. obovatum. The plantlets of T. grandiflorum x T. subincanum also developed, and were more vigorous than other T. grandiflorum trees, planted nearby. The original hybrids of the study of Addison & Tavares, 1952, on T. grandiflorum, T. obovatum, T. subincanum, T. sylvestre, and T. speciosum are still preserved at 'EMBRAPA Amazônia Oriental' and constitute the "Addison O'Neill" Theobroma collection. However, most of the information on this collection was never systematized or published, and unfortunately, some of the data was even lost, such as the original map and field notes. The collection "Basil Bartley" contains interspecific natural hybrids of T. grandiflorum and T. subincanum. A few decades ago, hybridization between species of the distinct sections was thought to be an extremely difficult, if not impossible. However, this hypothesis was refuted, when hybrids between species of the sections Glossopetalum and Andropetalum (T. angustifolium x T. mammosum; and T. simiarum x T. mammosum) (Cuatrecasas, 1964 and Reksodihardjo, 1964) and in 1966, the hybrid of T. grandiflorum (section Glossopetalum) and T. cacao (section Theobroma) (Martinson, 1966 and CRIG Newsletter, 1990) were obtained. Identification of hybrids: A theoretical key for the identification of interspecific hybrids was proposed by Silva & Venturieri, 1998, based on the identification of Brazilian species by Ducke, 1953, presented below: A1. Multi-branched tree, erect trunk, jorquette - see diagram with 3 branches, simple leaves. Ligules of the petals smaller than three times the length of the calyx, not spatulate. Leaves never cordiform. Single coloured fruit. B1. Staminodes subulate. Fruit with more or less distinct front and back, and a soft tomentum. C1. Ligule is broadly elliptic, sessile or subsessile. Ripe fruit is ellipsoid or subglobular, slightly pentagonal, with 5 weak ribs, not reticulated, 10 x 8 cm. D1. Flowers, in dense and multifloral cushions on the trunk, with a strong lemon smell. Young trees have flowers on the branches. Petals are dark red about 1 cm long. Stamens trigeminous with 6 unknown type: [[anthers]g1202. Fruit yellowish when ripe. There is no fruit [[abcission]g44], and it is common to see many fruits dry on the trunk, after harvest. T. speciosum Spreng. D2. Flowers mainly on small branches, some on the trunk, in small fascicles, scentless or with a weak vanilla scent. Ligule is rose or brownish red, of 2 mm length x 2,2 - 2,5 mm width. Stamens have 4 anthers. Ripe fruit is glaucous. There is no fruit abcission, and it is common to see many fruits dry on the trunk, after harvest, principally on the branches. T. sylvestre [T. spruceanum Bern]. D3. Flowers on branches and trunk, in small fascicles, with lemon scent, yet less intense than in T. speciosum. Ligule is red, 5-6 mm long x 7-8 mm wide. Staminodes are red, 10-12 x 2-3 mm. Calyx 6-7 x 3-4 mm. Stamens, geminate, have 4 anthers. There is no fruit abcission, and it is common to see many fruits dry on the trunk, after harvest. T. sylvestre x T. speciosum. B2. Staminodes petaloid. Stamens trigeminous with 6 anthers. Ligules triangular in keel form, tender unguis at base. Fruits are neither ridged nor reticulated. E1. New branches and petiole moderately tomentose, new leaves faintly tomentose on the underside. Staminodes are lanceolate, with acute and reflexed apex. Fruits are obovoid, brownish-yellow when ripe, with scentless sweet pulp. F1. Fruit with fragile husk, granular, 7 x 4 cm. Leaves rarely over 20 x 7 cm. Some fruits fall after maturity, but it is common to find them dry on the branches in large quantities after harvest. T. obovatum Bern. F2. Fruit with thick and resistant husk, not granular, but mildly tomentose, 10 x 6 cm. Leaves frequently over 30 x 15 cm. Some fruits fall after maturity, but it is common to find them dry on the branches in large quantities after harvest. T. subincanum Mart. F3. Fruit presents intermediate characteristics: obovoid as in T. obovatum, but not granular. Some fruits fall after maturity, but it is common to find them dry on the branches in large quantities after harvest. Presence of dimorphism on leaves: large leaves as in T. subincanum coexisting with small leaves as in T. obovatum. T. subincanum x T. obovatum. E2. Staminodes are lanceolate, abruptly acuminate on the apex. Young branchlets densely tomentose; leaf underside quite tomentose. Flowers and fruits are the largest of the genus. Fruits 24 x 12 cm, ellipsoid, subglobose or sub cylindrical, with thick woody epicarp covered with chestnut brown tomentum, acid pulp, with strong and pleasant scent. There is abscission of ripe fruits. T. grandiflorum (Willd. ex Spreng.) Schum. B3. Staminodes are red, oblong, inflexed, with acute apex and attenuated base, of 11-13 mm length x 3 mm width. Stamens with 3 anthers. Petal-lamina red, 8-10 x 4-6 mm. Caly0078 dimensions 4-5 x 4 mm. Fruit shows intermediate characteristics: ellipsoid shape; intermediate size (similar to T. obovatum), 78 x 44 mm; the texture of epicarp similar to T. obovatum, but without the typical granular surface; its chestnut brown colour resembles T. grandiflorum. General aspect of the tree is similar to T. grandiflorum, but smaller. There is abscission of ripe fruits. T. grandiflorum x T. obovatum. B4. Flower similar to T. grandiflorum, although smaller. Staminodes are red, oblong, inflexed, with acute apex and attenuated base, 10-11 x 3mm. Petal-lamina is red, triangular, in keel form, 8-9 x 4-5 mm. Calyx dimensions 5 x 4-5 mm. Fruit presents intermediate characteristics: format similar to T. grandiflorum, but size similar to T. subincanum. Abscission of ripe fruits when mature. T. grandiflorum x T. subincanum. Description of hybrids: Hybrids of T. grandiflorum and T. subincanum: These hybrids are more vigorous than the larger parental species, T. grandiflorum, suggesting the occurrence of hybrid vigour, though they are highly susceptible to witches' broom disease (Crinipellis perniciosa). The fruits from these hybrids are slightly larger than those from T. subincanum, but the format and colour of its tomentose pericarp is alike to T. grandiflorum fruits. In relation to the organoleptic characteristics, the seed-pulp scent resembles T. grandiflorum, although the taste is less acidic and reminds one of banana flavour. These hybrids have a pronounced characteristic of abscission of mature fruits, apparently a dominant character inherited from T. grandiflorum, because fruits from T. subincanum do not abscise. However, dry fruits hanging from the trunk of these hybrids are a common sight (as observed in T. subincanum), but it is not clear if this is a characteristic of the hybrids or due to the attack of witches' broom disease. The seed shape and size are more similar to T. subincanum, and when sown, the seeds germinate and produce seedlings. The general aspect of the tree is similar to T. grandiflorum, but some of the leaves are remarkably larger than in the parents (length and width). The flowers are very similar to T. grandiflorum (shape and colour), but it is possible to distinguish the flowers from the hybrid tree by the smaller petal size (petal lamina and calyx) and also by the staminodes that do not show an abruptly acuminate apex as in T. grandiflorum. The flowers of the hybrids present stamens with 6 anthers (Silva & Venturieri 1998). Hybrids of T. grandiflorum and T. obovatum: The general aspect of the tree is similar to that of T. grandiflorum. The best way to identify this tree is by flower characteristics, mainly the size and shape of the staminodes or by the intermediate fruit characteristics (Figure 2.3). Under the canopy projection of the hybrid trees, it is possible to detect supposed F2 seedlings, derived from the fruits, which crack open when dropping to the ground. Not all F2 seeds germinate, and those that do emit radicle and shoots, cease to grow at about 10 cm, before expanding the first true leaves. Some of the F2 hybrid seeds display a normal or nearly normal development, but develop leaf necrosis, as described for the F1 hybrids (Addison & Tavares, 1952). One of the hybrid trees presented a flower with 6 petals. This tree was highly productive and witches' broom resistant, and it should be considered a potential hybrid for further evaluation in yield tests (Silva & Venturieri, 1998). Hybrids of T. subincanum and T. obovatum: The flowers exhibit intermediate characteristics to those of the parental species. The hybrid can be identified by comparing the flower, which resembles that of T. obovatum (petal-lamina with an invagination), and the general aspect of the canopy , which is dimorphic: large leaves, as in T. subincanum, together with small leaves, as in T. obovatum; or by the presence of fruits with intermediate characteristics (Figure 2.4). The fruit shape is similar to that of the T. obovatum, but slightly elongated, smooth, and light-brown when mature, as in T. subincanum. Unharvested fruits dry on the tree. The tomentose underside of new leaves persists until leaf senescence (Silva & Venturieri, 1998). Hybrids of T. sylvestre and T. speciosum: The trees have a similar general aspect to that of the parental species, which are in turn very similar to each other. The hybrids present flowers on the trunk and branches (Figure 2.5). These flowers are similar, in format and colour of the ligules (red) to the flowers of T. speciosum, although smaller, and have anthers with 4 tecas as T. sylvestre has (T. speciosum presents anthers with 6 tecas, so the character 4 tecas of T. sylvestre seems to be dominant). The larger size of the hybrid flower, mainly of the petal lamina, when compared to that of T. sylvestre; the flower shape, similar to that of T. speciosum; and the characteristic lemon scent of T. speciosum, though less intense, are the clearest criteria to distinguish the hybrid from parental species. Some hybrid trees present flowers with 5, 6, or rarely 7 staminodes and 5 or 6 petals. For example, one of the hybrid trees presented nearly 50% of the flowers with 6 staminodes, besides flowers with all possible combinations of number of staminodes and petals. All hybrids found produced fruits similar to those of T. sylvestre. These fruits, even when ripe, remain attached to the tree and dry. Some young supposedly hybrid plants were found (T. sylvestre x T. speciosum) within variable distances from the adult hybrids, probably resulting from natural dispersion (Silva & Venturieri, 1998). |
| Reserve proteins and phylogeny; Molecular markers and diversity; Seed fatty acids and proteins; Final considerations. | Return To Table of Contents |
|---|
| Storage proteins and phylogeny: Theobroma cacao seeds contain 15 to 20% dry weight proteins, and nearly 50% fat (Biehl et al., 1982 and Spencer & Hodge, 1992). The albumin fraction represents the highest percentage (52%) of the total seed storage proteins, while globulin covers 43% (Voigt & Biehl, 1993). The total seed proteins of the Theobroma species are being analyzed. Preliminary results revealed a remarkable vicilin gene polymorphism (Figure 2.6), indicating a potential use in phylogenetic studies (Silva et al., 1998). InT. cacao, the genes of the largest albumin and globulin constituent were cloned and sequenced: the gene of the trypsin inhibitor (polypeptide of 21 KDa) and globulins (polypeptides of apparent molecular weight of 31 and 47 KDa) (Tai et al., 1991; McHenry & Fritz, 1992 and Spencer & Hodge, 1991 & 1992). The translated amino acid sequence based on the cDNA of the 21KDa polypeptide presented a 38% similarity to the alpha amylase/subtilisin inhibitor of barley (Tai et al., 1991) and a high similarity to trypsin inhibitors of the Kunitz type, which seems to be a protease inhibitor (Spencer & Hodge, 1991; Tai et al., 1991). Later, using a purified 21 KDa protein from cacao seeds, Dodo et al., 1992 confirmed that this was a trypsin inhibitor. The in vivo function of the trypsin inhibitor is uncertain, and it may act either as a natural protease inhibitor, hindering the digestion of seed storage proteins by insects, or function as storage source of sulphur, due to the high cysteine content, or a combination of both functions (Spencer & Hodge, 1992). Relevant fragments of molecular weight near 21 KDa were observed in other Theobroma and Herrania species (Figure 2.7) (Silva et al., 1998). The sequence of the cDNA of 67 KDa, cloned from immature mRNA from T. cacao seeds, presented considerable similarity with globulins of the vicilin class from dicotyledonous plants, especially with the cotton seed alpha globulin (McHenry & Fritz, 1992; Spencer & Hodge, 1992). Immunoprecipitation trials of in vitro translation products from unknown type: [[mRNA]g] of T. cacao immature seeds against immunoglobulins (IgGs), produced with purifications of the 31 KDa and 47 KDa polypeptides, identified the 67 KDa product as the precursor of the mature polypeptides of 47 KDa and 31 KDa (Spencer & Hodge, 1992). Proteins classified as vicilins include, besides the T. cacao vicilins, the precursors of pea (Pisum sativum) vicilin, bean (Phaseolus vulgaris) phaseolin and soybean (Glycine max) conglycinin. Some other Theobroma species presented bands of molecular weight and intensity near the 31 KDa and 47 KDa of cacao (Figure 2.6) (Silva et al., 1998). The usefulness of vicilin proteins in phylogenetic studies has recently been demonstrated by comparing the nucleotide sequence of the vicilin gene of the 67 KDa among 18 Theobroma and Herrania species. The phylogeny of this gene corroborated the evolutionary hypothesis proposed by Cuatrecasas (1964), established based on morphological characters (Whitlock & Baum, 1999). Molecular markers and diversity: Silva et al., 1998, used RAPD analyses to describe the genetic relationships between wild Brazilian Theobroma species; to estimate the genetic diversity established in germplasm collections of ´EMBRAPA Amazônia Oriental´ and the ERJOH/CEPLAC; and to confirm the origin of interspecific hybrids. Twenty-one RAPD primers and gene-specific amplifications with primers derived from cacao gene sequences were used to analyse species/populations of T. grandiflorum, T. subincanum, T. bicolor, T. sylvestre, T. obovatum, T. microcarpum, T. speciosum and supposed hybrids between T. grandiflorum x T. subincanum, T. grandiflorum x T. obovatum, T. subincanum x T. obovatum, and T. sylvestre x T. speciosum. The RAPD patterns presented intra and interspecific polymorphism, and some similar sized bands between species classified in the same or a correlated section were detected. The identification of the probable hybrids was confirmed by RAPD and morphologic data. The trypsin inhibitor gene was detected in all species and the presence of the protein in seed confirmed by electrophoresis in denaturing polyacrilamide gel (SDS-PAGE), however the gene exhibited restricted polymorphism at the DNA level. The vicilin storage protein presented broad polymorphism detected by SDS-PAGE. The phenogram obtained based on RAPD data (Figure 2.7), using the Dice similarity coefficient and the UPGMA cluster method, revealed: differences between T. microcarpum plants established at the CEPLAC´s germplasm collection; differences between plants of the same species, such as, for example in T. bicolor and T. speciosum, established in different collections (CEPLAC and EMBRAPA); and separation between T. grandiflorum with and without seeds (variety). Both individuals of Theobroma sp., morphologically identified as members of the section Oreanthes, grouped with T. speciosum and T. sylvestre and their interspecific hybrid, however in an isolated form, suggesting it to be an unclassified species. All species classified in the same section grouped together. The established P bootstrap values allowed significant grouping of the sections Telmatocarpus (100%); Rhytidocarpus (99.9%); Theobroma (99.3%); and Oreanthes (99.7%). The cluster composed by the Glossopetalum section presented a less robust P of the bootstrap value (70%). This fact seems to result of the more ancestral characters of this section, mainly in relation to T. subincanum and T. obovatum, since the group involving T. grandiflorum could also be significantly (P = 97.1%) separated (Figure 2.7). Thus, the molecular and biochemical markers confirmed the morphologic classification of the genus into sections, as proposed by Cuatrecasas (1964). Fatty acids and proteins in seeds: The composition of fatty acids in Theobroma and Herrania seeds suggested the existence of differential metabolic biosynthetic pathways in the diverse species. C22:0 was detected in Herrania species, considered a genus derived from a common ancestor with Theobroma. The composition level of this fatty acid showed a clear distinction between Theobroma species of sections considered more evolved (absence of C22:0), and those that belong to sections considered more ancestral (presence of C22:0). This fact corroborates the evolutionary hypothesis of the sections (Silva et al., 2002; the original book reference, Silva et al., unpublished, was substituted here). Price oscillations of cocoa butter on the world market have led to an intensive search for alternative fats. However, the existing substitute or equivalents ("cocoa butter equivalent" or CBE and "cocoa butter substitute" or CBS) are not completely satisfactory at meeting the requirements of manufacturers. Besides, there are legal restrictions for their utilization in several countries. On this background, the use of fats of the Theobroma species as potential substitute for cocoa butter in the chocolate industry or as raw material for specialty fats is a promising area. The percentage of total protein in defatted seeds was determined by the Kjeldahl method. Theobroma and Herrania species and hybrids presented a high percentage of total protein, while Theobroma bicolor presented the highest (27.6%) and T. microcarpum the lowest percentage (14.4%). The Theobroma species protein average was 22.8% and Herrania sp. presented in average 25.6% protein. Owing to their nutritive value, seeds of the species and hybrids of Theobroma could be used as source of protein in human nutrition. Final considerations: The importance of the genetic conservation of the Theobroma species lies in their potential use as gene reservoir for Theobroma cacao, the widely cultivated commercial species, and in the development of new products. For instance, the demand for T. grandiflorum (cupuassu) is attracting the attention of the world's tropical exotic fruit market and has increased its economic significance in the Amazon. The fatty acid profile of seeds of the various species indicates that these can be differentiated according to the expression of the metabolic pathways of fatty acid desaturation and elongation. This differentiation could suggest a potential use in the industry and/or for studies into the biosynthesis of storage lipids and regulation mechanisms of fatty acid compositions (Carpenter et al., 1994). Likewise, the purine alkaloids (caffeine and theobromine) were not detected in seeds of other Theobroma species, except for T. cacao, while tetramethyluric acid (theacrine) was detected in all but T. obovatum (Hammerstone et al., 1994). The recent medical concern about the excessive consumption of caffeine favours the development of products without this alkaloid. The use of seed storage proteins as indicator of diversity and phylogeny can be associated with the importance of these components in the development of the characteristic chocolate flavour (Voigt et al., 1995). Theobroma seeds are recalcitrant, intolerant to low temperatures and to dehydration, which makes the conventional ex situ conservation of genetic resources by seeds unfeasible. Germplasm should be conserved in ex situ genetic repositories and/or in situ, which, when properly managed, are able to preserve the evolutionary potential of the species (Debouck, 1993). However, the exact distribution and origin of the Theobroma species is not well assessed, restricted to reports on specimen collections in herbaria. The risk of extinction of certain populations is not recognized. According to Almeida et al., 1995, despite all the efforts by CEPLAC to prioritize the collection of genetic resources in populations in process of genetic erosion, there is a number of ecological niches of T. cacao that have been destroyed, as a result of the accelerated and, to a certain point, uncontrolled deforesting practiced in the Amazon region. If deforesting continues at the same rate since 1985 of 30.000 Km2/year, probably no rescue program for native species in the Amazon could ever keep abreast with such devastation. With the exception of T. cacao, only a few specimens of Theobroma species are kept in rare collections, or material stored in herbaria, collected by researchers during the last century. Therefore, the evaluation of the genetic diversity of the various species is necessary, as well as the establishment of in situ conservation areas and the organization of new expeditions to collect genetic material to establish the new specimens in germplasm collections. |
| Chapter 3. Origin and distribution of Theobroma cacao L: A new scenario. L.A.S. Dias. | Return To Table of Contents |
|---|
| Content: Introduction; Phylogenetic relationships; Origin and spread of populations; New scenario of origin and spread; Evidences for the new scenario - Anthropological, Historical, Paleogeological, and Biogeographical evidences, and Genetic population studies; Final considerations and prospects. Summary: Cacao is an important crop of Neotropical origin. Populations from Mesoamerica and northern South America make up the racial group called Criollo and those from the Amazon basin form the Forasteiro group. Three hypotheses, mutually excluding and conflicting, put forward explanations of these populations’ origin and spread: i) South-North spread; supposes that the cacao, coming from a centre of origin in the Upper Amazon basin, had been taken over the Andes towards Mesoamerica by humans. These would have transported only Criollo cacao seeds; ii) North-South spread; backs a Mesoamerican origin and spread in the opposite sense; the centre of origin would have been Mesoamerica and the Mesoamerican cacao trees would have also been transported to South America by Amerindians, and; iii) continuous spread; from the Amazon region to the South of Mexico. Anthropological, historical, paleological, and biogeographical evidence and genetic population studies have reinforced the South-North hypothesis and have resulted in a new scenario: in pre-Colombian times, presuming a wider centre of origin in the region of the Upper Amazon and Orinoco basins, Criollo and Forasteiro cacao would have been transported together along the Andean valleys by humans to Mesoamerica. Under the Mayan empire, the high chocolate quality produced by the Criollos initiated a conscious selection of this population, and the inadvertant extinction of the Amazon Forasteiro population. This scenario amplifies the understanding of the evolutionary process of the species, with significant impacts on conservation and genetic improvement; half a century after the South-North hypothesis had been proposed. |
| Introduction; Phylogenetic relations; Origin and distribution of populations. | Return To Table of Contents |
|---|
| Theobroma cacao L. is a relatively recently domesticated species, diploid (2n = 20), which belongs to the Sterculiaceae family. The generic nomenclature Theobroma, which means food of the gods, was coined by Linnaeus, in view of the indigenous belief in a divine origin of cacao. The word cacao is derived from cacahuatl, directly from the Mayan Nahuatl. Historical records point to Central America as home of the first cacao cultivations over two thousand years ago (Vavilov, 1951; Cuatrecasas, 1964; Purseglove, 1968; Bergman, 1969; Cope, 1976; León, 1987 e Hansen, 1991). Based on the characteristics of fruits and seeds and the geographic distribution, the many known botanic forms and types are, by mutual consent, classified in two large racial groups (Hall, 1914; Cheesman, 1944; Cuatrecasas, 1964 and Toxopeus, 1985): Criollo and Forasteiro, which contain distinct genes for production, disease, and pest resistance and for adaptation to different environments. It is appropriate to detach the nomenclature of these groups from the intuitive concept of geographic origin, according to which Criollo would be the native cacao type, and the Forasteiro a foreign one, although this was in fact the original idea of Morris in 1882, when he put forth the above cited classification (Cuatrecasas, 1964). As this classification, in spite of its dubiousness, is universally accepted, Cheesman thought it convenient to redefine it, instead of introducing a new one. Thus, "Nicaraguan Criollo", for instance, refers to the group of Criollo cacao from Nicaragua, and not to the native cacao of that country. The same applies for Forasteiro. Criollo populations were originally cultivated in Central America, from Mexico to the South of Costa Rica. The "non-Criollo" populations, on the other hand, also designated Forasteiro, are prevailing in the wild state in South America, in the Amazon basin region. Through the search for improvement of chocolate through better fruit quality, the domestication of the Criollo in Mesoamerica (Mexico, Guatemala, Belize, Honduras and El Salvador) induced profound changes in the group. Such changes affected the yield potential and rusticity of the tree, as well as the post-harvest processing and the chemical composition of its beans. The Criollos cultivated since the pre-Colombian era in Mesoamerica have large seeds, rounded and of white or light violet cotyledons when humid. The fermentation period is short, usually two days, which provides a superior chocolate due to its fine flavour and superior organoleptic qualities. Criollos are described as having a low yield potential and low rusticity. Unripe fruits are green or red, turning purple or yellow when ripe, and have a soft husk with ten deep furrows and a wrinkled surface. Forasteiro cacao tree cultivation, which began only three centuries ago in Equador, produces flat, purple seeds. These seeds ferment in five to seven days and the chocolate produced from them is relatively bitter. The Forasteiros present high rusticity, a high yield potential, green unripe fruits that turn yellow when ripe, with a hard husk, a smooth surface, and ten shallow furrows. Literature contains an attempt to subdivide the Forasteiro according to the fruit shape, based on the ratio length:diameter. There are four specific shape types among the populations and varieties: Calabacilo (ratio below 1.5:1), Amelonado (inferior to 2:1), Angoleta (2:1 without constriction) and Cundeamor (2:1 with constriction). This subdivision is, however, somewhat ambiguous since the same Forasteiro population, with one and the same fruit type, can express differences with regard to other traits. Taxonomists have therefore only partly accepted this subdivision, for example for the group of Amazon Amelonados due to its apparently uniformity in several other traits. Among the Criollos, some populations are classified on a geographic basis, such as those from Central America (Mexican and Nicaraguan Criollos), and those from northern South America (Venezuelan and Colombian Criollos), of which the latter present greater genetic variation. Among the Forasteiros, populations from the Upper and Lower Amazon are found, also based on geographic localization. Forasteiros from the Upper Amazon are vigorous, early bearing, and resistant to certain pathogens. In turn, those from the Lower Amazon, comprise a group of apparent uniformity, particularly in its amelonado fruit form, and are therefore known as Amelonados. Recent research reveals the existence of distinct types, possibly local races of one of the two groups. The wild cacao trees of French Guyana, for instance, seem to form a separate group within the Forasteiros (Lachenaud, 1997), similar to the ones from the Upper and Lower Amazon. The 'national' cacao from Ecuador also presents unique agronomical traits and molecular profile, apparently neither belonging to the Criollos nor to the Forasteiros (León, 1987; Lerceteau et al., 1997a; Lerceteau et al., 1997b and Crouzillat et al., 1998). Lastly, cacao populations found in the South of Mexico also have a different molecular profile from the Criollos (de la Cruz et al., 1995 and Whitkus et al., 1998). As a matter of fact, Cheesman, in 1944, had already foreseen the possibility that new forms and types that would not fit into the conventional classification would be reported. Spontaneous hybrid types between Forasteiros of the Upper and Lower Amazon and the South American Criollos appeared in Trinidad. They form the Trinitarian populations with a broad trait variation. The contribution of Central American Criollo genes to the Trinitarian, however, is somewhat uncertain. This hybrid group stems from natural hybridization when a disease decimated the Criollo plantations in mid-18th century, and the dead cacao trees were substituted by plants of the Forasteiro group, planted in the midst of the remaining Criollos (Cheesman, 1944). According to the taxonomy, the Trinitarians are part of the Forasteiro group, although their fruit and seed traits, especially, are intermediate between the Forasteiro and Criollo groups. It would therefore be rather temerous to classify the Trinitarian as a racial group, and would certainly undermine the traditional classification. Obviously, this traditional classification in Amazonian Criollos and Forasteiros is not ideal, mainly because of its ambiguity and the relative absence of genetic concepts. The Forasteiro group (non-Criollo), for example, comprises very distant populations, with broad variability. Besides, some of the determining characteristics are not exclusive to the Criollo group. Cheesman, 1944, claims that, despite being equally difficult to be defined, both are easily recognized. This is still is the best available classification, mainly because it satisfies requirements for breeding purposes. A critical revision of the origin and distribution of cacao (Bartley, unpublished) emphasized two aspects that contributed to the enhancement of this traditional classification: the fruit colour and the regional species distribution. We know that, generally speaking, the colour of unripe fruits is basically red in Criollos and green in Forasteiros and that all populations have their historical origin in only two areas of distribution, with the Andes chain as reference: the Caribbean chain in the west and the Amazon in the east, which are the distribution areas of the Amazon Criollo and the Forasteiro, respectively. Isoenzymatic analyses have also confirmed the geographic separation of the Forasteiros in South America from the Criollos in Central America (Ronning & Schnell, 1994). It is worth recalling that earlier pioneer work mentioned the absence of Forasteiros with red fruit and the differentiating role of the Andes (Cheesman, 1944). The definition of cacao distribution areas (Bartley, unpublished) was analogous to the one used for the definition of distribution centres of neotropical birds (Haffer, 1977): the trans-Andean and the cis-Andean; the first comprising the whole region on the West of the Andes to the Caribbean Islands and the second the entire Amazon. Analyses on molecular diversity are currently being used with the objective of examining the consistency of the traditional classification (Figueira et al., 1994; N'Goran et al., 1994 and Ronning & Schnell, 1994). Results are not yet conclusive however, due to a number of factors, such as the molecular marker type and the reduced sample size used in the analyses, as well as the constitution of the evaluated germplasm. (These factors are discussed in detail in the section on Biogiographical and Population genetic studies). Data on ribosomic DNA (rDNA) of 29 genotypes of T. cacao, from different racial groups and with different origins (Figueira et al., 1994), allow a partial classification of such genotypes in wild and cultivated states, similar to the classification of Harlan & de Wet, 1971 for cultivated species in general. Nevertheless, the precise differentiation between the wild and spontaneous cacao trees and the trees cultivated under natural forest conditions is not always an easy task (Bartley, unpublished). The broad genetic manipulation by man and the hybridization with many other populations derived from introductions make the identification of natural wild populations difficult. The expression wild or spontaneous refers to cacao trees undisturbed by human intervention while the opposite characterizes cultivated cacao trees. It is worth pointing out that Dias et al., 2001 used RAPD markers to separate traditional non-improved local Forasteiro cultivars ('Maranhão', 'Pará', and 'Parazinho') from the improved cultivars ('ICS 1' free pollination and commercial hybrid). Results indicate that RAPD markers distinguish genotypes with different degrees of improvement. Figueira et al., 1994, partially classified the same 29 genotypes by RAPD markers according to the traditional classification. As is generally known, the RAPD markers (Williams et al., 1992) cover the entire genome, including the encoding and non-encoding sequences, the repeated sequences, the single copies, and can therefore provide a more representative insight into the genetic differences between the genotypes (Figueira et al., 1994 and N'Goran et al., 1994). On the other hand, N'Goran et al., 1994 and Ronning & Schnell, 1994 obtained success at classifying 106 and 86 cacao genotypes, respectively, which also belong to different racial groups with distinct origins, according to the traditional classification. RAPD and RFLP markers were used in the first, and isoenzymes in the second study. Currently, even though the information volume on the evolutionary process of cacao and its domestication is increasing, the complete picture has yet to be assembled. This type of information is as essential for the improvement programs aiming at high yielding cultivars, as it is for the conservation and management of genetic resources of the species. The present chapter aims at reorganizing the available information concerning the origin and distribution of cacao (Theobroma cacao L), and makes possible inferences on the movement and formation of racial groups, and proposes a new scenario to the evolutionary process of cacao. Phylogenetic relationships: By the classic method of comparative morphology, the genus Theobroma comprises 22 species, divided into six sections (Cuatrecasas, 1964; see Chapter 2). All these species are diploid and have 20 chromosomes (Carletto, 1946; Muñoz Ortega, 1948 and Simmonds, 1954). There is a certain difficulty in obtaining viable hybrids with crossing (Addison & Tavares, 1951). As the chromosome number is equal in all species, the reason for the non-viability of some interspecific hybrids is of a more genetic than chromosomal nature. However, the same prevailing basic genome among the species ensures that the wild populations can donate useful genes to the cultivated species. All 22 species have seeds with white cotyledons (Soria, 1970b), except for T. cacao L., whose seeds are purple (see also Chapter 2 for a revision on the diversity in the genus Theobroma). It can therefore be assumed that the primitive cacao trees had seeds with white cotyledons and that the colourful mutant types appeared later. The increase in the gene frequency that defined the cotyledon colour was due to by some adaptive advantage, very likely the longer period of seed viability of the coloured types (Soria, 1970b). Analogically, this fact is observed in other crops. In beans, for example, coloured seed types present high emergence and greater plantlet vigour compared to the white types (Heiser, 1988). Besides, it is likely that a greater vigour and higher number of the coloured type ovules (Bartley, unpublished) could contribute to the predominance of the former. Just to simplify the problem, it worth bearing in mind that the colour of the cotyledons is a trait determined by an allele pair, and that the white is recessive in relation to the coloured one (Soria, 1978); see also Chapter 6). Nevertheless, the Criollo group of T. cacao L. has white seeds, with some types or variants showing graduations from white to light violet. The Criollos stem from mutations that appeared in populations along the periphery of the species' area of distribution. These were maintained and fixed by means of geographic isolation and selection (Soria, 1959 and Purseglove, 1968). Botanical expeditions, which collected cacao germplasm in the Brazilian Amazon region, confirmed the presence of white cotyledon cacao in natural Forasteiro populations of the Lower Amazon, in a higher frequency than expected (Barriga et al., 1985), while the Amazon Forasteiro has purple seeds. Cacao trees with white seeds were also identified among the Bahia Forasteiro populations of the Lower Amazon and classified as mutants, designated 'Almeida' (Bondar, 1938b) and 'Catongo' (Miranda & Silva, 1939), increasing the complexity of the scenario even more. Thus, Amazon Criollo and Forasteiro with white seeds could be classified as, initially, mutants preserved by man, with phenotypes reverted to the original types. The pleiotrophic mutants 'Almeida' and 'Catongo', however, do not fit into this earlier classification; in these, the albinism gene has pleiotrophic action, determining the white colour of the seeds, as well as of the flowers and new sprouts and does therefore not seem to be the same gene that determines albinism in the Criollo group (Soria, 1963). To find out whether it is in fact the same gene that controls the albinism traits in both groups, it would be necessary to conduct the allelic test or isolate and sequence the gene by molecular techniques. Origin and spread of the populations: Three hypotheses consider the origin and distribution of the T. cacao populations: i) South-North, proposed by Hunter & Leake, 1933, formally proposed by Cheesman, 1944, and extensively backed up (Schultes, 1984; Stone, 1984; Hawkes, 1991; Young, 1994 and Coe & Coe, 1996); ii) North-South, sustained by Vavilov, 1951, agreed by Mora Urpi, 1958 and Patiño, 1963 and corroborated by Harlan, 1976 and Sánchez & Jaffé, 1992, and iii) a natural broad geographic spread proposed by Nava, 1953, elegantly corroborated by Cuatrecasas, 1964, and corroborated by León, 1987. According to the first hypothesis (South-North spread), cacao is exclusively native from South America, and the supposed centre would lie in the headwaters of the Amazon basin, at the foot of the Andes, on the Ecuadorian-Colombian border. The expression 'centre of origin' refers to the area where a species' origin lies and from which it spreads out (Harlan, 1971). Botanical expeditions (Pound, 1938 and 1943) observed the existence of extensive variability in the wild cacao trees of Ecuador. Based on these observations, Cheesman, 1944, described this centre of origin as a 400 km-radius circle, in the confluence of the rivers Napo, Caquetá, and Putumayo, all tributaries of the Amazon River. Cheesman claims that the centre is more likely to lie in South than in Central America, because both, Criollo and Forasteiro, occur there, while there are no reports on the occurrence of Forasteiros in Central America. It is therefore assumed that the movement of cacao populations, despite its uncertainty, led to the formation of two sub-populations derived from a great ancestral population that thrived in this centre of origin. The absence of physical barriers to the north-east, east, and south-east direction of the supposed centre kindled the spread of the green fruit and the purple cacao bean by natural means; the reduced variability found downstream of the Amazon River confirms this route. Conversely, the Andes Cordilleras formed an insurmountable barrier for natural spread north and westwards. Only human intervention would have been able to have overcome this obstacle. It is therefore assumed that humans took a small group of cacao from the original eastern side population over the Andes. The two Criollo groups (central-American and northern South American), as known today, would have evolved from this group. It is believed that there were two independent migrations over the Andes: one that gave rise to the cacao 'Nacional' in western Ecuador and another that fathered the uniform Criollo type in the north of the country. DNA marker based data support this thesis (Lerceteau et al., 1997a). A few 'Nacional' cacao fruits were transported from one side of the Andes to the other, in an east-west direction, within the equatorial territory (Soria, 1966). The Andean Cordilleras acted as a first barrier that led to the differentiation of the Criollos, while the Isthmus of Panama formed a secondary barrier that gave rise to the differentiation of the Central-American Criollos. The mountain range allowed an independent differentiation of the Criollo traits on the occidental side of the Andes. Thanks to the Andes, part of the white seed populations remained isolated on the occidental side of the new mountains and brought forth the Criollo population (Soria, 1970a and Purseglove, 1968). In turn, the Forasteiros seem to have been differentiated in Upper and Lower Amazon Forasteiros in the Upper Amazon region. Most likely, nomads dispersed the cacao populations from the foot of the Andes, across northern Ecuador and/or southern Colombia to the south of Mexico. The cacao would have crossed the Andes thanks to humans, their most efficient dispersal agent, and developed a great part of its current traits under human influence. However, this hypothesis suggests that man had transported only seeds of white cotyledon Criollo cacao. Such seeds taste less bitter than those of the Forasteiro cacao, due to the low tannin contents. Consequently, the Criollos were dispersed in Central America and Mexico, where they could be developed under human influence. Conversely, the second hypothesis (North-South spread) is the antithesis of the first. It proposes that the Central American Criollos had a local origin in the far south of Lake Nicaragua, and were later brought to South America by Indians (Mora Urpi, 1958; Patiño, 1963; Sánchez & Jaffé, 1992). Vavilov, 1951, acknowledged the existence of a primary and another secondary centre for the origin of the cacao tree. The centre of origin and of primary cultivation would correspond to the seventh centre of origin (Central-American and South-Mexican) while the secondary centre would be included in the eighth centre (Brazilian-Paraguayan), specifically in the Amazon River valley. Soria, 1975, corroborated the hypothesis of Vavilov on the existence of two centres of origin. Ecuadorian expeditions to the basins of the Orinoco and Amazon rivers found similar types to the Criollo, at a surprisingly high frequency. All types had large white seeds and were related to the cacao populations 'Arriba' ('Nacional' cacao class) on the west Coast of Ecuador. Quantitative analyses of the amino acid contents of beans (Zarks, 1973, cited by Soria, 1978) also confirmed similar patterns of the Central American Criollo and the "Arriba" cacao, indicating a broad spread and common origin of the Criollo from these two continents. In a more conclusive form, RAPD analyses pointed to a South American origin of the Central American Criollos (de la Cruz et al., 1995). Lastly, the third hypothesis (broad natural geographic spread) assumes that, in remote times, a great Theobroma cacao population was dispersed from the central Amazon region to the South of Mexico. The geographic isolation of two subpopulations of this ancestral population by the Isthmus of Panama caused the appearance of the two subspecies: T. cacao ssp. cacao in Mesoamerica and T. cacao ssp. sphaerocarpum in South America; the first subspecies comprising the Criollos and the second the Forasteiros. The classification in subspecies is justified, because a) the two inter-cross and bring forth fertile and viable hybrids and b) the classic allopatric model, which implies speciation by geographic isolation and disjunction, supports the taxonomy. Within the subspecies T. cacao ssp. cacao there are three further forms: (i) T. cacao ssp. Cacao forma pentagonum called "cacau lagarto" (lizard cacao), known for its top quality products, supposedly originated from Central America, probably a mutant of T. cacao L.; (ii) T. cacao ssp. cacao forma leiocarpum, also a high quality cacao and a probable mutant of T. cacao L., only used for cultivation in Central America; and (iii) T. cacao ssp. cacao forma lacandonense cultivated in the primary forests of the northern State of Chiapas, Mexico, and known as probable ancestor of the Criollo cultivated nowadays in Mesoamerica. Other Neotropical genera also have a broad natural geographic distribution, as for instance Phaseolus vulgaris, the common bean, whose occurrence area stretches out from the north of Mexico to north-western Argentina (Gepts, 1991). The peculiarities and requirements of the wild cacao, nonetheless, let the hypothesis of Cuatrecasas, 1964, appear as the least likely. A tree that develops in the understory of tropical forests in hot, humid climate under dense shade could not spread out naturally over such a wide area. The recalcitrant nature of its seeds, the self-incompatibility, and the specificity of the pollinator that seems to have co-evolved with the tree, point in the same direction. |
| A new scenario for origen and distribution; Support for the new scenario. | Return To Table of Contents |
|---|
| New scenario of origin and distribution: Based on the three hypotheses on the origin and distribution of T. cacao L. as a starting point, it is possible to draw up a new scenario for its evolutionary process: From its centre of origin, consisting of the region of the Upper Amazon and Upper Orinoco basins (white and black stars, respectively, in Figure 3.1), in an area larger than was originally suggested, the cacao of the ancestral population would have evolved in Criollo and Forasteiro groups in the Andes region. Supposedly, in the midst of this now extinct large population that consisted of cacao with purple seeds and mutants with white seeds would have appeared. These mutants, supposedly in small number in the population on the oriental side, multiplied on the occidental side of the Andes. Due to geographic isolation, these mutants brought forth the Criollos, differentiating themselves from the Forasteiros. The elevation and formation of the Andes at the onset of the Tertiary epoch, 65 million years ago, was a decisive factor for this differentiation. The Forasteiro group seems to have split in to Upper and Lower Amazon Forasteiros in the region of the Upper Amazon (white star in Figure 3.1). Geoclimatological events such as the recurrent appearance of the Amazon Lake and the glaciation in the Pleistocene, about 4 million years ago, led to this differentiation. Subsequently, Criollos and Forasteiros were both transported by humans towards Mesoamerica and Mexico along two conceivable distribution routes: from the high Orinoco basin in Venezuela and/or through the Andes valleys across the tropical rainforest of the Pacific coast of Equador and Colombia. In view of the localization of the centre of origin, which involves the basins of the Amazon and Orinoco, both routes seem realistic and equally probable. Only few fruits were probably transported on each journey, since tree species were, historically, introduced by individuals in a small number. The short viability of the cacao seeds in particular limited their transport over long distances to small quantities. Once the area came under the influence of the Mayan (Mesoamerica) and Aztecan (Mexico) imperia, the high quality of the chocolate produced by the white Criollo cacao bean induced an intentional selection of this racial group and the inadvertent extinction of the Forasteiro populations. Chocolate thus came to be a driving force for the spread and domestication of cacao in Pre-Colombian America. Evidence for the new scenario: A tree with indehiscent fruits such as the cacao needs monkeys, rodents, and birds for seed distribution. Its seeds are recalcitrant and of short viability, which means that human intervention is necessary for a wider distribution. The influence man exerted on the evolutionary process of cacao is too meaningful to be ignored (Cheesman, 1944). In this context, the ethnic groups that lived, and are still living, in the distribution areas of the crop deserve a closer examination in studies on cacao evolution. Their paleoethnobotanical knowledge reflects the close relationship between man and plants, which is reflected in customs and ways of life. Therefore anthropological, historical, palaeogeological, and biogeographical evidences are analyzed, as well as the genetic population studies underlying the new scenario proposed in the previous section. The combination of useful information of the most diversified fields of knowledge is the recommended procedure to digress on the origin and spread of cultivated plants (Harlan, 1971). |
| Support for the new scenario; Anthropological. | Return To Table of Contents |
|---|
| No reports confirm the existence of peoples in the Amazon and Andes region that would have undertaken any attempt to select cacao trees. In other words, there is no evidence of cacao cultivation by the Indians of these regions - not in the past, nor in the present. The Amazon region gives home to numerous indigenous tribes, acculturate or not, which speak different languages, classified in at least 12 linguistic families. The knowledge on plant diversity of these indigenous people is remarkable (Schultes, 1991). However, the ethnic groups that actually live in the region of the Orinoco and Amazon rivers, supposedly remainders of the ancestral peoples, do not have a proper name for cacao ((Sánchez et al., 1989). This fact indicates a certain disinterest of most indigenous groups for the tree, although many have proper names for other Theobroma and Herrania species (the last mentioned is a genus similar to the first). The Ianomamis, by contrast, who use cacao wood to light fires, call the cacao tree Pojoroa, an indigenous word that means fire. A small cacao stick is rubbed by hand to and fro on a log of soft, inflammable wood, frequently also cacao, until it catches fire. This pre-historical technique to light a fire is also used by the Lacandons, surviving peoples of the Mayas that live in Chiapas, México (Annequin, 1978). Besides, the Ianomami have the habit of sucking the cacao seed pulp and prepare a light beverage of pulp, seeds, and water. Nonetheless, based on today's distribution, the occurrence of wild, supposedly native cacao populations in the Amazon region seems to have occurred independently of the distribution pattern of most of the indigenous populations that live in the Amazon nowadays, or have lived there in the past. This fact confirms the absence attempts to select the cacao tree by the natives of the Amazon region. Two great civilisations prevailed in the New World: the Andean, represented by the Incas and the Mesoamerican, represented by the Mayas and the Aztecs. Cacao is cited as a noteworthy example of a useful tree, well-known among the Mayan and Aztecan peoples, but apparently not known to the Incas before the arrival of the Spanish conquerors (Mangelsdorf & Reeves, 1939 and Bergmann, 1969) and even by the peoples that dwelled in the Amazon basin (Pio Corrêa, 1926; Nava, 1953; Bergmann, 1969; Cope, 1976 e Schultes, 1984), in spite of the nutritive value and the stimulating properties of the seeds. The Manuscrito Badianus, the oldest book on Mexican medicine plants (1552) reports the occurrence of Criollo cacao trees in Mesoamerica, compiled by two Aztecan Indians and translated to Latin by Juannes Badianus (Emmert, 1940, cited by Cuatrecasas, 1964). Illustration 68 of this manuscript shows six plants; number two represents the tlapalcacauatl, the coloured cacao (tlapal means coloured and cacauatl, as mentioned before, means cacao). Emmert adds, "this is the oldest illustration of cacao, Theobroma cacao L., the source of chocolate". In pre-Columbian South America however, some knowledge and primitive use of cacao is assumed among the Cuica Indians in Venezuela (Bergmann, 1969). There is at least one report on the use of cocoa butter in a religious ceremony of these Indians. In this case, the butter was prepared by triturating and cooking. After cooling and breaking it up, the solid white cacao butter was burnt to appease the demons. Still, this report does not necessarily imply cultivation. It is more likely, as stated by Cheesman, 1944, that some primitive use of the tree had merely been incentive enough to transport limited quantities of cacao seeds over the Andes Cordilleras. How cacao seeds could have crossed the Andes, in view of the inability to support dehydration and low temperatures, even for short periods, is to date a mystery. Cacao seeds exposed to temperatures of 2oC for only four minutes loose their viability completely (Boroughs & Hunter, 1963). The high degree of adaptation of cacao and its seeds to the hot and humid ecological conditions of Neotropical forests is well-known; only due to the recalcitrance, the viability of seeds is preserved under high humidity and elevated temperatures and only for a few weeks. Stone, 1984, believes that the plants or even stems with buds were transported, which is quite unlikely. There are two possible migration routes along which cacao could have eventually spread (Schultes, 1984). The first would have led across the rainforest of the northern Atlantic Coast of South America, through the territories of Guyana and Venezuela, until reaching the mouth of the Amazon River. The other course could have followed the Rio Negro, in view of its connection with the Orinoco in Venezuela. Moreover, a third route could be assumed along which a few fruits would have been transported by Indians through valleys of the Andes, until reaching the tropical rainforest of the Pacific Coast of Equador and Colombia. Whatever was the route of cacao spread - to assume the transport of exclusively Criollo type fruits with white, de-pigmented beans would only be justified if the peoples that spread the species in Central America had been familiar with the less bitter beverage obtained from its seeds - a fact that is not backed up by any historical report. In this case, only the Aztecs in the Mexico valley and the Mayas in Mesoamerica could have selected in favour of the Criollos. There is a thesis that wandering Mayan traders could have transported cacao from the North of South America to Mesoamerica, as well as in the opposite direction, just to revive themselves by sucking the pulp of the seeds, maintained fresh when transported within the fruit (Stone, 1984). This does not invalid the presupposition that the Forasteiros and Criollos had been transported indiscriminately to Central America and that the selection of the Criollo at the expense of the Forasteiros had begun on that continent only. One needs to remember that the colonization of the Americas including the Amazon forest happened over 10 thousand years before the trade of the Mayan merchants. Besides, sucking the pulp of the cacao seeds was customary before or, at the most, simultaneous to the discovery and manufacturing of chocolate. If the selection criterion for the transport of cacao trees in ancient times was only the taste of the seed pulp, it is not reasonable to assume that only white seeds would have been transported. The best tasting pulp, thickest and least acid, is not an exclusivity of this or that group. Scavina clones, for example, classified as Forasteiros of the Upper Amazon and possibly originated from the region of the river Ucayali (Figueira, 1997), have a thick and sweet tasting pulp. Plants have always been selected with regards to a stable yield, size and colour variations in seeds, and flavour quality (Harlan et al., 1973). The selections of white wheat grain for bread, of barley for food, for beer, and for animal food are some examples of the influence men exerted on plant selection. Traditional agriculture has always shown a sharpened intuition to increase the nutritive value of crops. The conscious selection of white or light-coloured bean seeds, for example, was practiced because the black and red beans presented an excess of tannins, which, in reaction with proteinase, reduce the digestibility (Heiser, 1988). The Mayas and Aztecs owned great knowledge and used the Criollo cacao beans to prepare a delicious and aromatic beverage of cooked and ground beans with maize and pepper, which they called chocolate. Among Mesoamerican natives, cacao beans circulated widely as currency for trading goods, as well as to pay taxes. This great importance gave rise to commercial routes between the Pacific side and the Caribbean of the Isthmus of Panama (Bergmann, 1969). Reports from 1576, recorded by Diego García de Palacio (Bergmann, 1969) inform that cacao cultivation in Izalco, El Salvador, was a ceremonial process in which the seeds were carefully selected and the planting of the trees done in conjunction with lunar phases. Cacao was also used in birth ceremonies and as a wedding gift. In Guatemala, there are reports on cacao cultivation under irrigation in the Pre-Colombian era (Bergmann, 1969). The cacao tree was of such importance to the Mayas and the Aztecs that they had a special god called Ek Chuah (Figure 3.2) to whom the responsibility for the production of plantations was ascribed (Mangelsdorf & Reeves, 1939). Based on these facts, it is conceivable that both the Mayas and the Aztecs practiced conscious, deliberate selection in favour of the Criollo, once the tree reached the domains of their empires. An analysis of the history of the New World, using the chronological standard of the Mayas, reveals that the beginning of agriculture in America can be dated back to 4 thousand years BC (Spinden, 1968) and that its diffusion through Mesoamerica occurred some 2.5 thousand years ago (Annequin, 1978). The plants domesticated by the Amerindians were developed far beyond the wild plant types, with an even greater intensity than that dedicated to the domesticated plants of the Old World. Obviously, this development spanned many centuries. Verifiable commercial interactions show that the Mayan civilization, in this connection, exerted its far-flung influence in the New World from the Yucatan peninsula to Colombia in the south and to New Mexico in the north. The introduction of the cacao tree into Mesoamerica from the north of South America is ascribed to trade (dated back to over 1600 years BC) of the Mayas or even of the ancient Olmecs (Stone, 1984). After the Mayan cataclysm, the abandoned regions were re-populated by a succession of South American tribes that settled on the Yucatan. Also in the times of the Aztecs, there was an active flourishing gold market in trade relations between the Isthmus of Panama and Mexico. The consequently intensive movement of tribes between South and Central America since remote times was verified, upholding the idea of an intense interchange of producer goods and plants between the two continents ((Mora Urpi, 1958 and Sánchez & Jaffé, 1992). In fact, there is strong evidence for the transport of plants, especially cacao and cassava, from South America to Mesoamerica (Schultes, 1984; Stone, 1984), taking into account that the migration of types of languages and peoples to the Amazon's interior lasted only 2500 to 4000 years (Frailey et al., 1988). Particularly, the importance of the Mayan civilisation for the domestication of cacao is such that it deserves a closer look at the history, culture, traditions, habits, customs, and way of life of this civilization. Up to a certain point, the history of cacao is told by the history of the Mayas. The Mayan civilisation: The Mayas, considered the Greeks of the New World, were discovered by accident on March 4, 1517, by Francisco Hernandez de Cordoba, a rich Cuban who was looking for slaves for his plantations. A violent and persistent storm forced the little Cuban fleet of three sailing ships to drop anchor on the northern margin of the Yucatan. Before the expedition left, the explorers stumbled upon the Mayan city. The inedited vision of the bold architecture composed of temples and pyramids full of idols and plenty of gold inspired the discoverers to name the city "Great Cairo". The reports of Díaz Del Castillo in 1568, cited by Spinden, 1968, are essential since they are an eyewitness account of this first contact with the Mayas. This incomparable civilisation had persisted for over a millennium (3rd to 14th century) in the Mesoamerican region between the South of Mexico (far north of the Yucatan peninsula) and Guatemala. It is estimated that this region had the size of present-day Italy where one to two million Mayas would have dwelled. The discovery of this universe has enthralled numerous historians and archaeologists to devotedly study the civilisation. The appearance, apogee, and mysterious decline of the Mayas are discussed at large in several works. This section is based on Spinden, 1968 and Annequin, 1978. The Mayan civilisation reached its peak during the so-called classic period, between 300 and 900 AD. Some traces of its culture are also found in the contemporaneous civilizations of the Olmecs in the Gulf of Mexico, the Totonacs in Veracruz, the Zapotecs in Oaxaca, and the Teotihuacans 50 km north from Mexico City. This entire cultural framework allowed a facies americana for the civilisations of the New World. The Olmecs (1200 to 600 AD) are believed to have been the pioneers of cacao domestication and chocolate processing, three thousand years ago, and that the ancestors of the classic Mayas inherited all this knowledge between 400 BC and 100 AD. (Coe & Coe, 1996). The pre-conquest chocolate was not only a simple mixture for drinking; it was rather a mixed complex of beverages, porridge and mush to be served on special occasions, such as wedding ceremonies, while the 'saca' prepared with cooked maize, water, and cacao was an every-day drink. The Mayan universe was characteristically of a unique and extraordinary polymorphism in the most diverse domains. In the pre-Colombian era, the Mayan peoples spoke 123 different languages, whose dialects had little to do with each other and no relation at all with Asian languages, indicating an independent origin. Data on the physical complexity of Mayan peoples show an expressive variability in terms of height, pigmentation, blood groups, and morphology. The environmental scenario inhabited by the Mayas was (besides being vast and relatively unfavourable) of a confusing diversity at the moment of the Spanish conquest. This mosaic of environments covered the high fertile soils of Chiapas and Guatemala in temperate climate; the forests of the central limestone region of Tabasco in Mexico and of Peten in Guatemala, under tropical hot and humid climate, low fertility soils and heights between 30 and 180 meters, which, paradoxically, served as the scenario of the boldest cultural boom of Mesoamerica. Dwellings in the dry marginal region of the Mexican States Campeche, Quintana Roo, and Yucatan were restricted to the surroundings of the "cenotes" - great wells formed by the collapse of overlying limestone crusts, opening the access to ground water or even to underground rivers. In sacred cenotes like the one in Chichen Itza, humans were sacrificed by drowning in honour of Chac, the rain god, to obtain relief of some scourge such as drought. Criollo cacao trees were recently found in three cenotes of the region (Gómez-Pompa et al., 1990). The differentiation between the cacao from Chiapas and that from the Yucatan has been attributed to these cenotes (Whitkus et al., 1998) since, theoretically, the cenotes would have represented introduction sites of plants imported from other regions to the dry soils of the Yucatan. If this thesis were true, the cacao trees found in the cenotes are survivors of the primitive cultivars used by the Mayas. The social organization, architecture, the arts, and the development of agriculture, writing, and the calendar are some of the main legacies of the Mayas. As agriculture developed, sedentary life forms replaced the nomad life and were wisely harmonized in rural towns. Cornfields, known as milpa, planted in the forest clearings after burning, were found in the surroundings of the urban architectonic complexes. It cannot be overemphasized how important corn was for the Amerindians and particularly for the Mayas, since it represented their basis for food security. If humankind of the Old World originated from soil, the Amerindians including the Mayas believed that man had come from maize. Before working the earth and sowing the maize, the Mayas fasted and practiced abstinence and offers to the sun gods for a rich harvest. The strategies in agriculture, vegetable gardening, and forestry of the Mayas were based on mixed cropping, the selection of adapted varieties to the different climatic zones and the use of the stratified vertical and horizontal space, as it occurs in nature, by the use of terraces and drained and irrigated fields (Barrera et al., 1977). Although the wheel was unknown, and without draught animals (oxen, horses, donkeys etc.) or metal tools, the Mayas cultivated common bean, cassava, amaranth, pumpkin, sweet potato, green pepper, cotton, tobacco, coca, rubber, and cacao. The beverage known as chocolate was prepared by cooking a mixture of maize flour and cacao powder, spiced with pepper in boiling water. The cacao beans, as mentioned earlier, were used as currency, for the payment of taxes and for trading goods. The Mayas buried the dead in earthen urns together with all their earthly belongings, which included dishes of maize and cacao. The religion of the people was present in all their activities. The dozens of worshipped divinities all had an ambivalent character, either protecting or demonic. The god of cacao Ek Chuah (Figure 3.2), with its black silhouette, thick hanging lips, and a scorpion's tail, also had a dualistic personality; this figure appeared with a pack on its back, worshipped by travellers and merchants, or as a god of war and bloody sacrifices. The Mayas invented a vigesimal numeric system, in which a dot represented the unit, a hyphen number five, and zero was expressed by a shell; the concept of zero, of fundamental importance, was only to be known in mediaeval times by the peoples of the Occident. With this system, they realized advanced astronomic calculations on sun and moon phases and eclipses. They described the movement of the planet Venus, for example, with an error of two hours over a period of five centuries. In their hieroglyphs, they left an inheritance that is witness to their customs, way of life, and beliefs written in codices (folding books written on deer skin). The three remaining codices are the Dresden (Codex Dresdensis) with 78 pages from the 11th or 12th century, the Madri (Codex Tro-Cortesianus) with 112 pages written in the 15th century, and lastly, the codex of Paris (Codex Peresianus), also from the 15th century with only 22 pages. Cacao is represented in the first two codices (Stone, 1984). The Mayan hieroglyphic writing consists of 450 to 500 different signs called glyphs. Glyphs are more ideographic than pictographic, basically, as they are symbols, and not graphic representations of objects. This writing is by now nearly deciphered, in a computer-assisted process, which has helped scientists understand more about the greatness of the Mayan civilisation (Coe, 1999). In terms of scientific advance, this author compares the deciphering of the Mayan writing to discoveries in space and the decodification of the genetic code. Moreover, traditions and ancient myths were translated into Latin in the ethnic books Chilam Balam and Popol-vuh, in an attempt to immortalize the history of this people. Finally, a tiny ethnic group - the Lacandons - still preserves the language, traditions, and costumes of the Mayas in their purest form and has been serving as a live laboratory for studies on the Mayan universe. At this point, it is necessary to return to the theories on the colonization of the New World. The most traditional and broadly accepted among them assumes that the first migrants crossed the Bering Strait - between Rússia and Alasca - coming from the northeast of Asia, between 15 and 9 thousand ago, to colonize North America. From then, they moved towards South America, crossing Central America and descending along the Andes mountain chain. These first colonizers, called Paleo-Indian or also Clovis people (in honour of the city of Clovis in New Mexico, where archaeological sites of these first Americans had been discovered) to which the origin of the Mayas was attributed. They lived in open, dry, and cold habitats, and were hunters of large animals. Support of this thesis comes from the fact that the faces of most of the Mesoamerican Indians present Mongoloid features. Analyses of the mitochondrial DNA (Bonatto & Salzano, 1997) and of Y chromosome data (Santos et al., 1999 and Pena, 1999) also strongly indicated that the actual Amerindians of the north and south had their origin in a single migration of Siberian peoples out of the central region, 12 thousand years ago in the Pleistocene epoch. Supporters of the traditional theory argue that the tropical rainforest environment was not inhabited by the Paleo-Indians due to the lack of hunting and edible plants. Nevertheless, a recent discovery (Roosevelt et al., 1996) showed the existence of a primitive culture quite different from that of the Paleo-Indians, though contemporaneous or even prior to it, that would have lived in the interior of the Amazon forest over 9 thousand km south of Clovis. This discovery calls the traditional theory of the settlement of the New World into question. The culture dwelled in the tropical rainforest and lived on fruit, fish, birds, hunting of small animals, reptiles, and on molluscs of the alluvial plain of the Amazon Forest. Excavations in the 'painted stone cavern', on the left bank of the Amazon River, in Monte Alegre, Pará, brought traces of this primitive human existence to light, such as remains of carbonized plants and animals, rock paintings, and over 11.2 thousand year old stone arrow heads. Monte Alegre would have had, at the time, around 300 mil inhabitants, with a high degree of knowledge manifested in granite paintings and the development of pottery (Roosevelt et al., 1996). The discovery of this site was based on observations recorded by Alfred Russel Wallace in 1849 (Wallace, 1979), co-author of the theory of evolution by natural selection together with Charles Darwin, on the stone paintings found 305 masl in the highlands of Monte Alegre. This new discovery shed light on the understanding of the spread of cacao. The plants identified in the cavern of Monte Alegre include Brazil nut, palm seeds, numerous fruits of common tropical trees in rainforests, besides the wood of the Amazon region, although cacao seeds were not depicted (Roosevelt et al., 1996). An important observation is the fact that the plant species known to the people of Monte Alegre in the late Pleistocene are still found in the Amazon rainforest, demonstrating their age and the ability of the primitive peoples to survive by adapting to local food. Likewise, the settlement of the Amazon rainforest by the Monte Alegre people proves that the tropical environment did not restrict the subsistence of primitive peoples. In fact, the tropical jungle was tamed by human intelligence in many occasions. In Mesoamerica, for instance, the actual floristic composition of the forest where the Mayas lived reflects the integrated management these peoples used to get their living out of the jungle (Barrera et al., 1977). There is no way around accepting the idea of primitive anthropic activity with an expressive environmental alteration, especially in the Pleistocene. This impact depends, basically, on the density of the local demography, the agricultural practice, even if nomad, the use of fire, and hunting abilities of the primitive man. All in all there were about 54 million Amerindians in the New World before the arrival of the Europeans. In the first century after this contact, the population was reduced to about 10% of this total, and only recovered to around 30% in the 18th century (Denevan, 1992). Population declines and slow growth like these give vegetation a chance to recover after having undergone alterations. Furthermore, the human history mingles with that of the fire, known to the aborigines since over 500 thousand years. Fire gave men multiple capacities to provoke fast environmental changes. Data collected in areas of the Amazon forest that were altered by primitive slash-and-burn agriculture revealed that it takes 60 to 80 years to recover the diversity of species If ever), and 140 to 200 years to recover the lost biomass (Saldarriaga & West, 1986). Dynamic balances, in which pollinators and seed dispersers interact intimately with the plants, characterize the tropical forest environment. With this background, the hunting of large herbivores can lead to changes in some species and affect the survival of several others. Many studies suggest the presence of primitive peoples in the Amazon forest seven thousand years ago, who practiced nomad agriculture of mainly maize and cassava (Cohen, 1977). Thus, the activity of these primitive peoples has doubtlessly affected the original forest vegetation despite the magnitude of these alterations remaining an open question. |
| Support for the new scenario; Historical and Paleogeological. | Return To Table of Contents |
|---|
| Historical evidence: Bondar, 1924 reports "when the Aztecs conquered Guatemala, chocolate and cacao cultivation were already known but the people had to put up with a wild cacao species, dark-coloured and with an acrid taste, called patluxe, while the rich and noble used the excellent soconusco, whose bean was so precious that it was used as currency". In Mexico cacao was also used as currency. In ancient times, Soconusco was an important cacao district (Bergmann, 1969). Amerindians in Nicaragua treated cacao with reverence in their religious ceremonies (Bondar, 1938a). These reports demonstrate that chocolate had already become a widely used beverage in a large part of Central America, even before the Spanish conquest, and was kept in such high esteem by the local indigenous peoples that its quality discriminated social classes. Furthermore, the same report confirms the presence of Criollo and non-Criollo cacao in Central America. Similarly, Sahagún, in 1950, described the existence of a cacao variety in Mesoamerica with reddish seeds, a typical characteristic of Forasteiros (Whitkus et al., 1998). In view of the divine origin ascribed to cacao by the Aztec and Mayan peoples, as well as the great value attached to its beans, broadly used as currency and as medicine against dysentery by these peoples, it is not likely that a tree considered so precious was indigenous (Purseglove, 1968). This is in contrast to the hypothesis of a Mesoamerican origin of cacao. On the other hand, it is not so surprising that Mesoamerica became the centre of domestication of cacao that was not indigenous to the region. There are many examples of tropical tree species whose areas of cultivation are distant from their centres of origin. In fact, there is no reason to suppose that the centre of origin of a given plant should be the region that offers the most favourable conditions for its cultivation. On the contrary, cultivation in a distant region from the origin is often free of the origin’s pests and diseases. Perhaps the superior intellectual input of the Mayas, in comparison to the native civilizations of the Andes and the Amazon, and the more favourable environmental conditions of Mesoamerica for cacao cultivation, explain the fact that Mesoamerica became the centre of domestication of cacao, which was not indigenous to the region. Paleogeological evidence: The actual distribution of the species seems to be a result of geological events that happened in ancient times. Therefore, knowledge on the development of the Amazon is fundamental to understand the actual conditions of the vegetation in this mega region. The region’s geological history is quite simple (Sioli, 1991). During the Palaeozoic era, 570 million years ago, the Amazon depression extended to the South American continent, limited in the North and South by the archean granite shields of the Guyanas and Central Brazil respectively. This depression was covered by the sea, forming a gigantic gulf open to the Pacific in the west, but closed to the east, and South America was then connected to Africa. During the Carboniferous period, about 345 million years ago, the sea receded. Later, in the Mesozoic era, 225 million years ago, the Amazon basin finally emerged. Its rivers flowed towards the Pacific, forming a proto-Amazon running in the opposite direction to that of today. The separation between South America and Africa began in the Rhaetic, between the Jurassic and the Triassic periods (225 and 190 million years ago), characterized by the dolerite intrusions that emerged everywhere. Apparently, there was a wide spread of the angiosperms in the late Mesozoic and Caenozoic eras, in the direction South to Central America and at a lower scale in the opposite direction; indicated by the great similarity of this plant class and the entire flora actually found on these two continents. During the Cretaceous period, 130 million years ago, this continental region was a continuous and isolated corridor for the terrestrial forms, when Africa and South America split and the latter partially converged, in some parts, with Central America (Rich & Rich, 1991). The Andes emerged nearly 65 million years ago in the Tertiary epoch. Based on this it is worth emphasizing that the last 60 million years were important for the development of the modern American flora and fauna (Rich & Rich, 1991), that is, in the era after the elevation of the Andes. The Andes chain not only blocked the runoff of the Amazon river network to the Pacific, during the Pliocene and the whole Pleistocene epochs (between five and two million years ago), but also retained the water mass, forming a huge lake. The entire Amazon depression was then filled with sweet water sediments from rivers and lakes. Communications with the ocean were only found in the West, as sediment layers of salt or seawater side by side with sweet water deposits show. Finally, the whole fluvial network was flowing eastwards and pouring into the Atlantic - resulting in the actual Amazonian river system. The sedimentary soil formed in the water dried up and was covered by the Amazon proto-forest, the precursor of the present day Amazon forest, while the rivers formed their beds and valleys. Cold climates appeared two million years ago for the first time in the tropics, during the Pleistocene, a phenomenon known as glaciation. This provoked, even more, the spread of mountainous species, particularly the expansion of the savannah species (Raven & Axelrod, 1975 and Rich & Rich, 1991). In the Glacial period, the sea level lay around 100 m below today’s level and the incline of the rivers from the Amazon basin was greater and their current stronger, so that the rivers formed wide and deep valleys into the tertiary sediments of the lower and middle Amazon valley. Later, at the end of the glacial period about 10 thousand years ago, the sea, whose level was continuously rising, dammed the water of the rivers back up into the central Amazon, inundating its valleys. During the glaciation period, 18 thousand years ago, the climate of the Amazon was cooler by 4oC and drier than today. It is believed that these arid climate periods altered the extension of the Amazon forest significantly, although we cannot evaluate exactly to what extent and how this phenomenon occurred. Studies on the actual distribution of some plant and animal species caused some biogeographers to propose the theory of forest ‘refuges’. During the driest climate periods, the Amazon forest would have retracted considerably, and been reduced to isolated forest ‘islands’, separated by great Savannah areas. These islands, designated refuges, became typical allopatric centres of speciation, of the associated plants and fauna. The differentiation of the isolated populations would have occurred due to the variation in the selection pressure and the stochastic forces in the refuges, and the numerous new species dispersed with the re-expansion of the forest during more humid periods. The alternation of arid and humid climates in the Pleistocene epoch would thus have influenced the differentiation of the flora and fauna of tropical South America. The theory of the refuges was used to explain the speciation process of Neotropical birds (Haffer, 1977) and the resulting refuges map was compared to other presumed refuges of reptiles (Vanzolini & Williams, 1970), butterflies (Brown, 1975) and trees (Prance, 1973). The relative concordance between the four maps of the supposed refuges was considered a proof to support the refuge theory. Sioli (1991), however, considered this theory a highly unlikely speculation, which needs very critical evaluation as the four superimposed maps practically cover the forest of the entire Amazon region. Furthermore, with the progressive drying of the climate, the dry belts moved towards the Equatorial line. Therefore, there would have been an advance of the savannah and a retreat of the forest in the peripheral north and south regions, principally, and not in the Central Amazon basin. In the peripheral regions, the forest would have endured in the form of riparian forest buffers along the rivers, as is still the case in the contact areas of forest and Savannah. The forests would have disappeared from the driest area and survived in those with greater precipitation. The latter, with precipitations above 2500 mm a year, lie in the Upper Amazon, at the feet of the Ecuadorian, Peruvian, and Bolivian Andes along the Atlantic sequence of Guyana, including the Amazon mouth, and along the Central American mountain chain (Haffer, 1977). The studies suggest that the separation between the Upper and Lower Amazon forests occurred in the last dry period, between 4000 and 2500 years ago (Bigarella, 1965), with a re-expansion in the humid period, which lasts until today. Despite all the criticism the refuge theory generates, it has been the basis for the classification of the wild cacao of French Guyana into a separate group among the Forasteiros (Lachenaud et al., 1997 and Lerceteau et al., 1997b). According to these authors the rivers of French Guyana do not belong to the Amazon basin, nor the highlands of Guyana, which is separated from the Amazon basin by several mountains that hinder the contact between the cacao trees of that region with the Forasteiros of the Upper and Lower Amazon. Savannas actually separate the Amazon and Guyana’s highlands, and it is likely that they had also been completely isolated from each other during the dry periods of the Pleistocene or even the Holocene epochs. The authors support their hypothesis of a supposed Guyanian refuge that should have existed in the central region of Guyana, which extended south eastwards and formed a network of micro-refuges. Most likely, however, the extensive fluvial net that separates the rivers of the Amazon basin from those of Guyana had differentiated the cacao populations of both regions by forming an effective dam against the genetic flux between them. The fact that most of the proposed refuges lay, supposedly, on the shores of the lake Amazon, Frailey et al., 1988 offered an alternative model to the refuge theory, based on geological evidence. The core of this model is the assumption of an Amazon Lake, which would have promoted the distribution of the species in isolated tropical forest fragments along its edges. In a broad revision, Marroig & Cerqueira, 1997, recognized the dynamism and complexity of the South American history since the Pleistocene, and the need to link hypotheses to explain the Neotropical diversification. These authors associated a recurrent Amazon lake, covering the entire Amazon basin, with the theories of the refuges and river barriers. The appearance of the lake would have pushed the spread of the forest and the associated fauna back to the periphery of the region, similarly to the refuges formed in the highlands of the Amazon basin during the dry periods. Thus the speciation would have been determined by the isolation through large watercourses. Differently from the theory of the refuge, however, the combination of these theories does not foresee the coalescence of the forests during the humid periods. When the level of the sea and the rivers sank, they were no longer barriers to the genetic flux and allowed contact among the populations. Generally, for all species, endemism and diversity are greater at the periphery, mainly on the highlands, which suggests a centripetal migration in the depression of the Central Amazon (Bush, 1994). The plant differentiation pattern also suggests a centripetal migration, i.e., from the highland periphery of the Amazon basin inwards (Marroig & Cerqueira, 1997). Such occurrences would only have been possible if the appearance of the lake had been recurrent. Paleogeological studies (Klammer, 1984) confirm that the Amazon basin formed a gigantic lake - the Amazon Lake - stretching out from the Andes to the Amazon delta of today, during the last four million years (Pleistocene) and that this lake appeared recurrently, owing to oscillations of the sea level. Because of the lake, the Orinoco basin was probably connected to the Amazon basin in this period. Specifically, in relation to these two basins, there are certain biogeographical aspects which underpin the existence of a connection between them in the past, owing to the dynamics between the watercourses of these two rivers and the Rio Negro (Frailey et al., 1988). It is therefore possible that the recurrent appearance of the Amazon Lake could have caused a partial extinction and the spread of the cacao population towards the periphery of the actual Amazon basin, limiting it principally to the Amazon and Orinoco basins, where the differentiation process into local subpopulations began. It is worth remembering that the differentiation of the cacao populations occurred in watersheds (Dias et al., 2000), with more or less distinct central types in every one. The hypothesis of differentiation in basins had been put forth by Pound (1938) on the occasion of his botanic expedition to the Upper Amazon, recently and provisionally confirmed by Dias et al., 2000. |
| Support for the new scenario; Biogiographical and Population genetic studies. | Return To Table of Contents |
|---|
| Biogeographical evidence: The statement that Forasteiros occur in a wild state in the Amazon is generally accepted (De Candolle, 1883; Myers, 1930; Duke, 1940; Cheesman, 1944; Nava, 1953; Cuatrecasas, 1964; Purseglove, 1968; Hansen, 1991 and Almeida, 1996). The occurrence of most species of the genus Theobroma in this region (Cuatrecasas, 1964 and Sanchez & Jaffé, 1992) supports this assertion. Of the 22 known species of this genus, 19 are found in South America, of which 13 are restricted to the Amazon region. Ten of these 13 species are endemic in the area of influence of the Orinoco and Amazon Rivers. Furthermore, cacao trees from the Upper Orinoco carry a striking ancestral trait: their seeds mature early and germinate inside the fruit. In conclusion, evidence does not prove, but strongly suggests that cacao’s origin lies by the sources of the Orinoco and Amazon Rivers, as similarly expressed in the past (De Candolle, 1883 and Hunter & Leake, 1933). Similarly, the hypothesis of a Mesoamerican origin of cacao (Mora Urpi, 1958 and Patiño, 1963) was recently revived (Sánchez & Jaffé, 1992). Assuming that Ianomami Indians occupied the region of the rivers Orinoco and Amazon since ancient times, they possibly dispersed cacao across the Amazon forest in the Pre-Colombian era and would have been responsible for the introduction of cacao cultivation in the Upper Orinoco region from Central America (Sánchez & Jaffé, 1992). Advocates of this hypothesis reason that the variability is considerably greater in Mesoamerican than in South American cacao populations. They further allege that all types of known cacao occur in Mesoamerica, but not all are found in the South American region (Mora Urpi, 1958 and Patiño, 1963). If this statement were confirmed, the great variability in Mesoamerican cacao trees could easily be explained. It could be ascribed to mutations, self-incompatibility systems, cultivation since remote times, and the inter-varietal and inter-specific hybridization accelerated by the introduction of cacao from a number of regions when domestication occurred in Mesoamerica. However, the samples of genetic diversity collected in South and Central American Criollo populations contradict this hypothesis. They disclose a reduction of the variability from Colombia towards Central America, not in the opposite direction, based on the phenotypic (Cheesman, 1944 and Motamayor et al., 1998) and genetic diversity at the DNA level (de la Cruz et al., 1995 and Whitkus et al., 1998). The latter articles support the assumption that the Criollo had a South American origin as well. Curiously, the cultivars 'Matina' from Costa Rica, 'Ceilão' from Guatemala and of México and 'Amelonado de Trinidad' from the Dominican Republic are very similar to the cultivar 'Comum' from Brazil (Soria, 1966), indicating a common origin (see Chapter 12 on cacao cultivars). The hypothesis that humans selected and transported exclusively Criollos from South America to Mesoamerica and continued to select it there was justified by the supposed absence of Forasteiro cacao in this latter region (Cheesman, 1944). However, the difficulty of identifying natural wild populations, the broad anthropic manipulation, and the hybridization with many other populations from introductions may have been the reasons for this absence. On the contrary, the selection in favour of the Criollo may have contributed to the possible extinction of the Forasteiros in Mesoamerica. Nonetheless, the previously mentioned report of Bondar, 1924, and the description of a cacao cultivar with red seeds, a typical feature of the Forasteiros in Mesoamerica, by Sahagún in 1590 (Whitkus et al., 1998), supported the idea of the existence of non-Criollo types; in other words, Forasteiros in Central America, in Pre-Colombian times. Moreover, molecular studies using wild cacao from the Yucatan, point to a broad diversity beyond the limits of the Amazon Criollo and Forasteiro. Cacao trees found in the Mayan cenotes were genetically distinct from the wild cacao of South America, from the Criollo, and from modern cultivars, appearing to be primitive cultivars of these people or survivors of such (de la Cruz et al., 1995 and Whitkus et al., 1998). There are also reports (Cuatrecasas, 1964) on the presence of small wild cacao groups (T. cacao ssp. cacao fsp. lacandonense) in the primary Mexican forests of Chiapas inhabited by the Lacandons, very distant from any other cacao population with small, rough, pointed, elongated fruits with 10 furrows. The lacandonense form is believed to be the likely ancestor of the Criollo cacao cultivated today in Mesoamerica. One should not lose sight of the intensive domestication process to which the Mayas and other Mesoamerican primitive, peoples before and after them, subjected the cacao over thousands of years. In this sense, the powerful influence of domestication causing differentiation in plants and animals must be underlined. Finally, a common cultural characteristic in practically all primitive civilizations should be highlighted: the planting and preservation of a stock of cultivars distinct from the original species. Rice in China and cassava in Brazil, for example, were cultivated by the ancient peoples of these countries using dozens of cultivars, markedly differentiated in adaptations to environmental factors and to their end products. Without referring specifically to cacao, Barrera et al., 1977 mention that the Mayas used a surprisingly high number of cultivars from among all ‘milpa’ species, one for each purpose, according to its peculiarity - as their survivors still do today. There is no reason to deny that the Mayas would have also kept their stock of cacao varieties, fulfilling several objectives. Some of these could cause variations in the chocolate flavour they appreciated so much. Other cultivars, however, may have been cultivated for health purposes or simply economic yield criteria. Since the cacao beans also served as currency, it is clear that only cultivars with superior yield performance were planted and preserved in the past. It is a complete mystery why research with primitive cacao cultivars includes so little information on these aspects. Hernández in 1651 (Schultes, 1984), described in some detail four cacao types in Mexico and their use as food, currency, and medicine against dysentery. Genetic population studies: The genetic population structure is for the most part controlled by the reproduction mechanism of its individuals. Consequently, changes in the crossing systems can accelerate the genetic divergence in natural populations (Stebins, 1957). In cacao, this subject is of great importance for the evolution due to the presence of self-incompatible multi-allelic systems in certain populations (Knight & Rogers, 1955 and Cope, 1962b; see Chapters 4 and 6 for subject in greater detail). It was assumed that self-incompatible cacaos were predominant in the specie's centre of origin, while the self-compatible types were dispersed along the marginal distribution areas (Posnette, 1945; Cope, 1962a and 1976). This fact was refuted by the presence of self-incompatible cacao found in marginal areas also (Soria, 1978). Nevertheless, this speculative supposition was backed, to some extent, by the scarcity of pollinators and the reduction in the cacao population density in the marginal areas of distribution. Such circumstances would have exerted a strong selection pressure in favour of self-compatibility. We know that the mating system in plants is controlled by a few genes which, when mutated, provide the necessary flexibility to ensure the species' reproduction. In this situation, self-compatibility would have had a great adaptive value. There is one more possible explanation for the predominance of self-incompatibility in the centre of origin and the self-compatibility in the marginal areas (Cope, 1962a). On the assumption that there had been a great number of different alleles of self-incompatibility in the original cacao population, the number of incompatibility alleles would have been reduced at every stage if only a few fruits had been transported during the period of human migrations across the Andes. Had cacao been planted in tree groups, self-incompatibility would not have impaired the fructification, as long as trees with cross-compatibility had been included as pollinators. Finally a stage would have been reached where there would be few different incompatibility alleles left in the population. Even though the incompatibility system could, theoretically, be maintained in populations with only two individuals, de Nettancourt, 1977, confirms that an allele loss due to genetic drift or selection would reduce the number of alleles in small populations to values below the minimum required for an adequate functioning of the system. In this respect, when the Aztecs and Mayas selected the Criollo based on the white beans for a better chocolate, they indirectly selected self-compatibility. In other words, the conscious selection of cacao with white seeds resulted in automatic, unconscious selection for self-compatibility. Conscious selection and automatic selection refer respectively to intentional human selection (Heiser, 1988) and changes in the characteristics associated to a trait under selection (Harlan et al., 1973). This may have been the other reason for the prevailingly self-compatible Criollo populations in Central America. This idea finds support in a theoretical study on mixed populations of self-incompatible and self-compatible cacao (Cope, 1962a), assuming absence of inbreeding depression. In the study, self-compatible cacao was prone to predominate due to greater fructification, while the self-incompatible tended to be eliminated, demonstrating that self-compatible and self-incompatible cacao genotypes cannot coexist in equilibrium in an isolated population. In contrast, field observations of both self-compatible and self-incompatible Trinitarian cacao populations abandoned in 1950 seemed to indicate that the frequency of self-incompatible trees increased in a single generation (Warren & Kalai, 1995). However, this result could also indicate that the reduction of self-compatible cacao in the mixed population was caused by inbreeding depression, expressed as the establishment of compromised seedlings. This is certainly a complex question that calls for further investigation. We know that numerous endemic varietal characters are found in the centres of origin (Vavilov, 1951). Certainly, some of the strongly expressed traits found in the Upper Amazon populations of the proper Upper Amazon region and also of the Upper Orinoco, compared to those of populations of other racial groups such as the marked self-incompatibility (Posnette, 1945), the elevated butter content (Pires et al., 1998), the early seed germination in the fruit as soon as the fruit is ripe (Sanchez & Jaffé, 1992) and the higher concentration of genes for resistance to witches' broom (Pound, 1938 and Cheesman, 1944) are surely some of the endemic ancestral characters that confirm the origin of cacao in these regions and which, consequently, can be used in the identification of wild populations. In present-day studies on the diversity in cacao, (Engels, 1983 and 1986; Enríquez et al., 1988; Raboin et al., 1993; Bekele et al., 1994; Dias & Kageyama, 1997a and 1997b; Dias et al., 1997 and Lerceteau et al., 1997a), used morpho-agronomical markers, (Lanaud, 1986; Ronning & Schnell, 1994; Warren, 1994 and Sounigo et al., 1997) used isoenzymatic and (Wilde et al., 1992; Fritz et al., 1993; Lanaud et al., 1993; Laurent et al., 1993a and 1993b and 1994; Russel et al., 1993; Figueira et al., 1994; N'Goran et al., 1994; Lerceteau et al. 1997a; Lerceteau et al., 1997b and Whitkus et al., 1998) used DNA markers (see also Chapter 10). It is hoped that these markers will provide complementary information on possible evolutionary routes of T. cacao. Since the differentiation between populations is primarily based on fruit and seed characteristics, the multivariate analysis for characters such as yield and its components and adaptive characters is essential in phylogenetic studies. The importance of the morpho-agronomical characters should be stressed since the natural selection influences the phenotypic frequencies directly (Stebins, 1957). Multivariate analyses of morpho-agronomical markers (Raboin et al., 1993; Dias & Kageyama, 1997a) showed a close relationship between the Trinitarian and the Forasteiro populations of the lower Amazon. The relationship between the geographic and genetic divergence appeared partial or absent. This means that the geographic distance may not be a solid index for inferences on the genetic distance. Finally, when grouping the cacao genotypes, the morphologic markers show partial agreement with the conventional classification. Since the morphological characters are phenotypic expressions, genetically distant populations may be morphologically similar. This limitation has led to the use of isoenzymatic and DNA markers, complementary to the morphological markers, to evaluate phylogenetic relationships. Molecular markers free of environmental influence offer the opportunity to examine the genetic relationships among accessions more precisely. Studies with these markers point out that the variability of the Upper Amazon Forasteiros contains a great part of the total species variability (Lanaud, 1986; Laurent et al., 1993a and 1994; Russel et al., 1993; Figueira et al., 1994; N'Goran et al., 1994; de la Cruz et al., 1995; Lerceteau et al., 1997b and Whitkus et al., 1998). This supports the thesis that the region of the Upper Amazon represents concomitantly both the centre of origin and of diversity of T. cacao. Finally, these studies confirm the differentiation between Criollo and Forasteiro (Laurent et al., 1993a, 1993b and 1994; Russel et al., 1993; N'Goran et al., 1994; Ronning & Schnell, 1994; Sounigo et al., 1997; Lerceteau et al., 1997b and Marita, 1998), with a higher level of polymorphism for the Forasteiros (Laurent et al., 1993b; Lerceteau et al., 1997b; Whitkus et al., 1998). It is worth recalling that controversial results produced by morphological analyses, in relation to those verified in molecular analyses are easily generated. In the common bean for example morphological analyses indicate an increase of the variability during and after domestication, while molecular markers suggest a decrease (Gepts, 1991). The apparent controversy may be attributed to differences in genes that determine the two kinds of phenotypes and to distinct evolutionary forces influencing them. The morphological characters that distinguish the wild from the cultivated types are controlled by genes with pronounced effects and subjected to artificial selection. Molecular phenotypes, in turn, are frequently determined by genes of little effect and are predominantly subjected to genetic drift. Moreover, the different molecular markers reflect different evolutionary forces (Mitton, 1994). In plants, this aspect causes the genomic DNA to evolve more quickly than the chloroplastic DNA (cpDNA) and this one faster than the mitochondrial DNA (mtDNA). In turn, the repeated DNA sequences evolve more quickly than the sequences of single copies. Therefore, those markers that present different DNA types and sequences differ in their ability to capture events that happened in antiquity, which is certainly an influencing factor in the definition of genetic relationships between racial groups. Controversial results are to great extent also due to the use of small or large samples from the cacao germplasm banks, which retain little or no genetic representation of the natural populations of the species. These samples commonly stem from plant material selected for its resistance to diseases in Peru, Ecuador, and the Brazilian Amazon, which make up the narrow genetic base of the cacao cultivated nowadays. Thus, studies with these materials do not allow inferences on the genetic structure of the populations of the species since they deal with corrupted or biased samples of natural populations, presumably in a Hardy-Weinberg imbalance. The obtained results are restricted to the studied samples (see Chapter 11 on the nature of mathematic model effects for analysis). |
| Final considerations and prospects. | Return To Table of Contents |
|---|
| The broadening of knowledge on the evolutionary process of cacao will have a strong impact on the conservation and genetic improvement of the species. Efforts to save the germplasm should concentrate on the areas where cacao is indigenous and in which the species has greatest variability. In this case, the collection should first begin in the Amazon and Orinoco River basins, which hold the greatest centre of diversity of the species. The collection strategy, however, should include other centres of diversity, such as the southeast of French Guyana or the North of Ecuador and the Chiapas forests in Mexico in order to amplify the representation of the species’ diversity. The in situ conservation of these genetic resources is another strategy that should be implemented concomitantly to the ex situ conservation in these areas, because of the greater concentration of diversity and the eminent threat of extinction to which the species’ populations are exposed. It is well documented that the Amazon region holds the planet’s greatest biodiversity with approximately 16% of the world’s fauna concentrated in a 1 million km2 area. The risks of erosion of its genetic resources are all too well known. This threat emerges in the form of regional mega projects development that involve the occupation of forest areas, the expansion of the agricultural frontier, the rural exodus of the natives from the region, and the acculturation of the remaining inhabitants, who had always maintained a close relation with local plants. In this context the proposal of the creation of an international centre of genetic diversity of the crops in occidental Amazon (Clement, 1991), the region with the greatest number of species, is essential for the genetic in situ conservation of the biodiversity and the local culture. The impacts on genetic improvement will be greater once the genes of new accessions of the different racial groups of the species of the genus Theobroma are incorporated in the modern cultivars. Wild cacao trees collected in French Guyana, for example, were found to be immune, under natural conditions, against the fungus Crinipellis perniciosa, causal agent of the disease witches’ broom (Lachenaud et al., 1997); this disease decimated the Ecuadorian cacao cultivation in the 20s and actually compromises the socio-economic viability of Brazilian cacao cultivation (see Chapter 1). Cacao might have well been the most valuable tree of the entire American continent before the Spanish conquest. Highly appreciated by the primitive peoples, be it as source for chocolate, as currency, as tool to light a fire or as drug, this tree cannot be studied without considering human influence. By tracing and mapping the surviving peoples, it will become possible to reconstruct the spread route of cacao populations. The discovery of primitive cultures that inhabited the tropical Amazon forest over 11 thousand years ago could provide information on how these peoples altered the original forest vegetation and influenced the spread of plants such as cacao. Molecular markers will play an important role in tracing evolutionary process and tracking down the species’ spread routes as they enable population analyses at a large scale. The accelerated, computer-assisted deciphering of the written hieroglyphs and the investigation of cenotes and archaeological sites left by the Mayan pioneers in the cacao cultivation in Mesoamerica will certainly shed light on the darkness that covers the origin and domestication of T. cacao. |
| Chapter 4. Ecology of natural populations. C.M.V.C. Almeida. | Return To Table of Contents |
|---|
| Contents: Introduction; Geographical distribution pattern; Local distribution; Reproductive system, Gene flow, Life cycle components; Natural dispersion agents; Wild and domesticated populations, and Variability of natural populations. Summary: The wild cacao populations of the Brazilian Amazon are broadly dispersed and well adapted to distinct physical and biotic environments. The distribution pattern of cacao is random or aggregated, in groupings of heterogeneous density and spatial spread. Such groups range from a few plants, restricted to clearly defined areas, up to thousands of plants scattered over wide areas, frequently forming large interconnected local sub-populations. The reproductive system has intermediate species characteristics, owing to the variations caused by different factors. It is believed that isolated plants or clumps in the forest represent elements of genetic links between sub-populations that coexist in the same geographic area. It is further assumed that the formation of clumps, the occurrence of adjacent plants that present the same genotype, the high seed reproduction potential, and the incompatibility system are adaptive strategies of the species to ensure its survival. As well as these aspects, the gene flow over short and long distances, the difficulties in determining the degree of domestication of the natural populations, as well as the level and distribution of the genetic variation in these populations are also emphasised. |
| Introduction, Geographical distribution pattern and Local distribution. | Return To Table of Contents |
|---|
| Introduction: The genus Theobroma is distributed across the rainforests of the occidental hemisphere between the latitudes 18oN and 15oS, stretching from the South of Mexico to the Amazon Basin. It is presumed that the onset of the Tertiary epoch separated the Theobroma populations that were already broadly dispersed by the elevation of the Andes. This fact favoured speciation by means of geographic isolation and gave rise to the racial groups Forasteiro and Criollo (Cuatrecasas, 1964; see also Chapter 3). The region of the rivers Napo, Putumayo, and Caquetá in the Upper Amazon basin, comprising part of the territories of Ecuador and Colombia, is considered to be the centre of diversity and probable origin of Theobroma species owing to the great phenotypic variability (Cheesman, 1944). It is believed that cacao spread in two directions from this region: the Forasteiro group extended eastwards and the Criollo northwards. This theory has found support in studies with molecular, isoenzymatic, agronomic, and morphological markers, which indicated a greater genetic variability in botanical material from the Upper Amazon than that from other regions (Figueira et al., 1994; Lanaud, 1986; Laurent et al., 1993 and 1994 and Whitkus et al., 1998; see also Chapters 2, 3, and 10). Among the 22 species of the genus Theobroma (Cuatrecasas, 1964), only T. cacao, represented by the cacao tree, is economically exploited for the manufacture of chocolate and derivatives among other uses, and accounts for all global commercial cacao production. Recently, the commercial exploitation of T. grandiflorum in the Brazilian Amazon, where it is called ´cupuaçuzeiro´, has gained economic importance. Its pulp is being used to manufacture juice, syrup, ice creams, liquors, jams, jellies and yoghurt etc. Also, the potential of its seeds is being developed for the preparation of a fine quality chocolate (regionally known as ´cupulate´) because its fat is highly digestible, similarly to that of cacao butter. (Venturieri & Aguiar, 1998 and Souza et al., 1996). Historical records show that cacao trees were cultivated since pre-Columbian times in an area encompassing southern Mexico and the present border between Costa Rica and Panama (León, unpublished), principally in the districts of Soconusco and Tabasco, in Mexico, which produced cacao on a large scale for the regional market (Bergmann, 1969. The first cultivars used, which belong to the racial group Criollo, were closely related to the culture and religious ceremonies of the Mayan civilization. The elite classes used the cacao seeds to prepare food and beverages (Mooleedhar et al., 1995). Therefore, this entire region, which includes Mesoamerica with its long history of human settlements and agricultural activities, is an important centre of domestication, diversification, and continuous evolution of T. cacao (see Chapter 3). The first contacts Europeans had with cacao occurred in July 1502, when Christopher Columbus´ ship, in its fourth journey to the Americas, found a native canoe on the northern coast of Honduras, in the Caribbean. It contained some local market goods, among them a large supply of cacao beans. At that time, the cultivation and use of cacao were already well-established cultural characteristics of the indigenous communities that lived in Central America (Bergmann, 1969). The Spanish later introduced its cultivation in South America. The domestication of T. cacao had already began when the Spaniards arrived, since at the time the Mayas were already familiar with the cultivation as well as the use of the seeds (León, 1968). Cuatrecasas (1964) believed that the diversity in the types found in Mexico and Central America was the result of thousands of years of cultivation and selection practised by autochthon communities. Kerr & Clement, 1980, mentioned that the domestication process occurred in the last 2000 to 4000 years, while Clement, 1990, considering the degree of genetic modification of the indigenous Amazon fruit trees, classifies the cacao tree as a semi-domesticated species. Owing to the great diversity of physical (climate, altitude, latitude) and biotic (set of living creatures in a determined region) environments in the occurrence areas of T. cacao, and the consequently differentiated selective pressures, the conclusion is drawn that the species is adapted to different ecosystems of tropical America, demonstrating no preference for a specific habitat. This fact contributes to the increased variability, since the populations become genetically distinct through the adaptation process to different environments. The genetic variability constitutes the raw material required for the improvement and evolution of the species. In plant species, according to Hamrick, 1983, this variation is influenced by the following characteristics: effective population size, geographic spread of the species, reproduction method, crossing system, seed dispersal mechanism, and community type where the species is frequent. The magnitude and distribution pattern of the genetic variation in tropical tree species has been investigated by means of these characteristics (Dias & Kageyama, 1991). Working with the data obtained in various provenance and progeny tests with tropical eucalypt and pine, these authors show that this magnitude is high and mostly concentrated within populations, as is the case in temperate tree species. Knowledge on biological mechanisms that affect the population structure (set of genetic and demographic characteristics) of indigenous T. cacao communities is an important auxiliary tool in improvement programmes and for the collection of genetic resources of the species, especially for the establishment of strategies, tactics, and procedures for cacao collection and its conservation planning (see also Chapter 5). The present chapter aims to compile knowledge on the ecological and evolutionary processes that influence the natural T. cacao populations. It is partly based on reports of the cacao collection expeditions to the Brazilian Amazon region promoted by CEPLAC, (Almeida et al., 1995), on specialized literature, and on the experience acquired by the author in collecting cacao germplasm. It is an upgraded and updated version of the article by Almeida, 1996. Geographical distribution pattern: The natural T. cacao populations of the Brazilian Amazon grow in the regions between, approximately, lat 3' 41" N (in Roraima) and 11' 47" S (in Rondônia) and long 44' 00" E (in Maranhão) and 72' 46" W (in Acre), including, also, the States of Amapá, Amazonas, Mato Grosso, and Pará, in the areas of the ´Hileia Amazônica´ (a name coined for the Amazon forest by the scientist Alexander von Humboldt). According to Ducke & Black, 1954, the ´Hileia Amazônica´ is delimited by the occurrence limits of certain forest taxa, such as the genera Hevea] (rubber), Theobroma and Bertholletia excelsa (Brazil nut). The climate of this vast region varies from humid to extremely humid, where the relative air humidity indices lie mostly above 80% and can drop to 60% in the dry season. The annual rainfall indices are high (1600 mm to 3000 mm), with a mean of 2500 mm. There is no dry season, or only a short dry season, which lasts between three and four months. The temperature is always high and uniform, oscillating between 22o and 30oC, throughout the year. The region also houses a vast and ramified net of watercourses, with an estimated extension of 25 thousand km of navigable rivers in the Amazon valley where the Amazon river, the main river of the region, and its effluents and sub-effluents ramify into numerous lakes, streams, ´furos´, and ´paranás´. The river has a drop of only 65 m between the Peruvian border and the ocean, over an extension of 3.000 km. The annual cycle of the water level varies more in consequence of the rains in the upstream parts of the rivers than because of local factors; it can vary between four and seven meters in the Lower Amazon regions, and 15 to 20 meters in the lower Japurá river, an effluent of the Solimões (Sioli, 1984; Miranda Neto, 1991 and Pandolfo, 1994). In this area of geographical distribution, cacao is a spontaneous and typical element in forests of upland and wetland plains, found in any area where there is tropical rainforest (Ducke, 1940 and 1953); estimated in about 3.6 million km2. This spread, however, is irregular and generally restricted to isolated areas (Bartley, unpublished). In the first habitat, its occurrence is common in low to high natural fertility soils, in reliefs that vary from flat to strongly undulated. There are reports on collections in Brazil that followed contour lines a few meters above sea level, as in the estuaries of the rivers Munim and Itapecuru in the State of Maranhão, and the Amazon River in the State of Pará, and up to 400 masl., in the northern State of Mato Grosso, and in Rondônia (Barriga et al., 1985). In the plains, cacao frequently dwells in rarely or little flooded parts (restinga” or “high plain”), at maximum heights of 80 m. The tree is not found in open uplands, such as in Amazonian fields or savannahs, nor in ´igapós´ forests, mangroves, wetlands or some upland formations of minor territorial significance. In Rondônia, from approximately 11’ 47" lat S, the spread of T. cacao was interrupted in south and south-easterly directions. This disruption or blockade is probably caused by the large area of ecologic tension (contact forest/savannah) (Barriga et al., 1985), impeding a successful species reproduction besides representing, probably, the “species’ frontier” area, according to the concept of Mayr, 1977. Large cacao populations are common in the regions of high natural soil fertility in the State of Rondônia. In some settlement projects, this fact was a reference for the choice of appropriate areas for cacao cultivation. In the same state, pedalogical studies carried out by CPRM specialists showed that the abundant occurrence of wild cacao trees was always associated to a high natural fertility soil. Ducke, 1953, studied the spread of the Brazilian species of the genus Theobroma and observed that the occurrence of wild T. cacao in abundance was always associated with soils of high natural fertility. Obviously these observations are result of the best possible environmental conditions available for the spread of seeds, and consequently, for the growth and development of the plants. They do, however, not indicate an ecological specialization of the species to a given particular condition of the environment, as these populations are also found in other soil types. Local distribution: Population spread is the way to colonize new or deserted areas by means of movement of individuals or their dissemination forms (seeds, spores, and others), within or beyond the population or population area. Wild cacao is found in the Amazon in three basic forms (Bartley, 1977): i) spontaneous - without human interference; ii) sub-spontaneous - wild cacao trees exploited by man, and iii) cultivated - plants that stem from seeds of wild cacao trees. In the spontaneous and sub-spontaneous forms, T. cacao is an element that belongs to the under-story of plants under 20 m - dwelling, generally, in dense shade and under highly competitive conditions with other species for water, light, and nutrients. The cacao trees can be found in the form of single plants (single trunk) or plants with several trunks at different development stages (“clumps”). Both forms represent a random or aggregate spread pattern. In this last case, they form “reboladas” or groupings with a heterogeneous population density and spatial spread. Such groupings can have a few plants - 20 to 30 - restricted to well defined areas, - up to thousands of dispersed individuals in very vast areas, often forming numerous interconnected local subpopulations, a fact that complicates the establishment of the exact limits of a given population. Allen, 1984, made similar observations. Reports on botanical expeditions in Brazil mention the existence of wild subpopulations of T. cacao in Rondônia in a region of nearly 10 hectares, in Mirante da Serra (Almeida, 1983) and even up to 78 hectares in Seringal Extrema (Soria, 1965). Lisboa, 1990, who studied the cities Presidente Médici and Costa Marques, reported the occurrence of T. cacao populations with a density of 142 individuals per hectare of forest, where this species is the most frequent. The Jamari, the principal river of Ariquemes, became well-known in the 18th century for an abundance of wild cacao trees on its banks. (Silva, 1984). Three different situations can be visualised in the cultivated form: i) home orchards or gardens - cacao trees planted haphazardly in variable numbers, mingled with other fruit or ornamental trees in the surroundings of farmers’ houses; ii) commercial plantations - technically established cacao trees, forming dense forests, grown from seeds of wild plants or from improved botanical material (hybrid cacao), in plantations of varied dimensions, and iii) isolated plants - cacao trees that grow on pastures or diverse plantations, formed by sprouts of cacao tree stumps that subsisted or escaped burning. T. cacao, as well as T. grandiflorum and T. speciosum are common in small orchards in the entire Amazon region. Commercial plantations grown from seeds of wild cacao trees are normally found on plains and islands, especially in the lower and intermediate Amazon region along the inferior courses of its main tributaries, and the lower Tocantins. They consist of plant formations in varied extension and of uncertain age, traditionally implanted by ancestors of the riverine settlers. Furthermore, reports mention small plantations in Rondônia - results of “experiences” of recently arrived rural producers to the region, especially through the settlement projects under the coordination of INCRA in areas foreseen for the agrarian reform. The form of the isolated plants occurs randomly. |
| Reproduction system, Gene flow, Life history components and Natural dispersion agents. | Return To Table of Contents |
|---|
| Reproductive system: The reproductive system is an important component of the genetic populations´ structure. The distribution of genetic variation within the progeny, among individuals of the population, and among population subdivisions is controlled by the species’ reproductive system (Hamrick, 1982). Although cacao has hermaphrodite as well as homogamous flowers, its pollination is limited almost exclusively to the intervention of some midge Forcipomyia species, although other insects such as thrips, ants and aphids can cause accidental pollination (Chapman & Soria, 1983). Supposedly, the coloured flower parts attract such midges, especially by the petal guidelines and the staminodes (Kaufmann, 1975). Soria et al., 1982, however, believes that the behaviour of these insects does not depend on the flower colour. The cacao flowers’ anthers are covered by elongated petals, in concave form, (sepals) and the ovary is surrounded by a circle of infertile staminodes. This complex flower structure requires the participation of insects for pollination and represents the adaptation of T. cacao to the activity of its main pollinator agent (Soria et al., 1975). The high degree of cross pollination is characteristic of this species, whose natural crossing rates range from 50 to 100%. This contributes to the high heterogeneity in populations formed from seeds without controlled pollination (Vello & Nascimento, 1971 and Toxopeus, 1972), or that local populations, which had even evolved under isolated conditions, represent a more or less heterogeneous group formed by heterozygotic individuals (Cheesman, 1944). The pollination type of cacao contributes to the fact that a pollen mixture of the actual and of neighbouring plants is frequently deposited on the stigma (Voelcker, 1940). However, the presence of incompatibility systems in the population can limit natural fertilization and restrict the gene flow, as well as the yield of self-incompatible plants, despite such plants often presenting compatible crossing among each other. This incompatibility mechanism is exclusive of T. cacao, since gametophytic and sporophytic factors control the phenotype for pollen incompatibility together during microsporogenesis. The proposed explanatory hypothesis assumes the presence of three gene loci: S, A, and B, which regulate the syngamy in the self- and cross- pollinations (Cope, 1962). Locus S presents a polyallelic series, with dominance and independence relations (S1 > S2 = S3 > S4 > S5), while loci A and B would have the function of producing unspecific precursor substances to activate the first alleles. Due to this mechanism, there are three self-incompatible plant types in nature, as regards self-pollination: i) those with 25% unfertilized ovules; ii) plants with 50% unfertilized ovules, and iii) completely self-incompatible plants with 100% unfertilized ovules. Additionally, crossings between certain self-compatible clones can produce partially or totally self-incompatible progenies. This degree of expression of self-incompatibility has also been confirmed with biochemical markers. Lanaud et al., 1987, for example, used isoenzymatic markers and observed that the percentage of self-fertilized seeds varied from 0 to 89% in self-incompatible clones, while Yamada, 1991 found endogamy rates in self-incompatible clones, which varied only from 3 to 8% (see also Chapter 6, which describes the heritability of self-incompatibility). Such information shows that the reproductive system of T. cacao presents variations, assuming intermediate species characteristics and that factors such as: self-incompatibility, pollinator population, seasonal and temperature effects, among others, contribute to determine the number of fruits a plant produces. Gene flow: The gene flow is the allele migration from one population to another and this is important to understand the formation of the genetic structure of populations. (Martins, 1987). Cross pollination in cacao generally occurs among neighbouring plants (Voelcker, 1940 and Posnette, 1950). However, there is a report of gene flow, via pollen, between plants over 45 m from each other (Cope, 1939), and it is known that air currents are probably essential to determine the spread of adult populations of the pollinator midges (Winder, 1977). The studies related to determining the maximum distance over which pollen grains can be transported from one plant to another by Forcipomyia are incomplete. Neither is there any experimental information on the gene flow via pollen and seeds between wild populations of T. cacao, or on modifications in the gene flow caused by variations in method, density, and population spread. However, it can be assumed that the plant form of a single plant or isolated “clump” in the forest represents an element of genetic connection, random or not, between the populations that coexist in the same geographic area, without natural barriers, and should contribute to the formation of new genotypes and the appearance of certain similar traits among them. Under these circumstances, it is presumed that these populations or subpopulations present a certain degree of parentage among each other. In relation to the gene flow via seed, it is known that the spread can occur by means of different agents: man, water, and wild animals, as much as between as within populations (see Natural spread agents). However, the rapid decline in seed germination in less than four days after extraction from the fruits (Keleny, 1968), as well as the attraction they have for certain wild animals due to the sweet taste of the mucilaginous pulp must limit this process substantially, in cases where the seeds depend merely on accidental factors and not on intentional anthropic influence to preserve the species. In view of these facts, the great difficulty for the formation of a natural seed bank and, consequently, for gene flow in time via seed, becomes evident. Nevertheless, the formation of a plantlet bank is possible, especially at a distance from the mother tree, although this condition is seldom found. Another possibility would be the gene flow between the species of the genus Theobroma. Among the 22 classified species of the genus, the following are found in the brazilian Amazon: T. cacao, T. bicolor, T. sylvestre, T. speciosum, T. microcarpum, T. grandiflorum, T. obovatum, T. subincanum, T. glaucum and T. canumanense (Cuatrecasas, 1964). Collection expeditions of cacao in the Brazilian Amazon have demonstrated that the first eight species cited coexist in the same habitat, in different regions, although collectors were not always aware of this fact. Considering that there is introgressive hybridization in the genus Theobroma (Addison & Tavares, 1951 and Cuatrecasas, 1964) it is presumed that natural interspecific hybrids growing in different Amazonian regions, although this statement would be difficult to prove, even for experienced botanical collectors. (see Chapter 2, which deals with interspecific hybrids). The formation of small geographically or ecologically isolated natural populations suggests that the base population is formed by few individuals, which represent only a small fraction of the total genetic variability of the parent population. In this case, it is assumed that the pollen spread remains restricted to its components, and therefore reduced genetic recombination and variation within populations. The individuals of this population must relate to common ancestors and the hypothesis of a high degree of endogamy among them is possible, although it is known that populations are seldom completely isolated and that a certain degree of migration does, inevitably, occur under natural conditions. The T. cacao populations of great dimensions prevalent in areas without natural barriers can possibly be formed from a variable number of individuals, requiring a great genetic diversity in their components, in function of the genetic recombination that occurs in the course of the various generations. Such observations are important in relation to the sample strategy under field conditions, since the existence of genetic homogeneity between the components of a given population allows smaller sample sizes while the existence of genetic diversity in the population requires more comprehensive sampling, based on a larger number of plants (see Chapter 5). Life cycle components: The life cycle components consist of characteristics involved with the population growth, such as reproduction rates, mortality, duration of the life cycle, reproductive power, and others (Hamrick, 1982 and Martins, 1987). Among these, the formation of wild “clump” cacao trees (Figure 4.1) represents an important component of the life history of the species. Such clumps are easily identified in the forest where several trunks with a twisted and inclined aspect grow from the same planting spot. Many recently formed stems can be identified alongside those with diameters of over 40 cm and dozens of years of development, which suggests a situation of age stratification and, consequently, of overlapping generations. The proportion of the adult to young trunks established in explorative expeditions carried out in Rondônia, in 1983 and 1984 (Almeida, 1983 and Almeida & Almeida, 1985) for example, suggested that the ratio of adult to young stems was 1.0:0.78. In these cases, a clump may seem a lot older than the apparent age of the oldest trunk. This situation ensures the population’s genetic continuity over time, due to the reproductive interconnections between generations (Shorrocks, 1980), and supports the perennial nature of the clump. The wild cacao trunk can reach a height of about 16 m, in consequence of the successive development of orthotropic branches just under the apex (León, 1968), or up to 25 m, as observed in wild populations in the Oiapoque river basin, in French Guyana (Lachenaud et al., 1997) and in the Mayan Mountains, in Belize (Mooleedhar, 1999). The tree originates from orthotropic branches, which grow out from the original trunk. Cases of clumps with over 20 adult trunks are not rare. Collections carried out in Rondônia showed that about 68% of the clumps chosen for the collection of botanical material had 1 to 5 adult trunks, 28% had 6 to 15, and 4% over 15 trunks, with a mean of 5.2 adult trunks per clump (Almeida & Almeida, 1985). Another variation of this reproduction process is the possibility that neighbouring plants or clumps belong to the same genotype (Allen & Lass, 1983; Almeida & Almeida, 1987 and Sanchez & Jaffé, 1992), due to tilted old trunks of a determined clump and the posterior growth of orthotropic branches out of these fallen trunks, giving rise to new plants. This kind of expansion from the canopy of the mother tree can occur at any time of the year. Both reproduction processes probably represent adaptation strategies of the species to the environment in which it lives, in order to warrant the success of reproduction and its own preservation. Another probable adaptation strategy lies in the high reproductive potential of the species resulting from the following mechanisms: i). The high flowering potential of the species - cacao flowering is generally abundant, especially in experimental conditions under full sunlight, where up to 125 thousand flowers/plant/year can be produced (Lachenaud & Mossu, 1985), although less than 5% grow to become mature fruits, considering the open pollination. It seems as if the abundant flower production in cacao is a means to ensure the attraction of the pollinator agents. In natural populations, and because the cacao tree grows in the understorey of the forest, under conditions of dense shade and high competition for water, light, and nutrients, as well as the presence of bryophytes, lichens and other trunk flora, the expression of flowering as well as the efficiency of natural pollination is less pronounced, this driven by low Forcipomyia population levels. However, the elevated plant height, offering a greater area for the emission of flower cushions, should partly compensate the restrictive action of the environmental factors and allow flowering that is more than sufficient to ensure pollen spread and, as a result, the species´ preservation; ii). The high seed potential of the species - the cacao flower has 35 to 70 ovules per ovary and produces around 14 thousand pollen grains. The fruit, in turn, is indehiscent, of the drupisarcidic bacoidal type, with a short peduncle (Figueiredo, 1986), which remains fixed to the plant until harvest, and thus needs the participation of animals to disperse its seeds. The seeds are recalcitrant (they must be stored in environments of high humidity and temperature to reach a prolonged longevity), and quickly lose their germination ability after the fruits break up. Monkeys, rats, or squirrels are some of the dispersal agents, as well as predators of the species. They usually drive a hole into the fruit to remove the seed and suck its mucilaginous pulp, and then drop the seeds onto the soil (Toxopeus, 1985). In reports on collections realized in the Brazilian Amazon, fruits with up to 67 seeds in perfect germination conditions and plants with over 100 fruits were registered. It is also known that cacao yield is a quantitative trait, highly influenced by environmental conditions, presenting a high variability among individuals and that, in natural Brazilian Amazon populations, according to observations of elderly inhabitants of the region, there is an alternation of high yield with low yield years. If one imagines a small natural population with only 20 clumps, five adult trunks per clump, 35 normal seeds per fruit and a mean productivity of five healthy fruits/trunk/year it would be possible to produce at least 17.5 thousand cacao seeds potentially apt to form plantlets, during a relative short production cycle. A hypothetical efficiency rate of 3% in fruit formation, that is, 100 flowers produced on each adult trunk, would result, on average, in 3 healthy mature fruits, and a mean of 40 ovules/flower, and that the population in question would produce at least 668 thousand ovules/year. These conditions would establish a chance of 1 in 38 that each ovule could turn into a healthy and mature seed. Considering furthermore that it takes 5 to 6 months to form the fruit, that the fructification period of natural T. cacao populations normally occurs in January through March, and that green fruits in an advanced development stage are also used for food by the dispersal agents/predators of the seeds, the wild cacao populations consequently constitute food niches, during at least 7 to 8 months/year, for frugivorous animals and seed predators. Besides, the seed production potential of such a population can be quite expressive in certain years, creating many possibilities to escape from predation. However, no report on explorative collections in Brazil (Almeida et al., 1995) mentions the presence of germinating seeds or of growing cacao plantlets in the natural explored populations or in their immediate surroundings, although such collections are generally done two to four months after the beginning of fruit maturation of the wild cacao trees. These reports record the occurrence of green and mature fruit damaged by wild animals. In collections carried out in Ecuador, Allen & Lass, 1983, reported the occasional occurrence of cacao plantlets in natural populations, while Sánchez & Jaffé, 1992, in collections realized in Venezuela, observed only plants formed from vegetative sprouts. The perennial nature of the cacao tree, characterized by the formation of clumps in natural populations with overlapping generations, reduces the chances for the growth of plantlets close by. Nevertheless, the hypothesis of the formation of a plantlet bank far away from the mother tree, which would allow a greater chance of survival of the same owing to a greater light availability and lower concentration of seed predating animals, compared to the conditions of the habitats of natural T. cacao populations, should be considered. This hypothesis is supported by the modus operandi of the black capped monkey (Cebus apella), one of the main dispersal agents of cacao seeds, which usually carries the fruits away to eat them elsewhere, as observed by Bates, 1979, which can give rise to cacao colonization in new habitats. In this context, it becomes evident that the action of the predator/dispersal agents is a broad one and that their diet seems to include as much seeds and fruit shells, as well as plantlets. It also becomes clear that wild cacao trees invest a maximum of energy into reproduction in order to produce a great seed quantity as a reproductive strategy. This is a result of the natural selection process, to compensate for the action of the predator/disperser agents and, consequently, the small chance of seed germination and plantlet survival in the forest environment. Finally, the incompatibility system inherent in T. cacao is also an important component of its life cycle, since it causes the outcrossing between individuals in the population that have different S alleles, and makes the formation of new recombinants possible, hence increasing the chances of the adaptation of plantlets to the habitats undergoing colonization. In this sense, the self-incompatible cacao trees seem to have a certain competitive advantage, compared to the self-compatible, since such plants are, on average, taller in commercial plantations (Lockwood, 1977). Natural dispersal agents: The action of spread agents such as man, water, and wild animals certainly have a predominant role in the formation of natural T. cacao populations (Bartley, unpublished), as well as in other species of the genus. Water must have participated significantly in the formation of river populations in virtue of its capacity to transport fruits, seeds, and plant fragments over great distances. On the uplands, water must have influence on the seeds’ spread along the banks of streams. Another possibility of spread by water is the phenomena called “terras caídas”, which are mud slides of river banks of the Amazon caused by floods, which would cause the transport of cacao trees from one place to another. As an illustration of the species´ spread along rivers it can be assumed that, in rivers like the Purus, in which the mean water flow is 2.57 m3/s in the months of floods (December through to March) which also coincide with the fruit maturation; ripe cacao fruits that fell into the water, can travel uninterruptedly over 3,210 km (extension of its course) in only 15 days. This time is short enough to preserve the viability of the seeds within the fruits. Along the banks of the Purus, cacao populations occur relatively abundantly. It is therefore quite clear that gene migration over long distances is possible under these circumstances and that, therefore, it is natural that the wild populations of the intermediate and inferior water courses of certain rivers presented a certain likeliness in relation to those of the upper water courses. Wild animals, such as: monkeys, rodents (mice, squirrels, ´caxinguelês´, and others) and some bird species (parrots, budgerigars, ´jandaias´, owls, and others) probably contribute to the formation of populations on the uplands, as well as on the river banks, and would contribute to the transport of fruits and seeds over a long distance to new colonization sites. Cacao fruits unite certain characteristics, such as: greater size, shiny shell, yellow when ripe, and a mucilaginous pulp with sweet taste that represent an edible attraction for specific dispersal agents. One of the first reports on the action of seed spread agents in natural T. cacao populations in the Amazon was written by Bates, in 1850, who explored the intermediate Amazon. Two monkey species participate in particular: the black capped monkey, which usually carries away a lot more fruit than it is able to eat and the common squirrel monkey (Saimiri sciureus), which normally eats its food where it finds it (Bates, 1979). Huber, 1909, presumed that the monkeys were the first or maybe the only dispersers of plants whose fruit have a somewhat resistant pericarp and seeds covered in a sweet pulp, for instance the species of Theobroma, Herrania e Inga. Nevertheless, man is considered the most efficient dispersal agent. The distribution pattern of the species is often linked to the pattern of human occupation or movements in each region. It is a well-known fact that certain autochthon communities used to cultivate native perennial species, including cacao. Their participation in this process should have occurred under three situations: i). The establishment of home orchards or plantations with seeds of wild trees and the use the product for food or medicine, on the market, to attract prey, and/or as firewood. Examples cited are the Amazon natives (Caboclos) (Frechione et al., 1989), the ancient settlers, and diverse indigenous communities in the Amazon, such as: the Kayapó, in Pará (Posey, 1984 and 1985), the Tikúna, along the river Solimões (Kerr & Clement, 1980), the Tukána, along the river Negro (Kerr & Clement, 1980), the Mundurukú, on the river Tapajós (Frikel, 1959), the Waiãpi, along the rivers Araguari and Jari (Gallois, 1981), the Iyearmâmi, of the mountain range between the Amazon and Orinoco (Sánchez & Jaffé, 1992), the Ka’apor, in the North of the State of Maranhão (Balée & Gély, 1989) and the Chácobo, on the river Ivón, an affluent to the Beni (Boom, 1989). According to Bartley (unpublished), the spread of the cacao trees along the rivers presented characteristics that indicate human intervention. The presence of cacao trees in regions very distant from their center of origin such as the upper Anajás, Marajó Island, in the State of Pará, is attributed to Indians that lived in the region in the past (Balée, 1989). Another result of the ancient cultivation is the presence of cacao trees on the right banks of the river Gurupi, near the river Cacaual and the confluence to the Gurupi-Mirim (Marques, 1970), as well as the large wild populations of high density, as observed in Rondônia (Bartley, 1977 and Lisboa, 1990). This form of seed spread probably established the first steps for the domestication of T. cacao in a given or in several regions of its geographic distribution. ii). The sucking of the pulp of fresh wild cacao seeds as a seasonal subsistence food and their random distribution in the forest or along trails (Sánchez et al., 1989 and Leal et al., 1998). This is a common habit of several tribes that dwell in the Amazon, used during hunting and fishing expeditions, long-lasting harvesting missions, and visits between communities, and; iii). The use of the fresh seeds of wild cacao to prepare fermented beverages (“cocoa wine”) and for food (Sánchez et al., 1989) and piling them up in the surroundings of their camps, where the seeds would quickly germinate and grow, owing to the greater availability of light. It is known that certain ethnic groups that live in the Amazon migrate frequently to renew their fields because of the low natural soil fertility, and come back to the original territory many years later. Such changes normally occur when the plantations lie over 1 or 2 km away from the village. Furthermore, villages also move on because of the rivalry among neighbouring tribes. These migration processes of autochthon communities probably led to the establishment of T. cacao populations in a given region, be it intentionally, or be it simply by the germination of seeds thrown away after use. To recognize populations formed under any of these circumstances, decades after the area was abandoned, is difficult, due to the fast recovery of the plant cover and the establishment of plants in a disorganized form, assuming spread characteristics by means of other agents. |
| Wild and domesticated populations and Variability of natural populations. | Return To Table of Contents |
|---|
| Wild and domesticated populations: In this chapter, the term “domestication” is used for the evolutionary process under human selection, in which the wild state changes slowly and gradually to the cultivated condition. The genuinely wild condition or that with some degree of domestication of T. cacao populations in their Brazilian Amazon habitat is difficult or even impossible to recognize, due to the inexistence of characters that would allow such a differentiation. A simple observation of plants considering only the vigour, production aspect, and morphologic traits of the flower, leaf, fruit and seed does not permit a distinction between primitive and semi-domesticated types. Besides, this kind of problem had already been mentioned by Huber, 1901, who studied the fruit trees that occur in the State of Pará. He believed that all the cacao between the Amazon estuary and Óbidos and Santarém, Pará, one of the oldest and most traditional cacao producing zones of the Brazilian Amazon, was no longer wild. Later, Baker et al., 1954 reported the same difficulty when collecting cacao in Colombia. A similar situation was mentioned by Bartley et al., 1988 when describing collections obtained in populations in the municipality of Alenquer, Pará, for which the authentic condition of these collections in nature, whether wild or cultivated, cannot be assured. Although some historical reports say that cacao was well known and used in Mexico and Central America before the arrival of the Spanish (Cuatrecasas, 1964 and León, unpublished.), there is no evidence that it had been cultivated in the Amazon region before this period. This is why the species is still found undergoing domestication in the Amazon, since, in evolutionary terms, a few centuries of cultivation practices and plant selection are not sufficient time to bring forth notable genetic alterations, apparent to the naked eye. Examples of alterations that occur in wild plants during the domestication process are cited by Ford-Lloyd & Jackson, 1986 and Heiser, 1988. Such alterations basically constitute: i) the loss of natural dispersal mechanisms; ii) uniformity and speed in seed germination; iii) simultaneous ripening of seeds or fruits; iv) loss of mechanical protection means; v) change in fruit or seed colour; vi) loss of physical properties, and vii) alteration in the reproductive biology. The larger size of cacao fruits and seeds that had been selected by humans over various generations is perhaps the only indicative aspect of the occurrence of domestication in the act of harvest. For example, Spix and Martius, when exploring the Amazon between 1819 and 1820, observed that the fruit of the “bravo” (wild) cacao tree was always heavier and more bitter compared to the “manso” (cultivated) tree, besides generally having smaller seeds (Spix & Martius, 1981). Huber, 1906, who was cited by Ducke, 1940, also verified that “the fruits of the wild cacao tree were a little smaller and had a thinner and softer husk than the cultivated trees; they had a lower number of seeds, whose form and dimensions differ little from the shape of those cultivated most frequently in the Lower Amazon”. At the beginning of the 20th century, the presence of spontaneous cacao trees was verified on the plains of the upper Purus, together with individuals of T. subincanum, T. speciosum, T. microcarpum and T. sylvestre. This fact let Huber, 1901, to assume that the condition of T. cacao in the region was genuinely wild. Ducke, 1940, in turn, believed that the cacao trees in the region of the river Branco, northeast of Óbidos, in the intermediate course of the Tapajós, the lower Trombetas, in the east of the Lago Salgado, and along the lower Javari and Solimões near Tabatinga, were wild. Furthermore, Ducke, 1953, reported that along the Solimões, the frequency of T. cacao increased upstream, and that the area of spontaneous cacao spread out to the States of Rondônia and Acre. A further indirect indication of the anthropic influence on wild T. cacao populations could be the occurrence of a great plant density in the same area, which indicated that in the past there had been a deliberate influence on its formation. This was the situation observed in harvests realized in the uplands in the State of Rondônia, in the Mirante da Serra region, where thousands of wild cacao trees were found within an area of about 10 ha (Almeida, 1983). Besides, the city Cacoal, in Rondônia, 480 km southeast of Porto Velho, was so named because of the great quantity of wild cacao found there by General Cândido Mariyear da Silva Rondon, at the beginning of the 20th century, when telegraph lines were being installed between Mato Grosso and the Amazon (Silva, 1984). Another important factor to be considered in an area to be studied is the existence of traces of human occupation, such as pottery fragments and tools, among others. Such remnants were found in the near proximity of some wild cacao populations in the region of Ariquemes, Rondônia (Almeida, 1982), suggesting a probable link between the primitive indigenous communities and the use, spread, and establishment of T. cacao populations. In Central America there is archaeological evidence of the cultivation practice and use of cacao seeds by the Mayan civilisation (Mooleedhar et al., 1995), whose “cenotes” were recently discovered on the Yucatan peninsula in Mexico, having wild cacao plants that are considered survivors of the ancestral cultivar (de la Cruz et al., 1995). Although cacao cultivation in the Amazon dates back to the 17th century (Alden, 1974) and the ancient settlers and Amazon natives as well as the Indians have certainly participated, in some form, in the manipulation of the species, the existence of natural T. cacao populations without human interference is also a possibility. According to historic documents, such populations would lie at a considerable distance from the Amazon river bed and its tributaries Xingu, Tapajós, Madeira, Negro, and Japurá, as well as from Tocantins; areas which did not come under European colonial influence in the 17th and 18th century (Ribeiro, 1983). It is possible that there are still genuinely wild T. cacao populations in remote regions of the States of Amazonas, Rondônia, and Acre, as well as in the valleys of the watersheds having natural obstacles, such as waterfalls or torrents - a fact which has surely impaired the explorative actions of the ancient colonizers, as for example the river Teles Pires, affluent of the Tapajós. Beliefs and myths of autochthon communities have certainly exerted an additional influence for the isolation of man in certain regions, maintaining the integrity of genuinely wild cacao populations. This might have been the case with the populations along the river Yaloupi, tributary to the Oiapoque, in French Guyana, owing to a local belief in the presence of dangerous and different Indians (“white” or “long-eared”) in the referred region, as Lachenaud et al., 1997, reported. Yet the latest ethno-ecological investigations reveal that certain indigenous communities of the Amazon are accustomed to manage and explore different ecological and microclimatic zones by the establishment of different botanic species of interest, making it impossible to assess the real impact of this influence on the forest and in the field (Posey, 1985). Such areas are considered promising ecological niches for human life, as well as important repositories of semi-domesticated or manipulated plants, as for example T. cacao, T. grandiflorum, and T. speciosum. It is therefore clear that a thorough prior investigation of the human occupation of the area in question is necessary to determine the exact biological condition (wild, semi-domesticated, or domesticated) of the T. cacao populations of a given region. Variability of natural populations: The variability potential in the natural cacao populations in the Amazon only became known after the explorative collections undertaken by Pound, 1938, in the region of the Upper Amazon, especially in the tributaries that flow through Peru and Colombia. In the Brazilian Amazon, the existence of a broad phenotypic variability in the wild cacao populations became evident when CEPLAC carried out the first scientific expedition to this region in 1965 (Vello & Medeiros, 1965). Later, this great variability was confirmed by several explorative collections realized in different years in the region. This information was summarized by Barriga et al., 1985 e Almeida et al., 1987 and 1995. More recently, the use of RFLP and RAPD molecular marker analyses has allowed a better evaluation of the genetic variability inherent in some genotypes collected in the Brazilian Amazon (N’Goran et al., 1994). The evaluation of some agronomic traits in a germplasm bank has also revealed an expressive genetic variability in the Amazon populations and the possibility to use this variability for the improvement of the species (Silva, 1999). It is certain that the variability detected in the exploited Brazilian populations does not encompass the entire range of natural variation in T. cacao, since, of the three racial groups Criollo, Trinitarian, and Amazonian Forasteiro, only the latter occurs in the Amazon. However, this variability is widely known for traits such as the presence or absence of anthocyanin pigments in different flowers, fruits, seed, and leaf parts as well as for fruit and seed characteristics, expressed in different magnitudes, as much inter as well as intra-populational. Despite such considerations, the scientific knowledge on these populations is yet incomplete, due to the low number of analyzed traits and because the exploratory expeditions in Brazil have reached only around 20% of the entire Brazilian Amazon area. The predominant traits observed in 13 cacao populations of the Brazilian Amazon are displayed in the Tables 4.1, 4.2, and 4.3. Some populations have certain traits in common, mainly when they co-exist in the same geographic region without natural barriers, so phylogenetic relationships among them can be established and conclusions drawn on evolutionary aspects of the species. In this case, the condition of a high degree of outcrossing of the species contributes to a reduction of the differences among the populations by the gene flow among adjacent populations. For example, one can cite the upland populations exploited in Rondônia, between Ariquemes and Cacoal (Table 4, regions 8, 9, and 10), lat 9' 30" N and 11' 47" S and long 61' 7" E and 63' 30" W. These populations are very similar to each other in relation to certain traits, giving basis to the hypothesis that they belong to a common gene pool. In other cases, some characters are exclusive of specific populations, due to the occurrence of a probable geographic or ecological isolation, particularly for the populations that grow along the banks of certain rivers. Under these circumstances, a predominant fruit type and a close association between the organization of the variability and the watershed is verified, a fact that has been corroborated recently (Dias et al., 2002). These authors tested and proved the complete differentiation between cacao progenies from four randomly chosen Amazon Basins (Japurá, Ji-Paraná, Purus, and Amazon). Probably, the isolation of populations in valleys separated by mountain chains or in specific watersheds has favoured divergent evolutions, owing to the gene fixation during the evolutionary history of T. cacao. As an example, the wild cacao from the river Içá (Table 4, region 5), affluent of the rivers Solimões and Amazon, can be mentioned; it is regionally known as “blue cacao”, owing to the bluish fruit colour, or the populations of the river Javari and its affluents Curuçá and Itaquaí and subaffluent Quixito (Table 4, region 7), which also belong to the Solimões watershed. Its fruits mainly have a thin and brittle husk and a mammiform apex (Machado, unpublished). Such traits are not common in the cultivated gene pool. The watersheds lie between lat 2' N and 5' S and long 68' E and 72' W. Another example of a probable occurrence of geographic isolation are the wild T. cacao populations found in French Guyana. Their morphologic traits are considered different from those observed in the subgroups known as the Upper and Lower Amazonian, which belong to the racial group Amazon Forasteiro (Lachenaud et al., 1997). The authors argue that the rivers of Guyana do not belong to the Amazon watershed and that the Guyana plateau is separated from the Amazon Basin by several mountain ridges, which reach a height of 3,015 m, such as the Pico da Neblina in Brazil, isolating the region geographically. This isolation probably occurred between 35.000 and 2.500 years ago, during the Pleistocene epoch. The evidence convinced the authors that there is no connection at all between the wild cacao from French Guyana and that of the Upper and Lower Amazon. Biometric data obtained during the collection missions (fruit weight and individual humid seed weight) of the cacao populations in the Brazilian Amazon indicate that the most expressive mean values were recorded for cacao trees sampled on the rivers Solimões, in the immediate vicinity of the Tefé and Içá, and Amazon (regions 3 and 5, respectively, Table 4). The variability in fruit weight was more expressive in samples from the vicinity of Tefé and the river Japurá (regions 3 and 4, Table 4). It is known that particular indigenous communities in the Amazon (Kerr & Clement, 1980), as well as riverine settlers usually apply mass selection techniques to obtain seeds of fruit trees and other edible species, which they use for the plantation of new areas. It is also known that the cacao fruits are easily identified in the forest interior when ripe, by the contrast of the yellow husk against the green vegetation. These fruits are extremely appealing for the autochthon communities due to their size, coloration, shape, and, in some cases, brightness. In view of these facts, it is assumed that humans have participated in the spread of T. cacao in these regions, giving preference to types with certain botanical-agronomic attributes and that this behaviour contributed to increase the diversity and the sampled fruit and seed size. Such information illustrates that the wild cacao populations probably evolved under two situations: i) restricted to specific environmental niches, due to geographic or ecological isolation and ii) spread over a vast geographic area, due to the absence of natural barriers and the active participation of natural dispersal agents, such as humans, water, and wild animals. Certainly both situations favour the appearance of ecotypes. Later evaluations under appropriate experimental conditions actually provide sounder botanical-agronomic information on the genotypes collected. However, a preliminary evaluation based on the in situ observations and data analysis realized during the collection should make some knowledge regarding the inter and intra-populational variability available, which is highly important for the genetic improvement programme of cacao. The use of analyses with isoenzymatic markers and DNA will supply, among other objectives, a more profound knowledge on the organization of the genetic diversity in sampled populations, which is also of great importance to establish strategies of genetic conservation in active germplasm banks, besides bringing forth information on the evolutionary relationships between the different racial T. cacao groups. |
| Chapter 5. Genetic resources. C.M.V.C. Almeida & L.A.S. Dias. | Return To Table of Contents |
|---|
| Contents: Introduction; Germplasm collection; Mission planning, strategies and adopted procedures (Sample collection, Mission planning, Order of priority of collection regions, Collections in specific watersheds, Timing of expeditions and Collection of shoots and fruits, Expedition planning - strategies and adopted procedures, Selective sampling and flow chart of the activities). Expedition costs, Establishment of germplasm banks; Germplasm evaluation and characterization, Morpho-agronomic evaluation and characterization, Molecular evaluation and characterization; Duplication of the genetic stock. Summary: Genetic resources represent the pool of potential genetic variability for improvement programmes of cultivated species. In turn, the existence of variation is a fundamental requirement for the improvement and evolution in the species. The Amazon’s genetic resources and their use are still little explored. It is the region with the most wide-spread area of wild cacao populations of the racial group the Amazon Forastero, the most cultivated worldwide, representing, therefore, an important source of new genes. It is necessary to emphasize that such resources are not renewable, so their conservation is of fundamental importance. In this Chapter, the steps needed for the planning and adequate organization of expeditions to collect genetic cacao resources are presented and discussed, and methods, procedures and strategies considered. The difficulties and restrictions inherent to these activities in the Amazon are discussed, as well as the costs of land and river expeditions. The actual knowledge status on the implantation, evaluation, and characterization of a germplasm bank in form of a live plant collection is presented. Lastly, the duplication of the actual genetic stock established in the Amazon is discussed, as a measure to protect and conserve the gene pool saved so far. |
| Introduction and Germplasm collection. | Return To Table of Contents |
|---|
| Introduction: Natural genetic resources provide the necessary raw material to create new cultivars that are more productive, more adapted to the crop’s region, and more resistant to economically important pests and diseases. Under these premises, the resources contribute to combat hunger and malnutrition in different regions of the planet, besides offering humankind an improved quality of life. In other words, genetic resources represent the pool of potentially useful genetic variability for the improvement programmes of cultivated species. Specialists unanimously agree that the Amazon region is still little explored regarding the use of its genetic resources and the immense plant richness that this region holds, composed of forest essences, medicinal plants, forage, and ornamental plants, resin, root, and tubercle producing plants, and plants that produce essential, combustible, and edible oils, among others. The Europeans first encountered cacao in the Americas in 1502. In 1526, in Veracruz, Mexico, soldiers of the Spanish conqueror Cortés saw cacao trees for the first time (León, 1959). Later, at the end of the 16th century and beginning of the 17th, cacao was already part of the trade of spices, also called “drugs of the sertão” (semi-arid area), together with Italian smilax, the lipstick tree, cloves, cinnamon, false anil and vanilla, among others. This commerce was one objective of colonization in the Amazon (Oliveira, 1983), since these products were of strategic value in the explorer countries. When Acuña, 1865, participated in the return expedition of the armed naval fleet of Pedro Teixeira from the city S. Francisco de Quito, in Peru, to Belém, in Pará, Brazil in 1639, he mentioned large numbers of wild cacao trees, in abundant fructification, on the riverbanks of the Amazon. This fact let Acuña to suggest this crop as very promising for the region, together with tobacco and sugarcane cultivation, and the exploitation of timber; “undoubtedly enough to make one and many kingdoms wealthy”. Cacao’s area of distribution lies in the rainforests between southern Mexico and the Amazon Basin, where representatives of the other 21 species of the genus are also found (Cuatrecasas, 1964). The area enclosed by southern México and the present border between Costa Rica and Panama is considered an important centre of ongoing domestication, diversification, and evolution of the species, since cacao had been cultivated there ever since pre-Colombian times by the ancient Mayas, who domesticated the first cultivars of the racial group Criollo (León, 1959; Bergmann, 1969 and Mooleedhar et al., 1995). In fact, the domestication of cacao is considered the most important contribution of pre-Columbian agriculture to the use of stimulating beverages (León, 1959; see also Chapter 3 on the Origin and spread of cacao). In this vast distribution area of the species, autochthon cacao populations are found threatened by genetic erosion. Among the reasons for the erosion are the depletion of forests by agro-pastoral activities, the construction of hydroelectric power plants and highways, and the exploitation of petrol and minerals, among others. Genetically representative samples of these populations must therefore be rescued (IBPGR, 1981; Barriga, 1982; López et al., 1984; Vera, 1984; Almeida, 1982 and 1983; Almeida et al., 1987; Almeida et al., 1995 and Barriga et al., 1985). Additionally, only a small part of the available genetic variability has been used in improvement programmes conducted in the different cacao-producing countries. These programmes are mostly based on the use of few genotypes per se or the hybrid parents, in the expectation of high short-term gains. This narrow genetic base, upon which breeding programmes are based, aggravates the crops’ genetic vulnerability and limits the progress in the control of the main diseases that attack cacao. Recapitulating, despite the fact that genetic diversity saved and stored in germplasm banks is small, in view of the species’ potential, only a small part of this diversity is effectively being used in the principal improvement programmes. The use of genetic resources from natural cacao populations in improvement programmes of the species aroused the interest of the international scientific community only after the explorative expeditions undertaken by Pound, 1938 and 1943) to the Amazon River and some of its tributaries in Peru and Ecuador, with the objective of identifying cacao resistant to witches’ broom (Crinipellis perniciosa). The notable genetic variability reported in these populations of the Upper Amazon and their potential value for a use in improvement programmes, along with the need for an amplified genetic base of the current improvement programmes in several countries, motivated various collection expeditions over the past five decades (Baker et al., 1954; Soria, 1970; Allen & Lass, 1983; Ocampo, 1984; López et al., 1984; Clement et al., 1988 and Lachenaud et al., 1997). In Brazil, CEPLAC has, since 1965, (Vello & Medeiros, 1965) carried out exploratory collections to the Brazilian Amazon in order to exploit the genetic variation in natural and cultivated populations of that region. It is well known that the region holds the most extensive area of wild cacao populations of the racial group Amazon Forastero, the most cultivated worldwide, representing, therefore, a large supply of new genes. Barriga et al., 1985, Almeida et al., 1987, and Almeida et al., 1995 gave an account of the results of these studies. The historical and analytical report presented by Almeida et al., 1995, after 29 years of germplasm collection expeditions in 36 watersheds of the Brazilian Amazon is worth mentioning. The collection of germplasm itself is only one of the stages in the formation of a genetic stock of the species, requiring further the establishment, evaluation, and characterization of the entire saved gene pool. The morpho-agronomical traits of these resources must be known in order to draw conclusions on the potential value for their use in improvement programmes. This Chapter’s objective is to compile the most relevant information on the collection methods, procedures, and strategies of genetic resources of cacao, as well as on their establishment, evaluation, and characterization in a germplasm bank. Moreover, it aims at an amplification of the knowledge on alternative experiences in genetic resources of cacao. It is predominantly based on the experience acquired in the Conservation Programme of Genetic resources of Cacao in the Brazilian Amazon developed by CEPLAC. Germplasm collection: The collections of genetic resources of T. cacao aim to save the natural or wild populations, the cultivated or semi-cultivated populations, related wild species, and primitive cultivars. Despite their potential for improvement, these resources are not adequately represented in germplasm banks. The gardens or orchards established by settlers living along the rivers or of farmers of settlement projects in the Amazon also represent a gene pool that has to be exploited. There, different plant species are protected against the genetic erosion process, T. cacao, T. grandiflorum, and T. speciosum being commonly found. It should be considered that collections offer the opportunity of delivering new genes that might provide genetic gain in future improvement programmes. However, a collector of genetic resources must explore the different alternatives and possibilities offered by the collection to their utmost, since the costs of such activities are extremely high and the difficulties of returning to a once explored region are high. In this Chapter, the following terms are used according to the concepts presented in the sequence: population - refers to a community of potentially inter-crossable individuals at a particular location (Mayr, 1977); wild or natural - a plant or population that was not affected by anthropic action in the form of domestication or systematic cultivation; cultivated - for a plant or population in systematic cultivation, originated from seeds of wild cacao; and semi-cultivated - for the previously defined condition, however, where little technology is used in the management. |
| Mission planning - strategies and adopted procedures (1). | Return To Table of Contents |
|---|
| Sample collection, Mission planning, Order of priority of collection regions, Collections in specific watersheds, Timing of expeditions and Collection of shoots and fruits. The undertaking of expeditions to collect genetic resources of cacao requires careful planning, in view of the high costs of the enterprise, the need for a quick removal of the collected material, and of the great geographic extension of the spread area of the target species. To warrant an expedition’s success, the planning requires the adoption of the methods, strategies, and procedures presented as follows. Sample collection: To ensure that any individual in a population has the same probability of being represented in the sample, mother plants, henceforth called matrixes, should be randomly selected as donors of seeds and/or shoots with buds. This is the principle of collecting representative samples of the existing variability in the explored populations, and has been adopted in explorative expeditions to the Brazilian Amazon since 1979 (Almeida et al., 1995). Observing this principle guarantees that the sample will contain common alleles of the population in their relative frequencies. The harmful effects of the genetic drift, which mean allele loss, are thus minimized. Random sampling and the minimization of the genetic drift in population genetics are treated under the concept of the effective population size (Ne), which measures the sample’s inherent degree of genetic representation. The expectation of the randomizing sampling and of the Ne is that a maximum of diversity be captured in a minimum amount of collected material. T. cacao is a monoecious perennial species, has hermaphrodite and allogamous flowers, and is pollinated by midges. The gamete union occurs randomly and, to a smaller extent, also includes self-fertilization. For species like this one, the basic expression of Ne for the quantification of the effect of genetic drift on finite samples is expressed by Crow & Kimura, 1970, as:  where n is the number of sampled individuals represented by seeds; K is the error of panmixia of the sampled matrixes, which is a negligible quantity in allogamous species; K is the mean number of gametes contributed to the sample by the matrixes, and; S2k is the variance of K. As N is the compound formed by all matrixes, we have K = (2n)/N. In turn, S2k, whose complete composition is presented by Vencovsky, 1978, has the parameters F (number of randomly sampled matrixes), and M (number of pollinator plants, whose male gametes participated in the generation of the n sampled individuals). This variance is also composed by the quantities u = F/N and v = M/N. By neglecting _ and substituting S2k by its components in expression (1), we obtain equation (2) below, derived by ]]Vencovsky, 1987]r551951]:  For the germplasm collection in natural populations of allogamous species such as cacao, M is an unknown quantity, large enough (as it is a denominator) to be neglected. Similarly, the quantities u and v are equal and tend to be zero, since F and M represent a very small fraction of the compound of the N plants of the species. Under these assumptions, a derivation of Ne in (2) establishes the expression (3):  Expression (3), however, ignores the control over the gametes contributed to the F matrixes to the generation of the n descendents. It could be assumed, for example, that 200 seeds (n = 200) had been harvested from 20 matrix cacao trees (F = 20). In this situation, Ne would be 57, in other words, the 200 sampled seeds represent, genetically, 57 matrixes of the original population. Note that Ne is defined in relation to the immediately preceding generation. Nevertheless, there is a situation where the gamete control is realized at harvest and an unbiased seed sample is collected for this purpose. In this situation, the adequate expression for Ne is (4):  where u = v = (aprox.) 0 and M negligible, the following expression is valid (5):  We now assume that the harvest of 200 seeds had been unbiased, taking 10 seeds from each of the 20 matrixes. For this situation, the substitution of these values in expression (5) results in Ne = 62, that is, an effective larger size for the same sample size. It is worth noticing that in the case of cacao, the harvest unit is the fruit that holds, on average, 30 seeds in natural population conditions. It is therefore easy to sample 10 normal seeds per matrix for an equitable sampling of the 20 matrixes. It is worth bearing in mind that it is the sample composed of F matrixes that must be random, and not the seeds collected from them. Under these premises, the strategy of unbiased sampling, also called gamete control, broadens the sample’s genetic representation. The numeric evaluation represented in Figure 5.1 points out that the largest possible number of matrixes must be sampled equitably. For example, 400 seeds harvested impartially from 80 random matrixes, result in Ne = 200. The simple fact of collecting the same sample in a non-equitable manner from the same 80 matrixes, causes a reduction of 11% in the effective size (Ne = 178) (Figure 5.1). Let’s suppose now that four times more seeds (1600) were collected impartially from only a fourth of all the matrixes, i.e., 20 matrixes, so the result would be Ne = 77, or 61.5% smaller. It is therefore strongly recommended that a smaller number of seeds be collected from a greater number of matrixes. How many matrixes are to be sampled, is a question that depends on factors as the percentage of the viability loss of the collected seeds, the quantity of accessions that can be preserved, and the rigour with which one wishes to maintain the alleles above a given minimum frequency in the collection. Literature defines, with some consensus, that Ne over 150 offers security for germplasm conversation. Again, under natural conditions, each population could be represented by an unbiased sample of 400 seeds collected from 80 randomly selected matrixes, with 5 seeds/matrix, which would guarantee Ne = 200. When collecting shoots, the Ne is equal to 1F, equivalent to the effective size for self-fertilized progenies. Therefore, if the same 80 matrixes were sampled by means of their buds, the sample would have the same genetic representation as 80 individuals of the population. Thus, only the genotype of the matrix can be collected by this kind of sample. On the contrary, the collection of seeds provides samples of the alleles and is therefore more efficient. In fact, since no trait can be efficiently selected under natural conditions in wild cacao trees, the collection of shoots becomes meaningless and is only justified in cases where fruits are rare and/or to complement the sampling by seeds. Sometimes it is necessary that the collection team, before proceeding to the collection itself, investigates the population of wild cacao trees, noting the geographic extension, the existence or absence of interconnected populations, and variations in the fruit, seed, flower, leaf etc. traits. This modus operandi adjusts the theoretical number of individuals to be sampled to the spread area under study. Obviously, in a next stage the matrixes are randomly chosen. However, it is not always possible to take a strictly randomized sample, since this could lead to the choice of matrixes with propagules (seeds and buds) in insufficient quantities for multiplication. Another great difficulty is the delimitation of the exact extension of natural cacao populations, mainly because these are continuously distributed over extensive areas. However, random sampling avoids collections biased by a breeder’s or botanist’s point of view. The breeder focuses his search on exceptional plants that could solve his improvement problem, while the botanist, frequently, prefers to collect a great quantity of botanic material to form a herbarium. It is imperative that the germplasm collection is looked at and practiced under a conservationist view and is free from any selection practice. Another relevant point in the sample process is the space that should be maintained between collection zones (Hawkes, 1980), which depends on the quantity of the environmental diversity (climate, soil type, vegetation, height, among others) in the region of exploration. An important factor for the sampling design, which deserves attention, is the peculiarity in reproductive biology of cacao in natural populations (see also Chapter 4, which deals with the ecology of natural populations). The first aspect refers to the possibility that plants or neighbouring clumps represent the same genotype. This is caused by tilted old trunks, from which new orthotropic branches grow. These have root primordia in their base and can form a taproot. This way, new plants that appear near the old one have the same genotype as the plant of origin. Another possibility could be the formation of a clump from seeds, which stick to each other or remain very near to one another in the spread process. In this case, a clump can represent two or more genotypes, as observed in populations of the upper Solimões, in the State of Amazonas (Barriga et al., 1985). Knowledge on both reproductive strategies of cacao must be taken into consideration in the sampling process, with the objective of maintaining the efficiency of the collection, in terms of the genetic representation of the populations and of the identity of the collected matrixes. Summing up, collectors of genetic resources should always bear in mind that the genetic diversity of natural populations is captured by the highest number of sampled matrixes per collection area, as long as these plants represent the existing diversity effectively. In each collection area, the greatest number of matrixes should be saved, in both vegetative and seed form, in the latter case, in an equitable or nearly equitable number of seeds per matrix. The second procedure, as shown, contributes to increase the value of Ne. We should bear in mind that vegetative sampling represents exclusively the matrix, while sampling via seed represents the matrix and its progeny. Mission planning: The regions of occurrence of cacao must be reconnoitred four to five months before the expedition, with the aim of compiling information of interest on the species of the genus Theobroma, especially of T. cacao. This information includes occurrence sites, quantity of trees in the region, expectation on genetic variability and fruiting period, possible support for the collector team, presence of official support institutions, means of transport, and the existing ways of access, conditions to evacuate the collected material, contraction of a local guide, and the existence of regional development plans that could threaten the integrity of the natural populations. This information will allow the planning of the necessary financial funds, the duration of the expedition, the sites for collection in a given agro-ecological zone, and the most appropriate areas for collection. Reports on expeditions organized by La Condamine, 1944 and Ferreira, 1983, in the 18th century; by Spruce, 1970, Coudreau, 1977, Bates, 1979, Wallace, 1979, Spix & Martius, 1981, Maw, 1989 and Moura, 1989, in the 19th century; and by Ducke, 1907; 1910; 1925; 1934; 1940; 1946; 1948; 1953, and Ducke & Black, 1954 and Huber, 1901 and 1909) in the 20th century, furnish a reserve of interesting information on the geographical spread of the species in the Amazon that could help choose the regions of potential interest for the collection of genetic resources of cacao. Order of priority of collection regions: The strategy to give priority to collection regions where the populations are in a process of genetic erosion depends as much on previous reconnaissance work as on the expedition itself. Previous information on governmental plans of regional development, for instance, must be gathered and taken into consideration. Ideally collection expeditions should precede the establishment of economic activities that expose the natural T. cacao populations of a given region to risks. Large agro-pastoral, mining, and settlement projects and the construction of highways or hydroelectric power plants are some of the activities that endanger the integrity of the natural ecosystems. Regions with a potential importance in geographic extension and expressive genetic variability of local populations should be granted priority as targets of collection expeditions. For this purpose, information should be recovered from old documents or reports on expeditions realized in the past, above all in relation to the geographic extension of the occurrence areas of the species and the expectation of the genetic variability. Collections in specific watersheds: It is believed that the natural cacao populations are restricted to particular valleys (Bartley, 1963). This point of view is backed up by the existence of populations that present different traits, associated to specific watersheds (Pound, 1938). This underlying belief and the fact that the complex Amazon watershed is basically the only access to certain regions, above all the most distant ones, has led collection expeditions to concentrate on specific basins where the Hiléia Amazon occurs. Such expeditions are predominantly carried out along the banks of navigable rivers, covering a belt of about 600 m or 1 km at the most, according to the presence of natural cacao populations. However, where there are passable roads or a well-ramified road net, expeditions are also carried out on the uplands, a long way from the riverbanks. This strategy has facilitated the analysis of the inter and intra-population diversity, the study of the evolutionary relationships among the prevalent cacao types found in each river system, and has also confirmed, in practice, the association between the distribution of the variability and the watershed. Lately, although for the moment in preliminary form, the hypothesis on the organization of the genetic variation of the cacao populations in basins was tested (Dias et al., 2002). Based on 15 fruit and seed traits, 64 progenies from open pollination were evaluated, represented by five cacao trees each, and randomly collected in four Brazilian Amazon basins (intermediate Amazonas, Japurá, Ji-Paraná, and Purús). The results of the univariate hierarchal variance analyses and the multivariate analysis corroborated the differentiation of populations in basins. This fact validates the germplasm collection process of cacao developed so far in the region. Furthermore, since the greatest part of the variation is concentrated in plants within basins and among basins, it becomes clear that the sampling procedure that optimizes the cacao germplasm collection is the sampling of the largest possible number of cacao trees, based on few progenies, yet collected from the greatest possible number of basins. It is worth emphasizing that the studies with isoenzymatic markers (Ronning & Schnell, 1994) and RAPD (Russel et al., 1993 and Marita, 1998) clearly discriminate the wild cacao populations, as much as in respect to the geographic origin as in the racial groups. Naturally, when speaking of germplasm collection strategies, there are many variations on the same theme. For example, the project LCTAP, carried out in the Equatorial Amazon in partnership with the national INIAP, adopted the strategy of a systematic collection in geographic strata that consisted of squares of half a degree latitude by half a degree longitude, in other words, in squares of 55 x 55 km (Allen, 1984). The author vindicates his sampling criterion with the expectation of covering the entire regional variability, and making it possible that each square would be visited at least once. Timing of expeditions: The expeditions realized during the peak fruiting period of the wild cacao trees offer the following advantages: i) a clearer idea of the existing variability in the natural exploited populations and ii) availability of the seed type of multiplication. This way results in a better genetic representation of the populations, greater durability of the collected material, and a greater efficiency in propagation. Yet, it must be remembered that the fruiting period of wild cacao can last around five months a year. If, on one hand, the elasticity of this fruiting period enables a more flexible realization of the expeditions and makes them practically possible, while on the other hand it impedes that the seed sampling would represent all the seeds produced by the plant; according to Schattan, 1984, this disagrees with one of the principles of the sampling theory. The natural T. cacao populations are found under very varied environmental conditions and certainly, a number of factors interfere in the cacao’s reproductive biology. This impairs the prediction of behaviour with regards to flowering, fruiting, and the emission of new leaves, making such phenomena variable between and within years. Over and again it should be remembered that these unforeseeable reproductive phenomena represent an adaptation strategy of tree species of tropical forests, in a way that gives the pollinating species and dispersal agents the opportunity to fulfil their roles. Nonetheless, the information gathered on reproductive biology of cacao in the Brazilian Amazon allows the inference that the fruiting period normally occurs between January and May, with fruit production peaks in February and March. On this occasion, it becomes easy to locate populations of wild cacao in the interior of the primary forest by the outstanding yellow-golden colour of the ripe fruits. In cultivated or semi-cultivated plantations on the wetlands and high islands of the lower and intermediate Amazon, grown from seeds of wild cacao, fruiting lasts from December/January to July. It is most intensive from April/May to June (Alvim, 1971 and Nascimento et al., 1975). Le Cointe, 1934 reports that in the Lower Amazon there are two fruiting periods: in January or February, which is called the summer or monkey’s harvest, and from April to August, the so-called main harvest. The following factors, among others, must have the greatest influence on the alterations in the flowering and fruiting periods, anticipating or delaying the occurrence of the same: i) rainfall intensity and distribution during the year, for the populations on the uplands; ii) intensity of flooding, for the populations in the wetlands and high islands of the lower and intermediate Amazon, and iii) intensity of the highs and lows of the tides for populations in the wetlands and high islands of the regions in Tocantins and the Amazon delta. Collection of shoots and fruits: Collection expeditions of genetic resources can gather shoots or fruits cacao or both. However, collecting both from the same matrix ensures the population’s representation in the sample. Once again, this strategy makes it possible to evaluate the per se potential of the matrixes and of its respective progenies simultaneously. Genes and genotypes are sampled simultaneously. While the shoots sample exclusively the matrix, the fruits sample the allelic contribution provided by the ample pollen of the plants near the matrix besides, of course, of the alleles of the matrix itself. Due to the high losses that occur through the vegetative propagation, the additional fruit collection assures the representation of the exploited populations (Table 5.1). In fact, the fruit from free pollination is the most valuable sample unit. This is especially true for populations in remote regions, of difficult access, where it may take up to seven days from the removal of the collected material to its establishment in the germplasm bank. In cases of a high incidence of witches’ broom and weevil borers (Conotrachelus sp.) in ripe fruits, the strategy of collecting half-ripe fruits can also be used because the seeds are already physiologically ready for germination. The greater efficiency of seed propagation can be seen in Table 5.1. Despite showing an efficacy of 2.5 to 6.2 more in relation to vegetative multiplication, the values reached in some years may be low due to the occurrence of witches’ broom and of borers in the collected fruit. In turn, the low indices obtained by vegetative multiplication are caused, especially, by the precarious physiological state of the available buds, owing to the excess shade in the under-story of the forest. The low indices may also be a consequence of the longer duration from the collection of the material to its multiplication. In the cases in question (Table 5.1) the time elapsed varied from 1 to 4 days for the collections carried out in Rondônia, since the reception base for the collected material in Ouro Preto do Oeste, Rondônia, lies within a radius of less than 150 km from the agro-ecological zones explored. For collections in the Amazon region of Tefé, this period varied from 2 to 7 days, because the reception base in Marituba, Pará, 2000 km away. Nevertheless, the expedition to the Amazonas achieved a better performance in the multiplication of botanic materials, owing to the better phytosanitary state of the cacao river populations, caused by the half-cultivated state of these plantations, resulting in a greater production of healthy fruits and buds in a good physiological state. |
| Mission planning - strategies and adopted procedures (2). | Return To Table of Contents |
|---|
| Selective sampling and Flow chart of the activities. Selective sampling: Selective sampling, only justified in cultivated or semi-cultivated plantations, is realized by collecting supposedly elite matrixes, which present agronomical attributes of interest or of rare occurrence. Such matrixes are previously identified by the farmer or by rural extension technicians during field surveys on the plantations. The adoption of this strategy must be carefully thought over, since the normally selected traits, for example bean yield and its components, have a polygenic heritability; this means that the environment has a great influence on their expression, and could hide genotypic differences (see Chapter 6 that deals with selection strategies and methods). The high number of fruits on a given cacao tree, at the moment of the collection, might be a consequence of the favourable environmental conditions where it is found and the random interaction genotype x year x locality, and not the genotypic superiority of the plant itself. The occurrence of plants that are apparently free of witches’ broom symptoms in a given infected cacao population could mean nothing more than an escape, and not genetic resistance or immunity. The infection rate also increases with the age of the plant, so apparently symptom-free plants at a certain age could manifest the disease later (Bartley, 1978a and 1978b). The selection of cacao trees for resistance to C. perniciosa can lead to the selection of plants presenting a vertical resistance to a given pathotype of C. perniciosa. This strategy cannot be recommended for the search for resistance sources of polygenic nature, which is realized to select the horizontal resistance, and not necessarily immunity (Kennedy, 1985), mainly because there are different pathotypes of the pathogen in question (Wheeler & Mepster, 1988). This type of selection can also provoke the elimination of genotypes of potential value for other traits by the simple fact of showing witches’ broom symptoms (Bartley, 1981). It is worth remembering that the efficiency of the phenotypic selection is based on the frequency with which the selected plants represent superior genotypes. It will be more efficient with the greater the estimate for the heritability of the target trait, which, in case of the trait for resistance to C. perniciosa, is still not well understood. It is worth emphasizing that with the use of accurate methods, such as the Best Linear Unbiased Prediction (BLUP) and Restricted Maximum Likelihood (REML), it is possible to practice plant selection by the genotypic value or even by the predicted genetic additive values (Resende & Dias, 2000), as also presented in Chapter 6. Over the past three years, CEPLAC in Rondônia, Brazil, has been putting efforts into identifying matrixes with field resistance to witches’ broom, on abandoned plantations with a high incidence of the disease. The matrixes with the lowest incidence are monitored during at least one year. In the 80’s, this methodology was used in a modified form by Andebrhan et al., 1991 with the same objective in the Amazon. Flow-chart of the activities: Time of collection: The duration of an expedition depends on the following factors: i) geographic extension of the exploration area; ii) facilities of access to wild cacao populations; iii) local climatic conditions, and iv) variability and quantity of the populations to be explored. In general, on expeditions carried out either by water as well as those by land, nearly half the programmed time is spent on getting to the exploration sites, preparation for the collection work itself, and packing and removal of the botanic material for the multiplication, among other activities. In some regions of the Brazilian Amazon explored for collections, it took nearly four days by motorboat to reach the furthest site. Under theses circumstances, the collection is realized on the way back, in order to improve the conditions for the preservation of the reproductive material. In overland collections, when possible, the support station must lie at a maximum distance of 40 km from the wild cacao populations to be explored, to avoid long daily transports that make this kind of activity onerous. The duration of the expedition must be long enough to: i) collect the greatest quantity of germplasm possible; ii) sample several distinct areas within each agro-ecological zone; iii) sample local variations of soil, climate, height, agricultural practices, and others, and iv) search for wild related species (Hawkes, 1980). In this context, short duration expeditions (10 to 15 days) allow a more intensive exploitation of a particular agro-ecological zone and can provide a greater efficiency in the preservation of the collected botanic material. These advantages depend on the availability of a nearby support station for a quick removal of the material. Long duration expeditions (20 to 30 days), however, are normally carried out to distant regions of difficult access and communication, where the efficiency in preservation of the botanic material is usually low. Collection team: The collection teams must be made up of the strictly lowest number of people needed to carry out the activities in order to minimize the costs, accelerate the operation, and simplify the transport and accommodation of the team. In expeditions over land, a team normally has three members: a graduate technician, who is the leader of the team and almost always a breeder; a technician with an intermediate degree, who is also the driver; and a rural worker. Preferentially, all three would have some experience in collections. On explorative collections by water, two more people are needed: a captain of the main boat and an auxiliary for transports in smaller boats (regionally known as “fliers”); these also assume the function of cooks for the team. Under any circumstances, it is necessary to contract a local guide, connoisseur of the area under study, and, in some cases, people of the local indigenous communities to facilitate getting certain important information for the success of the expedition. The participation of persons with experience in and knowledge of the region is essential. Preferably, the team leader should have knowledge on the life history of the species T. cacao (reproduction system, mating system, geographic distribution, spread mechanism of seeds and others) and its related species (the ecology of natural populations is discussed in Chapter 4). The leader should also have general information on the phyto-geography of the agro-ecological zone to be explored (vegetation type, climate, soil, agricultural systems in use, main rivers, and geographical features) and on the history and results of previous expeditions. Another relevant aspect for the success of the expedition is the training of the collection team to standardize the procedures, use appropriate sampling methods to collect as much genetic diversity as possible, and an adequate elaboration of field reports on the collected botanic material. Materials and equipment: Exploration expeditions of wild cacao to the Brazilian Amazon have shown that the items listed below represent the basic material for a collection in a given region: 1 - plastic bags to wrap up fruits and shoots; 2 - burlap sacks for field transports of the collected material; 3 - polystyrene boxes to pack and send the collected material; 4 - machetes, secateurs, pruning shears, and pocket-knives; 5 - field notebook to describe the botanic material, as in the model presented in Figure 5.2. Also include drawings of fruits of the other Theobroma species, to make field recognition easy; 6 - pens, pencils, eraser, adhesive labels and string; 7 - indelible ink markers and wax crayons; 8 - newspaper and paraffin wax to coat the shoots; 9 - compasses; 10 - binoculars; 11 - detailed political and ecological maps of the region; 12 - cameras; 13 - metallic tape measure; 14 - colour films; 15 - flashlights and batteries; 16 - basic first aid medicines. Vaccination against yellow fever and other tropical diseases may be required. The field information can be further enriched by the use of an altimeter and a GPS (Geo Positioning System), to localize the explored areas by means of the geographic coordinates obtained via satellite For further details on material and equipment see Hawkes, 1980. Field collection: In practical terms, on the overland expeditions to the settlement projects in Rondônia, shoots and fruit samples of 9 to 15 randomly selected matrixes were generally collected on each rural propriety visited. To ensure the representation of the sample and in view of the precarious conditions of the buds, besides the occurrence of witches’ broom and weevil borers in the fruits, and of losses during the greenhouse phase, around 12 to 15 shoots and two to four ripe fruits were collected from each matrix. This is considered enough to warrant samples with 50 to 80 buds and 60 to 120 seeds, respectively. This quantity of propagules is sufficient to ensure that each matrix is represented in the germplasm bank by at least five clones (grafts) and by a free pollination progeny composed of 25 cacao trees. After the expedition, around eight months pass from the nursery phase to planting in the germplasm bank. The shoots must have a diameter between 10 and 15 mm and be 15 to 20 cm long. The open pollination progenies must be obtained from seed samples of the fruits collected from each matrix. For the Criollos collected in Central America, the loss of collected material in the subsequent propagation and establishment stages reaches up to 30% (López et al., 1984). Certainly, these losses are a lot smaller than those verified in the collections of the Amazon Forasteros. The reason is simple: while the Criollos were domesticated by Mayas and Aztecs 3 thousand years ago, the Forasteros were only subjected to cultivation in the 18th century, explaining the greater rusticity of the latter group (see also Chapter 3). When a property had small sub-populations with less than 50 cacao trees, the sample was reduced to less than 8 plants. In such situations, another farm in the neighbourhood would be visited to complement the sample of that agro-ecological zone. According to the occurrence of the wild populations and/or to the physiographical diversity found, a mean interval of around 13 km was maintained between the collection zones, with a variation of amplitude of 5 to 25 km. In the diverse expeditions realized in the States Acre and Amazonas, basically by river, due to the smaller physiographical diversity and/or the existence of farms, ranches, and rubber plantations on the river banks, these collection zones were farther away (around 18 km) with an amplitude of 10 to 58 km. When collecting, main characteristics of the matrix and its habitat are identified and described (Figure 5.2). In the overland collections done in settlement projects, which have a road net that enables the access to any rural property, a team can collect 12 cacao accessions a day. This performance of the team depends on the availability of an appropriate vehicle, trafficable roads, favourable climate conditions, a support station near the occurrence area of the species, and precise information on the localization of wild cacao trees. It is thus possible to explore two properties a day, or around 500 ha. After a 30 day expedition with 15 effective collection days, having covered a distance of about 1.000 km, representative samples of the variability of the natural dispersed populations from an area of 5.000 km2 are obtained. In collections along navigable rivers, a collection team with a suitable boat, under favourable climate conditions, and with precise information on the occurrence of cacao, can explore two properties a day, (gardens, small holdings, rubber plantations, farms, etc) and rescue 11 accessions. After 30 days, considering 15 effective collection days, it is possible to cover a distance of 400 to 600 km of the river valley and obtain samples from about 1.200 km2. The term accession is used here to qualify any germplasm sample that represents the genetic variation of a given plant multiplied in vegetative or seed form. A return to a particular agro-ecological zone, considering the costs of such an enterprise, can only be justified in the case of a significant loss of sampled material, the realization of the collection at an inappropriate moment caused by uncontrollable factors, or a restricted sample collection in an area of expressive variability. Coding of the samples: The coding, preferentially alphameric, must be established for the sampled botanic material. In Brazil, the acronym of the exploited State is used, followed by a number in increasing sequential order as a provisional register in view of the losses at this stage. This provisory label is printed on aluminium tags, which are fixed to the matrixes selected in the field, for a future localization, if necessary; this tag is also used during the nursery phase, when accession losses occur and a reduction of the number of plants per accession. The definitive indexing occurs after the establishment of the accession in the field, using the acronym CAB – “Cacau da Amazônia Brasileira” - followed by a number with four digits, where the clones are represented by the numbers 0001 to 4999 and the progenies by numbers from 5000 upwards (Barriga et al., 1985). Preparation and packing of the sample: The collected fruits must be duly identified and packed in burlap sacks or polystyrene boxes to be sent to the reception basis. This procedure implies higher costs with plane fees and the possibility of introduction of new diseases and pests into the reception base. However, the procedure offers better conditions to realize a preliminary characterization of fruits and seeds, besides giving the collection team more time for the target activity. The perishable shoots must be wrapped in moistened newspaper and polystyrene boxes, taking care so as to protect the extremities with paraffin wax, to avoid dehydration. An excess of humidity in the packing leads to the appearance of saprophytic fungi and decay of the shoots. Report preparation: It is imperative that detailed notes on the exploration mission be done. Such accounts must be presented in form of a report and focus on: i) general aspects of the explored agro-ecological zone, such as climate and edaphic characteristics, vegetation type, main rivers, geographic incidences, and others. A description of each collection area must also be included in order to allow its future localization, if necessary, made easy by the use of GPS; ii) general aspects of the natural cacao population, such as: population quantity and density, spatial spread, action of seed dispersal agents, occurrence of diseases and pests, fruiting period, characteristics of fruits, leaves, flowers, and seeds, and others, and iii) report on the collected material and its description, based on the adaptation of the descriptors proposed for cacao (Engels et al., 1980). This is the way CEPLAC is working in the Amazon, using 24 traits, six of which refer to the plant, four to the flower, nine to the fruit, four to the seed, and one to the leaf (Figure 5.2). Pictures of the expedition should also be included, especially of ripe and green fruits of each accession, which should be appropriately identified. Additionally, the report can and should be enriched with biometric data on fruit and seeds and sketches of transversal and longitudinal sections of these sampled fruits. In these cases, to accelerate the collection and work under more appropriate conditions than those found in the field, these activities are preferably carried out at the support station. Another alternative would be to record the biometric data and sketches of the fruits at the reception area for the genetic material; this would imply additional costs with fruit transportation. Information such as the occurrence of indigenous communities in the past, traces of indigenous pottery, variant cacao types, among others, furnished by farmers or persons who regularly visit the collection area may be of great value for the researchers. It must be made clear, however, that more important than working out the report on the expedition, is its publication in a reputable scientific journal, so that all those interested can access it. This procedure gives the work of the collector team publicity, strengthens the support and sponsor institutions of the event, and opens the discussion among the scientific community on the methods and strategies applied. Not only the collector team, but also the promoting institutions gain credibility, augmenting the opportunities that national and international private and public organizations, which invest in successful groups and companies, will support future expeditions. All this information will be the preliminary documentation or the passport of the collection and will also provide knowledge regarding inter and intra-populational variability, of great importance for the future use of these genetic resources. |
| Difficulties and limitations and Mission costs. | Return To Table of Contents |
|---|
| Difficulties and limitations: The success of any collection programme of botanic material depends, naturally, on the knowledge of the life history of the species, the organization of the expedition, and the adopted strategies and procedures, among other factors. However, factors that cannot be manipulated by the organizers and executors of this activity also influence the success to a greater or lesser extent. In the cacao collections realized in the Amazon such factors are: 1). Coincidence between the periods of the expedition and of the fruiting peak of the wild cacao trees - as one of the adopted strategies is based on acquiring representative samples, it is of prime importance that these periods coincide, since fruits are the best indicators for variability in populations. Environmental factors can cause alterations in the expression of the fruiting peak and reduce the seed material available for the collector at the time of the expedition; 2). Difficulty of access to areas of species occurrence - the fruiting period of the wild cacao trees coincides with the rainy period of the region and with the greater water volume in the rivers, streams, natural channels, and ditches. In collections by water, such factors, normally, improve the navigation conditions, making the access to populations on the wetlands and high islands easier. However, certain rivers, especially the Purus and Juruá, bear serious risks for navigation through the phenomenon called “terras caídas” (mud slides) and the great quantity of wood dragged along by the current. In collections done in April, a quick discharge or flooding of the rivers, owing to sudden strong seasonal variations, can provoke grounding or other boat accidents. Another common obstacle for the river explorations are fallen trees, floating islands, and stones under the water surface. For collections over land, due to the precariousness of the roads, which is provoked by the heavy rainfalls in the region, any work carried out at this time of the year becomes extremely exhaustive and is subjected to unforeseen incidents, demanding appropriate vehicles; 3). Great extension of the areas of occurrence of wild cacao - the natural populations are found within clearly defined regions, but can also spread over vast areas such regions. In the latter case, long approach routes through the forest are common before one can proceed with the collection, mainly when there are several connected populations or subpopulations; 4). A low availability of botanic material for multiplication - despite the explorative collections being planned to coincide with the fruiting period, few healthy fruits are found in some regions, especially in the State of Rondônia because of predator agents and the occurrence of witches’ broom and weevil borers (Conotrachelus sp.). Years of high or low fruiting also occur, due to the interaction of genotypes with environments. Additionally, because of the dense shade in the lower forest stories, the vegetative buds frequently present inadequate physiological conditions for grafting. The flooded rivers can also cause the fruits of cacao trees in populations along the rivers to rot; 5). Difficulty of preserving the viability of the botanic material for multiplication - owing to the difficulties in the conservation of the viability of the buds and the recalcitrant nature of the seeds that have to be stored in highly humid conditions and under high temperatures to maintain their longevity and more frequent transport of the material to the site of introduction is necessary. In collections on the uplands, this is done every two or three days; in those by water, the transport can take over a week, mainly when regions of difficult access are being explored. One solution to this problem is the transport of the actual fruits, ensuring the conservation of the seeds over a longer period. However, this solution brings about additional difficulties in removing the fruits from the site of collection, especially when large populations are explored in the uplands, as well as higher airfreight expenses; 6). Imprecise or incorrect information on the occurrence of cacao - it is not rare that little or no botanic material can be saved in a day of collection, due to inaccurate or wrong information on the occurrence of the species. In recently colonized areas, this difficulty arises from the farmer’s poor knowledge about his own farm and on the species T. cacao. Farmers commonly mistake cacao for other Theobroma species. In older settlement projects, information gained two or three years ahead of the expedition can prove worthless, in consequence of the progressive deforestation. The trips are often long, costly, and inefficient in terms of saving material. Costs of the expedition: Generally, the costs of expeditions for collecting genetic resources of cacao are not mentioned in scientific articles on the subject, probably because they are extremely variable, mainly when such expeditions are realized in the Brazilian Amazon. These costs obviously depend on the circumstances under which the expeditions are realized, such as: duration time, number of participants, means of transport, the explored agro-ecological zone, and distance from the reception base of the collected material, among others. As an illustration, one can imagine an expedition to the interior of the State of Rondônia that would last two weeks, to smallholder settlement projects, with three participants: a researcher from another State, an agriculture technician, and a field worker stationed in Rondônia. It is furthermore assumed that the reception basis for the material would lie outside the region, in this case in Belém, Pará, and that the vehicle to be used was that of the institution, with only fuel and maintenance expenses. Under these circumstances, considering the cost of the daily expenses for three people, a plane ticket for one of them, maintenance of the vehicle, air freight of the collected material, the guide’s wages, and acquisition of expendable materials, the costs could add up to US$3535 of which the plane ticket and the daily expenses represent about 56%. In the case of an expedition by water, the expenses would increase by about 40%, compared to the inland expedition. We assume an expedition along the Amazon River banks, in a region close to Manaus, also of a fortnight and with five participants: a researcher from another State, an agriculture technician/and a local field worker, besides the captain of a ship and an assistant, both from the region, with a reception base in Belém and a ship of the support institution of the event or conceded by a regional organization. The expenses could run up to about US$4944. In this case, the maintenance of the ship and the crew’s wages amount to 31% of the final account. In both situations, assuming ten effective collection days in the two weeks of activity, it would be perfectly possible to save about 120 accessions. Under the previously defined circumstances, the cost of each saved accession would be US$29.00, on an inland expedition, and US$41.00 for those by water. Allen, 1984, mentions a total cost of US$120 thousand for 250 saved accessions in the three initial years of LCTAP, or nearly US$500.00 per collected accession. However, the account of the project LCTAP included employees’ wages and the documentation of the collection, which were not included in the examples cited. |
| Establishment of germplasm banks. Evaluation and characterization of germplasm: Evaluation and characterization morpho-agronomic. | Return To Table of Contents |
|---|
| Formation of the germplasm bank: The germplasm or gene bank consists of the physical infrastructure where the ex situ conservation of the accessions is realized, in other words, the conservation outside of its habitats, be it in the form of permanent collections of pollen, seeds, tissue culture, or of plant collections maintained in the field etc. In the case of cacao, the recalcitrance of the seeds requires that the genetic resources be maintained in field collections of live plants for the species’ conservation. Therefore, there is the need for a base station with a minimum infrastructure containing, at least, a laboratory to receive and describe the sampled fruits and seeds, nurseries covered with netting for the vegetative and seed multiplication, and, obviously, areas for the establishment of field collections. The vegetative material is generally multiplied by bud grafting, of the ‘T’ type, which is easy to do and has a high success rate. The success index of this technique is above 80% thirty days after grafting, when healthy buds are used. Rootstocks with different stem diameters should be available owing to the variability of the diameter of the collected shoots. The grafts are kept in the nursery for eight to twelve months, until their definitive implantation, when they should have at least six mature leaves. When the number of grafts/accessions is insufficient for a complete establishment of a plot in the germplasm bank, it is implanted in the nursery, where it receives special cultural treatments to stimulate the production of buds and ensure a posterior multiplication. The sampled seeds of the mother plants are sown and kept in nurseries for around six months. Where fruits had been collected, the opportunity to obtain agronomical traits of interest should not be missed out. Seed samples of the fruits are obtained from each matrix which will form the progenies of open pollination; for study purposes, they will be considered as half-sib progenies. At planting, a seedling sample of each progeny is selected, according to the availability of seedlings and/or the area where the collection will be installed. When sowing the progenies and grafting the clones of each accession, care must be taken to label the plastic bags, to avoid any mix up. A possibility is to write the proper alphameric code of each accession on the plastic sack of each seedling. The maintenance of seedlings in the nursery requires the use of routine cultural treatments for this phase, such as: irrigation, foliar fertilization, pest and disease control, elimination of weeds, and removal of side shoots. The installation of a germplasm bank must follow a system that separates the clonal from the seed accessions, owing to the different canopy architecture of both. Currently, in the Amazon, CEPLAC has adopted a system where the clones are established in rows of 5 plants, while the progenies are established in neighbouring squares, in plots with 5 to 25 plants also arranged in 5 plants per row. Some accessions are established in plots with replications. Most of them, however, are planted in a single plot, so as not to increase the cost of this activity even more. The problems that arose at processing the statistical analysis of the evaluation and characterization data caused by the absence of replications in the germplasm banks caused some research centres to adopt the strategy of implanting clonal material in plots of 5 plants in at least 2 replications. This allows a greater precision in the comparison of complex traits such as yield and its components and enables the achievement of estimates of variance and heritability coefficient components. This strategy has been adopted since 1998 by CEPEC, which plans to amplify the new areas of expansion of the germplasm bank to three replications. For a more profound approach to the subject read Chapter 11, which deals with experimentation in improvement. The commercial spacing for cacao is 3.0 x 3.0 m. However, a highly recommended practice in the germplasm banks is planting at a reduced spacing (1.0 x 1.0 m; 1.0 x 1.5 m) and with a smaller number of individuals per accession. Under these circumstances the agronomical evaluation is prejudiced, especially regarding yield and its components and disease resistance, traits strongly influenced by the environmental conditions. Frequently, banana (Musa sp.) is used as temporary shade and exceptionally, cassava (Manihot esculenta) as complementary shade for the sake of its growth speed and great biomass production for the protection of the cacao trees. The most commonly used species for permanent shade in the last decade have been eritrina (Erythrina poeppigiana), bandarra (Schizolobium amazonicum), and gliricidia (Gliricidia sepium). Generally speaking, the approved cultural treatments for cacao in the Amazon region according to Silva Neto et al., 2000 are observed for the maintenance of the field collections. These Theobroma collections are huge and occupy extensive areas, making the maintenance costly and the management difficult. Additionally, the environment in the humid tropical climate is extremely favourable for the development of fungal diseases and the excessive occurrence of weeds - factors that also contribute to increase the maintenance costs. Germplasm evaluation and characterization: The germplasm evaluation and characterization refer to the use of processes and methods to generate information on the genetic pool united in the bank. Such information addresses a broad public such as breeders, botanists, geneticists, biochemists, and others. It represents a guideline for the use of genetic resources in improvement programmes of the species and in correlated research. The evaluation and characterization can be processed using the traditional morpho-agronomical markers and/or biochemical molecular and DNA markers, as will be shown as follows. Morpho-agronomical evaluation and characterization: As mentioned earlier, this kind of evaluation and characterization is processed by the use of morpho-agronomical markers. In conjunction, these markers should describe each accession in detail, which is why they are called descriptors, and express the germplasm’s potential use for the different research lines. According to Engels, 1993, such descriptors are, predominantly, taxonomical, while the breeder needs detailed agronomic information. On the other hand, one should always bear in mind that too many descriptors are of little practical use or even of limited use. A data management system that allows the construction of a data bank, the processing of the statistical and genetic analyses, and that is efficient and practical to orientate decision making and monitor the efforts realized should also be available. The evaluation, frequently, boils down to the traits of economical importance such as yield and its components and disease resistance. These traits usually have polygenic control and are, therefore, under a strong environmental influence. To generate reliable and reproducible data, special ecological conditions are necessary (Simmonds, 1981 and Engels, 1993). In turn, the characterization consists in obtaining data, above all of qualitative traits, to describe and differentiate the existing accessions. In general, the main descriptors are grouped as describing the plant (height, shape, growth habit, and ramification); the leaf (shape, width, length, colour, border type and veins); the flower (shape, colour, and calyx type); the fruit (shape, colour, volume, and number of seeds per fruit), and the seed (size, shape, and colour) (Querol, 1993). According to Engels et al., 1980, the establishment of descriptors in cacao is necessary for the following reasons, among others: i) standardize the descriptive terminology to allow the interchange of information among researchers working with genetic resources; ii) make the elaboration of an inventory easier that would be available to all researchers and, consequently, define which available accessions should be duplicated elsewhere; iii) help the breeder to select the best accessions for the breeding programme; and iv) simplify the management and the maintenance of the collection. The standardization of the descriptors becomes particularly relevant when international projects of technical cooperation are realized, which create an opportunity to study the role of the environment in the expression of phenotypic characteristics in cacao systematically, as well as possible genotype x environment interactions linked to agro-economical and botanical-morphological traits (Bekele & Butler, 2000). A number of studies have been developed in the last five decades to define descriptors of interest in genetic resources of cacao. For example, Ostendorf, 1957, found evidence that the quantitative traits of the leaf and the plant growth habit are of no discriminative value. He proposed the use of the flower components as a more efficient means to identify cacao accessions. Thereafter, Enriquez & Soria, 1967b, analysed 27 flower traits, 16 of which were quantitative and 11 qualitative, and verified the existence of a great variability in most cases, concluding that any trait that presents highly significant differences among cultivars and has a variation coefficient below 10% must be considered in the description of the accessions. Besides, Enriquez & Soria, 1967a, began to use, in systematic form, qualitative and quantitative morphological descriptors besides agronomic information on disease resistance, to evaluate cacao cultivars. This study resulted in a catalogue that compiles information on the 67 most used cultivars at the time, in the main cacao producing countries of the Americas, involving over 60 descriptors of fruits, seeds, flowers, and inflorescences, among others.. Engels et al., 1980, proposed a provisional list of 87 quantitative and qualitative descriptors to assist the selection of discriminative traits of the genetic resources of cacao. This provisional register served as a base to formulate a list of cacao descriptors recommended by IBPGR (IBPGR, 1981). IBPGR suggested the use of 68 descriptors, of which 22 are part of a preliminary characterization at collection and 46 related to the evaluation and characterization in the germplasm bank. In the same year, the Collection Catalogue of CATIE (Engels, 1981) was published, with information on 294 cacao clones and 8 related species. This catalogue allows for 66 descriptors related to the yield, leaf, flower, fruits, seeds, self-compatibility, disease resistance, etc. Both descriptor lists came into use, with adaptations, in several research centres. One hundred accessions of the germplasm bank in the ICGT in Trinidad were evaluated for 54 morpho-agronomical descriptors (Bekele & Bekele, 1994). INIAP, in Ecuador, included the description of some of its hybrids and clones recommended for planting (Vera et al., 1984), adding agronomical information of direct interest for the breeder to the list. The characterization of cacao ‘Nacional’ of the country was also realized based on 38 descriptors of the flower, fruit, leaf, and seed (Quiroz & Soria, 1994). Nevertheless, the descriptor list does not seem to be consolidated yet. Frequently, the activity of evaluation and characterization of genetic resources of cacao is a tedious, troublesome, and extremely costly process. It can become so difficult because of work interruptions, or because a multitude of descriptors of little practical interest were considered, or then because of the huge number of accessions established in the field. CRU (Cacao Research Unit), in Trinidad, has called undergraduate and graduate students (Master’s of Science and Doctor’s students), closely assisted by their board of specialists, to help with the evaluation and characterization of cacao germplasm at ICGT (CRU Newsletter, 1995). This strategy, besides offering the students training in research, also enabled working with accessions on a large scale. In view of these facts, some research was done to establish a minimum descriptor list. Engels, 1993, for example, developed a statistical method that allows the selection of quantitative and qualitative traits that effectively do have discriminating power. Of the 39 descriptors analyzed, ten were selected for different criteria such as: score of reliability, discriminative value, agronomical and taxonomical importance, easiness of observation or measurement, and correlation among traits. Later, a list with 65 descriptors was also presented, optimized to 12 (Bekele et al., 1994), which used multivariate methods of quantification of the phenetic divergence among accessions. Bekele, 1993, found evidence that reproductive descriptors, especially those that refer to the flower, are more useful in terms of taxonomy than the vegetative. It is likely that this is due to the greater selection pressure and the better genetic control of the former during the evolutionary process. Therefore, a minimal list containing 26 descriptors of the flower, leaf, fruit, and seed (15 quantitative and 11 qualitative) was proposed. More recent studies (CRU Newsletter, 1995 and Bekele & Butler, 2000) assembled a minimal list of 23 descriptors (Table 5.2), which has been adopted since 1995 in the description of germplasm at the CRU, of the University of the West Indies, Trinidad (see Bekele & Butler, 2000, for further details and the data collection of these descriptors). Again, the minimal descriptor list does not seem to be completely consolidated. There are actually more cacao descriptors than necessary. Under this point of view, Dias et al., 1997, developed multivariate statistical procedures to reduce their number, without loosing significant information or the capacity to discriminate accessions. The authors studied 26 clonal cacao accessions: 15 of the series SIC and 11 of series SIAL, evaluated by 13 descriptors, of which 8 were of fruits and 5 of seeds. Cluster and main component analyses were used to quantify the phenetic divergence among the clones. The analysis of main components was processed based on the matrix of correlation between the trait means. The cluster analysis was processed by calculating the mean Euclidean distance and applying the algorithms of Tocher and the nearest neighbour onto the distance matrix. In both analyses, standardized data were used for the zero mean and the unit variance, hereby removing the scale effect by which the different descriptors had been measured. Of the 13 evaluated descriptors, four (31%) could be discarded due to redundancy, without loss of information or of the capacity to discriminate accessions. The visualization of the phenetic divergence among clones by means of multidimensional scaling graphs (Figures 5.3 and 5.4) shows that there was no distortion in relation to the number, and neither regarding the composition of the formed clusters, based either on 13 (Figure 5.3) or on nine descriptors (Figure 5.4). This methodology presents a potential that should be tested with a greater set of accessions evaluated by a large number of descriptors. Note that not only the clusters are maintained regarding number and composition, but the relative distances between clones in the bi-dimensional plan also seem unchanged (Figures 5.3 and 5.4). The number and composition of the clusters revealed by multidimensional scaling were identical to those obtained by the application of the cluster analysis of main components and the Tocher algorithm. Using 9 descriptors to quantify the divergence improved the visual dispersion of the clones within the clusters (Figure 5.4), although in this case the degree of the bi-dimensional representation measured by the stress was higher (9.4% in 5.4 and 3.8% in 5.3, after 50 iterations). The principles and application of the multivariate analysis mentioned here (cluster analysis, main components, and multidimensional scaling) can be referred to in Dias, 1998. Numerous studies have quantified the phenotypic diversity in cacao using morphological descriptors in multivariate analyses (Engels, 1983; Enríquez et al., 1988; Raboin et al., 1993; Bekele et al., 1994; Dias et al., 1997, and Lerceteau et al., 1997). Numerous research centres have used a limited number of descriptors for the characterization of their germplasm collections, according to the nature of the work and the availability of human resources, materials, and time. In a preliminary evaluation of the cacao collection of LCTAP, in Equador, Allen & Cabrera, 1988, considered only yield in the form of the weight of humid seeds and resistance to Moniliophthora roreri, C. perniciosa, and Phytophthora (under conditions of natural infection). At the station of Paracou-Combi, in Guyana, 13 qualitative and quantitative descriptors, including fruits, seeds, and ovules were used for the evaluation of selected clones in remaining populations on ancient cacao plantations of the country (Lachenaud et al., 1998). An enormous effort of collection and conservation of genetic resources of cacao has been made in Brazil (Almeida et al., 1987 and Almeida et al., 1995). Since 1963, the country has undertaken and participated in 17 collection expeditions to the Brazilian Amazon (Figure 5.5). The germplasm banks of the Brazilian Amazon conserve 1817 accessions, most of which are wild regional genotypes, represented by 940 clones and 877 families (Almeida et al., 1995). Parallel to the collection programme, the systematic evaluation and characterization of the accessions was developed, using leaf, flower, fruit, and seed descriptors (Castro & Bartley, 1983 and 1985 and Castro et al., 1989). Furthermore, studies were carried out to quantify the variability between and within wild populations of the Brazilian Amazon (Almeida & Almeida, 1987 and Dias et al., 2002), so as to collect and administrate the genetic resources of cacao efficiently. In the Brazilian Amazon, the evaluation and characterization of genetic resources of cacao date back to the 70s, when the base station in Belém, Pará, which belongs to CEPLAC, took up its activities. At the time, 14 clones from collections realized in 1965 and 1967 in wild and semi-cultivated regional populations were described, based on 9 quantitative and qualitative characteristics (Carletto & Costa, 1977). With the transference of the genetic stock to ERJOH (at km 17 of the highway BR-316, direction Belém-Castanhal), Marituba county, State of Pará, this characterization programme was interrupted, and was only reassumed from 1982 onwards. From this date the trait yield and its components were focused on, as well as the reaction of compatibility and resistance to witches’ broom, besides the morphological characterization of some accessions (Barriga et al., 1985). Giving continuity to the germplasm description in the Amazon, Bartley et al., 1988, analyzed accessions from the region of Alenquer, Pará. He discovered an expressive variability in traits of vegetative vigour, shape, leaf colour and size, productivity, and quantitative fruit and seed attributes. In the 90s, owing to the limited human and financial resources and the rich genetic stock established at ERJOH, these activities were restricted to the use of few quantitative descriptors, such as weight of the fruit, the husk, the wet seeds, and the number of seeds (Kobayashi et al., 1998). More recently, Silva, 1999, evaluated 143 open pollination progenies, assumed to be half-sibs, implanted at ERJOH. They stem from wild populations explored in the watersheds of the rivers Acre, Japurá, Ji-Paraná, and Solimões/baixo Japurá, all in the Brazilian Amazon, and were discriminated by the descriptors: total of collected fruits, weight of wet seeds, and tolerance index to witches’ broom. It was found that there is expressive genetic variability among the progenies from the different watersheds that can be explored in improvement programmes. |
| Establishment of germplasm banks. Evaluation and characterization of germplasm: Molecular evaluation and characterization and Duplication of the genetic reserve. | Return To Table of Contents |
|---|
| Molecular evaluation and characterization: The evaluation and molecular characterization of accessions is realized by biochemical markers (proteins, isoenzymes, and others) and DNA markers (RAPD, RFLP, AFLP, Microssatellite and others); it is neither more, nor less important than the morpho-agronomical characterization, but complements the latter. There is a tendency in literature to draw up comparisons between the information on variability among accessions generated by one or the other. This is a great mistake. Frequently, the morpho-agronomical and molecular analyses provide distinct results for the same set of accessions under study. This can be explained, basically, by the differences in the genes determining these two sorts of phenotypes and the distinct evolutionary forces affecting them (see the discussion in Chapter 3). However, the introduction of the molecular techniques can change the present panorama of improvement and genetic conservation of cacao. Biotechnology, a branch of biology that unites the set of techniques that enable DNA manipulation, can accelerate the gain of basic knowledge on biology and genetics of cacao. Biotechnology can provide the breeder with a quantitatively larger and qualitatively more profound evaluation of the available germplasm (Dias, 1995). In terms of biochemical molecular markers, the use of six polymorphic isoenzymatic loci in the study of genetic diversity of 28 “populations” (comprising 482 clones) of the international germplasm bank of cacao in Trinidad (ICGT) demonstrated that the greatest part of genetic variation (77% and 61%, quantified by the Nei and the Shanon index, respectively) is concentrated within populations (Sounigo et al., 1997). It is worth underlining that ICGT shelters about 80 “populations”. Ronning & Schnell, 1994 used eight loci to investigate the diversity in 86 clones of the USDA/ACRI collection, and also detected that over 90% of the genetic variation, quantified by the estimator GST of Nei, are found within populations. Morphometrical and isoenzymatical studies with diverse temperate and tropical tree species (Dias & Kageyama, 1991) corroborated these results and revealed a similar spread pattern of the genetic variation among and within populations. This means that a greater number of plants or subpopulations in few populations should be sampled for the germplasm collection of cacao. Lanaud, 1986, realized pioneer work with this marker type for cacao, while a revision on the introduction of the isoenzyme electrophoresis technique can be referred to in Dias, 1995. The great opportunity of using molecular analysis in the evaluation and characterization of the germplasm lies in genotyping (fingerprinting), that is, in the unmistakable identification and differentiation of each accession. As a matter of fact, Figueira, 1998, demonstrated, by means of RAPD markers, differences between the identification of cacao accessions from the germplasm banks from Brazil and from Malaysia (see Chapter 10 on the application of molecular markers in genetic conservation). The identification of accessions by molecular genotyping, contrary to the morpho-agronomical characterization, does not depend on the plant age, the developmental stage of the plant, or on the environment (year, place, or country) where it is cultivated. Certainly, the interchange of accessions and cultivars among countries will become a lot more efficient and secure when genotyping has become a routine practice in the germplasm banks of the crop. Genotyping will make the identification of duplicate accessions and those with doubtful registers possible. In the cacao germplasm banks of the producing countries, it is possible to find different accessions with the same codification or, on the contrary, the same accession identified by different codes. Because of this, a new methodology of alphameric codification for freshly collected accessions has been proposed (Sounigo, 1998). Such a codification would have three initial letters, identifying the country where the accession was collected or selected, followed by the year of collection (the two last numbers), differentiated by letters in increasing alphabetic order for collections in one and the same year. The year of collection would be followed by the identification of the matrix, indicated uniquely by a large case letter, from which fruits or shoots were collected. The matrixes would be identified by an increasing numeration of up to 26 (number of letters of the alphabet), in front of which a new letter, also large case, would be juxtaposed to the first. Citing an example, BRA05aAB31 would identify clone 31 obtained from the mother plant AB during the second collection expedition realized in 2005 in Brazil. The objective of this new codification is to provide a unique identification for each accession, which would be relatively easy to use and unite as much information as possible. For accessions selected in research institutes, a code indicating the name of the institute followed by a number is sufficient. Adjustments and enhancements of the new codification are expected so that it may come to be accepted and adopted in all cacao germplasm banks. The DNA marker types, on the other hand, used with cacao are mitochondrial and chloroplastic (Laurent et al., 1993b); RFLP (Laurent et al., 1993a and 1994; Figueira et al., 1994; N´Goran et al., 1994; Lerceteau et al., 1997); RAPD (Wilde et al., 1992; Russel et al., 1993; Figueira et al., 1994; N’Goran et al., 1994; Lerceteau et al., 1997 and Marita, 1998); AFLP (Perry et al., 1998 and Queiroz et al., 1998) and microssatellites (End et al., 1999 and Lanaud et al., 1999). In general, these studies reveal that the variability of the Forasteros in the Upper Amazon encompass the total variability of the species, reinforcing the hypothesis that the Upper Amazon region is in fact the centre of origin of T. cacao. They furthermore reveal that the Forasteros have a greater degree of polymorphism and are distinct from the Criollos. The use of RAPD markers especially (Williams et al., 1990), seems to be more adequate for genetic studies on cacao, due, mainly, to their relatively low cost, their simplicity, speed, and fast automatic operation with bulk samples (Penner et al., 1993 and Dias, 1995). The potential use of RAPD markers for the improvement of cacao, also for the prediction of heterotic parental combinations, has already been suggested (Wilde et al., 1992 and Russel et al., 1993) and demonstrated (Dias et al., 2003). On the other hand, the potential of the technique for genotyping cacao accessions deserves special attention. According to Wilde et al., 1992, a single primer allowed the undisputable characterization of 10 T. cacao L. accessions, one of T. microcarpum L. and two of Herrania L., a genus related to cacao. Likewise, Russel et al., 1993, demonstrated that three primers were sufficient to discriminate 25 cacao accessions that belong to the series IMC and PA collected in Peru and the series LCTEENS collected in Ecuador. This last study also validated that the greatest part of genetic variation (59%, according to Shanon’s diversity index) is concentrated within populations. Marita, 1998, stated that few PCR reactions on the thousands of accessions grown all over the world could identify those with a high probability of containing genes resistant to witches’ broom and other diseases. The author based her claim on the study carried out with RAPD markers applied to 270 clonal accessions of the germplasm bank of CEPEC. Two other important conclusions were drawn from that study. The first was that the inclusion of 90 accessions, chosen randomly from the bank, did not amplify the genetic variation found in the 180 accessions originally investigated. This means that the potential of available genetic variability in a germplasm bank cannot be evaluated by the number of accessions it contains. The second conclusion was that clone SGU 26, a hybrid from Guatemala, presented tolerance to witches’ broom and was genetically distance from the accessions of the Upper Amazon, which are, as we know, sources of resistance genes to witches’ broom. This fact indicates that the Upper Amazon is not the only region with sources of resistance genes to this disease, so the collection of these sources must not be limited to that region only. The RAPD technique is certainly the most economical for countries in development. In these countries, research funds are scarce and the financial aspect is of vital importance (Dias, 1995). Furthermore, from the methodological point of view, RAPD markers have proved to be adequate for studies on the cacao genome, since the greatest part of its loci presents Mendelian inheritance (Ronning et al., 1995). Therefore, more should be invested in the technique, and a standardization of the laboratory procedures aimed at, so that results become comparable and reliable. Besides genotyping, genomics also offer the possibility of gene introgression from wild species into commercial cultivars. The construction of molecular maps opened up the possibility to seek not only a desired phenotype among the numerous accessions available in germplasm banks, but the very favourable genes present in those accessions (Tanksley & McCouch, 1997). Surely, this change of paradigm - to search for the gene and not for the phenotype - will occur more as mapping becomes a routine in the research centres. The interspecific molecular maps of the tomato plant (Lycopersicon esculentum), for instance, enabled the introgression of QTLs (Quantitative Trait Loci) of wild species Lycopersicon hirsutum into elite lines. The introgression of these QTLs increased the production, the content of soluble solids, and the fruit colour by 48, 22, and 33%, respectively (Tanksley & McCouch, 1997). The possibility of introgression of genes monitored by the genetic genome is a promissing research line for cacao. Molecular maps of the species are already available (Lanaud et al., 1995 and Crouzillat et al., 1996) and various attempts are being made for the mapping of specific QTLs for yield components, earliness, and disease resistance, among other traits (Lanaud et al., 1995; Phillips-Mora et al., 1995; Crouzillat et al., 1996; Clement et al., 2000 and Risterucci et al., 2000). Some Theobroma species present traits of interest for the cacao tree (Theobroma cacao L.), (see Chapter 2) as for example: smaller canopy and witches’ broom resistance, present in T. microcarpum, the un-pruned canopy of T. speciosum, and fruit abscission in T. grandiflorum. QTLs for these and other traits, after identification, could be inserted into commercial cultivars, with large benefits for cacao cultivation. Duplication of the genetic stock: The establishment of the entire genetic stock of cacao collected so far in a single germplasm bank and in the form of a field collection of live plants, as presently in use, is subjected to threats such as fires, floods, and hurricanes, among others. Additionally, cacao plantations have a high combustibility owing to the thick layer of residues on the ground, from the continual shedding and accumulation of leaves, branches, twigs, and fruits. The possibility of a disaster becomes more threatening, as there is a drought period of three to four months a year in the Amazon when cacao plantations are often destroyed by fire. With this background, the urgent need to replicate or duplicate such a collection in another place or other places, as a security measure, becomes evident. The need to evaluate the genetic stock in relation to resistance or tolerance to C. perniciosa, causal agent of the principal disease of the Brazilian cacao cultivation, which presents local pathotypes or isolates with different pathogenicity levels (Wheeler & Mepster, 1988) is also apparent. Therefore, an evaluation of the resistance at different sites is imperative to study the possible effects of genotypes x sites more in depth. Such duplication causes problems and significant costs, which will hardly be wholly covered by research institutions. Partnership institutions and national and international organizations must be sought that can make the referred duplication possible. At present, the duplication of part of the genetic stock established at ERJOH, which comprises 1.694 clonal and seed accessions (Silva et al., 1998), is being done. This duplication includes the establishment of a third of this collection at CEPEC, in Ilhéus, Bahia State, the main cacao producing region of Brazil, one third at the Experimental Station Paulo Dias Morelli, Rodovia TransAmazon, Pará State, the second cacao producer, and the rest in ESEOP, in Ouro Preto do Oeste, Rondônia, the country’s third producer state. |
| Chapter 6. Selection strategies and methods. L.A.S. Dias & M.D.V. Resende. | Return To Table of Contents |
|---|
| Content: Introduction, Variability, Yield and its components and Association of yield components. Character heritability; Plant characters, Fruit characters and Bean quality. Character heritability; Disease resistance. Estimates of genetic parameters. Predictive biometric genetics. The fundamentals of prediction, Prediction of additive and genotypic effects, Best linear unbiased prediction (BLUP) and Restricted Maximum Likelihood (REML). Application of REML/BLUP procedures. Multi-character selection index, Bayesian inference and Heterosis prediction. Selection of commercial hybrids and Participatory improvement Summary: An evaluation of the accumulated knowledge on genetic improvement of cacao is presented, in order to realign future research in this area. Predictive techniques of biometric and biotechnological genetics and the importance of their application in improvement programmes are emphasized. For the first time, the application of mixed models to estimate genetic parameters and predict additive and genotypic genetic values in such programmes is presented. The use of such models will allow greater efficiency in cacao improvement. The basic objectives to be considered in these programmes are the yield increase, with a rise in productivity and reduction of production costs, and the output of a superior quality cacao to create differential prices practiced by the chocolate industry. |
| Introduction, Variability, Yield and its components and Association of yield components. | Return To Table of Contents |
|---|
| Introduction: Genetic improvement of plants, understood as the science that manipulates plants towards the interests of humanity, is part of the socio-economical environment. However, society and governments are not always aware of this fact. In many countries, especially developing ones, there are no well-defined, permanent agricultural policies, nor a flow of resources that would ensure the continuity of improvement research. The idea that one or more improved cultivars could reduce the end cost of the product at the consumer’s level, as a result of enhanced productivity in space and time, is still not well understood by society and by governments. The commercial product of cacao is its seeds, duly fermented and dried (henceforth called beans owing to the death of the embryo during post-harvest processing). Brazil is among the five greatest cacao bean producers of the world (see Chapter 1) and had the worldwide highest mean productivity of this important commodity (750 kg/ha/year) in the late 70s. To reach this productivity level, it was necessary to install a long-term project, bringing cacao farmers and technicians together in an institution for cacao politics, called CEPLAC. The programme began in the 60s with the foundation of CEPEC, arising from a diagnosis of the main limiting factors of the country’s cacao sector. Cacao culture in the State of Bahia, which was then practically solely responsible for the total Brazilian production, obtained a productivity of 450 kg/ha/year. The determining factors of this low productivity were ascribed to the advanced age of plantations, severe pest and disease attack, excessive shading of the trees, lack of soil correction and fertilizer application, and the use of local, unimproved cultivars. However, it is estimated that the global and Brazil’s cacao productivity occupy the same level at 450 kg/ha/year. This situation triggered the elaboration of a genetic improvement programme for cacao aiming at high-yielding genotypes that would be resistant to the disease known as black pod and, simultaneously, bear large seeds, to improve the industrial output (see Chapter 12). The studies were set in motion by the now extinct IAL and ICB. The selection involved over 300 high productivity cacao matrices from commercial plantations of the States of Bahia and Espirito Santo, which were subsequently cloned (Vello et al., 1969). Alongside the local selections, clones from the Brazilian Amazon and from other producer countries were introduced. Obviously, clones with some proven superiority in agronomical traits of interest were given preference in the introduction process. After establishing the germplasm banks with these selected and introduced clones, a biclonal hybrid programme was chosen, crossing clones in pairs. The exploitation of heterosis in cacao opened the perspective to combine favourable traits, in their highest possible phenotypic expressions, in a single individual. Thereafter, hybrids offer the possibility to obtain superior genotypes in the short term, optimizing the accumulation of favourable gene interactions and the production of hybrid seeds in a large scale. Certainly, the greatest contribution to the productivity increase of Brazilian cacao is basically a consequence of the development of superior hybrids (Dias & Kageyama, 1995 and Dias et al., 1998). However, the parental clones of hybrids differ in combining capacity for yield and frequently this productive capacity per se is not associated with the performance as parents. From the beginning it was clear that some productive clones could generate low-yield progenies, while the opposite is also true (Bartley, 1967 and 1969; Purseglove, 1968; Bradeau, 1970 and Dominguez, 1975). Other clones, however, had a positive association between their performance per se and as parents (Adeleke, 1982 and Lockwood & Pang, 1994). The selection of the parental hybrid clones began on the grounds of combining ability tests. Clones of high combining ability for the trait of interest that stem from genetically divergent populations were inter-crossed. This practice eliminated a lot of the empiricism and randomness on which the exploitation of heterosis had been based so far, despite the selection of hybrid parents still being an open question. Clearly, knowledge on the performance per se of the parental clones and the general clone combining capacity among and between racial groups (see the discussion on these groups in Chapter 3) is fundamental for the selection process for high yield. Knowledge on the yield components that best express the yield is also essential. This chapter evaluates the accumulated knowledge on genetic improvement of cacao yields to realign future research projects in this area. Predictive techniques of biometric and biotechnological genetics are emphasized and the importance of their application in cacao improvement programmes. For the first time, mixed prediction models of additive and genotypic effects are presented and applied to these programmes. Basic objectives on which these programmes must focus on are the yield increase by means of increased productivity, together with a cost reduction in production, and the production of top quality cacao, to differentiate the prices offered by the chocolate industry. Variability: Economically, cacao culture ranks among the most important crops of the Neotropics. It is rated as one of the 15 most profitable agricultural commodities. Ecologically, it is a self-sustained perennial crop, which, in countries like Brazil, offers a certain protection for remaining tropical forests and in particular for the Atlantic coastal forest. The species is adapted to a broad variety of tropical humid environments and is intensely cultivated between lat 20' N and S, from near sea level to heights of 1000 m. Several articles mention the broad genetic variability assessed in the species, in plantations as well as in natural wild populations of Ecuador and Peru (Pound, 1938 and 1943) and the Brazilian Amazon (Barriga et al., 1985 and Bartley et al., 1988), in terms of yield and other traits. Yield and its components: It is a difficult task to develop high-yielding cacao cultivars considering that dry bean yield is a complex trait, strongly influenced by the environment and therefore of quantitative nature and generated from several components (Soria, 1977, see also ‘Estimates of genetic parameters’). First studies with yield components were carried out in the 30s (Pound, 1932a, 1932b, 1932c and 1933 and Cheesman & Pound, 1934). These authors suggested an evaluation of the cacao yield potential by means of two inversely related indices: the fruit (FI) and the seed index (SI). FI would express the number of fruits required to produce one pound of dry cacao, (now an international standard of 1 kg of fermented and dried beans), while SI would represent the mean weight of the dry bean, also obtained as weight of 100 fermented and dry beans. The studies were carried out with trees of the racial group Trinitario. Previously, other yield components were proposed with the aim of evaluating the superiority of the tree with greater accuracy. A new component - designated fruit value (FV), which is the mean number of seeds per fruit times the mean weight of the dry beans without the testa (Atanda, 1972a and 1972b and Jacob & Atanda, 1973). Superior cacaos would present higher relative FV values. Although used little, as its determination is more labour intensive, FV advantageously combines the information of two components - the number and the weight of the seeds. Another suggested component was the efficiency index (EI), which was used for the comparison of cacao cultivars from the racial groups Amazon Forastero and Amelonados. EI relates the fruit weight to the dry seed weight. It took 15.1g of Amazon Forastero fruit and 13.0g of Amelonados to produce 1 gram of dry bean (EI). This result demonstrated, concurrently, the superiority of the second group’s cultivars to convert dry matter and the discriminative power of the yield component EI. Another suggested component was the conversion factor from wet seed weight to the weight of dry seeds. This factor is genetically controlled (Carletto et al., 1983), and varies among cultivars (Are & Atanda, 1972 and Carletto et al., 1983). When establishing criteria for the selection of cacao hybrids in Bahia, the use of the conversion factor, in combination with other yield components, was considered satisfactory (Carletto et al., 1983). This factor, together with SI and FI, can increase the selection efficiency, since superior genotypes are also discriminated by the efficiency of the conversion of dry matter. The experience with improvement, however, reveals that the number of fruits per plant, the number of seeds per fruit, weight of wet or dry seeds per fruit and per plant, and the mean seed weight are some of the main yield components (Soria, 1977 and 1978). The path analysis classified the number of healthy fruits per plant and the weight of dry beans per fruit as main yield components, while the number of seeds per fruit and SI were considered secondary components (Almeida et al., 1994). The yield itself refers to the weight of dry beans per hectare, which is difficult and tedious to evaluate. One should bear in mind that environmental factors exert marked influence on the cacao yield and its components (Pound, 1932b; Ruinard, 1961; Bartley, 1970; Are & Atanda, 1972; Atanda & Jacob, 1975; Lockwood, 1976; Toxopeus, 1985; Dias & Kageyama, 1997b and Dias et al., 1998). The weight of wet seeds per fruit and also per plant are the most commonly used components for an indirect evaluation of the cacao yield (Dias & Kageyama, 1995 and Dias et al., 1998), although they can be subject to error caused by environmental influence (Bartley, 1969). Moreover, the cultivars differ in percentage of moisture in their seeds, reducing the trait’s effectiveness and discriminative power (Bartley, 1967). Notwithstanding, this component appears to be a good yield estimator (Pound, 1932a and b and 1933, and Esquivel & Soria, 1967) since it is easily measured and positively associated to fruit weight. A strong association between large fruits and large seeds has been observed (Cheesman & Pound, 1934) and this variation in the same sense leads to a cost cost reduction at harvest, less fruit breaking and to a higher value of the commercial product. As highlighted by Ruinard, 1961, the chocolate industry’s demand for seeds requires a mean dry weight of seeds of over 1 gram. An extremely variable component is the number of seeds per fruit, which depends directly on the number of ovules and their percentage fertilization (Pound, 1932b; Kuppers, 1953; Enriquez & Soria, 1966; Jacob & Atanda, 1973, and Atanda & Jacob, 1975). The concepts of fruit and seed weight and size in cacao confound and seem to reflect the same variation, although weight seems less significant than size. Soria, 1977 and 1978, states that seed and fruit size are among the most favourable yield components for improvement. From an economic point of view, it is more beneficial to produce beans from a lower number of large fruits, with less husks and bearing large seeds. However, in the last instance, the seed quality is the yield component that does or does not allow a definition of the genotype’s superiority (Urquhart, 1963). Economically speaking, dry seeds should reach a mean weight of 1.07 g (equal to 93 seeds for 100 grams), with 10 to 12% testa and a fat content in the dried cotyledon of over 55% (Toxopeus, 1985). The number of fruits and FI are among the most important yield components. Fruit yield can be considered a reliable criterion to estimate the productive potential of a given cultivar (Glendinning, 1963; Esquivel & Soria, 1967; Soria & Esquivel, 1969; Atanda & Toxopeus, 1971; Atanda, 1972a and Atanda & Jacob, 1975), being easily measurable in a large number of candidates for selection. In turn, the FI provides details of the yield characteristics (Atanda, 1972b) and can be used to refine the selection, after screening for fruit yield. The total number of fruits (healthy and diseased) per plant or per area represents the yield potential, while the number of healthy fruits represents the real yield, both being broadly used components (Mariano & Bartley, 1981; Almeida et al., 1994 and Dias & Kageyama, 1995). The ratio between the number of diseased fruits and the total number of fruits represents the percentage of diseased fruits. This last component, despite being strongly affected by the environmental influence (Dias & Kageyama, 1998b), can be helpful for the discrimination of tolerant genotypes to a given regional disease. Some other factors, despite not actually being part of the category of yield components, do act as such. These are incompatibility and stand density, which can bring about considerable alterations in the yield levels. Self-incompatibility as well as inter-incompatibility can drastically reduce the yield. Commercially, the seeds destined for planting come from at least five hybrids from biclonal crosses. This hybrid combination or mixture is basically to avoid problems of self - and inter-incompatibility. The manipulation of the planting density can also alter the final yield substantially. Dense stands (over 1111 trees per hectare) give a higher dry bean yield, at least in the first years of production, before the plant reaches its full physiological maturity in the period called the yield climax (Freeman, 1929 and Dias et al., 2000). In the post-climax (from the eighth or tenth year of planting), the situation is inverted and stands of low density surpass those of high density (Freeman, 1929). Special care must be taken with dense stands in regions under witches’ broom attack (Dias et al., 2000), especially in clonal stands. Finally, it should be mentioned that the use of one or another yield component in cacao depends, basically, on the objectives and stage of research, on the nature and quantity of the evaluated plant material, the prevailing environmental conditions, the cost/benefit compatibility, the ease of measurement and the component’s reliability, and also on their relative importance. If an improvement programme is at the initial stage and a great number of genotypes to be evaluated are available, then not more than one or few easily measured yield components should be used, such as fruit yield and FI. When the programme has reached a more advanced stage, and a small number of superior genotypes is being evaluated, a greater number of components should be assessed, in order to discriminate such genotypes precisely. Association of the yield components: The association between components is important because the selection of a trait with low heritability, of difficult and costly measurement, becomes possible by the underlying selection of a second trait of easy evaluation and higher heritability, associated to the first. In this situation, selection becomes more efficient in consequence of the simultaneous improvement of various components. In the only study that applies the path analysis to traits evaluated in hybrid cacaos, Almeida et al., 1994, showed that the number of healthy fruits per tree is genetically correlated to the dry bean weight per tree, in other words, to the yield per tree (Table 6.1) and that direct selection for the first component can therefore maximize the selection for yield. Despite the cited study having a fixed effect and that it evaluated a reduced number of hybrids over a short time, it can help understand better how the cacao traits are associated. Other studies also found high positive phenotypic correlations between the number of fruits produced and dry bean weight (Esquivel & Soria, 1967) and between the first and the weight of wet seeds (Glendinning, 1963; Esquivel & Soria, 1967; Atanda & Toxopeus, 1971; Atanda, 1972a). In this case, it must be explained that phenotypic correlations only quantify the association between the pair variables, without, however, considering their causes, which can be various. On the other hand, the participation of the genetic correlations based on direct and indirect effects on a variable considered principal, as in the path analysis, quantifies its relative importance, besides specifying the cause. The path analysis (Almeida et al., 1994) showed that yield per tree was mainly determined by the combination between the number of healthy fruits and dry bean weight per fruit, regardless of the negligible correlation between the latter component and the yield (Table 6.1). The simultaneous use of these two components as yield predictors, however, is to be used with care. The reason is that in practice, an increased fruit yield frequently leads to a reduced number of seeds, seed index, dry bean weight per fruit, fruit weight, and conversion factor (Table 6.1) Consequently, the dry bean weight per tree is reduced due to the internal competition for photosynthates. The number of seeds per fruit and the seed index presented low direct effects on the yield per tree, suggesting that these two components are inefficient when the objective is yield maximization. However, these two last components were linked to the dry bean weight per fruit and therefore were considered to be secondary yield components. Furthermore, an association between the trunk diameter at 30 cm from the ground, measured between the second and third year in the field, and yield and its components was observed, although this association is reduced with the increasing age of trees Soria, 1964 e Mariano, 1966). Furthermore, a high and significant correlation between the trunk diameter and the jorquette height and between the latter and the number of lateral branches was observed (Moses & Enriquez, 1981). Early selection for fruit yield by means of the correlation to the trunk height of young trees is possible (Soria, 1964). In fact, trunk height has a high positive genetic correlation to the number of healthy fruits per plant (Table 6.1). More vigorous trees, expressed by a greater diameter and trunk height, shade the neighbouring trees, which may be the reason why these traits were strongly correlated to the yield. Actually, the two latter traits (trunk height and diameter) are of limited value for the process of indirect selection, mainly because they are affected by the momentary physiological state of the plant at measurement. It is worth emphasizing that path analysis works basically with yield components and that the inclusion of traits that are not classified as such in the analysis, e.g., trunk height and diameter, should be avoided. Likewise, since one is dealing with multiple regression path analysis, a distorted situation may arise when yield components, obtained by the combination of other components, are processed. Seed weight per fruit, for example, which is that obtained as a product of the seed index and the number of seeds per fruit. Finally, there is consensus in literature that improvement for seed size furnishes better results than improvement aiming at the number of seeds per fruit (Kuppers, 1953; Ruinard, 1961; Engels, 1983). Fruit size is genetically determined by the seed size, though there is a negative correlation between the size and the number of seeds (Engels, 1983). At any rate, selection programmes of matrixes for yield, which involve thousands of plants, must begin by assessing the number of fruits per tree. At a second stage, once the low yield matrixes are eliminated, the selection should be based on FI and SI, if matrixes with high fruit yield and large seeds are to be retained. |
| Character heritability; Plant characters, Fruit characters and Bean quality. | Return To Table of Contents |
|---|
| Character heritability: The heritability of the yield components and of other important traits for cacao improvement is neither well known nor well studied, which supports the idea that such improvement has been practiced on empirical grounds from the very beginning. Results of studies on cacao heritability exist, though they are limited in number and to conclusions and observations made in crosses aiming at the selection and the production of commercial hybrids. Plant characters: Incompatibility reaction: Self-incompatibility is the incapacity of a hermaphrodite plant to produce zygotes by self-pollination. Consequently, the genetic flow and the heterozygosity within the specie’s populations are broadened. Regarding genetic improvement, the importance of self-incompatibility lies in the possibility of making hybrids, without requiring manual crosses. In terms of commercial production, the yield capacity of cacao depends on the genotype(s) used in the stand and the pollination efficiency. This last aspect, especially, is a function of the activity of the pollinating midge Forcipomyia. However, even if these two factors are met, the mechanism of self-incompatibility found in the specie’s populations can affect the productivity of cacao. It is known, for example, that self-compatible cacaos tend to produce larger fruits and dry bean yield than the self-incompatibles (Lockwood, 1977 and Morera et al., 1994). To avoid the restriction of self-incompatibility, a hybrid combination from a mixture of seeds of five or more hybrids is distributed for planting. Self- and inter-incompatibility in cacao were first recorded by Harland, 1925 and Pound, 1932c, who verified that some trees in Trinidad did not fruit with their own pollen, nor with the pollen of others. It was Voelcker, 1937, however, who initiated a detailed investigation of the phenomenon. It was later verified that the self-incompatible cacaos from the Upper Amazon, introduced in Ghana from Trinidad, showed inter-compatibility Posnette, 1945). It is worth remembering that both self-compatible and self-incompatible trees produce hermaphrodite flowers. However, the spatial distribution of the flower parts impairs the free access of the pollen to the stigma. There is a crown formed by five staminodes] around the pistil. Analogically, the anthers are protected by the basal petal part in shape of a hood, known as the sepal or calyx. Besides this, the pollen, due to its viscosity, clumps in a mass, so the wind cannot transport it. This peculiarity of the cacao flower requires the visit of the midge for pollination (see Chapter 4 on pollinating agents). Self-incompatibility in cacao is complex, unique among higher plants and still poorly understood, and localized in the embryo sac, under both sporophytic and gametophytic genetic controls (de Nettancourt, 1977). In the gametophytic control, incompatibility is determined by the pollen genome. In the sporophytic control, the maternal diploid genotype determines the incompatibility. The growth of the pollen tube in incompatible pollination is comparable to the normal compatible pollination, where the antherozoids reach the embryo sac normally. However, the anomaly arises in the embryo sac through the prevention of the syngamy: the fusion of the male with the female gametes carrying the same dominant allele. Therefore, 25%, 50%, and even up to 100% abortions can occur in incompatible pollinations (Cope, 1958 and 1962), demonstrating the graded expression of the phenomena. The system is controlled by a simple multi-allelic S locus. Locus S comprises five alleles, expressing actions of dominance (indicated by >) and of independence (indicated by =) in the male and female organs, according to the following relationship: S1 > S2 = S3 > S4 > S5 (Knight & Rogers, 1953 and 1955). Inter-incompatibility among self-incompatible genotypes occurs when these genotypes carry S alleles in common, as for example, S1S2 x S1S3 (also represented as S1.2 x S1.3), or when one genotype has independent alleles and the other a dominant allele, as for example in the cross S2.3 x S3.5. The cross of S1.2 x S1.3, however, leads to only 25% abortion, equal to the self-fertilization of S1.2, once the gametes carrying the S allele1 represent ¼ of the total gametal combinations. Self-fertilization of S2.3 results in 50% abortion, since ¼ of the gametes carry allele S 2 and the other ¼ carries allele S 3. Finally, the self-fertilization of S2.2 or the homozygote cross of the type S2.2 x S2.2 leads to 100% abortion. In these cases, the ratio of the unfertilized ovules:fertilized ovules is, respectively, 1:3, 1:1, and 1:0. According to Cope’s theory, this ratio would be 0:4 in self-compatible clones, therefore, with 100% fusion when self-pollinated, though different grades of self-compatibility have been observed. In practice, the self-pollination of some self-compatible clones resulted in 25% abortion (Carletto & Soria, 1973) and even 75% abortion (Coral, 1970), that is, ratios of 1:3 and 3:1, respectively. Owing to the occurrence of self-incompatible hybrid progenies from the cross between two self-compatible clones, as for example the ‘ICS 1 x ICS 45’, two other complementary loci A and B have been proposed, independent of each other and of locus S (Cope, 1958, 1962 and Purseglove, 1968). Loci A and B act before meiosis, expressing dominance action and producing an unspecific precursor for the S alleles after meiosis. The homozygotic recessive genotypes for at least one of the three loci (A, B, and S) are self and inter-compatible. Cope, 1958, still aggregated allele S 6, conditioning compatibility, to the allelic series of locus S. The segregation standard of the loci A, B, and S, and the distribution of phenotypes after crosses among self-incompatible and self-compatible cacao, in all their possibilities, were presented by Bartley & Cope, 1973. Studies on the reaction of self-incompatibility are done by means of controlled self-pollinations. Controlled crosses among self-incompatible genotypes are carried out to determine the cross-compatibility. Finally, the determination of the incompatibility genotypes is realized by crosses with a differentiating clonal series, carrier of the known incompatibility alleles. The evaluations are generally done three, seven, and 15 days after the pollinations (Terreros et al., 1982; Yamada et al., 1982 and Yamada & Bartley, 1984). Note that a single pollination generates between 40 and 45 seeds and that, in incompatible pollinations, the flowers are aborted within 3 to 4 days. Two criteria are commonly applied to test the incompatibility reaction of a given genotype. In the first, the trees that present retention of flowers above 5%, on the 15th day after the self-pollinations are considered self-compatible. Those with less than 5% are self-incompatible (Lopes & Carletto, 1995). By the second criterion, the minimum number of flowers retained in the self-pollinations and crosses, with one or more controls to express the compatibility, is estimated using Χ2 to test the proportion 1:1, at P< 0.05 and 1 degree of freedom (Terreros et al., 1982). The first criterion seems to overestimate self-compatibility more than the second (Lopes & Carletto, 1995). For each tree, 15 to 40 flowers are used for self- and cross-pollination. The mean optimum number was 26 (Lopes & Carletto, 1995). The methodology of artificial pollination (Carletto, 1946 and Hardy, 1961) defines that in the afternoon of the day before the controlled pollinations, the flower buds that are to open up the next day, recognizable by the swelling, should be protected by small plastic tubes that are fixed to the tree with paste and closed on the other end by a fine cloth or gauze. The pollination is carried out in the morning of the following day, when the flower has opened and its stigma is receptive. The points of the staminodes and stamen are cut to make the access to the stigma easy and the pollen grains are rubbed lightly onto the stigma with up to three anthers. Each manually pollinated flower is identified by a pin inserted beside the flower cushion and, eventually, by a tag, in case of controlled crosses in a small scale. In case of loss of the identifier or the appearance of other flowers on the same flower cushion, the manually pollinated flower is recognized by the cut staminodes and stamen. The tubes are removed after one to three days, since fertilization occurs between 10 and 24 hours after pollination. The confusing and contradictory results obtained in the studies on the incompatibility reaction clearly show the deficiency of procedures adopted to define the phenomenon. Flower abscission, for example, can be caused by the inefficiency of the pollination mechanism as much as by physiological and seasonal reactions that go on in the reproductive organ. The criteria adopted for declaring incompatibility are also questionable. There are variations from 7 to 15 days in the post-pollination period for this declaration, which hampers the comparison among results found by different authors. The sample size used in these studies is also highly variable with some authors carrying out only 15 self and cross-pollinations, while others do up to 40. Since incompatibility is also characterized through the application of the test X2, the reduced sample size affects the results greatly and compromises the inferences. As if this were not enough, the various theories proposed for the reaction are complex, genotype specific, and are therefore not applicable to all situations. Pandey, 1960, recognized the physiological nature of the incompatibility system in cacao, but criticised the unnecessarily complex genetic interpretation given. To justify the occurrence of self-incompatible hybrids produced by self-compatible parents, this author used the competitive allelic interaction, discarding the postulation of the independent loci A and B proposed by Cope (1958 and 1962). The competitive interaction among S alleles that impairs the complete expression of any one of them alone, would explain the situation. A cacao tree with two S alleles interacting competitively would not produce enough of the specific incompatibility substance for pollen growth for either one of the two alleles and would be completely self-compatible. Nevertheless, the progeny of a cross between two such trees could then be self-incompatible. Consequently, the progeny of the cross between two of these cacao trees could be self-incompatible. As self-compatible and self-incompatible trees in the progeny could appear in the proportions 3:1, 1:1, and 1:3, this hypothesis of Pandey would plausibly explain the occurrence, based only on the S alleles. The dominance of one S allele over the other and the action of independence of both alleles have already been adequately used in the theory of Knight & Rogers (1953, 1955), to explain how the alleles of locus S work. As a matter of fact, studies on incompatibility (Carletto & Soria, 1973) have demonstrated that the phenomena in T. cacao is explained better by the theory of locus S of Knight & Rogers (1953, 1955) than by the theory of the independent loci A, B, and S of Cope (1958, 1962). The greater importance of self-incompatibility is its use for the synthesis of biparental hybrids by open pollination in isolated fields, known as clonal gardens. Since forest improvement came to be the new paradigm for improvement in cacao (Dias, 1993; see also Chapter 13), the expression adopted for these gardens will be ‘biclonal seed orchards’. The biclonal hybrids distributed to cacao farmers, be it for the implantation of new areas or the renewal of decadent areas, have either one or two self-incompatible parental or both self-compatible clones. On this basis, the knowledge of the incompatibility reaction allows a planning of the biclonal crosses and a prediction of the occurrence and proportion of incompatible cacaos in the hybrid progenies, so as to ensure a high index of cross pollination as much in the seed orchards as in commercial stands. CEPEC invests efforts into the study of incompatibility in clones of potential importance for the genetic improvement programme (Carletto & Soria, 1973; Yamada et al., 1982; Yamada & Bartley, 1984; Lopes & Carletto, 1995 and Lopes & Yamada, 1995). Basically, this programme uses hybrids that stem from the cross between selected clone pairs that belong to the same racial group or, which is more common, to different racial groups. Cloning of the parental genotypes induces earliness in flowering and a reduction of the plant height, making the hybridization work and harvest of the hybrid’s fruits easier. The hybrids are distributed to the cacao farmers in form of seeds, produced in seed orchards through manual pollinations or the use of self-incompatibility. In these orchards, alternating lines of the self-incompatible clone (from three to eight pollen receptor female rows) are planted with a self-compatible row of the clone (pollen donor male row). Evidently, pollination is natural in this situation and the hybrid seeds are collected exclusively from the rows of the self-incompatible clones. Orchards based on open pollination, without flower protection, require planting in alternating rows of two different self-incompatible clones, with hybrid seed production in both. On the other hand, in orchards that use self-compatible clone rows, manual pollination with or without flower protection is necessary, however with the use of a genetic marker which can be easily identified. Vello et al., 1972, proposed a solution for the case of an emergency such as a high demand of hybrid seeds. They suggested using an orchard composed of a mutant homozygous clone of white seeds, in this case the cultivar Catongo, as pollen receptor, and of alien clones with purple seeds, as pollen donors. The hybrid seeds, produced by manual pollination without flower protection, are harvested in the ‘Catongo’ rows and recognized by their purple colour. In this scheme, the mean contamination rate with white seeds was 12%. These were obviously eliminated since they were a product of Catongo self-pollination. The strategy adopted in the seed orchards of hybrid cacaos is broadly accepted in the cacao-producing countries. However, it is known that the self-incompatibility mechanism does not ensure absolute security for cross pollination (Glendinning, 1960; Opeke & Jacob, 1969 and Lanaud et al., 1987), while the incompatibility genotypes of most parental clones used in the production of hybrid seeds are unknown as well (Terreros et al., 1982). Variable percentages of seeds come from self-pollination (Glendinning, 1960, 1962 and 1972 and Edwards, 1969). The studies involving the pollination of self-incompatible cacao have shown that the genetic incompatibility can be overcome by the application of a pollen mixture of a tree with cross-compatibility with itself (compatible allopollen) with pollen from the tree itself (incompatible autopollen), resulting in a presumably low percentage of self- pollination (Glendinning, 1960; Opeke & Jacob, 1969 and Lanaud et al., 1987). The tree used as compatible pollen donor is homozygous for the marker gene of axil spot, a trait already expressed in the plantlet stage. Soon, plantlets that do not present the axil spot are selected as self-pollinated progenies of the self-incompatible clone. Remember that seeds from self-pollination may be result of pollinations from among different flowers of the same tree or from different flowers from trees of the same clone. Incompatibility can also be overcome by prior application of pollen of Herrania, a genus related to T. cacao, followed by the application of incompatible autopollen. Another method uses CO2 to overcome self-incompatibility (Aneja et al., 1994). When self-pollinated flowers of self-incompatible genotypes are stored in plastic bags for six hours, indices of pollen germination of 95%, close to the 100% verified in compatible pollinations are obtained. The effect obtained by the accumulation of CO2 in these plastic bags also stimulates fruiting, and guarantees the production of viable seeds with 95% germination. Experiments realized by Lanaud et al., 1987, showed that under natural pollination conditions in the orchards involving various biclonal combinations, the mean percentage of self-fertilization is high (51%) and very variable (0 to 97%). Among the different factors that influence this variation is the paternal effect, as well as some self-incompatible clones that are eventually more fractious to self-fertilization than others, ‘IMC 67’ for example. The other factors are the effect of harvest period and the local ecology of the orchard, in terms of climate and pollinator population. An application of the mixture of the compatible allopollen with incompatible autopollen in different proportions does not only allow incompatibility to be overcome but also demonstrates that the reaction is quantitative (Glendinning, 1960; Opeke & Jacob, 1969; Lanaud et al., 1987). The reaction is intense when only incompatible autopollen is applied, with a nearly total abortion of the flowers up to three days after pollination. However, only 50% of the flowers drop after the third day when a balanced mixture of compatible allopollen with incompatible autopollen is used. These experiments (Lanaud et al., 1987) explain the contamination that occurs in the orchard, since pollination in a cacao population can occur by the deposition of the tree’s own pollen mixed with pollen from neighbouring trees. An identical situation can occur in commercial stands and compromise the productive potential of biclonal hybrids. Therefore, knowledge on the mechanisms of self-incompatibility is fundamental for the distribution of hybrids with maximum cross pollination to farmers. Studies on pollination biology and evolution of the species will be expanded with this knowledge. Crown growth: Cacao trees basically show erect and branched habits. The clones ‘Alto Amazônicos’ SCA 6 and SCA 12, which have a branched growth habit, carry this trait forward to their progenies in crosses with various other clones, indicating dominance action (Soria, 1977). However, the branched habit is not desirable, since it requires frequent pruning to correct the canopy architecture and avoid self-shading. Branching height: The canopy normally forms at a height of 1.2 m from the ground, when the main stem of the cacao tree loses the apical dominance and the lateral buds develop into a crown formed by four or five plagiotropic branches. This trait is determinant for the plant height and the presence of a low canopy favours phytosanitary pruning. On the other hand, a high canopy makes the passage of small machines and tractors in the plantation possible. Clones UF 613, CC 42, and Catongo, which have a low canopy, pass this trait on to the progenies in crosses, suggesting some degree of dominance of the low over the high canopy (Soria, 1964 and 1977). Axil spot: The reddish pigmentation at the base of the leaf petiole, easily identified in plantlets and known as ‘axil spot’, is controlled by two complementary genes (A and B) (Harland & Frechville, 1927). The colour is expressed when both genes meet in dominant homozygosis. This trait represents an important morphological marker in genetic studies and has been successfully used for the investigation of the incompatibility reaction (Opeke & Jacob, 1969; Lanaud et al., 1987). The presence of axil spot is associated with the presence of red fruits. Cacaos with both dominant AB genes have red fruits (Harland & Frechville, 1927). Fruit characters: Colour: The colour of immature fruit, whether green or red, is under the control of gene R, with two alleles (R and r); the red pigmentation (RR and *R_) is dominant over the green (rr) (Pound, 1934; Soria & Esquivel, 1969 and Soria, 1977). The monogenic control, however, may be complemented by the action of modifying genes, as there are colour variations from green to red. The colour of immature fruit is highly important to discriminate racial groups (see details on these groups in Chapter 3). Red fruits have not been observed yet among the Amazon Forasteros, but are found in the Criollos and Trinitarios (Bartley, unpublished). * In R_ the underscore ( _ ) represents the absent allele. i.e. RR or Rr. Convolutions: This is a trait of quantitative nature, varying in expression from smooth to convoluted. The former expression is recessive in relation to the latter (Soria, 1977). Seed size: The heritability of this trait is still controversial. In Trinidad (Bartley, 1966), Ghana (Amponsah, 1969), Costa Rica (Soria & Esquivel, 1969), and Brazil (Pereira et al., 1987) the clone SCA 6, for instance, carrier of small seeds when crossed with Trinitario and Forastero clones from the Lower Amazon, transmitted this trait to the progenies, which suggests some degree of dominance for small seeds. However, in Ecuador, large seeds proved to be dominant over small ones (Alvarado & Bullart, 1961) and in Brazil, Trinitarian clones such as ICS 8 with large seeds were recommended for crosses with Forastero clones from the Lower Amazon for transferring high weight and seed size to their hybrid progenies (Pereira et al., 1987). Seed colour: The seed colour of cacao varies from white to faint pink in the Criollo racial group to purple in the Amazon Forastero group. Between these extremes there are the purple seeds of the Trinitario (see Chapter 3 for the characterization of these racial groups). Generally speaking, the Criollo fit into the category of fine cacao, while the Forasteros are classified as basic or ordinary cacao. Naturally, the Trinitario, from the hybridization of Criollos and Forasteros, is classified as semi-fine cacao. The content of the anthocyanin pigmentation seems to be determinant for the final cacao quality. The more anthocyanin, the more purple is the seed, and the worse is its industrial quality. In a study on the heritability of the seed colour done with ‘Catongo’ cacaos, a recessive homozygotic mutant for the absence of pigmentation, showed that this trait is controlled by a pair of alleles with complete dominance (Vello, 1972). Bean quality: Dry unfermented cacao beans present high and variable butter content (45% to 60%), second only to coconut. Almost the entire production of cocoa butter is earmarked for chocolate fabrication. The improvement practiced for the increase of the dry bean yield is also effective to increase the fat content (Pires et al., 1998). The heritability of fat content is quantitative, with an additive allelic interaction (Alvarado & Bullard, 1961 and Pires et al., 1998). |
| Character heritability; Resistance to diseases. | Return To Table of Contents |
|---|
| In spite of being a priority in many cacao improvement programmes, the genetic resistance to diseases has not really been worked upon and advances in this area are incipient. Reasons for this state of affairs, according to Zadoks, 1997, are inconsistent policies, lack of funds to finance research, discontinuity in the specialist staff, no standardization of resistance test methods, limited information and germplasm exchange, false concepts of resistance and absence of predictive tests. Considering that the selection of resistant genotypes is a complex enterprise, most programmes adopt the conception of accessing resistance indirectly, via selection for high yield. The breeder knows which are the main pathogens that occur in his/her local production zone and aims at resistance as protection against yield loss, considering that resistant genotypes are the most productive ones, with the smallest yield loss (see also Chapter 13). This strategy, however, does not exclude the selection practice for durable resistance (Van der Vossen, 1997; see Chapter 7 on resistance improvement). Conceptually, durable resistance is that which maintains its efficiency in a given cultivar which is widely cultivated for a long time in a disease-favourable environment (Johnson, 1983). Concepts of horizontal and vertical resistance originated from the study of the potato/Phytophthora infestans pathosystem (Van der Planck, 1968). Horizontal resistance proved to be polygenic, non-race specific, partial, and durable. On the contrary, the vertical resistance was oligogenic, race-specific, complete, expressed as a hyper-sensibility reaction and ephemeral. However, caution is advised for any extrapolation of these concepts of durable and non-durable resistance to other diverse pathosystems (Van der Vossen, 1997). This author alerts that the argument that highly resistant genotypes to all races of a given pathogen must be discarded based simply on the concept that they do not have durable resistance may be invalid for other pathosystems. There are examples of oligogenes that control durable resistance in some pathosystems. There are also reports that durable resistance was obtained by the pyramidization of Mendelian genes. Van der Vossen worked out an elegant revision on the strategies of incorporation of durable resistance into cacao, based on well-founded knowledge on the subject in other crops and cacao itself (see also Chapters 7 and 8 on resistance and biochemical and physiological bases of resistance). It is estimated that, on average, 20% of the worldwide production of cacao are annually lost to the five main diseases (Paulin & Eskes, 1995) of which the resistance heritability of three of the most important for Brazil are described as follows. (See also Chapters 7 and 8 which deal specifically with disease resistance). Pod rot: This disease caused by Phytophthora sp. can provoke yield losses from 5% to 90%, depending on the fungal species involved and the area of occurrence. The species Phytophthora palmivora, henceforth designated Pp-p, is of pantropical dissemination and causes, on average, a 40% worldwide yield loss of cacao. Phytophthora megakarya (Pp-m) is more aggressive, but its distribution is restricted to East Africa (Nigeria, Cameroon, Togo, and Ghana), causing about 10% annual production loss. In Papua Nova Guinea, the resistance to Pp-p seems to be polygenically controlled by genes of predominantly additive effects, and no genotype with complete resistance has yet been identified (Tan & Tan, 1990). The number of genes that control resistance to Pp-p in leaves was assessed at four (Warren & Pettitt, 1994). Witches’ broom: The occurrence of Crinipellis perniciosa, causal fungus of witches’ broom (WB), is restricted to the Neotropics and causes annual production losses of 20% to 90%, in more severe cases. The pathogen is endemic in the Amazon Basin, where it seems to have co-evolved with wild cacaos. Actually, it is found disseminated throughout the Amazon countries (Venezuela, Colombia, and Ecuador). In Brazil, it reached the State of Bahia in 1989 and affected the production of over 600 mil hectares of cultivated cacao (see Chapter 13 for a more ecological approach to the subject). This disease, due to its devastating impact, has been a major concern for cacao breeders for over half a century. Pound, 1943, undertook a botanic expedition to the Amazon Basin in search for WB-resistant genotypes. The genotypes collected stimulated the cacao improvement programme in Trinidad (Kennedy et al., 1987). Basically three methods are used to identify resistant genotypes: i) application of agar blocks with WB basidiospores to the meristems of plantlets or cuttings, as used in Trinidad (Laker et al., 1987); ii) use of calli formed in tissue culture to identify the extremes of susceptibility and resistance, in England (Muse et al., 1996); both methods work with a small number of plantlets; iii) selection of WB-resistant genotypes on a large scale, under greenhouse conditions in Brazil, using a semi-automatic system with inoculation of basidiospores, as proposed by Frias & Purdy, 1995. Although the exact relation between resistance expressed in plantlets and the resistance in adult plants is unknown, this last system has brought forth satisfactory results. Countless plantlets selected by this system remained symptom-free in field conditions when exposed to high natural infection rates over more than two years. An evaluation of 565 accessions of the CEPEC germplasm bank subjected to natural WB incidence (Pires et al., 1999) indicated a high variability for resistance, ranging from broom-free trees to trees with over 90 brooms. Mean and high resistance levels were observed in these accessions, including clone SCA 6 from the Upper Amazon, the supposed centre of origin of cacao (see Chapter 3). The observation of segregant families of crosses with SCA 6 as the parental clone (Soria, 1977) suggests that resistance is controlled by oligogenes, which is encouraging for WB resistance improvement. In Trinidad, the resistance level of SCA 6 has been maintained for over 60 years (Purseglove, 1968), however, it broke down in Ecuador. There are two pathotypes which cause WB. Pathotype A occurs in Ecuador, Bolivia, and Colombia, while B is found in Trinidad and Brazil. Curiously, pathotype A occurs in the area of the supposed centre of origin of cacao, from where it presumably co-evolved with the plant. Monilia pod rot: This is also a devastating disease, caused by the fungus Moniliophthora roreri, with an area of occurrence which covers all the western side of the Andes up to Central America. It is intriguing to note that this pathogen’s distribution area corresponds exactly to that of Criollo cacao. Maybe this fact explains why the pathogen has not yet attacked the Brazilian Amazon region, where the Amazon Forastero cacao is found. Together with Crinipellis, this disease destroyed the Ecuadorian cacao culture at the beginning of the 20th century. The fungus attacks the fruits, causing rot and produces spores with a viability of two to three months. Although the disease does not occur in Brazil, its dissemination across the Andean countries adjacent to the Brazilian Amazon, aggravated by the long viability of the pathogen, threatens the Brazilian cacao cultivation. As in WB, the control of Monilia by phytosanitary pruning and the application of fungicides is anti-economical and the only effective control is the development of resistant genotypes. There seems to be some association between WB and Monilia pod rot. Clone EET 233, for example, shows resistance to both diseases in Ecuador (Aragundi et al., 1988). Literature offers only scarce information on the heritability of Monilia resistance. It is assumed that it is incomplete and of quantitative nature, according to the revision of Van der Vossen, 1997. There is considerable variability for resistance to the main cacao diseases. Therefore, cloning has the leading part in the process, since it allows the production of resistant genotypes at any stage of the breeding programme (see clonal improvement in Chapter 9). Crossing for recombination of resistant genotypes or even backcrossing for the standard cultivar will broaden the parental resistance level. Eventually, the disease resistance, when duly linked up with other aims of the breeding programme, can result in an early, high-yielding cultivar, with vigour and a bean size and quality that surpass industrial requirements. |
| Estimates of genetic parameters. | Return To Table of Contents |
|---|
| As presented and discussed in Chapter 13, cacao breeding has to harmoniously integrate the asexual improvement (cloning) with the sexual (population improvement and hybrid). The new paradigm to be followed is that of forest improvement (Dias, 1993). This develops along three strong general lines: i) selection for the general combining capacity; ii) selection for the specific combining capacity; and iii) clonal selection (Resende, 1997). The first aims at the identification of individuals with higher genetic additive values, with a view to sexual propagation, by means of recurrent intra-population selection. The second is based on the total genotypic values of the individuals; a function of the additive values, the dominance deviation, and the epistatic effects, via reciprocal recurrent selection. The objective is, in this case, to improve the inter-populational hybrid, i.e. the heterosis resulting from the cross between two populations. Lastly, clonal selection capitalizes on the total genotypic values. Cacao, as in other perennial tree species, presents very particular physiological, reproductive biological, and agronomical aspects. The long productive and reproductive cycles, overlapping generations, the presence of genetic self-incompatibility in many of the populations, the long-term phenotypic expression of diverse traits with temporal alterations in their genetic control and the possibility of sexual and asexual reproduction differentiate the breeding of perennial tree species such as cacao. Furthermore, the requirements of vast experimental areas and long-time evaluations expose the trees to climatic adversities and to the attack of pests and pathogens, affecting the individuals’ survival rates. These aspects hamper the estimation of genetic parameters normally obtained in improvement of annual species. More accurate selection methods are therefore required for the improvement of perennial species such as cacao (Dias & Resende, 2002; Resende, 1999a; 1999b; 1999c and Resende & Dias, 2000). These authors evaluated the necessity for successive and consecutive evaluations over time in the same individual, the comparison of individuals of different generations under different environmental conditions, selection with a view to propagation by seed, selection aimed at vegetative propagation and selection that favours the tree at the expense of its group mean, by means of the prediction of its genetic additive and non-additive values. Finally, the use of accurate selection methods is also relevant for the treatment of unbalanced data, which is common in cacao research. For the first time in the genetic improvement of cacao, estimates of genetic parameters for different traits were compiled based on the results available in literature, (Table 6.2 and Table 6.3; the latter from data by Cilas et al., 1988; Lockwood & Pang, 1993 and Raboin, et al., 1993). The ratio of the dominance over the additive variance was included due to the importance of the hybrid for commercial cacao cultivation. The estimates presented in Table 6.2 were compiled from two trials conducted in Costa Rica: i) a topcross installed in La Lola in 1965, with 48 full-sib families, involving the cross among eight Trinitario clones, with large fruits and seeds, and six Amazon Forastero clones, with small fruits and seeds. The experimental design was a rectangular quadruple 7 x 8 lattice (including 8 clones), in four replications and 16 trees per plot, spaced 2 x 2 m (Soria et al., 1974; Moses & Enriquez, 1981; Enriquez & Soria, 1999a e 1999b) and ii) a complete 7 x 7 diallel installed in 1972 at two sites (La Lola and Turrialba) with 42 full-sib families obtained from the cross among seven parental clones from different genetic origins. A balanced lattice design was used, in four replications and six trees per plot, spaced 2 x 2 m (Lopez, 1984; Ramirez, 1987; Ramirez & Enriquez, 1988 and 1994 and Enriquez & Soria, 1999a and 1999b). To study the seed traits, this 7 x 7 diallel was reduced to a complete 4 x 4 diallel, for economical reasons, involving contrasting clones for these characteristics, especially for seed size (Lopez, 1984 and Lopez & Enriquez, 1988). In this last case, the heritability estimates were obtained at the plant level. This same diallel was reduced to another complete 4 x 4 diallel to study the components of industrial bean quality (Pardo & Enriquez, 1988). The theory on diallels is dealt with in depth by Hallauer & Miranda Filho, 1981; Vencovsky & Barriga, 1992 e Cruz & Regazzi, 1994). The estimation of genetic parameters in cacao is uncommon and subjected to restrictions (shown in Tables 6.2 and 6.3). A first limitation is the questionable genetic representivity of the clones used. The parental clones used for the study of full-sib families are often selected samples, which would allow the estimation of mean square components (Dias & Kageyama, 1995) but never the estimation of variance components (Hallauer & Miranda Filho, 1981). A second is that the process of estimation is not born from the needs of improvement programmes, nor have any adequate designs been drawn up to meet this goal. On the contrary, it was rather developed to meet academic objectives. It is obvious that estimates of genetic parameters must be obtained routinely, to realign the diverse stages of improvement programmes. A third restriction is the poor skills of cacao breeders in dealing with biometric genetics. This aspect is responsible for confusing and sometimes incoherent results. See, for example (Table 6.2), that the narrow-sense heritability for fresh seed weight was 0.89, based on one annual harvest, and 0.17, based on three accumulated annual harvests. In fact the opposite should occur. Multiple observations throughout several annual harvests would remove a great part of the environmental effects and generate heritability estimates of a greater magnitude than the ones obtained based on a single annual harvest. Finally, the few articles that mention estimates of genetic parameters of cacao do so without clearly defining at which level they had been obtained: at the plot level (total or means of plots), or the family means, or even at the level of individuals. The post-planting age is also rarely mentioned, which is an important factor because, in perennials principally, the genetic control of yield components is age-dependant. A less attentive reader might conclude that most traits in cacao are highly heritable and controlled by additive gene effects and therefore they would be easily manipulated by breeding (Table 6.2 and 6.3). This is wrong and delusive. Most of these estimates of genetic parameters were obtained at the plot mean level, which is the reason for the high magnitudes observed and the broad as well as restricted heritabilities. Nevertheless, it is sufficient to observe that both heritability estimates obtained at the plant level for seed traits (fresh weight, dry weight, number, length, width, and thickness, in Table 6.2) were low, compared to estimates at the plot level obtained for the same traits. For these same seed traits evaluated at the plant level (Table 6.2), the dominance errors were more important than the additive effects. This demonstrates the poor practical applicability of the estimates obtained at the mean level for drawing conclusions on the genetic control of a given trait. Besides, estimates of genetic parameters at the mean plot level limit the comparison among studies that use different numbers of replications and plot sizes. For these cases, and also for the prediction of genetic and genotypic values, the estimates of heritability and repeatability at the plant level are ideal. Recently in Brazil, heritability and repeatability estimates at the plant level have been obtained for the fresh yield and its components, using, for the first time in cacao, mixed models (Table 6.4). Data from an almost complete 5 x 5 diallel were used for the estimation of these parameters (Dias & Resende, 2002; Resende & Dias, 2000) by the evaluation of 20 full-sib families. The experiment was set up in 1975 in Linhares, State of Espirito Santo, in complete randomized blocks in four replications and 16 plants per plot. Accumulated annual yield data from 1981/82 to 1984/85 (sixth to tenth stand year) were used for the analyses. The narrow and broad-sense heritability estimates were of low magnitude. The dominance errors were significant, confirming the heterosis manifested earlier in these cacao hybrids (Pereira et al., 1987). Besides, how the relative importance of the non-additive overrides the additive genetic effects for the yield components in cacao, such as the number of harvested and healthy fruits and the wet seed weight per plant, was demonstrated by Dias & Kageyama, 1995, based on the square means of the effects. The estimates of repetition obtained were modified (Table 6.4). Repetition estimates of the same order were obtained for the same yield components in 43 randomly selected trees on a commercial plantation in Linhares, ES, under cultivation for over 50 years (Dias & Souza, 1993; see Table 6.4). In this last study, data from three successive production years (1986 to 1988) were used to estimate the repeatability. |
| Predictive biometric genetics. | Return To Table of Contents |
|---|
| What is notable among the main contributions of biometric genetics to the improvement of perennial plants is the possibility to predict the actual additive genetic values (as with sexual propagation) and genotypic values (regarding vegetative propagation) of all individual selection candidates. The quantities mentioned are random variables, unknown and unobservable, which breeders could predict by functions of observable variables or phenotypes. |
| The fundamentals of prediction, Prediction of additive and genotypic effects, Best linear unbiased prediction (BLUP) and Residual Maximum Likelihood (REML). | Return To Table of Contents |
|---|
| The fundamentals of prediction: As the phenotypic observations of a given quantitative trait in a given environment are functions of the genotypes and the environment and as the genotypes are changed by segregation due to sexual propagation, this highlights the need to adopt robust genetic models for an accurate evaluation of the alleles, in terms of their effects, and of the plant genotypes subjected to selection. Such models permit the derivation and utilization of sophisticated genetic-statistical methodologies, which, in essence, allow the elimination of environmental and segregation influences associated with the phenotypic expression of the traits. Generally speaking, the non-use of this approach in selection would make improvement extremely inefficient and less probable. In the following, based on Resende, 1999b, some basic principles for the prediction of additive and genotypic values are presented. The phenotypic value (y) of a given quantitative trait in a given panmitic population, ignoring epistasis, is given by: y = µ + a + d + and (1) where µ = genotypic mean, a = additive genetic effect, d = dominance effect, e = environmental effect, µ + a = additive genetic value, µ + a + d = genotypic value, and g = a + d = genotypic effect Based on model (1), the phenotypic variability of the population can be split in: σ2y = σ2a + σ2d + σ2e (2) where σ2y = phenotypic variance σ2a = additive genetic variance σ2d = genetic dominance variance σ2e = environmental variance σ2g = σ2a + σ2d = genotypic variance Evaluating the quantitative trait more than once in the same individual, model (1) is expanded to y = µ + a + d + p + et (3) where e = p + et p = permanent environmental effect et = temporary environmental value µ + a + d + p = permanent phenotypic value In model (3) applies σ2e = σ2p + σ2et, where σ2p and σ2et, refer to the variances of permanent and temporary environment, respectively. In association with the models (1) and (3), the following individual narrow-sense heritability parameters (h2) are defined
and the individual repeatability (ρ)
When the families are implanted in plots containing several individuals, model (3) is extended to y = µ + a + d + p + et + c (4) where c expresses the permanent plot effect, with variance σ2c. Based on model (4), the intraclass correlation parameter between individuals is defined (c2), which is due to the common environment of the plot
and repeatability is defined as
The quantities of primary interest for the breeder are the additive genetic values, when the objective is the sexual propagation of the selected individuals, or of the genotypic values, when the objective is asexual propagation. Since these quantities are unknown random variables, they have to be predicted by phenotypic values. In general, the principle of linear regression is used for the prediction of additive (a) and genotypic (g), effects, i.e., the technique of regression allows inferences on the proportion in which y explains a and g, according to the predictors presented in the following, which are associated to mass selection
where ßa,y and ßg,y are the regression coefficients of the additive and genotypic effects on the phenotypic value y, respectively, which are equal to the individual narrow and broad-sense heritabilities, respectively. If repeated measurements (m) are realized for each individual, the predictors of a and g, based on the phenotypic mean value (yˉ) of the various measurements are given by
This confirms the importance of the parameters h2, h2a e ρ for the prediction of the genetic values. In this context, parameter c2 is also relevant, especially when experimental designs with several plants per plot are used, as seen in the following items. Prediction of additive and genotypic effects: Considering the biological peculiarities, particularly the reproductive system and the methods of cacao improvement, it can generally be seen that the basic experimental material for selection refers to tests with full-sib families obtained by factorial or diallelic crosses. The two main steps that must be taken in the prediction of genetic values are the adjustment of data to the identifiable environmental effects [E(y)] and the posterior multiplication by a regressor (see ‘The fundamentals of prediction’). Following this same principle and considering the complete randomized block and factorial cross experimental design, the following predictors are established for the additive and genotypic effects, according to Resende, 1999a, based on the definition of the following phenotypic quantities: Yˉ = general mean fi = (Yˉi… - Yˉ…) = effect of the general combination capacity of female parent i mj = (Yˉ.j… - Yˉ….) = effect of the general combination capacity of male parent j fmij = (Yˉij.. - Yˉi... - (Yˉ.j… + Yˉ….) = effect of the specific combination capacity among the parents i and j eijk. = (Yˉijk. - Yˉ..k. - (Yˉ..k. + Yˉ….) = effect of plot ijk dijkl = (Yˉijkl - Yˉijk. = effect of the individual within the plot yˉi… and yˉ.j… = means of the full-sib families of the parents i and j, respectively yˉij.. = means of the full-sib families from the cross between the parents i and j yˉ..k. = mean of block k yˉijk. = mean of plot ijk yˉijkl = individual value The multi effect index for the prediction of additive effects of the individuals of the experiment, assuming n plants per plot, b block, m male and f female parents, according to Resende, 1999c, is expressed by â = b1fi + b2mj + b3fmij + b4eijk + b5dijkl where
The additive effects of the female parents can be predicted using the index: â = b6( yˉi... - yˉ….) where
The additive effects of the male parents can be predicted through the same expression as â, but substituting m by f and yˉi... by yˉ.j... For the prediction of the individual genotypic effects, the following index must be used ĝ = (b1 + b7) fi + (b2 + b8) mj + (b3 + b9) fmij + (b4 + b10) eijk + (b5 + b11) dijkl where
The multi-effect indices presented for the a and g prediction explore, besides the general and specific combination capacity of the parents, Mendelian segregation within the progenies, allowing a selection of superior individuals within the experiments. The indices presented can be applied to data of only one harvest, to the mean yield of several harvests or to the total accumulated yield of several harvests, if the parameters (h2, h2a e c2) linked to the respective situations are considered. Best linear unbiased prediction (BLUP): The predictors described in ‘Prediction of additive and genotypic effects’ are only adequate (unbiased) in situations where the data are balanced (same number of crosses per parent, same number of plants per progeny, same number of harvests per individual, etc.). However, since the survival rate of individuals in experiments with cacao is normally not 100% and the cross designs are unbalanced, alternative procedures of genetic analysis are needed. The procedure denominated BLUP was developed by Henderson, 1949, and later formally presented by Henderson, 1973. Such a procedure allows a simultaneous adjustment to the identifiable environmental effects and the unbiased prediction of the genetic values. Only in the late 1980s BLUP came to be used in animal improvement practices, on a worldwide level, with the progress of computer technology. In plant improvement in Brazil, it began to be applied from 1993 on in forest improvement (Resende et al., 1993) and is as yet little used beyond this area (Resende & Dias, 2000). However, in view of its notable theoretical and practical properties, it is likely that BLUP will come to be a routinely used methodology in cacao improvement. BLUP presents the following desirable statistical and practical properties: Maximization of the correlation between the predicted genetic values and the true ones, i.e., maximization of selective accuracy; Minimization of the difference square between the predicted genetic values and the true ones, i.e., minimization of the prediction error; Maximization of the probability to select the best from any two individuals; Maximization of the probability to select the best amongst various individuals; Maximization of the genetic gain per selection cycle, and; Unbiased prediction of the genetic effects. For the situation described in ‘Prediction of additive and genotypic effects’, Resende, 1999a, presented the equations for the prediction of the additive (â), the dominance (^d), the plot (ĉ), and the block effects (^b) by the individual BLUP procedure, given the matrixes of incidence X, Z, and W for b, a (d), and c, respectively, with y as data vector, as expressed below:
where
In the balanced case, the equations of the mixed model in (5) lead to the same results as the predictors â (item 6.7.2). Furthermore, the equation ĝ = â + ^d is valid. Consequently, the multi-effect index presented in 6.7.2. is the BLUP for the balanced case. When several evaluations per individual are carried out, the linear mixed model is given by y = Xf + Za + Zd + Wc + Tp + c (6) where p refers to the permanent environmental effect and T to the matrix of incidence for the cited effect and f is the vector of fixed effects containing the block and measurement effects or the combinations block-measure. The equations of the mixed model (EMM) can be expressed as
where
Restricted maximum likelihood (REML): The prediction procedures of genetic values assume that the genetic parameters are exactly known, and the properties of the predictors are only assumed under these conditions (Henderson, 1984). However, in practice, such parameters are estimated and therefore require the greatest possible accuracy of estimation to adequately substitute the parametric values. In this context, the recommended procedure for the estimation of variance components is REML, described in detail by Searle et al., 1992. The REML method allows an efficient utilization of the data obtained in the improvement programmes, which, by default, are usually unbalanced. With the arrival of the methodology of the mixed models or individual BLUP, a great change occurred in thoughts about the estimation of variance and genetic parameter components. Earlier, the covariances among parents had been estimated and interpreted in terms of their mathematical expectations (equalling them to the observed values), giving rise to the variance components. Actually, the variance components can be estimated directly by the variances of the random effects of the mixed linear model. In this context, the REML method as an individual model came to be the standard model for the estimation of variance components and genetic parameters based on unbalanced data. This method is preferred for the sake of its superior statistical properties compared to the properties of the least-square and maximum likelihood estimators (Searle et al., 1992). The application of REML depends, essentially, on the equations of the mixed model associated to the linear model and the algorithms associated to the estimators. Considering the linear mixed model presented in the item BLUP - ‘Best linear unbiased prediction’, associated to the evaluation of full-sib families (obtained in diallelic crosses) and their parents, with several measurements per individual, the REML procedure is described in the following: The initially attributed values to the variance components in (7) lead to the prediction of the effects a, d, c, and p. Calculating the variances of these predicted effects establishes the additive genetic variances (^σ2a),dominance variance (^σ2d), and variance among plots (^σ2c) and of the permanent environment (^σ2p). These will probably differ from the initial values used in (7), which means that the initial values were not plausible or verisimilar. Thereby, the EMM must be solved again, using these calculated variance components. Proceeding successively like this, the values of the variance components convert, in other words, it is found that the values used in the EMM become equal to the proper variances of the predicted effects. This means that the values used in the EMM became plausible or verisimilar to the data set. This is the principle of likelihood. The REML estimators to obtain the random effect variances, using the EM algorithm of Dempster et al., 1977, are
where tr = matrix trace operator r(x) = posto of matrix X N-r(x) = degrees of freedom of the error q = number of individuals s = number of plots N = total data number where C22, C33, C44, and C55 are derived from
Together with the mixed model at the individual level, the REML method estimates the variance components directly with the variances of the random effects of the model, but considers moreover all parentages (matrixes A and D) among the individuals in evaluation. The systematic or fixed environmental effects of the model are also considered in the estimation so that the variance components are valid within and among each level of fixed effects. Summing up, the repetition of the variance components (variance of the predicted random effects) in the equations of the mixed model allows a verification of the likelihood of particular numeric values of the parameters, in other words, allows to prove the plausibility that the data had been/were sampled from a population with evaluated parameters. Based on the models of the presented predictors and estimators, the following parameters are estimated : individual narrow-sense heritability (h2)
individual broad-sense heritability (h2a)
individual repeatability (ρ)
Correlation due to the common environment of the plot (ĉ2)
Correlation due to the permanent environmental effect (^p2)
Determination coefficient of the effects of dominance associated to the full-sib families (^d2)
|
| Application of REML/BLUP procedures. | Return To Table of Contents |
|---|
| In order to illustrate the application of the REML/BLUP procedure in the routine of a cacao improvement programme, data from individual measurements relative to the component number of healthy fruits per plant, evaluated in the 5 x 5 diallel (described in ‘Estimates of genetic parameters’) were used. In total, 5,888 observations were utilized. The results presented here, and published by Resende & Dias, 2000, include the 20 full-sib families and three of the respective parents, represented by their correspondents of generation S1. Means of the number of healthy fruits per plant trait: Cacao presented highest productivity means in the first and second harvest years (1981/82 and 1982/83). In the third year (1983/84) there was an expressive productivity drop, which was largely recovered in the following cropping season (Table 6.5). Estimates of genetic and phenotypic parameters: The estimates of narrow-sense (^h2) and broad-sense heritability (^h2a), of the correlation due to the common environment of the plot (^c2) and the determination coefficient of the dominance effects together with the full-sib families (^d2) (Table 6.6) for the component number of healthy fruits reveals that the greatest expression of additive genetic variability (highest h2) occurred in the years of greatest productivity (1981/82 and 1982/83). Wide annual oscillations in the expression of the genetic variation evaluated by means of the genotypic determination coefficient between means of cacao cultivars have also been observed in Linhares, ES, (Dias et al., 1998). In these cases, estimates of high magnitude of this coefficient were obtained in climatically favourable production years. The presence of allelic dominance was detected in the loci that control the trait number of fruits, since the h2a were greater than the h2 estimates (Table 6.6 and Table 6.7). The allelic dominance for the yield component- controlling loci of cacao, such as the number of harvested and healthy fruits and the weight of wet seeds per plant was also demonstrated by Dias & Kageyama, 1995, based on the mean square effects. Based on the data of four consecutive evaluation years (Table 6.7) it is concluded that the trait number of healthy fruits per plant presents low heritability and dominance (h2a > h2), a fact that contributes also to the presence of heterosis. The trait repeatability presents a relative high magnitude, close to the one obtained by Dias & Souza, 1993, indicating that it is not necessary to evaluate each individual during many years for the selection effect. If the efficiency of the maximum utilization of m successive harvest years is given by [m/[1+(m-1)ρ]]1/2 instead of 1, efficiencies of 1.13; 1.18, and 1.21 are obtained when taking 2, 3, and 4 years into consideration, respectively. Hence, the selection based on the use of 4 successive harvest years should provide about 21% more gain than selection based on only 1 year. On the other hand, considering 4 years instead of 3 brings only 3% more (24%) genetic gain. Three years of successive harvests seems to be a reasonable number to work with, allowing for about 90% accuracy in selection. Accuracy, in this case, is expressed by [mρ/[1+(m-1)ρ]]1/2. The repeatability coefficient estimated by the variance analysis, by the factorial model, and in experiments is also used to determine the minimum necessary period to evaluate the selection potential of superior cacao hybrids in yield and its components. For the south of Bahia, this period was estimated at 2 years of successive harvests, with an accuracy of 90%, when evaluating hybrid trees of 11 field years (Dias & Kageyama, 1998b). It was also verified that the repeatability estimates as correlation between successive measures of the same genotype assume equal values, whichever the applied model for its estimation (Carvalho, 1999). The values 0.10 for h2 and 0.38 (~= 0.40) for h2a (Table 6.7) allow the inference that the selection experiments must contain about 60 individuals per family in progeny tests and around 15 ramets per clone in the clone tests, according to evaluations Resende, 1995 carried out. Note that, when dealing with progeny tests, a minimum of 64 trees per family is routinely included in cacao improvement (Pereira et al., 1987 and Dias & Kageyama, 1995). In the following, the estimated parameters ^h2, ^h2a, ρec2 were used for the prediction of the additive, dominance, and genotypic effects. Predicted additive, dominance, and genotypic effects: Among the 20 best individuals for sexual propagation (sorted by â), only nine were found among the 20 best for vegetative propagation (regulated by ^g) (Table 6.8). This fact illustrates the importance of prediction and sorting in function of the propagation system to be adopted. The 20 best individuals for sexual propagation do not include any of the parents (Table 6.9 and Table 6.10). Genetic gains associated to different selection/propagation systems: Selection aiming at vegetative or clonal propagation is the most efficient, generating genetic gains of over 100% (Table 6.11). Selection of the superior hybrid (biclonal cross) results in practically the same gain as the selection of the 10 best individuals of the experiment aiming at sexual propagation. Although the latter system does not explore the specific combination capacity, the selection of the 5 best individual results in a higher gain than selection of the best hybrid (Table 6.11). Therefore, the best way to exploit the specific combination capacity is by clonal selection through genotypic values, which is a large advance over the other two systems. In fact, the inferior confidence interval limit of the improved population mean indicates that the ideal numbers of individuals to be selected in each system are 20 individuals for sexual propagation, three hybrids for biclonal crosses, and 10 individuals for asexual or clonal propagation. It was also verified that if the selection is based on the inferior confidence interval limit, biclonal hybrid selection turns out to be the worst of the system. Additive genetic values are also useful in the planning of crosses for an evaluation in the next selective cycle. In this case, the individuals with the highest genetic values must participate in a larger number of crosses, adopting, for example, the 4 x 6 rectangular factorial scheme, where the four best individuals would be crossed six times and the six subsequent individuals in the evaluation would be crossed four times. Advantages of the use of the mixed model methodology: The REML method together with the mixed model methodology (BLUP) is a flexible instrument for the estimation of genetic parameters and presents the following advantages: It can be applied to unbalanced data; It allows the simultaneous use of a large quantity of information provided by different generations, sites, and ages, providing more precise estimates; It does not require data obtained under rigid experimental structures, but can be applied to data obtained normally in the improvement programmes, not necessarily linked to experimental designs, requiring only information on the genealogy of the individuals; It allows of an estimation of the dominance and epistatic effects as well as the additive effects as it uses a higher number of parental relations; It allows for the estimation of selective accuracy and of the confidence interval of the predicted genetic value for each individual; It allows for the adjustment of the various alternative models, so the one that adjusts best to the data can be chosen, while it is also parsimonious (has a lower number of parameters); It permits the development of national programmes of genetic evaluation, where the individuals that belong to different improvement programmes can be compared, a fact that allows a unique organization for each country, or better, organization for each environment, but encompassing all individuals under evaluation in the country, even those that are not being directly evaluated in the specific environment. This allows for an elevation of the global improvement rate of the species. In view of these advantages, the estimation of genetic parameters in cacao should be based on the REML/BLUP procedure under the individual model. The application of mixed models will certainly mark a new era in cacao improvement, since data sets that were never considered before can now be evaluated, competing for a greater efficiency of improvement programmes. This greater efficiency is result of the better access to information on genetic parameters and a more precise prediction of the additive and genotypic effects, a fact that will lead to a greater selection accuracy and, therefore, a higher rate of genetic gain. |
| Multi-character selection index, Bayesiana inference and Heterosis prediction. | Return To Table of Contents |
|---|
| Multi-character selection index: The importance of simultaneous improvement for various traits in cacao to achieve “improvement for total quality” is described in Chapter 13. In this case, the diverse traits must necessarily be considered according to their respective weighting or economic values. According to the classical theory of selection indices, the various traits must be considered simultaneously according to their phenotypic values, their heritabilities, correlations, and economical weightings. However, with the advent of the prediction procedures, the actual predicted genetic values for the various (n) traits can be chosen directly for their economical weightings (b), so that the aggregated genotypic value (H) is obtained which, in this case, is given by…. <^H = b1â1 + b2â2 + ... + bnân. Once ^Hi is established for each individual i, a new ordenamento of the individuals must be realized for the selection effect. It is important to mention that when the predicted genetic values are used, the heritabilities of the various traits had already been taken into consideration. The correlations among the traits are only useful when certain traits can assist the selection of others. When auxiliary traits are used in the improvement, the gain in precision can be quantified by the ratio of the multivariate (h2m) and univariate heritabilities (h2). In situations where gains in precision are significant, the multivariate prediction of the genetic values must be chosen. In selection using auxiliary traits, the accuracy is defined by hm, i.e., the square root of multivariate heritability. The multivariate heritability is given by (Resende and Rosa-Perez, 1999):
where h2x and h2y refer to the heritabilities of the auxiliary trait (x) and the target trait (y), respectively, and rAxy and rFxy refer to the genetic and phenotypic correlations between the two traits, respectively. In this case, the efficiency or gain in accuracy with the multivariate selection is given by the ratio between the multivariate and univariate accuracies, i.e.:
This expression confirms that (Resende & Rosa-Perez, 1999): if hx/hy rAxy = rFxy, the efficiency is 1 and the multivariate is equivalent to the univariate analysis; When the heritabilities are equal (hx = hy), the efficiency depends only on the difference between the genetic and phenotypic correlations. When the phenotypic correlation is equal to rFxy = rAxyhx hy + rExy[(1-h2x)(1- h2y])1/2, where rExy is the environmental correlation between the two traits, then rAxy - rFxy = rAxy - { rAxyhx hy + rExy[(1-h2x)(1- h2y])1/2. This last expression verifies that the maximization (which means maximizing the efficiency of the multivariate selection) of the difference between rAxy and rFxy, can only occur through the maximization of the difference between rAxy and rExy. Thus, a gain in accuracy depends on the absolute difference between the genetic correlation and the environmental correlation among the traits; When the heritabilities are equal (hx = hy) and the genetic and environmental correlations are also equal (rAxy = rExy), the multivariate analysis is equivalent to the univariate analysis, too; For the cases in which the genetic correlation is superior to the phenotypic correlation, the gain in accuracy is the higher, the greater the heritability of the auxiliary trait in relation to the heritability of the target trait of the improvement. Bayesian inference: A deficiency of the procedure REML/BLUP for the estimation/prediction of variance components/genetic values lies in the fact that the REML method provides only approximate confidence levels for the genetic parameters, by means of approximations and suppositions of asymptotic normality. This is because the distribution and variance of the estimators are not known and questions regarding the efficiency of the selection can therefore not be answered with certainty. The Bayesian analysis is based on the knowledge of the distribution a posteriori of the genetic parameters and allows a construction of the exact confidence levels for the estimates of the genetic parameters. Hence, this analysis provides a more complete description of the reliability of the genetic parameters than the REML method (Resende, 1997 and 1999b). The statistical Bayesian inference is based on the conditional distribution of parameter (Θ) given the data vector (y), that is, in the distribution a posteriori of the parameter, given the phenotypic observations, which is equal to:
where: f(yΘ) - probability density function of the conditional distribution of an observation (y) given Θ (called likelihood function or data model). f(Θ) - probability density function of the distribution a priori, which is also the marginal density of Θ. This function shows the knowledge degree accumulated on Θ, before the observation of y.
marginal or predictive distribution of y regarding Θ, where R is the amplitude of the distribution of Θ. It is hence confirmed that the distribution a posteriori is proportional to likelihood x priori, i.e., the likelihood function connects the a priori to the posteriori using the experimental data (observations). Through this, the distribution a posteriori observes the previous knowledge degree on the parameter (Θ) and also the additional information provided by the experiment (y). For the situation in which the distribution a priori as well as the distribution of information are normal, that is, Θ ~ N (u, r2) e (Y|Θ) ~ N (Θ, σ2) with known σ2, the distribution a posteriori of Θ is also normal, in other words (Θ |Y = y) ~N (u1, r21) where:
Using this a posteriori distribution, it can be demonstrated that the basic fundaments of prediction of genetic values are essentially of Bayesian nature, as initially presented by Robertson, 1955. The genetic additive effects are defined as errors and a population of genetic effects presents zero as mean and σ2a as variance. The best prediction of the genetic effect of a given individual without any information, taken randomly from the population, is the population mean u, which can be taken as the a priori estimator, whose variance is σ2a. Taking an information of an individual, a second estimator of the genetic effect is the phenotypic error (y-u) in relation to the population mean, which has the variance σ2e = σ2f - σ2a. These two independent estimators can be combined linearly in the best possible form, taking the reciprocals of the respective variances as weights. Under the Bayesian point of view, the expectation of the distribution a posteriori corresponds to the mean weighed by the precision, the means a priori and of likelihood. That is:
as under the focus of the classical statistic inference, where h2 is the heritability. This is also (considering y as the mean of the selected individuals) the estimator for the genetic gain mass selection. In the Bayesian context, the equations of the mixed model (designated mixed model equations of Robertson) are given by:
where r1 = E(f) and 0 = and (a). Taking the information a priori on the fixed effects as non-informative (expressed as S → ∞ and thus S-1 → 0) , the resulting equation is equivalent to the equations of the mixed model (EMM):
where: R = Iσ2e; G = A σ2e e S = Var(f). The basic problem of the implementation of the Bayesian analysis lies in the numeric integration (in the parameter space) of the probability density function a posteriori. Such an integration is, by analytical methods, impossible in practice. Hence, the procedure of stochastic statistical simulation has been used, designated Gibbs sampling (Gs), to make the Bayesian estimation viable (Sorensen, 1996). Besides the assumed (normal) distributions for the random effects (a) in the classical linear model and for the likelihood observation vector (y), the Bayesian approach requires attributions to the a priori distributions of the fixed effects and variance components. In general, non-informative or uniform a priori distributions are attributed to the fixed effects and variance components as a form of characterizing vague a priori knowledge on the referred effects and components (Gianola & Fernando, 1986). When using a priori non-informative distributions for the fixed effects and variance components it is found that the modes of the marginal a posteriori distributions of the genetic parameters correspond to the estimates obtained by REML and generate variance estimate components by the variance estimation from integrated likelihood (VEIL), presented by Gianola & Foulley, 1990. The great advantages of the Bayesian analysis, in this case, are the exact standard errors and confidence levels for the genetic parameters, as well as more precise estimates. Resende, 1999a, using the approaches REML and GS, obtained estimates of 0.32 ± 0.06 and 0.34 ± 0.005, respectively, for the restricted-sense heritability of the trait cylindrical volume in eucalyptus, confirming the great precision of the GS method. Regarding the estimation of the fixed effects (block and years of measurement effects) and the prediction of the random effects (genetic values), it is found that the means of the marginal a posteriori distributions of the site parameters (fixed and random effects), given the dispersion variance components or parameters, are equivalent to the solutions of the equations of the mixed model of BLUP as long as: non-informative prioris are attributed to the fixed effects, normal prioris to the random effects, and normal likelihood to the observations vector. In this case, the advantage of the Bayesian approach is the achievement of genetic values predicted with smaller standard errors. Summing up, the Bayesian analysis gives more precise estimates of variance components, genetic parameters, genetic values, and genetic gains. Additionally, the Bayesian estimation allows an exact analysis of samples of finite size. This last aspect is very important, especially for improvement programmes based on sets of unbalanced data, where this approach elegantly analyses the finite sample, which would not be possible by the classic methodology of mixed models. Heterosis prediction: Cacao improvement practiced by the main producing countries made expressive gains with breeding for hybrids (particularly Brazil) and recently, on a smaller scale, with asexual improvement (especially Malaysia and Indonesia). Paradoxically, the improvement of populations was neglected. It is not a surprise that the anxiety to synthesize new hybrids has come down to a frenetic and empiric search for new parental clones that could produce new hybrids, which when crossed, would be better than the former. However, for the success of hybridization there is still one open question: the choice of the parental clones. From the point of view of genetic improvement, this question can be approached from two angles: the a priori and the a posteriori choice of parents (Baenziger & Peterson, 1992). The first is based on the per se performance of the parents and includes selection methods based on the mean parent values, on the parentage coefficient, the complementation of the phenotypic expression of the traits, on the geographic distance, and on the analysis of the parental distance. The second possibility chooses parents by the evaluation of their progenies, using data from 1 and F2 generations. Owing to its potential of prediction and viability pointed out earlier, only the a priori choice based on the parental distance, calculated by phenotypic (Dias & Kageyama, 1997a) as well as molecular data (Dias et al., 2003) will be discussed here. The predictive nature of this approach avoids carrying out hundreds of undesirable crosses. Thus, only the predicted promising crosses are realized, saving time, cost, and materials. The association between genetic divergence between parents and heterosis in their hybrids is theoretically well-established, illustrated for the case of a simple gene (Falconer, 1989). By the biometric expression of the heterosis (H), considering the joint effects of all loci as being the sum of its individual contributions, we have:
This shows that the heterosis quantity in the F1 hybrid between two inter-crossed populations depends on the square of the difference of the gene frequency (y), which is a measure of divergence between these two populations and of the dominance errors (d) for the considered trait. So, if the association divergence-heterosis occurs at a high degree, it is possible to use the divergence estimate as a solid criterion for the selection of parents and, subsequently, for the synthesis of heterotic hybrids. Dias & Kageyama, 1997a, quantified the degree of association between the multivariate genetic distance (DM) (Figure 6.1) among five parental cultivars, evaluated for five yield components during five consecutive years, and the mean and heterotic performance of their corresponding hybrids. The frequency of the heterotic hybrids and the greater magnitude of the heterosis for the evaluated yield components were higher in the crosses between moderately divergent parents. The values of the Spearman rank correlations between the parental distance, estimated by means of the D2 statistics of Mahalanobis, and the mean performance of the hybrids for the weight of wet seeds per plant (PSUP, in kg/plant) and weight of wet seeds per fruit (PSUF, in g/fruit) were positive and significant. The heterosis percentage for PSUP and PSUF was also linearly associated to D2 (Table 6.12). Accordingly, it is possible to use this strategy to identify heterotic clone groups, even before realizing the crosses. It is, however, necessary to test the potential of this predictive methodology in diallels involving a greater number of parents. Simultaneously, it was demonstrated that this divergence presents a relative degree of stability over the years and can be calculated using data from a single year under crop-favourable climate conditions (Dias & Kageyama, 1997b). The statistics of multivariate distances that use the matrixes of residual variances and co-variances, as for example D2 and the Euclidean distance obtained based on canonic variables, were shown to be the most robust and adequate to calculate the parental distance aiming at the prediction of heterotic hybrids (Dias & Kageyama, 1998a). In all these studies, the data used to calculate the genetic distance were obtained from a 5 x 5 diallel, analyzed in detail in terms of genetic mean components ((Dias & Kageyama, 1995). Following the same line of research, the genetic distance calculated from RAPD data (DG) (Figure 6.2) obtained from the same five parental cultivars was used to predict the mean and heterotic hybrid performance (Dias et al., 2003). The distance matrix (DG) was calculated using 130 RAPD bands. The correlation between DM and DG was 0.67 (P < 0.03). Therefore, the two matrixes of parental distance (DM and DG) were linearly related to each other and also to the mean performance of the hybrids for PSUP and PSUF. The heterotic performance for these same components also appeared correlated to the parental distance (Table 6.12). For the first time in cacao, the viability of the joint prediction of heterosis by molecular RAPD data and phenotypic data of yield components was demonstrated. Certainly, this is a research line to be intensified, and, if all its predicted potential were confirmed, it should be applied on a large scale for the parental clone selection stage. |
| Selection of commercial hybrids and Participatory breeding. | Return To Table of Contents |
|---|
| Selection of commercial hybrids: In the selection of the best hybrids for planting in a given region, not only the superior per se performance must be taken into account, but also its yield stability over time (Dias et al., 1998 and Carvalho, 1999). Any analysis methods of adaptability and stability can be applied to evaluate the temporal stability of the cultivars (see Chapter 12). There are indications that the improvement realized for yield increase also contributes to a greater temporal stability of the cultivars. Information on temporal stability is important for the cacao farmer interested in the yield stability of the crop on his farm, using a cultivar with stable productivity over the years. This same information will also be important in the future, when the recommendation for cultivars in a given region considers the heterogeneity of the series of years (Dias et al., 1998). Participatory improvement: The task of recovering from the Brazilian post-witches’ broom cacao cultivation period requires changes in direction and attidude (see Chapter 1). So far, all improvement has been practiced only by CEPEC breeders, where the cacao farmer was a mere spectator, completely dependent on the former. The breeder produced the hybrid seeds that were handed out for planting to the producers, who did not even know what was being distributed. All orientation on field management was offered by technical advisors who, in turn, were trained by these same breeders. On the other hand, the breeders only cared about their own world, ignoring significant progress in improvement in other diverse crops. A notable example was the constant disinterest shown towards cloning. This extremely useful technique in perennial crops, which allows the perpetuation of genotypes of interest, was ignored by the cacao breeders for decades. This model has become obsolete and must be revised. The recovery of the Brazilian cacao cultivation is not a task for a single institution. Errors and success must be shared with other institutions and researchers. Neither can the farmer be left out of the process. He/she must be involved and committed in the various stages of the improvement programme. Human and financial resources are not enough that one could do without the farmer’s support. A participatory improvement structure must be created, where part of the work would be realized by the cacao farmers. Their tasks could involve cloning, selection, or even yield trials. An attempt of this kind brings farmers and breeders closer together, awakens an active participation of the farmer to notice the differences between genotypes in a trial, and offers a guarantee that the programme’s end product will not be rejected by the farmers. Currently, with the arrival of witches’ broom (WB), a unique opportunity exists to practice participatory improvement, following the example of other countries (Witcombe et al., 1996). There is a broad variability for WB resistance, especially on plantations with hybrids recommended since the 70s. These combinations, because they have the genetic population structure F2, segregate a lot, and in recombination with the local cultivars have amplified the crop diversity. Thus, the cacao farmers can select trees that appear WB symptom-free, clone and test them, and renew their stands with disease-tolerant material. CEPEC is orientating an experimental net on fields, where each tolerant clone is represented by 10 trees, with 3 replications. Note that a partnership system has been adopted for coffee research in the State of Espirito Santo that would fit into the concept for participatory varietal selection proposed by Witcombe et al., 1996, where coffee farmers produce the cuttings from genotypes validated by the local research company (EMCATER). Currently, there are 160 clonal gardens for the production of coffee cuttings in the entire State of Espirito Santo, in partnership of producers with cooperatives, commercial associations, and municipalities. Moreover, the producer can yield a return by selling the cuttings (A. F. A. Fonseca, personal communication). It is hoped that a similar proceedure will be adopted in the area of cacao research in Brazil. |
| Chapter 7. Breeding for disease resistance. R.A. Rios-Ruiz. | Return To Table of Contents |
|---|
| Contents: Introduction and Diseases; Resistance breeding; Objectives and pre-selection, Sources of resistance, Reaction of cacao to pathogens and durability and stability. Resistance tests; Test criteria, Componentes of resistance, Field evaluations and selection tests and Test standardization. Evaluating and selecting germplasm for resistance. Breeding strategies for resistance and future prospects. Summary: Diseases have become the main limiting factor of cacao production. The most important control measure has been the use of resistant genotypes, which are more economical, stable, and environmentally sound. A few successful improvement programmes for resistance, developed in the 30s and 40s in Trinidad, are presented here and the reasons for the slow progress in the development of resistant cultivars are discussed. The large germplasm collections in the world form a large pool of potentially important genetic resources for resistance to diseases. In the last 15 years, several accessions have been identified as resistance sources. For this reason, improvement for resistance has become a priority in many countries. Cacao populations show ample variability for resistance, probably of polygenic nature. In view of this, improvement should aim at searching for adequate levels of partial resistance in studies of resistance components that ensure greater durability and stability. Field resistance data, obtained under experimental conditions, are useful to initiate selection programmes of resistant genotypes. The artificial inoculation tests are equally important. A revision of the latest studies on the evaluation of genotype reactions and the identified resistance sources is also presented. Certainly, the greatest progress in the development of high-yielding and resistant cultivars can be achieved by recurrent selection, and much of this progress will be realised in the next decade. |
| Introduction and Diseases. | Return To Table of Contents |
|---|
| Introduction: Cacao is an important crop of industrial and nutritional value. Diseases, however, depreciate the product quality, reduce its yield and increase the costs. It is estimated that they cause global losses in the order of 40%, but reach up to 90 or 100% in cases of diseases such as witches’ broom and monilia pod rot in certain locations (Andebrhan et al., 1998 and Rios-Ruiz et al., 2002). In view of the great negative economical impact of the diseases, great effort has been made to develop chemical and non-chemical management strategies. Cultural practices, when applied properly, provide a certain degree of control. Variable and only partially effective results have been achieved by protection through fungicides. The combined practice of phytosanitary pruning with adequate fungicide application is being used in the control of most diseases. However, such practices raise the production costs and are therefore frequently economically unviable. The best solution for disease control has been the replacement of susceptible cultivars by others with durable resistance to the pathogens. Thus, one of the great objectives of cacao improvement programmes is to develop tolerant or resistant cultivars that could be used in the management or control of diseases. In this sense, many studies are being conducted at different cacao research centres, with a view to finding genotypes resistant to the different diseases. This chapter presents a critical approach to improvement for disease resistance in cacao. Diseases: During the first century of cultivation in America and Africa, cacao remained relatively disease-free, or presented a low incidence in many cases. However, diseases now occur throughout the entire development of the crop cycle and attack all plant parts. Cacao diseases are caused by fungi, viruses, and nematodes (Wood & Lass, 1987), despite only a few being of global importance; most of them are only regionally or locally important. Five main diseases can destroy up to 40% of the global production annually: black pod (BP), caused by Phytophthora spp. (P. palmivora, P. capsici, P. citrophthora, P. heveae, and P. megakaria); witches’ broom (WB), caused by Crinipellis perniciosa; cocoa swollen shoot badnavirus - CSSV, caused by virus; vascular streak dieback - VSD, caused by Oncobasidium theobromae, and monilia pod rot (MO), caused by Moniliophthora roreri (Van der Vossen, 1997). BP, WB, and MO are the main prevalent diseases on the American continent. WB has drastically reduced the cacao yield in Brazil. Diseases such as cacao wilt (Ceratocystis blight), caused by Ceratocystis fimbriata; Verticillium wilt, caused by Verticillium dahliae; and pink disease, caused by Corticium salmonicolor, are of local occurrence, although they are sometimes of relative economical importance. Symptoms, taxonomy, origin, dissemination, transmission, and disease control in cacao are described by Wood & Lass, 1987. |
| Resistance breeding; Objectives and pre-selection, Sources of resistance, Reaction of cacao to pathogens and durability and stability. | Return To Table of Contents |
|---|
| Resistance breeding: The progress in improvement for disease resistance in cacao has been modest, compared to the achievements in many other crops. The reasons are various and include: i) rareness of studies on resistance sources in wild germplasm ii) absence of efficient screening tests; iii) ignorance of the genetics of the host resistance mechanisms; iv) absence of effective improvement strategies that consider more than one selection cycle (Van der Vossen, 1997); v) relative late initiation of improvement programmes for the crop; vi) few trained and active breeders in the area until recently; vii) discontinuity of the few research groups; viii) short-term programmes with lacking or interrupted financial resources; ix) lack of an international research institute; and x) absence of research programmes in an international context (Zadoks, 1997). A lot of the information on cacao improvement in general and on improvement for resistance in particular was initially produced in Trinidad. In this country, the clones ‘Trinitario ICS’ (Imperial College Selection) were first selected for high yield but showed a high susceptibility at later stages. The search for and selection of resistance genes realized by Pound in the region of the upper Amazon in the 30s and 40s led to the initiation of an improvement program where upper Amazon clones were crossed with local selections. New clones were then selected from these hybrid progenies. This programme was run for about 40 years in Trinidad and gave rise to the high-yielding clone selections of the series TSH (Trinidad Selected Hybrids). This has been a successful improvement programme for resistance (see also Chapter 12). Another notable attempt of improvement for resistance began with Posnette, in Ghana, with genotypes introduced from Trinidad. As the yield factor was given greater emphasis as the selection criterion, improvement for resistance did not become a priority until the early 80s, when studies in this area were finally intensified (Luz et al., 1996 and Andebrhan et al., 1998). At this time, the idea was prevailing that it was necessary to achieve a high resistance level (complete resistance), preferably conferred by simple dominant resistance genes (Purdy & Schmidt, 1996; Zadoks, 1997 and Andebrhan et al., 1998), so interesting genotypes from the point of view of partial resistance were ignored (Bartley, 1981). Therefore all programmes in search of resistance sources worked on a narrow genetic base, predominantly concentrated in the germplasm collected by Pound, in the Ecuadorian as well as the Peruvian Amazon. In the last 15 years, numerous cultivated, semi-cultivated, and wild cacao accessions have been identified as resistance sources against diseases (Silva et al., 1987; Rios-Ruiz, 1994; Rios-Ruiz & Rivas, 1997; Luz et al., 1999a and 1999b; Phillips-Mora, 1999a and 1999b; Pires et al., 1994 and 1999a; Suarez-Capelo, 1999 and Rios-Ruiz et al., 2000). Improvement aiming at resistance is currently considered to be a priority by the diverse cacao research institutes around the world. Also, different research programs of international scope have been set up more recently (Eskes et al., 1998 and Knight, 1998). These have focussed on the main cacao diseases. However, improvement for diseases of regional or local importance such as WB and BP in Brazil (Fonseca et al., 1999b; Luz et al., 1999b and Pires et al., 1999a), MO and WB in Ecuador (Suarez-Capello, 1999) and Peru (Rios-Ruiz et al., 2000 and Rios-Ruiz et al., 2002), MO and BP in Costa Rica (Phillips-Mora, 1999a and 1999b), CSSV in Ghana (Adu-Ampomah et al., 1999) and VSD in Malaysia and Indonesia (Bong & Lee, 1999) also receive attention. In both cases, the objective is to develop selection strategies with a view to incorporating resistance as desirable agronomical trait into the new cultivars. In Brazil, for instance, the survival of the cacao culture in Bahia depends, to a large extent, on the release of resistant cultivars. Objectives and pre-selection: The main selection criteria in genetic improvement of cacao are: high vigour, early yield, high productivity, adequate bean size and quality, resistance or tolerance to diseases and pests, and a good adaptation to local conditions (Toxopeus, 1987). Traditionally, this breeding is a slow process because of the tree’s long juvenile phase. In the past, the improvement programmes for resistance and productivity were conducted separately. This procedure is acceptable in the case, for example, of a site where the incidence of a certain disease is high, up to the point where the economical viability of the production is affected. In such cases, a special programme for resistance should be organised (Zadoks, 1997). Actually, there is consensus that disease resistance should be a programme’s prime objective, though thoroughly integrated with the other objectives (Paulin & Eskes, 1995 and Van der Vossen, 1997). However, it is worth repeating that disease resistance should only be focus when there are safe methods of evaluation; hence the importance of developing early tests. Such tests should help to select promising resistant genotypes and to discard susceptible ones. The lack of consistent methods for an early evaluation of resistance to C. perniciosa, for example, has limited the identification of new resistance sources to this pathogen and, consequently, the incorporation of genotypes of interest in the improvement programmes in Brazil (Andebrhan et al., 1998). This demonstrates the need to develop more efficient and relatively simple tests of pre-selection or to consolidate some of them for most cacao diseases. Pre-selection tests are also useful for the genetic analysis of resistance and in the study of resistance mechanisms against the pathogen’s invasion into the host’s tissues. Sources of resistance: One of the first and most important phases of improvement for disease resistance is the identification of resistance sources. In cacao, as in other crops, it is desirable that such sources are found in the genotypes most intimately related to commercial cultivars, as in traditional and or wild cultivars. Much of the variability found in the crop, including resistance genes, comes from South America, especially the Amazon Basin, the supposed origin of the species (Paulin & Eskes, 1995; see also Chapter 3). Improvement for resistance in cacao was initiated by Pound (1938 and 1943), in the 30s and 40s, with the search for supposedly witches’ broom-resistant cacao in the region of the upper Amazon. Despite the limited number of supposedly resistant trees found (particularly the clones SCA 6 and SCA 12), these plants were the most widely used sources in the improvement programmes of Trinidad and Ecuador. Later, ‘SCA 6’ and ‘SCA 12’ expressed susceptibility to witches’ broom in Ecuador (Bartley, 1981), Brazil (Andebrhan et al., 1998 and Fonseca et al., 1984) and Peru (Rios-Ruiz, 1989). In the following decades, more sources were being saved by collection missions in the areas of natural cacao occurrence (Almeida et al., 1995). It is expected that the improvement program for resistance, in particular, will benefit from the due exploration of these genetic resources. The cacao collections maintained around the world list some 7000 cacao genotypes, including cultivated and wild material. Besides the universal collections of ICGT in Trinidad and CATIE in Costa Rica, large collections with local genotypes also exist in Brazil, Colombia, Ecuador, French Guiana, Mexico, in and Venezuela. Collections have also been established in the producing countries of eastern Africa and south-eastern Asia, including local selections and accessions introduced from Latin America (Eskes et al., 1998). New and modest collections are also found in other countries with a less significant cacao production. Altogether, these collections represent the greatest pool of genetic resources for resistance to diseases and insect plagues, and for other traits of interest. However, the reaction to diseases of most of these accessions is to date unknown, and few wild resistance sources are actually used in improvement programmes. This is the case in Brazil, where the Amazon genotypes collected in large number in the decades 60, 70, and 80 (Almeida et al., 1995) have found little use. Certainly, a systematic screening of all theses accessions would greatly benefit improvement for disease resistance. This became evident from the reports on more recent results that indicate that some genotypes collected in the upper Amazon presented resistance or tolerance to witches’ broom and/or to pod rot. Reaction of cacao to pathogens: Improvement for resistance is generally considered to be more complex than for agronomical or morphological traits, as long as the efficiency of the disease resistance is not static, and is simultaneously influenced by the traits of the host (magnitude of the efficiency etc.) and by the genetic variability of the pathogen (aggressiveness of the isolate, genetic selection for an increase of virulence etc.). The concept of physiological races, identified by the pathogenic response to differentiating genotypes of the host, allows the understanding of the pathogenic variability in relation to improvement for resistance determined by oligogenes (oligogenic or vertical resistance). Oligogenic resistance, however, has not been found for the main cacao pathogens. On the other hand, evidence of horizontal resistance (polygenic or partial) are mentioned in literature. In this latter, cacao genotypes tested under uniform conditions express amplitude of response, varying from those very infected to those that are relatively immune (Luz et al., 1999a; Phillips-Mora, 1999b; Pires et al., 1999a; Suarez-Capello, 1999 and Rios-Ruiz et al., 2000). The variability of pathogenicity in cacao is in general poorly understood, and a lot of research will be needed to optimize the improvement work. Economically acceptable partial resistance is that resistance level which makes the disease manageable, yet without eliminating the pathogen (Sadoks, 1997). Partial resistance is used in many annual crops against important pathogens (Parlevliet, 1989 and 1993). It presents a quantitative nature, usually polygenic, although it can be monogenic in some cases (Parlevliet & Zadoks, 1977), and is a predominant phenomena in perennial crops, including cacao (Zadoks, 1999). To characterize partial resistance, the components resistance against entrance, development, and the reproduction of the pathogen are evaluated. Many of these components have already been evaluated in cacao, for instance the latent period, period of symptom development, production of inoculum, and infection frequency, among others (Andebrhan et al., 1998; Luz et al., 1999b; Silva et al., 1998; Rios-Ruiz et al., 2000; see also Chapter 8). However, many others could be systematically evaluated to determine their real value and permit the selection of the most representative in each interaction host-pathogen. The quantitative character of horizontal resistance is apparent in the slow progress of the disease, in the reduction of the disease quantity, and consequently, the reduction of the inoculum level (Van der Plank, 1963). Different countries are already including an approach to resistance under this epidemiological context in evaluations of cacao accessions. Durability and stability: Actually, improvement programmes aim at incorporating durable and stable resistance in cultivars. Resistance is durable when it remains effective in a given cultivar throughout a long cultivation period in a disease-favourable environment (Johnson, 1979). Resistance stability, in turn, refers to the reproducibility of the host response to the pathogen genotype, in field conditions. On the other hand, durability refers to the continuity of resistance over time in some particular space, and stability is defined as the continuity of resistance in space (Zadoks 1997). Durability and stability can be modified by extreme factors, such as climate conditions (modifying the adaptation of the pathogen, host resistance or the interaction host-pathogen) and by alterations in the pathogen genotype (increase of the frequency of virulent races, mutation frequency from avirulence to virulence, the quantity of virulent races when the cultivar was released, and the multiplication rate of virulent races (Kiyosawa, 1982). In literature, polygenic resistance is addressed as synonym of durable resistance (Jacobs & Parlevliet, 1993). However, there are several cases in which the performance of the monogenic resistance also behaved like the durable and stable resistance (Prabhu & Morais, 1993 and Van der Vossen, 1997). Neither durability nor stability are to date well documented in cacao. Clone SCA 6, WB-resistant, expresses durable resistance in Trinidad (Laker et al., 1988a and 1988b), but not in Ecuador, where another pathotype of the fungus is predominant (Wheeler & Mepsted, 1988), nor in Peru, where it presents relative high infection levels (Rios-Ruiz, 2000). There is also some evidence, yet to be confirmed, which indicates that clone IMC 67 is resistant to Ceratocystis wilt of cacao (Ceratocystis fimbriata) and expresses regional stability. The phenomenon of constant ranking of the cultivars for resistance is considered an indicator of horizontal resistance (Van der Plank, 1963 and Robinson, 1976) and for durability and stability (Van der Vossen, 1997 and Zadoks, 1997). Even when not clearly determined, the ranking of some black pod resistant genotypes was observed to generally prevail unaltered in time and space. There is no way to predict durability and stability, which can only be identified by testing ex post facto over long periods and at multiple sites, respectively. Inherently, such particularly important studies require signed agreements of international cooperation that allow testing for (partial) resistance stability of the different potential genotypes at national and international levels. Only more recently have these studies been included into international projects of utilization and conservation of cacao germplasm (Eskes et al., 1998). Improvement for resistance in cacao should aim at adequate levels of durable resistance. Despite that new results indicate the ever-increasing importance of the major genes for their control, the best way to obtain durable resistance is still the accumulation of minor genes. In cacao, the studies realized up to now suggest that the species has broad variability for polygenic resistance. As a result, the systematic germplasm evaluations, the application of appropriate tests, and finally the identification, selection, and crossing of resistant parents give rise to the expectation of accumulated resistance, followed by transgressive segregation. Durable resistance is expected in this case. |
| Resistance tests; Test criteria, Components of resistance, Field evaluations and selection tests and Test standardization. | Return To Table of Contents |
|---|
| Tests for resistance: Plants use diverse defence mechanisms, which can be classified as: evasion, resistance and tolerance. Generally speaking all are exploited in plant improvement (Parlevliet, 1989). Evasion refers to disease prevention by planting at times or in areas when or where the inoculum is inefficient, rare, or absent. Tolerance refers to the inherent or acquired capacity of a plant to support pathogen attacks without significant production damage occurring. However, the defence against pathogens by evasion and tolerance have less importance. Resistance, the host’s ability of reducing pathogen growth and or development, is by far the most significant defence strategy (Parlevliet, 1989). This can be: i) absent, such that the infected plant quickly expresses symptoms; ii) complete, so the plant does not express symptoms; and iii) intermediate. Complete resistance can be found frequently in annual crops against fungi that develop physiological races; it is normally monogenic and race-specific. In tree species, however, complete resistance is rare, although some reports point at its existence in the case of some leaf pathogens, such as apple scab (Venturia inaequalis) of the apple tree, a disease known as CBD (Coffee Berry Disease, caused by Colletotrichum coffeanum) of coffee beans (Van der Vossen, 1997 and Zadoks, 1997), and rust (Puccinia psidii) of eucalyptus (Junghans et al., 1999). Many crops present intermediate or partial resistance against the most pathogens. Partial resistance is genetically determined, usually polygenic and accessible by component analysis, quantifiable at any stage of the infection cycle. These components are evaluated by the resistance tests. Test criteria: Resistance tests are necessary to represent the situation of a given disease that is infecting cultivated cacao or to offer the cacao farmer hybrid or clonal cultivars that meet minimal resistance standards, the latter being the primary objective. Various tests are being used for cacao, from those less sophisticated in the field, followed by in vivo tests that use plant parts in a controlled environment, up to the most sophisticated biochemical and molecular tests. Desirable resistance tests are those that unite easy management, low cost and high efficiency. Before being chosen a test should comply with certain attributes, which are: i) Representativity - ideally, field tests are the most representative. Although they occupy a lot of space, take many years, give a lot of work and are not always easy in the interpretation of the results, frequently demanding genetic, agronomical, and epidemiological knowledge, field tests are indispensable. Some recent progress in improvement had its origin in the identification of resistant genotypes under field conditions (Aragundi et al., 1988; Laker et al., 1888a and 1988b; Pires et al., 1994, 1999a; Morera & Mora, 1996 and Rios-Ruiz et al., 2000). Resistance tests in the field are very useful in confirming the reactions that are expressed in the development of screening tests; ii) Reproducibility - a resistance test is valuable if its results are reproducible. The problem of reproducibility becomes critical when the test is applied at different places, at different times and/or by different people. Field tests are quantitative tests, which require repetition in space and time. Tests in vitro are supposedly qualitative and highly reproducible (Zadoks, 1997). In cacao, comparisons of tests regarding reproducibility have not yet been done. Zadoks (1997) recommends that the reproducibility of a test should be determined by the elaboration of a simple, standard, and universally accepted protocol. iii) Predictability - predictive tests save time and space demanded by field tests which evaluate mature trees. Ideally, results of such tests should predict the future performance of a genotype for the farmers’ conditions. The predictive value of a test can be proved when a group of tested genotypes scores the same ranking in tests conducted separately by two or more researchers (Zadoks, 1997). Not only in cacao, but also in other crops, this methodology is not a common practice among phytopathologists, although it should be promoted. The experience of some countries with the tests shows that extreme values found were more consistent and reproducible than average values in situations in which the tests show small discriminatory capacity (Porras et al., 1988). In this sense, the great number of genotypes that are neither susceptible nor highly resistant are, for the most part, manipulated according to the breeder. The component analysis is important and recommendable in this situation since it would allow and aid the precise selection of the material under study. Components of resistance: The process of monocyclic infection consists in a succession of morphological, physiological, and biochemical steps, designated infection cycle components. Each step is measurable and quantifiable (Zadoks & Shein, 1979). The reduction in pathogen growth at each step can be considered a form of partial or incomplete resistance. Partial resistance is based on one or more resistance components that may or may not be associated. If one component is highly correlated to another, or if one component is preponderant, this can be used as selection criterion in the improvement. However, if different resistance components are under distinct genetic controls, the recombination of host genes by hybridization can lead to new combinations of components with a relatively high resistance (Zadoks, 1997). Many phytopathogenic fungi are effectively controlled by partial resistance. This has a favourable epidemiological effect, because, if its level is adequate, the quantity of the pathogen’s inoculum is reduced to very low levels. In the case of cereal rusts, the latent period of infection and the infection frequency were highly correlated to the inoculum production (Parlevliet, 1979). In the case of WB, the only parameter related to the final production of the inoculum was the infection frequency (Andebrhan et al., 1998). The reduction of the inoculum is proportional to the level of partial resistance, the size of the area planted with the genotype bearing partial resistance, and to the age of the plantation (Zadoks, 1997). The evaluated components of partial resistance for BP are percentage of infection (resistance to penetration), incubation period (resistance to lesion formation), diameter and/or area of lesion (resistance to lesion expansion), latent period, and sporulation quantity (resistance to sporulation). The diameter and/or, the lesion area is the parameter that best estimates the reaction of the material (Luz et al., 1999b and Phillips-Mora, 1999a). For WB, the following components of partial resistance evaluated are: percentage of infection, latent period, broom size and type, period from the appearance to the death of the broom, period of dormancy (from death of the broom until fruiting), percentage of sporulated brooms and number of basidiocarps formed. The resistance of the vegetative parts is epidemiologically more important than the resistance of fruits or of flower cushions (Fonseca et al., 1984 and Andebrhan et al., 1998). The following components are evaluated for MO: percentage of infection (resistance to penetration), reaction type, and incubation period (resistance to the development of a lesion), latent period and area of lesion with sporulation (resistance to sporulation) and area of internal lesions (resistance to the development of lesions) (Phillips-Mora, 1999b). Field evaluations and selection tests: The final test for resistance to some pathogens is doubtlessly the measurement of the reaction of a given genotype exposed in the field, over a long time. However, early tests are essential to speed up improvement for resistance. In the field, various types and methods of evaluation are generally performed according to the infection agent, using natural infection or artificial inoculation. Various and interesting early tests are available for cacao, which evaluate germinated seeds, plantlets, roots, detached fruits with or without injury, detached leaves, and leaf discs. Furthermore, there are the biochemical and molecular tests. Resistance tests that evaluate natural infection in the field bring good results in terms of resistance and susceptibility, be it for BP (Pires et al., 1994), for WB (Fonseca et al., 1999b; Pires et al., 1999a; Rios-Ruiz et al., 2000) or for MO (Aragundi et al., 1988 and Rios-Ruiz & Rivas, 1997), principally when applied in germplasm banks. However, a precise evaluation of partial resistance in large areas, with the use of a statistical design is necessary. It must be stressed that the method may be of little value in isolated observations, in view of the fact that the infection severity varies according to the seasons of the year. However, if trained staff gather observations all year round and in consecutive years, it is possible to determine the best performance material for a determined area. The evaluation assesses the number of brooms per plant, the percentage of infected fruits, and the curves of disease progress. The application of the WB evaluation methods has confirmed the fact that factors unrelated to the host are sometimes largely responsible for the field performance of cultivars. As is known, a successful infection requires synchrony between the host, pathogen, and environment, in which the high environmental humidity, which allows for the formation of infectious basidiospores, coincides with the presence of young tissues of the host. In turn, the presence of inoculum potential in a determined area depends on multiple factors, such as the proximity of the plantation to highly susceptible populations, the seed material, plot size, and the management. Together, these factors mask the plants’ genotype. The greater vigour of the plants favours field infections, be it for the genetic constitution, physiological behaviour, development conditions, or their age. Finally, some agronomical practices may prove inadequate, such as pruning at the beginning of the rainy season, which stimulates sprouting and favours infection; excessive fertilization, favouring the formation of succulent tissue; and poor drainage, giving rise to the infection. A unification of these factors will make the results representative and comparable. Despite that evaluations of the natural incidence of the disease in different genotypes provide information on the relative susceptibility or resistance levels, understanding of the reactions is achieved by studies with artificial inoculations, involving different tests and plant organs. The analysis of BP resistance tests was described by Lawrence, 1978b. According to this analysis as well as later studies, the artificial inoculation of detached or non-detached and completely developed fruits (4 to 5 months old) is the most broadly used and accepted screening test (Rios-Ruiz, 1994; Luz et al., 1999b and Phillips-Mora, 1999a). More recently, an inoculation test on young leaves and/or leaf discs in Trinidad (Iwaro et al., 1999) and Costa of the Marfim (Nyassé et al., 1999) was developed. The results revealed a strong association between the two organs (fruit and leaf), in the post-penetration stages. Consequently, resistance in fruits can be predicted by the reaction of the leaves, and early screening in the plantlet stage can help reduce the maintenance costs of large populations up to maturity. The application of agar blocks containing basidiospores of C. Perniciosa inoculated onto tender cacao shoots in the field or onto growing points of young plantlets has been the most commonly used inoculation method so far. This method has allowed studies on the pathogenicity of isolates and the discrimination of resistant genotypes and of resistance components (Fonseca et al., 1984; Rios-Ruiz, 1994 and Andebrhan et al., 1998). More recently, the University of Florida developed a semi-automatic inoculation method to measure the WB-resistance of plantlets (Purdy et al., 1997 and 1998). Standardization studies of spore concentration, plantlet age, environmental conditions during the infection and period of incubation, time of symptom evaluation after inoculation, and good parameters of evaluation have been developed at CEPEC (Silva et al., 1998). One awaits the confirmation of the method in such a way that similar results in the field be obtained. The evaluation method in the case of MO is applied only to non-detached fruits as the incubation period is long (2 to 3 months) and the detached fruits rot before showing any symptoms. Flowers are pollinated and the resulting young fruits covered with plastic bags. Sixty-day-old fruits are inoculated by atomization or with dry conidia and covered with plastic bags. The reaction is evaluated by the external as much as the internal incidence and severity on a grading scale (Sanchez, 1982; Porras et al., 1988; Sanchez et al., 1988 and Phillips-Mora, 1999b). In situations where the method requires mature fruits and is time and space-demanding, it is worth to investigating the possibility of developing an early test based, for example, on inoculation onto plantlets. Test standardization: As was evident in the previous item, several tests are used in cacao, many of them with relative satisfactory results, allowing the reaction variability in many genotypes to be found. However, there are also unsatisfactory and erratic results, even when compared at different moments at the same site. In this case, the genotype reaction was not correctly evaluated. Likewise, the use of different methodologies by researchers and/or countries has not allowed a definitive comparison of results, hampering the progress of the improvement programmes. Some revisions and an attempt of consolidation of tests have been carried out by Lawrence (1978b) for BP, for example, allowing some progress in the past. Similarly, several other experiences of development of adequate test methods were also realized (Sanchez, 1982; Luz & Yamada, 1984; Sanchez et al., 1988 and Purdy et al., 1997). However, there is still a great need to standardize these tests. It would be convenient to approve of and use an internationally standardized protocol for the application of tests to the different pathosystems. Several attempts to elaborate protocols have been made in the past by researchers united in the different cacao fora. Nevertheless, only recently the project of conservation and utilization of germplasm was established in a workshop realized in 1998 (Eskes et al., 2000) with the elaboration of protocols for test standardization as a priority of its work plan. About 25 standardized selection and evaluation procedures for cacao regarding the diseases WB (4), BP (5), and VSD (2), besides the phenological and yield evaluations, were approved at this meeting. In principal these protocols will be for use by the 10 member countries of the cited project: Brazil, Ecuador, Venezuela, Trinidad, Nigeria, Ghana, Ivory Coast, Cameroon, Malaysia, and Papua Nova Guinea. The acta of this workshop was also sent to other researchers who work in other countries so they could evaluate and propose modifications to the protocols. Therefore, one hopes that in the near future well-defined tests will be made available that can help the efforts of improvement for resistance, and offer time gains as well. In future, other similar efforts should allow the elaboration and testing of new protocols for other diseases taken into consideration in the project cited. |
| Evaluating and selecting germplasm for resistance. | Return To Table of Contents |
|---|
| Researchers from Trinidad were the first to use genetic resources to develop disease-resistant cacao cultivars in the 30s. They selected plants with low WB infection indices, developing the clonal series TSH, which is considered resistant to the disease. Large scale evaluations of the TSH resistance indicate that this material remains resistant in the germplasm as well as on the plantations in Trinidad (Laker et al., 1988a and 1988b). In spite of this rapid success with the exploitation of the variability of the host plant, improvement for disease resistance was not given priority until the beginning of 1980. The increase in production costs, the inefficiency of chemical control methods for an effective protection against the diseases that infect fast growing meristematic tissues, the availability of new accessions, and the emergence of new diseases infecting the main production zones in the 80s led the institutes to deal with improvement for disease resistance in the different countries as a matter of priority. The evaluation of the accessions for resistance on a greater scale led to the identification of numerous resistant genotypes to diverse pathogens. Considerable efforts have been made to identify resistance sources for black pod due to its international importance. Such resistance sources were identified over three decades ago (Rocha, 1965 and 1974; Soria, 1974; Lawrence, 1978a; Luz et al., 1986 and Silva et al., 1987). However, as pointed out by Lawrence, 1978b, the reaction of these genotypes can only be capitalized on in areas where they had been tested against the predominant Phytophthora species of those areas. Because various species of Phytophthora are responsible for black pod (Campelo & Luz, 1981 and Zentmeyer, 1988) and the most adequate methods need to be determined to test resistance to Phytophthora, the evaluations of cacao genotypes in the existing germplasm and in the assays of hybrid progenies have been intensified. In recent years several resistance sources have been described but no immune cultivar was identified (Rios-Ruiz, 1994; Iwaro et al., 1999; Kebé et al., 1999; Nyassé et al., 1999 and Phillips-Mora, 1999a). About 529 clones of CEPEC’s germplasm collection were evaluated for resistance under natural infection. The Amazon clones 21, EET 45, TSA 654, TSH 1188, CEPEC 40, UF 36, TSH 565, CEPEC 541, and PA 300 were identified as having high resistance potential. Pre and post-infection resistance mechanisms were also detected in this collection (see also Chapter 8) and a greater frequency of resistance factors in the far south of the natural species diversification area identified (Ecuador, Peru, and the Brazilian Amazon). In the past the series PA, P, NA, EET, CJ, RB, MA, CA, and TSA had been selected from this area (Pires et al., 1994). Many of these clones behaved as resistant when tested subsequently by artificial inoculation (screening test) with P. capsici, P. citrophthora, and P. palmivora (prevailing species in Brazil). From this group, where 21 clones showed adequate resistance levels and were therefore selected as promising, PA 30 and PA 150 were the most outstanding (Luz et al., 1996 and Luz et al., 1999b). When the progenies of some of these clones used as parents in comparative assays of hybrids were evaluated for black pod, they also presented low disease levels (Luz et al., 1999b). In the hybrid combination trials conducted at the experimental station of CEPLAC, in Medicilandia, State of Para, the hybrids with the clones PA 150 and PA 121 as female parents also stood out for their low infection levels (Francisco Neto et al., 1999). Variability was also found between Amazon and non-Amazon clones when their reactions to P. palmivora and P. capsici were tested in field and laboratory conditions in Peru. The clones UF 613, H 24, H 44, and U 24 presented the smallest lesion diameter and were considered promising for improvement (Rios-Ruiz, 1994). Up to 1996 350 cultivars were evaluated at CATIE, Costa Rica, by the paper disc method for their reaction to P. palmivora. Based on the lesion diameter, only 36% of the cultivars were classified as resistant and moderately resistant, distinguishing APA 4, Catie 1000, CC 42, CC 71, CC 83, CC 214, CC 225, CC 232, CC 240, EET 59, EET 156, ICS 44, ICS 89, POUND 7, BR 41, TSH 812 and UF 703 (Phillips-Mora, 1999a). Of the new 109 clones characterized between 1998 and 1999, 3C, ARF 30, PMCT 35, and PMCT 37 were classified as resistant (Phillips-Mora & Castillo, 1999). Of all the CATIE germplasm evaluated to date a high percentage (22%) presented resistance or moderate resistance. This broad diversity of reactions could be studied complementarily through some other resistance components, in order to classify the partial resistance potential more in detail, as well as to prove its efficiency under field conditions since the resistance is polygenic (Partiot, 1975; Tan & Tan, 1990; Warren & Pettitt, 1994; Iwaro et al., 1999 and Phillips & Castillo, 1999). Five QTLs related to P. palmivora resistance and inserted into five of the 10 linkage groups of the genetic map of cacao were identified, furthering the confirmation of the polygenic nature of resistance to this fungus (Frits et al., 1995). More recently, Risterucci et al., 2000 detected eight QTLs linked to resistance against P. palmivora, P. Capsici, and P. megakaria. Three of them were significant for more than one species of Phytophthora, which is of great interest for an application of resistance tests in different countries (see also Chapter 10). The reaction of around 500 accessions of CRU in Trinidad showed wide genetic variability for resistance against Phytophthora. The distribution of accessions within populations presented widely dispersed resistant genotypes among populations. Clone SCA 6 and its progenies presented a high resistance level against pre and post-penetration indicating that resistance, which is polygenically inherited, can be combined with other resistance forms to achieve greater gains in improvement. This suggests, moreover, that recurrent selection could be the appropriate improvement method (Iwaro et al., 1999; see Chapter 12). The predominance of the additive gene effect for yield and other traits found in the studies also suggests that the performance of the progenies can be predicted by the genetic potential of the parents (Cilas et al., 1999 and Iwaro et al., 1999). When not more than two Phytophthora species are present in a country and the isolates of these species differ in their growth rates these can show a clear distinction between resistant and susceptible genotypes in one and the same ranking. Such results were found in Trinidad and Peru for the species P. palmivora and P. capsici where the first is predominant (Rios-Ruiz, 1994; Iwaro et al., 1999). This indicates that selection based on a more aggressive and predominant isolate would also be effective against the others. It has been reported in literature that certain genotypes classified as resistant in one country are also resistant in others. The fact of finding of material with resistance stability and making it available, together with a painstaking assessment of the reaction of the tested genetic material, would allow for progress in improvement programmes. Over 50 clones and 40 hybrid progenies are currently being evaluated for BP resistance in an international project involving countries of America, Africa and Asia, coordinated by IPGRI/CIRAD (Eskes et al., 2000). Results of this study are expected to validate previous findings. Except for some older populations developed in Trinidad the intensive evaluation of cacao for WB-resistance has only been undertaken recently. WB resistance was discovered in the Amazon Basin on Pound’s expeditions in the 30s and 40s when the clones SCA 6 and SCA 12 were selected. Resistance in this material was described as immunity and later as low disease levels in Trinidad. The research status into resistance improvement at young stages was revised by Bartley, 1981 and a list of cultivars considered resistant was supplied by Enriquez & Soria, 1981. However, in the light of novel results, the information from this last article can at best be considered referential. With the arrival of WB on the plantations in the southeast of Bahia and its consequential devastating effects there has been a strong demand for productive and resistant cultivars to face the problem and revert the situation. With the objective of identifying WB-resistance sources, a germplasm collection of CEPEC is being evaluated for its response to natural infection. Data collected from brooms removed between 1995 and 1996 indicated a large variation for the number of vegetative brooms (0 to 99 brooms/plant). Sixty clones, among which clone SCA 6 and its descendents (EET 376, 392, 397, IAC 1, and the series TSA and TSH), as well as other Amazon clones (series Cruzeiro do Sul, RB, CAB), especially from the State of Acre and from Peru presented the highest resistance levels. The existence of numerous clonal series with low levels of WB infection included in the list manifested the presence of expressive genetic variability for resistance (Pires et al., 1999b). In a study with RAPD markers a low genetic similarity between important groups such as Scavina, C Sul, RB, CCN, and CEPEC was found (Marita, 1998). This confirmation amplifies the possibilities of success of an improvement program for WB resistance. With the objectives of confirming the field results and obtaining a method able to discriminate resistance levels rapidly the genotypes identified as WB-resistant were tested under artificial inoculation. Despite some contradictory results the inoculation tests generally confirmed the field tests (Purdy et al., 1998 and Silva et al., 1998). The search for apparently WB-resistant trees on plantations in Bahia has also been practiced in recent years with promising results (Pinto et al., 1999a). The evaluation of the existing germplasm in CEPEC (Pires et al., 1999b) and the clonal assays (Pinto et al., 1999b) demonstrated low levels of infection in the Scavina progenies, and intermediate levels in the hybrid progenies derived from them, such as the series TSH and TSA; materials generated and described as resistant in Trinidad. CEPEC thus confirms the results found earlier in Trinidad for the conditions in Bahia. These evaluations enabled CEPEC to release hybrid SCA 6 x ICS 1 (‘Theobahia’), and the clones TSH 565, TSH 516, TSH 1188, EET 397, and CEPEC 42, and more recently the clones TSA 792, TSA 654, TSH 774, and TSH 656 (Pereira, 1998 and Pereira et al., 1999; see also Chapter 9). These results confirmed the great merit of the development programme of the cultivars of the series TSH and TSA of Trinidad. It is still expected that a broader field evaluation of all these materials on the different Brazilian plantations will confirm its resistance potential. Presently CEPEC has a series of ongoing studies of identification of new resistance sources, evaluation of clones and progenies, and studies of the general combining capacity, in order to form base populations for yield and resistance improvement through recurrent selection (Pires et al., 1999c). One may expect that resistant material will be developed and released in the medium term. During the years when WB was only found in the Brazilian Amazon, research with hybrids focusing on productivity and resistance was developed by CEPLAC of Belém and Rondônia. Observations of various genotypes under natural infection in the field indicated differentiated reactions in terms of quantity of infected fruits and tissues and type and size of the hypertrophies. These variations were also observed by artificial infections at an experimental level. Initially, only the incidence in fruits was evaluated, not allowing the observation of the reaction of the existing materials in terms of leaf flushes. Later, the parameter number of infected leaf flushes per plant was used in the resistance evaluations. Although this index increases with the plant age, the relative differences between resistant and susceptible genotypes tend to persist, even in highly infected areas (Andebrhan et al., 1998). At the onset of the epidemic the most important inoculum for infections emanates from the canopy and therefore a selection of genotypes with low infection indices in the canopy should be one of the goals to be pursued. However, parameters such as smaller broom size and number of basidiocarps per broom seem to be related to higher resistance levels (Andebrhan et al., 1998 and Rios-Ruiz et al., 2000), which are important from the point of view of the epidemiological management of the disease. In 1995 521 accessions of ERJOH in Belém, Pará, were evaluated for WB-resistance in the field. Approximately 30 clones of the series CAB, RB, and AS showed resistance as they consistently presented low disease incidence. For the genotypes evaluated the greater proportions between collected clones/resistant clones were found in the states of Rondônia and Acre (Fonseca & Albuquerque, 1999). The majority of the accessions originating from wild populations that were visually selected for their apparent WB-resistance (absence or presence of few brooms in the canopy) presented susceptibility reactions (high number of brooms in the canopy) when introduced in the germplasm collections and tested for resistance. This fact demonstrates the inefficiency of the visual selection for field resistance in wild populations. Nevertheless, some selections that stem from these populations were identified as resistant (Allen & Cabrera, 1988; Fonseca & Albuquerque, 1999 and Rios-Ruiz et al., 2000). In the most recent evaluation of biclonal hybrids realized in ESEOP, in Ouro Preto do Oeste, Rondônia, the hybrids that included clones of the Brazilian Amazon were, in general, those that presented a smaller WB incidence (Fonseca et al., 1999a). However, an evaluation of the incidence of vegetative infections for many years, as well as the potential for production of basidiocarps of the inoculum sources should allow a more complete evaluation of the clone and hybrid field performance regarding WB. In 1998 and 1999 the field reaction of 120 accessions of the international collection (IC), of the Huallaga Collection (HC), and of the Ucayali Collection (UC) - the latter composed of local Amazon clones of a recent collection - were evaluated in the germplasm bank of UNAS, Peru. Wide variation was found for the number of vegetative brooms and of flower cushions (0.6 to 236.6 brooms/plant, in UC and IC, respectively). UC had the lowest infection levels; clones U 6, 10, 11, 12, 14, 19, 28, 25, 30, and 51 presented the highest potential for WB resistance improvement. HC presented intermediate to high infection levels. High infection levels were found for IC, including the clones SCA 6 and SCA 12, which formed 23.6 and 24.5 brooms per plant, respectively (Rios-Ruiz et al., 2000). These levels were similar to those found in Ecuador, though distinct from those found in Trinidad and Bahia, Brazil. This variation in the SCA 6 reaction is apparently due to genetic differences in the pathogen populations of the different countries. The first experimental evidence of the existence of genetic variation in C. perniciosa isolates from different countries was shown by Wheeler & Mepsted, 1988. These authors acknowledged the existence of two pathotype groups: the group of isolates from Ecuador, Colombia, and Bolivia and the group from Brazil, Venezuela, and of Trinidad. However, broader studies on genetic variation that would explain the differences found in some old and new resistance sources are badly needed. An attempt in this sense is actually being undertaken in the project CFC/ ICCO/IPGRI (Eskes et al., 2000). Sources with a high MO resistance level are not available for improvement, which is why the development of MO resistance is being emphasized in Ecuador (Suarez-Capello, 1999), in Costa Rica (Phillips-Mora, 1999b), in Peru (Rios-Ruiz, 1998) and in Colombia (Cardenas & Giraldo, 1986). Some of these sources have already been identified. As of the last decade, CATIE established an evaluation programme for MO resistance (Sánchez, 1982; Brenes, 1983; Phillips-Mora, 1999b and Phillips-Mora & Castillo, 1999). Up to 1999 nearly all of the 441 evaluated cacao clones were susceptible; only an extremely low percentage (2.3%) was classified as resistant or moderately resistant, such as ICS 43, UF 273, 712, CC 240, CC 252, EET 75, ICS 95, SPA 7, and BE 8 (Phillips-Mora, 1999b; Phillips-Mora & Castillo, 1999). Up to the present, little is known about the heritability of MO-resistance. Some ongoing studies in Costa Rica (Phillips-Mora & Castillo, 1999) indicated that M. roreri as well as P. palmivora resistance are apparently polygenic; although in the first case resistance is principally manifested as a recessive trait, while it appears as a dominant character for P. palmivora. This explains, at least partially, the high clone frequency in the germplasm collection with a high level of resistance against P. palmivora, in contrast to the low frequency for M. roreri. This also accounts for the relative high frequency of resistant progenies produced by the crossing between a P. palmivora resistant and susceptible clone. When an identical crossing is realized for M. roreri, all progenies show susceptibility. This fact justifies the search for novel resistance sources and/or the adoption of a genotype selection strategy with partial resistance that would delay the internal development of the fungus and reduce sporulation. The evaluations for MO resistance in Peru were initiated in 1995 with the cacao germplasm of UNAS when the disease severity increased. Similar results to those of Costa Rica are being found, i.e., high susceptibility of almost all evaluated clones. However, studies are aiming at a partial resistance of the sporulation component. Some clones, such as ICS 95 and others of the Ucayali series, have the ability of reducing the lesion development and as a consequence reduce the sporulated area, which are both very important parameters for disease control from the epidemiological point of view (Rios-Ruiz & Rivas, 1997). Derived from a cross between a national clone and an unknown hybrid, cultivar EET 233 in Ecuador presented consistent low disease incidence (Aragundi et al., 1988). More recently some cultivars were selected in evaluations from among the national cultivar collection of the ‘Centro de Cacao’ (see Chapter 12 on traditional cacao cultivars) for their resistance characteristics and are therefore being introduced into the improvement program of Pichillingue (Suarez-Capello, 1999). |
| Breeding strategies for resistance and future prospects. | Return To Table of Contents |
|---|
| Strategies of improvement for resistance: In spite of the great effort of screening the resistance sources and for transfer of resistance genes to agronomically favourable genotypes, few cultivars with good agronomic traits and high disease resistance levels have been developed by cacao improvement programmes. Maybe such efforts were not applied adequately. However, the absence of good resistance sources for most of the diseases might be responsible, at least in part, for the slow progress in the development of resistant cultivars. Although cacao breeders are aware of the risk run by using resistance sources with major genes, as are other breeders of other species, they also make use of the sources available for cacao. Still, there is no sound experimental evidence regarding the identification of major genes for disease resistance in cacao. Many resistance sources for fungi of the aerial part present varied resistance or tolerance levels. This partial resistance is, presumably, controlled by polygenes and similar to horizontal resistance (Fry, 1982). There are some practical difficulties in incorporating this type of resistance into germplasm with favourable agronomical characteristics, such as large fruits, high yield, high fat content, and superior organoleptic characteristics. Most resistance sources in cacao come from the Amazon basin and generally have low yield and undesirable fruit and seed size. The combination of Amazon genotypes with Trinitarios presents good possibilities to explore heterosis, achieving reasonable standards of fruit and seed characteristics in a simple selection cycle. This is the case of the cross between SCA 6 and Trinitarios (Pires et al., 1999a). However, in the selection of plants with best agronomical characteristics among crosses involving resistance sources, the resistance levels can be reduced. Nevertheless, a real evaluation of the components of this resistance would allow the use of these cultivars. In this sense, the use of such cultivars requires a good practical management of the partial resistance in order to reduce the disease severity and reproduction of the inoculum simultaneously. This way, the coexistence of host-pathogen is possible, in an economically viable, stable, and environmentally balanced manner. When resistance to multiple diseases is required it is difficult to accumulate sufficient polygenes to achieve good resistance levels against all diseases, especially if the resistance-controlling genes are inherited independently. Attempts to incorporate polygenes for resistance against two diseases can lead to the loss of resistance against the first when carrying out the selection for the second. Exceptions are possible when the same gene provides resistance to more than one disease, as may be the case in cacao. During the development of the WB-resistance, variation in the reaction to BP was also found (Pires et al., 1994 and 1999b and Luz et al., 1999b). Similar observations were made in evaluations for MO-resistance; several MO resistant genotypes were also resistant to WB (Aragundi et al., 1988 and Rios-Ruiz et al., 2000). Still, much research needs to be done to understand these resistance mechanisms. In many research centres there is also interest to select simultaneously for resistance against more than one disease. One strategy is to select for resistance against one disease amongst germplasm selected for resistance to another disease. For example, some Peruvian Amazon genotypes of the most recent collections, initially identified as bearers of partial WB resistance, also showed partial resistance against MO. Data from the first years of trials conducted in Peru showed the viability of this resistance strategy. In this sense the selection of resistant genotypes for one or more diseases should be realized based on the different levels of partial components of the resistance. Germplasm with varied levels of resistance to the main diseases are being used in Brazil (Pires et al., 1999c), Ivory Coast (Paulin & Eskes, 1995) and in Malaysia (Lockwood & Pang, 1994),= in an attempt to develop high-yielding cultivars with resistance to more than one disease by means of the recurrent selection method. The justification for the adoption of this procedure is the preponderance of the additive portion of the genotypic variance for some important agronomical traits. However, the accentuated genetic divergence of the subpopulations (see Chapter 6) also increases the chances of identifying transgressive hybrids. This process should consequently give rise to good results (Van der Vossen, 1997). Many of the improvement programmes for resistance have the objective of a pyramidization of resistance genes, based on the intercrossing of resistant genotypes and the conduction of recurrent selection cycles. In the specific case of Brazil, with the priority of developing WB-resistant genotypes, the improvement programme profits from the availability of a great pool of genetic resources in the CEPEC germplasm bank (Pereira et al., 1999) on which some information is already available. With the objective of developing clonal and seed cacao cultivars CEPEC set up a series of exploratory and regional competition trials of clones and hybrids, which are being evaluated for their reaction to WB. Crosses between promising genotypes are being obtained to form populations for the subsequent selection cycles or to be tested as clonal cultivars. Generally speaking, an improvement program should contemplate both improvement procedures in integrated form to derive seed as well as clonal cultivars (see Chapters 6, 9, and 13). However, since cacao has the advantage of easy vegetative propagation, clonal selection after recombination would be the recommended method to achieve progress in the short term (see Chapters 6, 9, and 13). Clonal selection in cacao is very old and it was and is still being intensely explored in countries such as Ecuador and Malaysia. Many expansion programmes of cacao are giving this method priority. A successful clonal improvement program is that of eucalyptus in Brazil. As a matter of interest the practice of clonal improvement in the country was presented in 1993 (Dias, 1993), where it was suggested that forest improvement should be the new paradigm for the improvement programme of cacao in Brazil. Future prospects: Much progress with improvement for disease resistance is to be expected in the next decade. Many disease resistance sources amongst wild and cultivated germplasm are being identified and some are currently found in use in improvement programmes. Improvement for disease resistance has become a priority in many cacao producing countries. Considerable attention is also being given to the use of wild cacao genotypes for disease resistance. The high level of partial resistance against WB, MO, and BP found in these materials can be exploited to generate high-yielding and resistant cultivars. It is believed that potentially resistant wild material, adequately evaluated and crossed, would allow for the accumulation of resistance. Finally, in various cacao research centres a number of researchers are developing methodologies for the use of molecular techniques in order to support cacao improvement in general and that of resistance in particular. Studies on genetic divergence among accessions, genetic linkage maps, and the identification of QTLs among others are being developed (see Chapters 6, 10, and 13). Such studies will permit accelerated improvement programmes for resistance, an enhanced use of germplasm and consequently higher genetic gains. In the future protocols of transformation and of regeneration of cacao plants could be readily available, similarly to what is beginning to happen with other perennial and forest crops. |
| Chapter 8. Biochemical and physiological bases of disease resistance. M.A.G. Aguilar & M.L.V. Resende. | Return To Table of Contents |
|---|
| Contents: Introduction; Resistance mechanisms; Pre- and post-formed structural factors and Pre- and post-formed biochemical factors; Biosynthesis of compounds involved in resistance and Factors and mechanisms of induction and accumulation; Resistance induction; Biotic and abiotic agents; Manipulation of environmental factors; Mineral nutrition and Final considerations and prospects. Summary: Various physiological, biochemical, and molecular events involved in the induction and the biosynthesis of secondary metabolites in higher plants are described with an emphasis on cacao. Additionally, evidence of the existence of structural and/or biochemical components that are part of the pre and post-infection resistance mechanisms in this species is given. Among these the phenols and procyanidins stand out because their quantification could be used as additional selection criterion for resistance against certain diseases and/or as complementary criteria for the classification of pathogenicity and virulence. The recent discovery that confirms the participation of phytoalexins as a post-infection defence factor in cacao opens up ample perspectives for their use as biochemical markers in the characterization and selection of resistant plants. An overall view of some chemical characteristics and the main pathways of biosynthesis is given, emphasizing the participation of important enzymes that can also be used as biochemical markers. Theoretical models of cellular communication are also presented. The role of induction molecules and/or induction agents that stimulate the host defence system is discussed, with examples and possible practical applications for cacao, emphasizing induced resistance and the manipulation of environmental factors. Finally, the lack of complementary research in certain areas of knowledge and the urgent need for the utilization of the existing and readily available conventional and non-conventional technologies are highlighted. The joint use of these technologies in improvement programmes will allow a better understanding of the mechanisms involved in the cacao-pathogen interactions, besides allowing for the selection and achievement of new sources of resistance in the short term. |
| Introduction; Resistance mechanisms; Pre- and post-formed structural factors and Pre- and post-formed biochemical factors. | Return To Table of Contents |
|---|
| Introduction: Their lack of mobility keeps plants from protecting themselves against the adverse environmental conditions and the attack of numerous potential predators and pathogens simply by not being able to move. Therefore, during the evolution process, particularly by the action of mutations and natural selection, higher plants developed several mechanisms which, by means of specific metabolic pathways, allowed them to synthesize, accumulate and secrete a great variety of secondary metabolites utilized in the formation of barriers of physical and chemical nature and whose main functions seem to be dispersion, protection, and/or self-defence (Taiz & Zeiger, 1998). Concomitantly, some pathogens co-evolved with their hosts, overcoming these barriers and causing diseases in the same (Ouchi, 1983). On the other hand the concentrations of these compounds, being considered undesirable for human consumption, were reduced or even eliminated from most economically important crops by means of the improvement and artificial selection practiced by humankind. Consequently, these crops became more susceptible to pest and disease attacks (Harborne, 1993). However, in most cases the plant defence mechanisms are more efficient than the biochemical attack mechanisms of the pathogens, so that resistance is more frequent than susceptibility, especially in natural ecosystems (Vidhyasekaran, 1988). In the specific case of cacao, due to its being a perennial crop of relatively recent domestication and therefore with few selection cycles practiced by man, the concentrations of these secondary metabolites would have been less affected, particularly in certain Upper Amazon Forastero genotypes. In the Upper Amazon region, supposed centre of origin and diversity of the species (see Chapter 3), there are wild or half-wild cacao populations which show rusticity, vigour, and disease resistance/tolerance (Toxopeus, 1985). Some secondary compounds are partially responsible for certain organoleptic proprieties and/or stimulants in chocolate (Griffiths, 1960; Jalal & Collin, 1979 and Cross et al., 1988), which should also have, at least in part, contributed to preserve the concentrations of the same in some genotypes, theoretically favouring the manifestation of resistance. In cacao, however, there are several diseases of economic importance, in particular witches’ broom, caused by the basidiomycete Crinipellis perniciosa, black pod, and monilia pod rot, caused by the fungi Phytophthora spp. and Moniliophthora roreri respectively and Verticillium wilt, whose causal agent is the fungus Verticillium dahliae. Actually, witches’ broom is considered the most important cacao disease in Bahia, in the Brazilian Amazon and in countries such as Bolivia, Ecuador, Colombia, Peru, French Guiana, Venezuela, Suriname, Trinidad, and Granada. It is estimated that the losses in the cacao bean yield caused by these and other fungal diseases reach 720 thousand tonnes annually (Holderness, 1993) or about 27% of the global production. In this respect numerous attempts have been made to obtain novel resistance sources against diverse diseases of fungal, bacterial, or viral origin (see Chapter 7 on improvement for disease resistance). Notwithstanding, the biochemical and physiological bases of the resistance mechanisms of cacao to diseases are very poorly known and studied. Resistance mechanisms: The resistance of plants to phytopathogens is a genetically determined, highly dynamic and coordinated process whose manifestation and efficiency depends on the existence of constitutive mechanisms and/or on the activation of the responsible genes by the induction of post-formed defence systems at the appropriate moment, place, and magnitude, after pathogen recognition by the host. It therefore consists of a system of multiple mechanisms, in which the level of expression is result of the sum of the individual contributions of each one plus the environmental influence (Pascholati & Leite, 1995). In this manner higher plants can react active or passively to the attack of phytopathogenic agents, using mechanisms or factors of pre-formed or pre-infectional resistance (passive, constitutive), pre-existing in the plants before the contact with the pathogens, and post-formed or post-infectional (active, inducible), inexistent or present at low levels before the infection, which are synthesized or activated in response to a tissue invasion by the pathogen (Misaghi, 1982 and Pascholati & Leite, 1994). In both cases, the factors can be of physical or structural and/or biochemical nature, while some of them ambiguously fit into both categories. Evidence for the existence of pre-formed and principally post-formed resistance in cacao factors is mostly circumstantial, inconclusive, and relatively rare in literature. However, the study, characterization, and systematization of these factors are of extreme importance for a more profound understanding of the resistance mechanisms, with a view to their application and utilization in adequate form in selection and improvement programmes, wether by conventional methods or through techniques of molecular biology and plant transformation. Pre-formed structural factors: The existence of pre-formed structural barriers in cacao becomes apparent by the higher concentration of epicuticular wax found in fruits of the Phytophthora-tolerant clones PA 150 and SCA 6, in relation to the more susceptible clone SIC 2, while a higher concentration is found in the mature leaves of the clones TSA 516, STAHEL, SCA 6, SIC 2, and ICS 1(Sena Gomes et al., 1995). Higher epicuticular wax concentrations were also observed in young and in mature leaves of the clones PA 150 and SCA 6 than in the common cultivar (Sena Gomes & Mached, 1994). On the other hand the low stomatic frequency (Sena Gomes & Rocha, 1995) and greater cuticular thickness (Sena Gomes & Mached, 1994) in SCA 6 seem to be more important structural factors than the stomata size for this clone‘s tolerance to certain pathogens. These anatomical and morphological traits obviously contribute individually to the global defence system of cacao against Phytophthora and other pathogens, but the combined effect of the cuticular thickness and stomatal frequency and size have a higher degree of correlation with the resistance (Iwaro et al., 1994). The high trichome density on the buds can also be a passive structural factor that can delay or hamper the entrance of the pathogen by impairing the deposition of the fungus’s reproductive structures onto the tissues (Tovar, 1991). Post-formed structural factors: The occurrence of anatomical and histological modifications after the penetration of the fungus Crinipellis perniciosa is a post-formed structural defence form, apparently manifested in resistant clones by the formation of vesicles and hardening or early lignification of the tissues (Evans & Bastos, 1980). Nevertheless, this type of reaction was observed in susceptible (‘Catongo’ and ‘ICS 43’) as much as in the medium tolerant (‘ICS 60’) genotypes and is supposedly a consequence of the pathogen-induced abnormal growth of the cortical cells. This is followed by the vesicles’ rupture, exudation, and subsequent darkening and necrosis of the cells exposed to substances within the same, besides inducing the formation of new cells that did not present plasmodesmata in their walls (Dabydeen & Sreenivasan, 1989). Later, histopathological studies showed that hyperplasia in the cortex and the vascular tissue, as well as hypertrophy in the cortex and secondary phloem occurred more intensely in witches’ broom-susceptible clones. On the contrary, the formation of a physical barrier as a result of a greater cambial activity and the darkening with posterior cell necrosis, limiting the development of the fungus, gave evidence of the existing inducible mechanisms of compartmentalization and hypersensitivity in tolerant genotypes (Laker et al., 1991), with a probable co-participation of phenolic phytoalexins. The occlusion of the xylem vessels by vascular gels or gums with high sulphur concentrations is also another form of histological defence induced after Verticillium dahliae infection in cacao (Cooper et al., 1996 and Resende et al., 1996). The gels formed consist of pectin and hemicelluloses and can be stabilized either by the formation of cross-linkages with phenols or impregnated with antimicrobial substances such as the actual sulphur (Cooper et al., 1996). Pre-formed biochemical factors: Studies realized in the 50s showed that the mycelial growth of Phytophthora palmivora in cacao tissues with lesions varied according to the clone utilized, which is probably associated to the resistance level of the tested genotypes (Orellana, 1954). Years later, Spence, 1961, suggested that the enzymes in the polyphenoloxidase system were important for the manifestation of resistance to P. palmivora. The activity of this enzyme was higher in clone SCA 6 than in ICS 1, considered resistant and susceptible, respectively. Turner, 1965, however, did not find significant differences in the polyphenoloxidase activity among resistant, tolerant, and susceptible plants, and concluded that this activity is not a factor that confers cacao fruit resistance against the pathogen. On the other hand, Prendergast & Spence, 1965, suggested that the activity of the polyphenoloxidase is a determinant factor of resistance because it acts by producing oxidised phenolic compounds that quickly limit the pectolytic activity of P. palmivora, avoiding fungal proliferation. Rocha, 1966, investigated the resistance mechanism of cacao to the fungus P. palmivora, with a view to study the content and fungitoxic activity of the phenolic substances present in the fruit tissues of the clones UF 613, UF 29, and UF 221. According to the author, the resistance mechanism in cacao fruit is of a physiological nature and depends fundamentally on the presence of certain specific partially identified polyphenolic substances in the epicarp able to inhibit the germination, formation, and development of the germination tube of the zoospores of P. palmivora. In view of the importance of certain phenolic compounds such as components of chocolate colour, taste, astringency, and flavour, several studies were realized in order to isolate, quantify, and identify these substances in cacao beans (Forsyth, 1952 and 1955; Forsyth & Quesnel, 1957 and Griffiths, 1958 and 1960). The results obtained confirmed the presence of anthocyanins, leucocyanidins, catechins and of some phenolic acids. The procyanidins and catechins represented about 70% of the total polyphenols in the cacao bean (Forsyth, 1955). Later studies also confirmed the presence of these substances (Jalal & Collin, 1977 and 1979; Cross et al., 1985 and Porter et al., 1991), besides other hydroxycinnamic derivatives and a coumarin - esculetin (Cross et al., 1988). In respect to other organs the pioneer study of Griffiths, 1958, demonstrated the distribution of phenols in different plant parts, among other objectives, to obtain basic clues on the probable participation of these compounds in the disease resistance mechanism of cacao, in view of that “there is evidence of intense metabolic activity involving polyphenols produced in the tissues of Theobroma caction after an invasion by the fungus Marasmius perniciosus”, the original name proposed for Crinipellis. Thus, it is quite likely that this is one of the first reports that tried to relate these compounds to witches’ broom-resistance in cacao, though more concrete results in this sense were apparently never published. Subsequent studies were able to associate certain constitutive phenols with the resistance of some genotypes against Ceratocystis fimbriata. Capriles de Reyes et al. 1966 found higher contents of polyphenols in plantlets of the resistant hybrid IMC 67 x TSH 644 than in the susceptible clones OC 61 and Porcelana; the content of chlorogenic acid in IMC 67 clones, although in concentrations below those required for fungitoxicity in vitro, was twice as high as in ‘OC 61’. Later, Capriles de Reyes & Reyes, 1968, found other phenolic compounds in higher concentrations in ‘IMC 67’, among which an ester of the gentisic acid was detected at fungitoxic levels in the trunk of this clone. Follow-up studies showed that an inoculation of cacao fruits with the fungus Phytophthora megasperma would also accumulate more phenols and tannins in a shorter time in a resistant genotype (‘La Providencia 63’) in relation to the susceptible (‘Porcelana’ and ‘IMC 67’), besides maintaining higher levels of these compounds for a longer time (Reyes et al., 1981). This type of response in which certain pre-formed compounds increase after infection is designated semi-constitutive or partially induced resistance (Ingham, 1973 and Resende, 1996). In the interaction cacao-V. dahliae, for instance, the content of condensed tannins may almost double after the inoculation with the fungus (Cooper et al., 1995 and Resende et al., 1996). Condensed tannins or procyanidins have fungitoxic proprieties and are very abundant in shoot tissues of cacao, which can contain up to 10% (p/p) of these polyphenols (Brownlee et al., 1990 and Brownlee, et al., 1992). For these reasons, some authors suggested that the quantification of the same could be used as an additional selection criterion for resistance to witches’ broom (Brownlee, 1990 and Scott, 1991) and/or as a complementary criterion for the classification of the pathogenicity and virulence of isolates of the fungus (Andebrhan et al., 1995). Besides the fungitoxic action of the procyanidins, the inhibitory effect of other phenolic compounds and of cacao leaf extracts on the germination of basidiospores was confirmed by Bastos, 1988. With the aim of clarifying some pre-infectional biochemical mechanisms of cacao resistance to C. perniciosa, Aguilar et al., 1999, observed a higher content of soluble phenols in healthy leaf tissues of clones that presented some level of disease resistance, especially in the clone TSH 1188, considered to be among the most resistant under field conditions. Post-formed biochemical factors: The principal mechanism or post-infectional biochemical resistance factor are the phytoalexins, normally defined as “antimicrobial compounds of low molecular weight, which are synthesized de novo by plants and accumulated in the plant cells in response to infections by phytopathogenic microorganisms” (Kuc´ & Rush, 1985; Harborne, 1993 and Pascholati & Leite, 1995). However, other stresses can also induce phytoalexin synthesis (Ebel, 1986; Goodman et al., 1986 and Dixon & Paiva, 1995). In fact, plants also catabolize phytoalexins, so that their accumulation level is determined by the ratio of the catabolism and the biosynthesis rate (Swain, 1977 and Érsek & Kiraly, 1986). Additionally, some pathogens have detoxification mechanisms (Van Etten et al., 1989) and/or synthesis suppression mechanisms (Yoshikawa, 1983), which can furthermore influence their final concentrations. The phytoalexins have been profoundly studied in many plant species (Kemp & Burden, 1986 and Grayer & Harborne, 1994) ever since their participation in the pathogen-host interaction was suggested by Müeller & Borger in the early 40s. In cacao, nevertheless, the participation of the same as a post-infectional defence factor has only recently been proved (Resende, 1994; Cooper et al., 1995 and 1996 and Resende et al., 1996), opening new and wide perspectives for the realization of basic and applied studies in this area. The participation of phytoalexins as one of the main components in the system of induced resistance of Theobroma spp. was first suggested by Daguenet & Parvais, 1981 who associated the occurrence of some substances that emitted blue fluorescence to an increased resistance of T. grandiflorum to P. palmivora. Later, by using biotic and/or abiotic agents, it was possible to induce the formation of fungitoxic substances that developed violet fluorescence and conferred resistance to Phytophthora spp. in T. caction, T. bicolor, and T. grandiflorum (Ibarra et al., 1985). In cacao, the presence of a golden fluorescence in the cells around the stomata invaded by C. perniciosa was also observed, which was more intense in the resistant genotypes and seemed to inhibit the colonization of the adjacent tissues (Frias et al., 1991). These substances were probably phenols, although they have not been isolated and identified. The participation of phytoalexins as a post-infection factor in the defence of cacao has only recently been proved. For the first time, it was possible to isolate and identify four phytoalexins that attained concentrations of maximal inhibition 10-15 days after inoculation of plants of the clone Pound 7 with V. dahliae, giving one triterpenoid (arjunolic acid), two phenolics (4- hydroxyacetophenone and 3.4-dihydroxyacetophenone), and an elementary sulphur-rich compound (Figure 8.1) (Resende, 1994; Cooper et al., 1995 and 1996 and Resende et al., 1996). Baker & Holliday, 1957, report the occurrence of a hypersensitive response (HR) in Trinitario clones and in crosses where one of the parents was a highly witches’ broom-resistant Amazon Forastero. More recently, Resende & Bezerra, 1996, induced HR in cacao using a C. perniciosa isolate of Solanum stipulaceum. The hypersensitivity reaction leads to the quick and localized death of the neighbouring host cells at the site of the pathogen penetration, restricting the development of the phytopathogenic microorganism. In the interaction incompatible pathogen-host (pathogen x resistant plant), this reaction normally occurs together with a marked synthesis and accumulation of phytoalexins (Misaghi, 1982). Compounds of the phytoalexin type have already been found after the ‘SCA 6’ inoculation with C. perniciosa, but have not yet been identified (Andrade et al., 1999). |
| Biosynthesis of compoundes involved in resistance and Factors and mechanisms of induction and accumulation. | Return To Table of Contents |
|---|
| Biosynthesis of compounds involved in resistance: The metabolic pathways and enzymes of secondary metabolism are generally poorly studied in cacao. However, the general biochemical pathways presented in this Chapter are common to most of the higher plants including cacao. The biochemical compounds responsible for the resistance mechanisms are quite a heterogeneous group, since they belong to different chemical classes. Most of them are phenolic compounds of the phenylpropanoids’ pathway that include phenols of the flavonoid type; others are isoprenoids, a class that covers the steroids and terpenoids (monoterpenoids-10C, sesquiterpenoids-15C, diterpenoids-20C, triterpenoids-30C, etc.); and derivatives of fatty acids such as the polyacetylenes and furan-acetylenes (Misaghi, 1982; Ebel, 1986 and Pascholati & Leite, 1995). Besides these substances that frequently need at least one free hydroxyl radical in the molecule to manifest their activities (Paxton & Groth, 1994), Mayama et al., 1981 also detected the presence of phytoalexins that belong to the group of compounds that contains nitrogen, which include alkaloids, cyanogenic glycosides, and glucosinolates. The chemical structures of the only phytoalexins identified so far in Theobroma cacao L. are shown in Figure 8.1. The secondary metabolytes are biosynthesized along three basic metabolic routes: acetate-mevalonate, acetate-malonate, and acetate-shikimate (Érzek & Kiraly, 1986), also simply called the mevalonic acid, malonic acid, and shikimic acid pathway, respectively (Taiz & Zeiger, 1998). However, several phytoalexins can be synthesized by the joint participation of more than one pathway (Kuc' & Rush, 1985). Figure 8.2 displays a schematic representation of these three biosynthetic routes which also shows the existing interrelations among them and the primary metabolism. In higher plants the shikimic acid pathway occurs in plastids and there is also evidence for its presence in cytosol (Hrazdina & Jensen, 1992). This important metabolic pathway begins with the phosphoenolpyruvate (PEP), derived from glycolysis, and erythrose-4-phosphate derived from the route of the monophosphate pentoses and of the Calvin cycle, resulting in the biosynthesis of the amino acids phenylalanine, tyrosine, and tryptophan (Salisbury & Ross, 1992). Hermann, 1995, considers that, in reality, the final product of this pathway would actually be chorismate, the last common precursor of these amino acids, whose three terminal routes of biosynthesis use this compound as the first substrate. In the first step of the shikimic acid pathway, the PEP and the erythrose-4-phosphate are condensed to a heterocyclic compound of 7 carbons, the 3-deoxy-D-arabino-heptulosonate 7-phosphate (DAHP). In the next stage dephosphorylation occurs and one oxygen is substituted by carbon 7 of the DAHP to form the 3-dehydroquinate. The five remaining steps introduce a lateral chain and two of the three double bonds that convert this cyclehexane into a benzene ring giving rise to the chorismate through the action of chorismate synthase (Hermann, 1995). Once the chorismic acid is formed the mutase chorismate enzyme, first enzyme of the terminal route that involves other enzymes that give rise to the biosynthesis of phenylalanine and tyrosine, can become active, or otherwise the anthranilate synthase, the first specific enzyme in the biosynthesis route of tryptophan. (Schmid & Amrhein, 1995). The DAHP synthase (DS), the first enzyme of the shikimic acid pathway, has been purified in several species including carrot (Suzich et al., 1985) and potato (Pinto, 1984 and Pinto et al., 1986), which appears to be a homodimer activated by Mn2+ and tryptophan whose subunits have a molecular weight of 53 to 55 kDa. (Hermann,1995). Some isoenzymes have Mn2+ as cofactor (DS-Mn) while others require Co2+ (DS-Co) or other bivalent cations (p. ex. Mg2+) (Hrazdina & Jensen, 1992). According to Hermann, 1995, this enzyme is essential since it controls the carbon flux in the shikimate pathway, which can be regulated at the transcriptional and/or translational level. Other important enzymes such as shikimate dehydrogenase and 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase take part in the next stages of this pathway. Once phenylalanine or tyrosine are formed these amino acids can enter into the so-called general route of the phenylpropanoids, which give rise to diverse types of phenolic compounds. The pathway of the phenylpropanoids begins with the convertion of phenylalanine or tyrosine to cinnamic or p-coumaric acid, respectively, in a reaction catalyzed by the phenylalanine ammonia-lyase (PAL) in dicotyledons, and apparently by the tyrosine ammonia-lyase (TAL) in monocotyledons (Jones, 1984). These enzymes are two “key” regulatory enzymes and one of the main control points of the secondary metabolism of plants (Bell, 1981). PAL is a tetrameric enzyme with two active sites per molecule (Bolwell, 1988). Apparently the isoenzymes isolated individually presented Michaelis-Menten kinetics, while together they would present negative cooperativity (Whetten & Sederoff, 1995). Its activity is influenced by several external and internal factors such as: hormones, nutrient levels, light, infection by pathogens, and injuries. Fungal invasion for example induces the transcription of the mRNA that encodes this enzyme, hereby increasing its de novo synthesis and consequently stimulating the production of phenolic compounds (Jones, 1984). In cacao, preliminary results of determinations realized in leaf extracts indicated a greater activity of PAL in clone SCA 6, resistant to the C. perniciosa fungus, than in the susceptible clone ICS 39 (Lopez, 1995). After affecting injuries or inoculating with P. palmivora, a 100% increase in the PAL activity was observed in stem extracts of the black pod-resistant clone IMC 67, compared to extracts of healthy stems (Okey et al., 1997). IMC 67 also presented a greater activity than the clones ICS 1 and TSH 1188, classified as moderately resistant, and TSH 1076, SCA 6, and P 18, considered susceptible to the disease. The greater activity of PAL observed in the black pod-resistant clone seems to be one of the factors responsible for the higher production of phenols and lignin, which in the last instance would be the compounds that confer resistance to P. palmivora (Okey et al., 1995a and 1995b). The post-transcriptional control seems to play an important role in the regulation of PAL activity and can be associated to the phosphorylation of this enzyme or the inhibition by cinnamic acid, a product of the reaction catalyzed by the same (Dixon & Paiva, 1995). Diverse simple phenylpropanoids are produced from cinnamic acid via a series of reactions of hydroxylation, methylation, and dehydration such as p-coumaric, caffeic, ferulic, and sinapic acids and the simple coumarins, besides salicylic, benzoic and p-hydroxybenzoic acid, which also originate from cinnamate and p-coumarate in spite of having lost the lateral carbon chain (Dixon & Paiva, 1995). Lignin and suberin also are formed by the mixture of simple phenylpropanoids, varying in composition according to the species (Lewis & Yamamoto, 1990 and Whetten & Sederoff, 1995). Thereafter, the reaction catalyzed by the chalcone synthase (CHS) generates the skeleton of 15 carbons of the flavonoids, from which countless stress-induced phenylpropanoids are synthesized (Dixon & Paiva, 1995). Besides the PAL and CHS, some of the main enzymes involved in the common pathway of the phenylpropanoids are the following: cinnamate 4-hydroxylase, coumarate 3-hydroxylase, O-methyl-transferase, ferulate 5-hydroxylase and hydroxy cinnamate CoA ligase or 4-coumarate-coenzyme A ligase (Boudet et al., 1995). Some compounds of secondary metabolism also belong to the class of the isoprenoids, terpenes, or terpenoids, secondary compounds mainly derived from the mevalonic acid pathway, which begins with the joining of 3 CoA acetyl molecules whose product undergoes the subsequent action of the HMG-CoA synthase and HMG-CoA reductase, forming mevalonic acid (6C). The latter, by means of pyrophosphorylation, decarboxylation, and dehydration reactions produces isopentenyl pyrophosphate (IPP), which, together with its isomer, constitutes the basic activated subunits (5C) for the formation of the terpenes (Chappell, 1995). The biosynthesis of the terpenoids can thus be divided into four main stages: i) conversion of the CoA acetyl into activated IPP subunits; ii) activity of several prenyltransferases forming, from this IPP precursor, geranyl pyrophosphate, pharnesyl pyrophosphate, and geranylgeranyl pyrophosphate, which are the largest subunits of terpenes formation; iii) condensation of these subunits by the action of synthases creating the monoterpenes, sesquiterpenes, diterpenes, etc; and iv) finally reactions of oxidation, reduction, isomerization, conjugation, or other secondary transformation which confer the structure and proper characteristics of each terpenoid. Some enzymes of this pathway such as the acetyl-CoA acetyltransferase and HGM-CoA synthase use Fe2+ and quinone as cofactors, while the prenyltransferases (homodimers) require Mg2+ or Mn2+ in a proportion of two ions per catalytic site (McGarvey & Croteau, 1995). Another recently discovered important biosynthetic pathway of terpenoids involves intermediates of glycolysis and the photosynthetic carbon reduction cycle (PCR), however little is known about its details (Lichtenthaler et al., 1997). Factors and mechanisms of induction and accumulation: The de novo synthesis of the host’s chemical weapons suggests that the plant has to recognize certain signals emitted by itself or the pathogen which, in last instance, enables it to express its disease resistance (Yoshikawa, 1983). Thus, the term “elicitor” is used in a broader sense to designate the molecules that induce these active host responses (Dixon, 1986). The elicitors are normally classified as biotic and abiotic elicitors (Kuc' & Rush 1985; Ebel, 1986 and Stoessl, 1986). Dixon et al., 1994, consider elicitors to be molecules that can be released before or during the intrusion of the pathogens and are an integral part of the microorganism or plant surface and need enzymatic activity to be released. Darvill & Albersheim, 1984 suggest that the abiotic elicitors are not directly involved in the host-pathogen interactions and propose that the abiotic factors activate the biotic elicitors present in inactive form in uninfected plants. The released elicitors have specific receptors in the plasmatic membrane, which permit a selective perception of these signals at specific binding sites (Yoshikawa & Sujimoto, 1993). Some of the main reception sites of the elicitors are binding sites of glucans, of quitinas and of glycopeptides (Boller, 1995). Yoshikawa & Sujimoto, 1993, suggested that the specific binding site for the mycolaminaran derived from Phytophthora spp., an intracellular ß-1.3-glucan that elicits the accumulation of glyceollin, would be a thermolabile protein or a glycoprotein. The signal process developed by plant cells to notice and respond to intrinsic and/or extrinsic impulses to the plant, despite the different structural and functional characteristics, presents a certain analogy with that of animals and can be subdivided in three basic stages: i) perception of the signal realized by specific cell receptors that recognize a determined signal; ii) transduction of the signal that consists in the transmission of the same to its site of action within the cell, which can occur direct or indirectly (via secondary messengers, alterations in the phosphorylation of proteins, and by G proteins) and; iii) translation of the signal, which consists in the conversion of the signal into specific cell responses (Côté et al., 1995), as for example the activation of genes that induce the synthesis of certain enzymes responsible for the production of pathogenesis related phytoalexins and proteins (PRP’s). Some biochemical events that occur during the elicitation were summarized in a hypothetic model presented by Mehdy, 1994. This model suggests that the receiver protein of the elicitor was associated to the synthesis of active oxygen radicals (“oxidative explosion”) and to protein G so that when the elicitor binds to the binding site the GTP binds to subunit alpha of this protein, which is activated and released from the protein complex to induce the opening of the calcium channel, increasing the intracellular concentration of this ion. Calcium would activate the kinase protein which would in turn activate, by phosphorylation, a membrane-bound NAD(P)H oxidase that induces the formation of O2-. Concomitantly, the occupation of the receptor would activate a peroxidase associated to the plasmalemma, also resulting in the synthesis of O2- which, spontaneously, would undergo dismutation to H2O2, which is able to transverse the membrane. The active O2- and H2O2 compounds would provoke the death of the pathogen, while part of the H2O2 would take part in the formation of oxidative cross bindings in the cell wall and of the peroxidation of lipids of the plasmatic membrane, resulting in the synthesis of jasmonate and in the regulation of the expression of the host’s resistance genes. Recently Hammond-Kosack & Jones, 1996, Low & Mérida, 1996 and Dixon & Lamb, 1997 suggested other more complex models involving other factors, which would also result in the increase of the cell wall resistance in the localized and systemic induction of the defence genes, in the activation or biosynthesis of enzymes and phytoalexins, and in the hypersensitive response. It must be underlined that, whichever the most appropriate model to explain the induction of the defence mechanisms, there are a few details that differ among species. However, in general, the participation of the elicitor, product of the pathogen avirulence gene (gene avr) and of the receptor, product of the host resistance gene (gene R), in gene-to-gene interactions seems to be obvious (Hammond-Kosack & Jones, 1996; Knogge, 1996; Dixon & Lamb, 1997 and Taiz & Zeiger, 1998). Jasmonic acid can be one of the secondary messengers that participates in the signal transduction in the process of infection by pathogens or wounds, inducing localized resistance (hypersensitive response-HR) in plants (Reinbothe et al., 1994). Boller, 1995, reports that the application of this acid in cells treated with elicitors of several vegetal species induces phytoalexin accumulation in the same. Salicylic acid is also another factor that seems to participate in the induction mechanism of systemic acquired resistance (SAR) in plants (Raskin, 1992). The addition of salicylic acid to carrot cells, for example, results in the induction of the quitinase, a PR protein that induces the formation of elicitors (Müller et al., 1994). Quitinase has already been isolated, characterized, cloned, and sequenced in cacao. A considerable increase in the concentration of this protein (29 kDa) was observed after treating the vegetal tissue with autoclaved C. perniciosa mycelium or after inflicting injuries (Snyder, 1994). However, the mechanisms that involve the participation of salicylic acid in the transduction process were not yet totally clarified. In the case of cacao, the number of lesions, the injured area, and the percentage of the leaf area infected by P. palmivora in the clones ICS 1 and IMC 67 was significantly reduced after an application of salicylic acid onto the seed or soil or sprayed on the plant (Okey & Sreenivasan, 1996). However, the application on the seed reduced the disease symptoms to a greater extent than the other treatments. These results suggest that salicylic acid induces resistance to P. palmivora in cacao, so it might be one of the factors involved in the transduction of the signal responsible for the manifestation of the defence mechanisms in this plant-pathogen interaction. In general, the participation of jasmonate and salicylic acid as endogenous signals is justified by the capacity of both to induce resistance to the phytopathogens, by the induction of the expression of resistance genes and by the increase of the levels of these compounds in tissues close to or distant from the infection site (Yoshikawa et al., 1993). Similarly to jasmonate and salycilate, the treatment with calcium and cyclic AMP also promotes an increase in the accumulation of phytoalexins (Kurosaki et al., 1987). In cacao, some elicitors of biotic and/or abiotic origin have been efficiently used to induce resistance against Phytophthora spp. (Partiot, 1981; Ibarra et al., 1985 and Okey & Sreenivasan, 1996), Crinipellis perniciosa (Aguilar et al., 1998; Ram & Castro, 2000 and Resende et al., 2000), and Verticillium dahliae (Cavalcanti & Resende, 2000) as well as resistance against the swollen shoot virus or CSSV (Hughes & Ollenu, 1994). |
| Resistance induction; Biotic and abiotic agents; Manipulation of environmental factors; Mineral nutrition and Final considerations and prospects. | Return To Table of Contents |
|---|
| Resistance induction: The induced (IR) or acquired resistance consists basically of the activation of post-infectional mechanisms that had been inactive or latent by the use of biotic or abiotic agents without altering the genome of the plants (Pascholati & Leite, 1995; Romeiro, 1999 and Stadnik, 1999). IR is characterized by the absence of toxicity of the induction agent on the pathogen, by the need of a time interval for resistance expression after the application of the induction treatment, by the suppression of IR by specific inhibitors of the expression of the host defence genes, by the non-specificity of protection by being dependant on the plant genome, by the absence of response to increasing levels of the inductor, and by being local and/or systemic (Steiner & Schönbeck, 1995). The use of this phenomenon in cacao has been little exploited. Nevertheless, mainly in recent years, basic studies with quite promising results have been developed and applied in this area. Biotic agents: Studies on the use of biotic agents for resistance induction in cacao are relatively scarce in literature. However, some interesting results have been obtained by the use of cross protection with less aggressive races or isolates of certain fungi or viruses. The inoculation of cacao with a less aggressive race of P. palmivora induced resistance against another more virulent race of this fungus, leading to survival of 88% to 100% of the plantlets (Partiot, 1981). The duration of the induced protection was longer in plantlets that had been pre-inoculated several times, lasting up to 17 weeks. Ibarra et al., 1985, also managed to induce resistance to P. palmivora in cacaos of the Amelonedo group, using mycelium disks and zoospores of less aggressive P. palmivora and P. megakarya isolates. Similar results were obtained with the same fungus by Daguenet & Parvais, 1981 using clone P7, indicating a probable phytoalexin production after the resistance induction. The protection induced by mild virus isolates of the cacao swollen-shoot badnovirus (CSSV) had already been used since the early 40s as reported by Hughes & Ollenu, 1994. In spite of the initial success the pre-immunization programme was interrupted for a long time for unknown reasons. However, in the late 80s, due to the severity of the problem, especially in Ghana, pre-immunization was studied again as an alternative in disease control. In spite of the satisfactory experimental results the practical application of pre-immunization in cacao seems to depend on a weak strain with a greater capacity of protection and on an efficient method to obtain a 100% immunization of the seedlings (Rezende et al., 2000). In relation to witches’ broom, the pre-inoculation of plantlets with basidiospores of C. perniciosa, originated from the wolf apple (Solanum lycocarpum) reduced the disease incidence by 64% in relation to plantlets inoculated with the cacao isolate only (Resende, 1998). In general, despite the very encouraging results obtained by the use of biotic agents, their use for disease control under field conditions seems to be difficult to apply in the short term. In the case of fungal elicitors a mixture of biotic elicitors not yet chemically characterized has been sprayed onto plants, which on some occasions produce contradictory results. Results will surely be more consistent after the elicitors have been isolated and characterized. Abiotic agents: At a biochemical level, recent studies have focused on the mode of action of products synthesized in the laboratory from salicylic acid, characterized as activators of plant resistance, mainly BTH or acibenzolar-S-methyl that belongs to the chemical class of the benzothiadiazoles and presents very low toxicity (Moraes, 1998; Morris et al., 1998 and Osswald et al., 1998). This induction generally occurs by the activation of genes that encode a series of pathogenesis-related proteins, such as glucanase and chitinase, and enzymes involved in the phytoalexin and lignin synthesis, such as phenylalanine ammonia lyase (PAL), cinnamyl alcohol dehydrogenase (CAD), besides polyphenoloxidases and peroxidases (Reglinski et al., 1997 and Vidhyasekaran, 1988). The first study on the use of BTH as an inductor of resistance to C. perniciosa in cacao reports 61% less infected plantlets in the witches’ broom-susceptible Catongo cultivar when the product was applied at a concentration of 10 g(ai)/100L water 7 days before inoculation (Aguilar et al., 1998). Later, Resende et al., 2000, observed that this compound induced a reduced disease incidence that varied from 33.5% to 84.5% when sprayed 3, 15, or 30 days before inoculation onto cv. Catongo plantlets. The resistance inductor was more effective at reducing the disease when sprayed 30 days before the inoculation and at a dosage of 150 (ai)/100L water. The systemic effect of BTH in protecting new shoots was also demonstrated 30 days after resistance induction in the hybrid Theobahia (‘SCA 6 x ICS 1’). In another experiment, the effect of the BTH for the protection of seedlings of the cv. Catongo was compared to the effect of copper oxide and tebuconazole, all sprayed 15 days before the inoculation. BTH reduced the disease incidence by 60.2% compared to the inoculated control, demonstrating a better performance than tebuconazole and a tendency to outperform copper oxide, which did not differ statistically from either one. In trying to explain part of the inductor action mechanisms, the total content of soluble phenols and the polyphenoloxidase and peroxidase enzyme activities were evaluated at 3, 15, and 30 days after the pulverization of the plantlets with BTH. No significant difference in the phenol contents and the polyphenoloxidases activity was detected after pulverization; a significant increase in the peroxidase activity was verified in all periods evaluated, though. Under field conditions BTH reduced the number of brooms per plant and the number of infected fruits significantly when applied at 60 day intervals in doses of 0.5 to 1g ai/ plant (Ram & Castro, 2000), which showed a lot of promise for use in the system of integrated witches’ broom management for cacao. Spraying BTH 15 days before inoculation with Verticillium dahliae in plantlets of the Theobahia hybrid reduced the severity of the disease by 55.4% and increased the plantlet growth significantly (Cavalcanti & Resende, 2000). The activity of enzymes related to the phenolic metabolism was greater in plantlets treated with BTH, compared to those which were not treated, which indicates the probable participation of the phenols in the defence process (Cavalcanti et al., 2000). Another important abiotic inductor is the actual salicylic acid which, as previously reported (item ‘Factors and mechanisms of induction and accumulation’), has been used successfully to induce resistance against P. palmivora in cacao (Okey & Sreenivasan, 1996). Manipulation of environmental factors: Some environmental factors associated to edaphoclimatic conditions such as: humidity, temperature, light, wind, nutrients, soil pH, etc., can influence the manifestation of a particular disease negatively or positively (Bedendo, 1995). Among these mineral nutrition and soil correction (see Chapter 1) are those easiest to manipulate aiming at a more effective disease control. The disease control measures actually recommended recognize the need for an integrated management involving methods of cultural, chemical, biological, and genetic control. In this context the genetic control by the use of resistant or tolerant genotypes of high productivity and quality is of fundamental importance, since it permits a better cost/benefit relationship. Nevertheless, apart from human interference, the expression of a particular disease is result of the interaction between a plant, the pathogen, and the environment. Thus, although resistance is genetically controlled, it can also be influenced by environmental factors. Mineral nutrition, as one of the main environmental components of plants, seems to be one of the factors that most affects plants’ resistance or susceptibility to diseases and/or the virulence of the pathogens (Huber, 1994). In this way the possibility of manipulating mineral nutrition aiming at its use as a component of integrated disease management in cacao seems an interesting practice to induce or increase the resistance of this species against certain diseases. The participation of mineral nutrition in the manifestation of some cacao diseases will be outlined in the next sub-item Mineral nutrition: Many macro and micronutrients are considered important in relation to the incidence or severity of the diseases since, besides direct effects on growth and productivity, they influence nutritional, biochemical, physiological, and anatomical aspects of the plants (Marschner, 1995). These mineral elements participate in diverse events responsible for defence mechanisms such as cofactors, activators, inhibitors, and modulators in several reactions of the primary and secondary metabolism of plants (Zambolim & Ventura, 1993 and Marschner, 1995). Besides, some of them are considered to be resistance inducers as in the case of certain heavy metals (Pascholati & Leite, 1995). In this context Mn by its physiological, biochemical, and nutritional importance, has frequently been associated to the alterations observed in the resistance or tolerance of plants against certain diseases. In the case of cacao recent studies have highlighted the importance of an adequate nutrition to provide more favourable conditions for a manifestation of the diverse resistance mechanisms, especially emphasizing the participation of certain micronutrients such as the actual Mn for example (Batista et al., 1998; Nakayama et al., 1998a and 1998b; Aguilar, 1999 and Nakayama & Andebrhan, 2000a). Nakayama et al., 1991 stated that in cacaos attacked by witches’ broom Mn was exactly the micronutrient found in lowest quantity in the branches with brooms, which, according to the authors, was due to the low availability of the element in the soil and to its low mobility in the plant. Also, the cacao nutrition can be altered by the fungus (Basts & Pereira, 1994 and Nakayama, 1995), complicating the establishment of relationships of cause and effect in this pathosystem. Later it was verified that a Mn leaf application reduced the percentage of disease-infected plants (Nakayama et al., 1998b). Adding Mn in a concentration of 2.5 µM to the nutritive solution also reduced the percentage of plants infected by the fungus, though only in the hybrid Theobahia (disease-tolerant), while in ‘Catongo’ (susceptible) the presence of this micronutrient caused no difference (Aguilar, 1999). The author verified that the lowest infection observed at a concentration of 2.5 µM Mn coincides with a higher production of starch, sugars, and soluble phenols, especially in the first days after inoculation with the fungus. Besides this higher photosynthetic rates were also observed in the mature leaves. To explain the role of Mn in the induction of disease resistance, Graham & Webb, 1991, proposed various mechanisms: i) increase of the lignin synthesis; ii) participation in the biosynthesis pathways of soluble phenols and lignin; iii) inhibition of the induction of aminopeptidases caused by the pathogen for the production of amino acids required for its growth; iv) inhibition of exoenzymes such as pectin methylesterase, produced by certain fungi for the degradation of the host cell walls; and v) direct inhibition of the fungal growth by an increased Mn2+ concentration, attaining levels toxic to the pathogen. Besides these aspects, Römheld & Marschner, 1991, highlighted the participation of manganese as an activator of important enzymes such as deoxy-D-arabinoheptulosonate-7-phosphate synthase (DS) of the shikimic acid pathway, and the kaurene synthase of the isoprenoid pathway, as well as a constituent of the phytoene synthase, another enzyme of the isoprenoid pathway. Mn-superoxide dismutase (Mn-SOD), although it does not participate in the biosynthetic route of the phenolic compounds, is also activated by manganese, participating as a prosthetic group of this enzyme. According to Bowler et al., 1992, Mn-SOD must also be involved in the defence mechanism against pathogens due to, among other aspects, the probable participation of the radical superoxides and the hydrogen peroxide in the reactions of hypersensitivity of the host cell. The use of silicone (Si), considered to be a beneficial element for plants (Marschner, 1995), has also produced satisfactory results with a promising potential for the use in witches’ broom control in cacao. Recent studies verified that over 200 mg of Si/kg leads to the germination inhibition of basidiospores of several fungus isolates, and that it is possible to increase or induce resistance in some cacao clones by the use of this element (Nakayama & Andebrhan, 2000b). The diminishing effect of Si on the disease incidence is probably associated with the activation of defence genes responsible for the production of enzymes that participate in the biosynthetic routes of secondary compounds, such as phenols, lignin, and phytoalexins, as well as in the process of silication of the cell walls and of the epidermis which increases pathogen penetration-resistance (Marschner, 1995 and Lima Filho et al., 1999). Spraying with 3% urea also significantly reduces the incidence of witches’ broom in cacao fruits. It also causes the inhibition of basidiocarp production when applied on dry brooms in concentrations between 2% and 6% (Bastos, 1998). Final considerations and prospects: In the cycle of the pathogen-host relationships the progressive establishment of the diverse stages of the infection process induces a series of events at the molecular, biochemical and physiological level, which result, among other aspects, in the synthesis of pathogenesis-related proteins (PRP’s), phytoalexins, and other metabolites conventionally designated “secondary metabolytes”. There is an unquestionable lack of information on the subject for cacao. However, the discovery of the participation of the phytoalexins in the interaction cacao-V. dahliae allows the sight of a more encouraging panorama that could open more profound insights into the mechanisms responsible for resistance and help, among other aspects, establish a more adequate and effective disease control. The use of these substances together with procyanidins as biochemical markers of resistance for selection and improvement targets is also noteworthy. The use of biotic and abiotic agents for resistance induction is a form of disease control which appears very promising. In relation to environment manipulation, the adoption of practices that aim at the increase of the nutrient contents that participate as defence gene inductors and as cofactors and/or as enzymatic activators in the biosynthetic routes of secondary compounds seems interesting as a complementary measure so the plant can express its resistance potential in an adequate form. The use of grafting for a quicker establishment of graft/root-stock combinations with a greater degree of resistance and desirable agronomical characteristics represents another easily accessible technique which, together with cuttings, has produced excellent practical results over the lasts years. On the other hand, for more advanced technologies such as micropropagation, tissue culture, isoenzymes, molecular markers and plant transformation (Chapters 10 and 13), for the most part, there are already sufficient scientific subsidies available and accessible for application in cacao, with diverse objectives. In this sense, the principles, several theoretic aspects, and practical applications of most of the technologies previously cited were carefully revised and analysed by Dias, 1995. However, despite specific protocols being already available, some of these techniques were apparently not adequately explored. The isoenzymes, for example, have principally been applied to study enzyme systems of primary metabolism, leaving the enzymes of secondary metabolism in second position. Perhaps this is one of the reasons for the low level of polymorphism found in various studies realised in certain tree species (Forrest, 1994) and in cacao itself (Blaha, 1988 and Lanaud et al., 1993). Additionally, the availability of transformed plants able to synthesize enzymes, phytoalexins, and other secondary metabolytes at higher levels is also another important goal to be attained aiming at a more effective control of certain diseases. Finally, the combined use of conventional and non-conventional techniques in the improvement programmes should permit a deeper understanding of the mechanisms involved in the cacao-pathogen interactions, besides allowing for the selection and establishment of new sources of resistance in the short term. |
| Chapter 9. Clonal improvement. A.B. Pereira. | Return To Table of Contents |
|---|
| Contents: Introduction; Improvement programme; Clonal cultivars; Synthesis of clonal cultivars; Proposal for clonal testing; Clonal cultivars in cacao; Synthesis of clonal cultivars and Commercial clones; Cloning techniques; Cuttings and Grafting and Final considerations. Summary: Vegetative propagation or cloning is a common means of reproduction in diverse plant species and is particularly well used in eucalyptus improvement. It is an important strategy in improvement programmes since it is the quickest form of capitalizing on genetic gains. Once the superior genotypic combination is identified it can be fixed and multiplied by cloning. The continuity of the success in the implementation of this strategy still depends on a competent and well-designed genetic improvement programme which is able to furnish clones of good productivity, uniformity, and resistance to pests and diseases for commercial plantations. This linking of cloning to the sexual improvement programme is necessary since vegetative propagation is an end-of-line strategy and with its use a maximum genetic gain can be obtained in a single generation while no additional gain is obtained thereafter. As a consequence cloning must be inserted in a recurrent selection scheme where the progress in the increase of the frequencies of the favourable alleles is capitalized in the clones. It is with this focus that the present Chapter was conceived. |
| Introduction; Improvement programme; Clonal cultivars; Synthesis of clonal cultivars and Proposal for clonal testing. | Return To Table of Contents |
|---|
| Introduction: Cacao cultivation represents one of the rare cases among cultivated perennial species in which sexual reproduction replaces vegetative propagation (Dias, 1993). The latter was developed in the 30s (Pyke, 1933) and used on a commercial scale in Trinidad in the 30s and 40s. Later, however, many of these areas planted with clones were replanted with hybrids multiplied through seed (Purseglove, 1968). Vegetative propagation was also used in Brazil in the late 50s for the implantation and renovation of unproductive cacao stands (Yungtay, 1958). Bud grafting was predominantly used in Brazil while cuttings were preferred in Trinidad. Cacao improvement in Brazil has significantly contributed to the productivity increase on plantations (see Chapter 12). Due to the ease of implantation and low cost of seed plantings (the hybrid seeds were donated by CEPLAC) the use of hybrid seed cultivars inhibited the commercial utilization of vegetative propagation during at least four decades (1960 to 1990). As of 1989, with the dissemination of witches’ broom in the cacao region of southern Bahia (Pereira et al., 1989), cloning became strategic for the salvation of the country’s cacao cultivation. Cloning has since been recommended in the region not only for the implantation of new stands but also for the substitution of susceptible plants of existing plantations. Improvement programme: The good adaptation of the germplasm, the favourable biological characteristics in the sense of being an allogamous plant with high seed production and easily feasible artificial crossings and vegetative propagation point to the success in the implementation of new strategies in cacao genetic improvement programmes. The uniting of these favourable conditions ensures that such programmes can become ever more efficient at creating cultivars that meet the actual demand. In the case of Brazilian cacao cultivation this demand focuses principally on the offer of witches’ broom-resistant cultivars. Figure 9.1 presents a general scheme (Resende, 1997) that should guide the genetic improvement programmes of any perennial species, also valid for cacao (see Chapter 12). Basically, these programmes involve successive recurrent cycles and several population types. Starting from a given base population, the selection would be implemented at different intensities aiming at the constitution of production and breeding populations. The production population can consist of seed production orchards, clonal gardens, or hybridization fields, according to the breeder’s interest. Higher selection intensities can be adopted for the constitution of this population with the aim of exploiting a maximum of free genetic variability and capitalize on the immediate genetic progress. In the case of seed populations the selection may be intense since there are no great restrictions regarding the number of individuals to be recombined (Resende, 1994). On the other hand the constitution of the improvement population aims, in the long term, at the continuous and progressive increase of the frequencies of the favourable alleles through the various selection cycles. Genetic gain in the long term depends basically on the potential genetic variability, that is, on the variability that is maintained throughout the cycles and is released through recombination at the end of each cycle. The safe approach (without risk of loss of favourable alleles) towards the achievement of the selective limit of the populations requires the maintenance of the compatible effective population size (Ne). This way, the genetic gains in the improvement population must be maximized for the condition of restriction to the effective population size (Resende, 1994). The recombination procedures are related to the selective efficiency in two ways: i) the recombination stage generally coincides with obtaining progenies for evaluation in the subsequent cycle and, as certain crossing designs favour a more accurate selection, this stage is related to the selective efficiency of the subsequent cycle; ii) recombination can be realized in an unbalanced manner with greater emphasis on the individuals with higher genetic values. Thus, besides the maximization of the gain based on the utilization of accurate selection methods (Resende & Dias, 2000; see also Chapter 6), additional gains can be achieved by the use of the best individuals in greater proportion in the production population and with the selection of crossings in the breeding population (Resende, 1994). Clonal cultivars: Despite the vegetative propagation of cacao having been developed in the 30s (Pyke, 1933) and used commercially in the two following decades, the clonal stand was replaced by seed reproduction using biclonal seed hybrids from the late 50s onwards. This situation lasted over the following four decades. There are various reasons for the failure of clonal propagation. The principal one according to Dias, 1993, would have been the low yield performance of the clones compared to the seed hybrids. Clonal selection practiced in Brazil was inefficient. Even though such selection had only been practiced after two or more years of records on successive monthly harvests of the matrix, not one subsequent progeny test was conducted. The phenotypic selection practiced had no effect since dry cacao bean yield is a complex trait formed by some yield components of quantitative inheritance and of low heritability (see Chapter 6). Finally in the 80s countries such as Malaysia and Indonesia began using cloning in cacao on a commercial scale with great success (Toxopeus, 1985 and Kennedy et al., 1987). Since 1998 efforts have also been made in Brazil to renew the cacao plantations in the south of Bahia with clonal cultivars imported from Trinidad. It is worth emphasizing that in improvement programmes, cloning by means of bud grafting has long been used to propagate superior parents in seed orchards by cuttings. It reduces the height of the cacao trees making the hybridization and the harvest of the hybrid seeds easier. Moreover, as already mentioned, it is a strategy able to fix genotypes at any programme stage with the exploitation of the entire genotypic variance (additive genetic variance, dominance and epistasis). Cloning will represent a great leap in cacao improvement if manipulated correctly. However, if badly used it will result in great risks represented by the genetic vulnerability of excessively homogeneous clonal stands. Synthesis of clonal cultivars: The cultivars of some species of economic importance are commercially exploited based on vegetative propagules or apomictic seeds. These cultivars may come from introductions, from hybridization among genotypes, or from the mutation of a given genotype (Fehr, 1987). The steps that should be followed in the synthesis of a clonal cultivar are i) development of a base population with genetic variability; ii) evaluation of cloned individuals of the population; and iii) multiplication of the vegetative propagules of the new cultivar for commercial use. Notwithstanding, whatever may be the stages needed to obtain a clonal cultivar, it is indispensable to link asexual improvement (cloning) to sexual (improvement for hybrids and improvement of populations), as emphasized in Chapters 12 and 13. In this situation, the recurrent selection method is the most adequate (see Chapter 12 on the proposal for the method). Most of the traits of economic importance are under the control of numerous genes. The challenge is to increase the frequency of the favourable alleles of all these genes and consequently improve the expression of the trait. The difficulty is that the greater the number of genes the smaller the probability that all of them can be manipulated. This difficulty, together with the environmental effect on the expression of these traits demands that the improvement be carried out in cycles (Hallauer, 1992 and Ramalho, 1994). Due to this fact the frequency of the favourable alleles can only be increased continuously and stepwise by means of successive selective cycles, that is, by means of recurrent selection. Once again, recurrent selection aims at increasing the frequencies of the favourable alleles via continuous and stepwise selection and at broadening the genetic variability by means of intercrossing of the selected families. The success in the use of recurrent selection lies in identifying and selecting the superior genotypes when the plants and/or progenies are evaluated and simultaneously maintaining the population with sufficient variability to go on obtaining selective gains in the subsequent cycles (Ramalho, 1994). The method of reciprocal recurrent selection (RRS) was proposed by Comstock et al., 1949 and it aims at the selection of one population in function of its combining ability with another population and vice-versa. In this scheme two heterogeneous and preferably genetically unrelated base populations must be involved to meet the original recommendation to select simultaneously for the general and specific combining abilities (Allard, 1960). Therefore, the procedure also emphasizes heterosis, which is function of the genetic divergence among populations and of the dominance (Falconer, 1989). The various methods of recurrent selection can be classified in two groups: intra- and inter-populational recurrent selection. In the latter the improvement of a population is realized aiming at an improvement of its combination with another population. Yet the intra-population improvement is the most used and consists in the improvement of the population per se (Hallauer & Miranda Filho, 1981). Recurrent selection can also be realized at the individual level, also designated phenotypic or mass selection, as much as at the family level. In relation to the latter endogenous families S1, S2, etc. or non-endogamous, or families of half or full-sibs can be used. Except in the case of mass selection all the other methods of recurrent selection include three important phases: i) sampling of the individuals in the population that one wishes to improve to obtain the families; ii) evaluation of the families in experiments with replication preferably in various environments; iii) intercrossing of the superior families to form the population of the next selection cycle (Paterniani & Miranda Filho, 1987 and Hallauer, 1992). Actually some cacao farmers have been using mass selection to select the best individuals from their plantations, which are cloned and used in new stands or grafted to substitute crowns in the existing stands. In this case the selection aims at identifying productive matrixes which are resistant to witches’ broom caused by the fungus Crinipellis perniciosa. Despite this selection being based on the data of two or more harvests the genetic gain realized should not be substantial due to the low heritability for yield and to the also supposedly low heritability for resistance to witches’ broom. However, the progress could be meaningful if clonal test of these matrixes were made later on. As commercial cacao plantations were installed by means of seed planting, especially of hybrid mixtures, it is reasonable to suppose that there is a great quantity of genetic variability available for the selection of different traits. Other strategies for a greater efficiency of the improvement programme are the evaluation of clones in multi-site trials for an investigation of the clones x localities and clones x years interactions and the development of studies into early selection, as suggested by Dias, 1998, to make the recommendation system for new clones become much more dynamic. The Ivory Coast (Paulin & Eskes, 1995) and Malaysia (Lockwood & Pang, 1994) have practiced the improvement of cacao by reciprocal recurrent selection since 1990. In the Ivory Coast crossings among parents of the Lower Amazon and Trinitario groups were used to form the base populations, the latter for bearing genes for a high cacao quality. In Malaysia the programme is conducted by private initiative and uses parents of the Upper Amazon group (Dias, 1998). Schemes of reciprocal recurrent selection were also proposed for Brazil (Pires et al., 1999a; see Chapter 12). Proposal for clonal testing: Before being released commercially clonal cacao cultivars must undergo rigorous tests that evaluate their entire potential. As one deals with a perennial plant with a long production cycle, an erroneous recommendation of a given cultivar can cause the producer great losses. This precaution applies to any cultivar type. Note for example the case of the hybrid SCA 6 x ICS 1 renamed as “cultivar” Theobahia and recommended by CEPEC for planting in the south of Bahia. This hybrid when planted in a smaller spacing (high density) showed complete susceptibility to attack by the fungus Ceratocystis fimbriata (Dias et al., 2000). Plants to be evaluated in clonal tests can be obtained from three distinct sources: i. - Commercial areas - selection of matrixes in commercial plantations; ii. - Introductions - matrix selections in germplasm banks. Such banks consist of collections of live plants from different origins, evaluated for traits of economic importance such as productivity, disease resistance, fruit and seed traits, fat content and fat fusion point among others; iii. - Hybrid programmes - matrix selections which originated from controlled crossing plans aiming at the association of genes that control pest and disease resistance and of genes that determine higher levels of cacao productivity and quality. To allow greater selection intensity and ensure a higher genetic gain the clonal test is divided in two evaluation stages (I and II). At least three hundred new clones per year are implanted aiming to identify at least three superior clones at the end of each selective cycle in these tests (Figure 9.2). In the first stage the principal objective is to discard clones of minor potential, that is, practice negative selection, and not necessarily select the best. In this sense, the clones are evaluated with less rigour. Thus the adopted design will be the triple lattice in two replications and single row plots of five plants. A selection intensity of 10% will be applied after the second successive harvest year of the clones, although harvest monitoring should go on until the sixth year to confirm stage II. The traits to be evaluated in this phase are described in Figure 9.2. Although no specific results for clones are available the minimum period for an evaluation of the productive potential of cacao hybrids in the south of Bahia was calculated to be two years of successive harvests based on the coefficient of repeatability (Dias & Kageyama, 1998; Chapter 6). For the time being this period can be extrapolated for the evaluation of clones. The selected clones are evaluated again in stage II aiming at a better discrimination among those selected previously by means of stricter evaluations. In this case a greater number of replications and larger plots must be used to diminish the competition among clones. In this way the tests are implanted in lattice or randomized complete blocks with three replications and plots of 20 plants (four rows of five plants), with a the centre part of 10 plants (two rows of five plants) evaluated in two years of successive harvests and at three sites. The evaluation results of the tests at stages I and II at different ages are considered for the final selection of clones. In stage I the measurements of two harvest years are used and in stage II six years of successive harvests. This allows a better comparison of the materials’ performance including that between different years of planting. The selection of the best clones will be realized by means of the utilization of a selection index that takes productivity, disease resistance, height, plant vigour, quality, and other traits for each one of the evaluation stages into consideration, as described in the clone selection flowchart (Figure 9.2). To incorporate new gains the highest possible number of clones must be used at each evaluation stage. Moreover, the continuity of the genetic progress depends necessarily on the conduction of intra or inter-population improvement programmes (see Chapters 12 and 13), which are able to amplify the chances of generating and identifying new genotypically superior combinations. |
| Clonal cultivars in cacao; Synthesis of clonal cultivars and Commercial clones; Cloning techniques; Cuttings and Grafting and Final considerations. | Return To Table of Contents |
|---|
| Clonal cultivars in cacao: Alternatives for a commercial multiplication of cacao are available. The decision whether to plant clones or seed hybrids and cultivars depends on genetic and technological aspects. For traits and sites where the importance of the additive genetic variance is greater the two last alternatives should be equally appropriate. However, in situations where the genetic variation of dominance has more weight the use of clones is the most attractive alternative (Mori, 1994). On the other hand, in view of the difficulties of the implantation and management of clonal stands, these seem to be more adequate for plantations with a high standard of technology (Dias, 1993). The option for technically less equipped plantations would be the Theobahia “cultivar” which shows tolerance to the fungus Crinipellis perniciosa at satisfactory levels (Monteiro et al., 1995). Theobahia is propagated by seed and its seeds have already been widely distributed to the producers of the cacao region in Bahia since 1997 (Ferreira, 1997). The multivariate quantification methodology of genetic diversity by the D2 Mahalanobis statistic was also suggested to orientate the recommendation of clones for commercial stands (Dias & Kageyama, 1997). The use of this methodology ensures that only clones with a high performance and a large relative diversity are recommended for every region. This same statistic was also proposed by Dias et al., 1996 to process cultivar differentiation analyses, an interesting strategy for the protection of clonal cultivars (see details in Chapter 12). Cloning of superior hybrids as a way of fixing genotypes and eliminating irregularities would be the most promising alternative (Dias, 1995). In fact, Resende, 1999 and 2001) demonstrated that the best way to explore the effects of specific combining ability in its totality is the clonal selection of individuals within the best hybrids. Moreover, according to Dias, 1995, the strategies of vegetative propagation can accelerate the improvement of perennial crops that present long juvenile periods, as in the case of cacao, in shortening the selection cycle. Lastly, it can be affirmed that the witches’ broom-resistant clonal cultivars with high productivity and adequate bean quality play an important role in the actual context of the Brazilian cacao cultivation. They represent a significant reduction in the production costs as shorter plants with disease resistance are less labour demanding for broom removal and cause less expenses with fungicides besides generally making the management easier (Pinto & Pires, 1998). Synthesis of clonal cultivars: To control the witches’ broom incidence in the Brazilian cacao regions, CEPEC imported clone collections of Crinipellis-tolerant cacao from Trinidad. Among the known and widely used collections in the cacao producing countries of Central America, those of the series Scavina (SCA), together with other collections introduced from the Amazon region presented a greater tolerance to the disease in the test realized by CEPEC (Ferreira, 1997). Besides the Scavinas originating from Peru new witches’ broom resistance sources were identified in diverse clonal series, especially Cruzeiro of the Sul and RB, originating from Acre, MA, the Amazon, NA and Pound (P) of Peru and CCN of Ecuador (Pires et al., 1999b and 1999c). Controlled crossing programmes are being implemented in the search for an association of resistance-controlling genes with genes that determine higher yield and quality levels in cacao. These programmes aim to combine resistant but little productive types with those highly productive but with little resistance to witches’ broom. Progenies and commercial clones of the generation in question are obtained from the combinations among the population types, while the parents for the next generation are selected within each population based, principally, on the traits for which each population stands out. With this structure one desires: the association of resistance genes and increases in the stability level of the trait, increases for other traits of economic importance; achievement of clones and progenies without excessive genetic similarity, and exploitation of the hybrid vigour (Pinto & Pires, 1998). Preliminary data on the percentage of progeny plants infected with witches’ broom by natural infection under conditions of high inoculum pressure indicate that in combinations where both parents were susceptible only 30% of the plants were healthy. On the other hand this percentage is above 80% in the descendant combinations from SCA, a value surpassed in diverse descendants of the SCA combinations with resistant clones, such as those of the series Cruzeiro do Sul, CEPEC, Ca, Rb, Ma and CCN. This is the material that constitutes the second clone generation that CEPLAC should make available to producers in the future (Valle, 1999). Commercial clones: Five witches’ broom-resistant clones are being distributed by CEPLAC to renovate the plantations aiming at grafting and seedling production by cuttings: TSH 1188, TSH 516, TSH 565, imported from Trinidad; EET 397 from Ecuador; and CEPEC 42, developed by CEPLAC (Ferreira, 1997 and Pinto & Pires, 1998). These clonal cultivars have clone SCA 6 as the resistance source. Although they are not immune to the disease they present high resistance in the environmental conditions of the south of Bahia, as well as excellent productivity (Pinto & Pires, 1998). These five cultivars are self-incompatible but inter-compatible among each other so that clone TSH 1188 pollinates all the others. Cultivar ‘CEPEC 42’ and ‘EET 397’ pollinate ‘TSH 516’, ‘TSH 565’ and ‘TSH 1188’. ‘TSH 516’ and ‘TSH 565’ pollinate ‘TSH 1188’, ‘EET 397’ and ‘CEPEC 42’. When planting or grafting in the field it is recommended to mix the clones to ensure the plantation’s success (Pinto & Pires, 1998). Giving continuity to the renovation process of cacao plantations in the south of Bahia, CEPLAC is releasing two more seed “cultivars” (Theobahia 1 and Theobahia 2) and four new clones (TSA 654, TSA 656, TSA 792 and TSH 774) for the producers, amounting to three seed and nine clonal cultivars. However, the genetic improvement and the generation of new clones must be continual processes that make new planting material available to the farmer, thus amplifying the genetic base of the crop and reducing the risk of genetic vulnerability of the clonal stands as a result of the narrow genetic base of the clones presently in use. Cloning techniques: The conventional cloning techniques are of the cuttings and grafting types. Naturally, the propagules used depend on the type of cloning used, varying from buds to small shoots. The objective however is the same; that is - the fixation of the matrix genotypes. Cloning includes another great advantage which is the reduction of the juvenile period, accelerating the beginning of the productive phase of the cacao tree (Pereira & Souza, 1999). Cuttings: The need to multiply witches’ broom-resistant clones commercially in the cacao region of Bahia, where to date only multiplication via seeds had been in use led to the resumption of the cuttings technique. The technique of rooting cuttings was developed by Pyke, 1933, and optimized by Evans, 1951. Plants produced from rooted cuttings are considerably shorter than plants grown from seeds, which makes the field and harvest work in the first years difficult. But these difficulties can be eased by formation pruning during the development phase of the trees (Wood, 1985). The success of rooting cacao cuttings depends on a large number of factors; temperature, light and moisture where the cuttings are planted being the most important conditions. However, the various factors that influence the rooting process in increasing order of importance are the following: Cacao tree type: Wide variation has been found in the rooting ability among the diverse selected genotypes. The majority of the selections of the Amazon Forasteiros root easily (around 90% take), while the Criollo genotypes present around 60-70% rooting (Hall, 1963). When analysing only the clones of the series ICS among the Trinitarios many root easily while others do not. Seasons of the year: The percentage of rooting diminishes during the dry season although there are variations among the clones in relation to this susceptibility. Amazon clones are little affected by this factor while the reduction is evident in Trinitario clones (Toxopeus, 1970). Clonal gardens: Clonal gardens for a large-scale cuttings production are shaded to allow only 20 to 50% of the total sunlight to fall on the matrixes. As the trees age the cuttings lose vigour and every six years the clonal garden must be replanted. In Ghana the number of cuttings taken from a selection series planted in the clonal garden in a 2.6 x 1.3m spacing was evaluated. It was verified that Amazon selections produce 100-160 cuttings per plant/year, while Amelonado selections produce only 20-40. Clones that root more easily also proved to produce more cuttings per plant (Hall, 1963) and a similar result was observed in Nigeria (Toxopeus, 1970). Age of the shoots: For the success of rooting the shoots must be in the half-woody stage when the cuttings are taken. This stage is characterized by the light brown colour of the outer surface of the sprout (Wood, 1985). Hormone use: Tests with various hormones have shown that the best treatment is the immersion of the cutting base in a solution of 4-5g of naphthalene acetic acid (NAA) plus 4-5g of indolbutyric acid (IBA) in one litre of 50% alcohol (Evans, 1951). Temperature, light and moisture: For adequate rooting conditions low incidence of light, a fresh and humid environment (humid chambers or greenhouses with misting) are necessary during the rooting period (Wood, 1985), as shown in Figure 9.3. Conditions for rooting: Optimum air and substrate combinations must be established at the cuttings’ bases for quick rooting. Optimum conditions are also necessary for the growth and vigour of the root system, principally in the first thirty days (Wood, 1985). Grafting: The increase of the efficacy and efficiency of the grafting processes depends on diverse procedures that must be taken into consideration. These begin with the preparation of the rootstocks and matrixes, continue during the grafting itself and go on until the field planting of the grafted seedlings or the formation of the new canopy on the adult rootstock (Pereira & Souza, 1999). The shoots used are selected from the plagiotropic branches of the second and third flushes; light brown coloured and free from any kind of lesion. The diameter of the shoot depends on the type of graft to be done. Leaves are reduced to a third. The sectioned shoots are wrapped in newspaper or moistened burlap bags, and transported as fast as possible to the place of grafting. The shoots should be protected from winds and direct sunlight. Care must be taken to identify the clone and the date of collection so as not to mix them up. When the shoots are transported or used up to 3 days after collection it is necessary, besides a fungicide treatment, to seal the ends with wax and put them in a humid environment until the moment of grafting (Rosa, 1998 and Pereira & Souza, 1999). The shoots are divided into two sections: the larger one of a brown colour, which corresponds to the second and third leaf flushes which will furnish grafts or buds for top- and side-grafting (Figure 9.4 and Figure 9.5) or bud grafting (Figure 9.6); the smaller one of a green colour with young or partially mature leaves, corresponding to the first leaf flush, provides shoots for hypocotyledon grafting (Figure 9.7). Top full-cleft grafting (Figure 9.4) and side-grafting (Figure 9.5) are realized on two rootstock types: nursery seedlings and basal sprouts of adult plants. The grafting on basal sprouts has presented best results in the task of substituting canopies of adult cacao trees. The yield of the operation is 150 grafts in the case of seedlings and of 100 grafts in basal sprouts, both in a day’s work and with take indices of 80% and 90%, respectively (Rosa, 1998). Window-grafting is also called bud plate grafting, inverted “U” or closed window bud because the bark covers the bud for some days, protecting it (Figure 9.6). The time of duration of the process from the sowing of the rootstock until field planting varies from 10 to 12 months. The yield of this type of grafting is 200 grafts in a day’s work, with a take index of 80% (Rosa, 1998). Hypocotyledon grafting by side-grafting (Figure 9.7) is a process that already aims at maximizing the exploitation of propagules and reducing the time and cost for the cloned seedling production. It is also designated “green grafting” or “early grafting”, due to the use of grafts of young leaf flushes of the cacao tree grafted onto also very new rootstocks. According to the characteristics presented, this process requires that the shoots (later the grafts) are always protected from winds and direct sunlight so as not to lose moisture. The time of preparation of the seedling from the sowing of the rootstock to field planting varies from 5 to 6 months. The yield is 300 grafts in a day’s work (Rosa, 1998). Final considerations: The main advantage of using clones is that they maintain the integrity of the cloned individuals’ genotype including the traits of disease resistance and high productivity. Clonal seedlings formed from cuttings are destined for the formation of new plantations or replanting. Grafting directly onto fruit-bearing trees aims at the substitution of the witches’ broom-susceptible canopies. It is done in the trunks or basal sprouts, making use of the existing root system, which makes the plantations commercially profitable already from the second year after the grafting. Very soon it is to be expected that micro propagated seedlings can also be used on a large scale thus increasing the offer of plantlets (Chapter 13). Clonal cultivars are already being largely employed in the cacao region of Bahia. There is now a nursery (project “biofabrica”) for a regional large-scale production of these cultivars. The goal for 1999 was an output of 2,800,000 grafts and 1,200,000 seedlings. The clones TSH 1188, TSH 565, TSH 516, CEPEC 42, and EET 397 are presently being distributed. Other clones are in the multiplication process to be distributed as fast as possible. It is worth alerting that the use of clones needs be linked with a sound improvement programme that aims at the incorporation of other witches’ broom-resistance sources, thus broadening its genetic base and contributing to the increase of heterosis which is best capitalized on in clonal stands. |
| Chapter 10. Molecular markers in breeding. A.V.O. Figueira & J.C.M. Cascardo. | Return To Table of Contents |
|---|
| Contents: Introduction; Genetic markers; Genetic markers in cacao and Nuclear and chloroplast DNA. Applications of molecular markers; Diversity evaluation, Diversity in cacao in south Bahia and Determination of the centre of diversity. Applications of molecular markers; Genomic mapping. Applications of molecular markers; Genomic mapping - Considerations in cacao mapping. Applications of molecular markers; germplasm characterization, Genotype identification and Pathogen diversity; Conclusion. Summary: Markers are used as much for genomic mapping as for the evaluation of the genetic diversity in plant improvement. Cacao naturally has a high genetic variability that permits a broad manipulation of the species by improvement. Actually there is a range of clones and hybrids with the most varied traits originating from conventional improvement techniques. Nevertheless, the actual situation of cacao cultivation requires the release of new planting material to supply the deficiencies found in terms of more productive material, resistant to various biotic or abiotic stresses, or even material with a particular quality of industrial interest. The relative long cycle of cacao makes its improvement all too slow and onerous. Molecular markers can help the strategies of cacao improvement directly. Tracking down the traits of interest by means of markers permits acceleration of the process, as traits that would otherwise take a long time to be detected can promptly be evaluated. Markers furthermore allow that various traits of interest are introduced into a single genotype by monitoring this transference. The evaluation of the genetic diversity among several cacao accessions permits that crossings are planned using genetically contrasting materials. This chapter deals with the various aspects that involve molecular markers and their potential uses in the cacao crop, focussing on the plant and also its pathogens. |
| Introduction; Genetic markers; Genetic markers in cacao e Nuclear and chloroplast DNA. | Return To Table of Contents |
|---|
| Introduction: Success in the genetic improvement of plants by means of selection depends on the capacity of distinguishing genetically heritable factors from purely environmental ones. Since genetic markers are simple heritable units, they can increase the selection efficiency by identifying heritable differences when associated to the traits of interest (Staub et al., 1996). Genetic marker-assisted selection allows an early identification of the most desirable genotypes, reducing the population size to be evaluated and increasing the efficiency of the improvement process. The introgression of genomic regions can be accompanied and selected with greater efficiency when associated with molecular markers that also permit the pyramiding of resistance genes, adding barriers of phenotypic evaluation in the presence of more than one genetic resistance factor. The analysis of genetic marker segregation permits the construction of genomic maps by means of calculations of recombination between the various loci. Additionally, the co-segregation of the traits of interest in relation to the various genetic markers permits assisted selection. The variation in the allele frequency of the genetic markers also permits the characterization of the genetic diversity of individuals and populations, independent of the environmental effect. The description of individuals by means of genetic markers has application in the characterization of germplasm: identifying duplicate or related accessions; in the identification of cultivars and in the selection of parents to be used in crossing programmes. The characterization of the genetic diversity of populations by means of genetic markers has been used principally in; the analysis of germplasm collections, identifying duplications, optimizing the establishment of core collections, the evaluation of the collection and maintenance system through the identification of geographic regions of greater diversity and in basic studies of taxonomy and evolution. Another application of genetic markers in improvement consists in the characterization of the diversity of the main pests and diseases that affect the crop, so that adequate strategies for the selection of resistant cacao cultivars can be established. These various applications of molecular markers in cacao improvement are discussed in this chapter. Genetic markers: Previously, the ability of mapping genes depended on the synthesis of gene pools containing multiple mutations (morphological, cytogenetic and physiological), but as a result had a reduced viability and, therefore, the number of systems with such a flexibility was extremely small, rarely involving elite cultivars (Burr, 1994). The capacity of identification of polymorphisms in wild alleles, with the utilization of biochemical (isoenzyme) markers, amplified the possibility of mapping since in terms of survival these were neutral genes (Burr, 1994). However, the degree of polymorphism obtained was not satisfactory for the construction of saturated linkage maps. The development of molecular markers based on polymorphisms at the level of DNA size and sequence allowed the construction of saturated genomic maps for various cultivated species, which permit the detailed genetic analysis for a determination of the localization of genes that affect simple or quantitative traits (Paterson et al., 1991). The use of Restriction Fragment Length Polymorphism RFLP was initially described for humans (Botstein et al., 1980). RFLPs are markers inherited in a mendelian manner, derived from alterations in the DNA fragment size resulting from variations at the nucleotide sequence level. These markers occur in a theoretically unlimited number and are codominant and independent of the development stage and of the environment, with minimal pleiotropic effects. The random-amplified polymorphic DNA - RAPD, however, represents another type of molecular marker that uses the polymerase chain reaction. It presents the same genetic traits as RFLP, except for its dominant inheritance. RAPDs are anyway less complex than RFLPs since they dispense with previous knowledge on sequences, DNA cloning, digestion with restriction and transference enzymes for membranes and have therefore become broadly used. The polymorphism of RAPD markers has originated as much from the alterations of a single base, leading to the failure of initiation pairing, as by deletion of the initiation site or by insertion, altering the amplified DNA size (Williams et al., 1990). A disadvantage of these two marker types consists in the lack of polymorphism for some systems. RAPDs present problems of stability, and polymorphic fragments associated to traits of interest can be used as RFLP probes or else be sequenced for the development of specific primers for the Sequence Characterized Amplified Regions, known as SCAR (Rafalski & Tingey, 1993). To reduce the necessary operations in the RFLP test, a marker denominated Cleaved Amplified Polymorphic Sequence - CAPS was developed. This marker is based on the information available on the sequence of the locus of interest for the construction of specific primers, which are used for PCR amplification, followed by the digestion of the amplified products with various restriction enzymes to reveal RFLPs between individuals (Rafalski & Tingey, 1993). Another type of marker denominated Amplified Fragment-Length Polymorphism - AFLP, proposed by Vos et al., 1995, which also produces dominant but more stable fragments, combines the advantages of the use of the technique of PCR of the RAPD with the exemption from cloning demanded by RFLP. This method involves the DNA restriction by two enzymes, one of which cuts rarely and the other frequently, followed by the linkage to specific adaptors for the terminations produced by the enzymes, and the amplification of these fragments with primers of homologous sequences to the adaptors, restriction sites and with the extremity 3' with single differentiating sequences. When using radioactivity-marked primers, 100 to 150 fragments are produced and detected in acrylamide gel. Considering that there are some genome regions that present more polymorphisms than the single copy sequences (Burr, 1994), specific molecular markers were developed for these regions. The variable number of tandem repeats - VNTR found adjacent to the single copy sequences in plant genomes present a high level of polymorphism, with alleles that have a variable number of base repeats (Akkaya et al., 1992; Morgante & Olivieri, 1993 and Thomas & Scott, 1993). The variation in the number of repetitions seems to stem from the unequal crossing-over between chromatid sisters or DNA polymerase slippage during the replication process. According to the size of the repeated base sequence, these markers are classified in minisatellites (Repeated base sequence of 10 to 45 base pairs) or microsatellites (repeated sequence of 2, 3, or 4 bases, such as AT, ATT, AC, etc), and are being used as markers for one or various loci (Akkaya et al., 1992; Morgante & Olivieri, 1993 and Thomas & Scott, 1993). A probe of homologous oligonucleotides for one of these repeat base sequences reveals a complex of bands when hybridized with digested genomic DNA, representing various specific loci for each individual (“DNA fingerprinting”). For a better resolution of the potential of this marker type the sequences that flank these repeated regions, which can be used as primers in PCR reactions to amplify single sequence repeats- SSR are used (Akkaya et al., 1992). The loci distribution seems to be uniform for the entire genome. Markers of a single microsatellite locus can be obtained by sequences available in gene sequence banks and by the generation of specific primers (Akkaya et al., 1992 and Morgante & Olivieri, 1993), or by means of the isolation of positive clones of such a genomic library, after hybridization, with a probe of the basic sequence, followed by the sequencing of the fragments adjacent to the repeated sequence to generate specific primers (Thomas & Scott, 1993). The great disadvantage of the SSR markers lies in the cost of generating the specific sequences of the primers. Alternative primers were proposed to avoid the cloning and sequencing. Zietkiewicz et al., 1994 proposed the utilization of primers anchored by one to three bases in the terminations 5' or 3' of SSR (CA)n, used for the DNA amplification of various organisms and separated in acrylamide gel, producing information of various loci per reaction and segregating in a mendelian manner. This marker based on inter-microsatellite amplification - IMA or on the inter-simple sequence repeat - ISSR permits the revelation of polymorphism without previous knowledge of sequences and therefore has a great potential for mapping and characterization of germplasm. The microsatellite regions are found through the entire genome of eukaryotes and are very polymorphic in length (Charters et al., 1996). Genetic markers in cacao: In the case of cacao, only two simple genetic morphological markers have been described: axil spot (Harland & Frechville, 1927) and cotyledon colour (Wellensiek, 1931). The need of large areas to maintain segregant populations and the long juvenile period of cacao together with the presence of self-incompatibility systems and the difficulty of overlapping generations restricted the acquisition of more genetic information. These morphological genetic markers were used in crossings to investigate the self-incompatibility system and development of haploid cacaos in the Ivory Coast (Dublin, 1972 and Lanaud, 1987a). More recently, Falque, 1994, used the axil spot marker to detect the development of haploids based on pollination by irradiated pollen by the technique denominated in situ parthenogenesis. Morphological traits, principally those related to fruits and flowers, have traditionally been used to evaluate the variability among clones. The great disadvantage of this practice is that these traits are frequently expressed at plant maturity and are greatly influenced by the environment, needing large samples to obtain precision in the evaluations. However, the morphological markers have not been considered in inheritance and segregation studies and can therefore not all be considered to be genetic markers. Only Engels, 1986 assessed the inheritance of the qualitative morphological markers: leaf sprout colour, cotyledon colour, basic colour of the immature and mature fruit and shape of the fruit apex. For this investigation, the author employed a diallel with 7 cultivars and proposed the genotype for these parental traits. Recently, in the development of the genomic map of cacao, the presence of anthocyanin colouration was evaluated in sprouting leaves, in flower staminodes, and in cotyledons using an F1 ‘Catongo x Pound 12’ population and a backcross population of ‘Catongo x (‘Catongo x Pound 12’)’ (Crouzillat et al., 1996). ‘Catongo’ is a natural albino mutant without anthocyanin pigmentation in flowers, leaves, and cotyledons (Miranda & Silva, 1939). The absence of anthocyanin in leaves, seeds, and flowers did not segregate in the F1 plants, but segregated 1:1 in the backcross progeny suggesting that one is in fact dealing with a simple (Anth) gene with dominance for the presence of anthocyanin. Isoenzymes are proteins able to be distinguished by their weight or size that catalyze the same biochemical reaction and are allelic variants. These present numerous advantages over the morphological genetic markers, because they present simple and codominant inheritance, are less affected by the environment and can be evaluated as early as the seedling stage. In cacao, a restricted number of isoenzymatic systems presented polymorphism, essential for genetic and mapping studies. In a preliminary study, between 24 enzymatic systems tested for leaf tissues, 17 demonstrated activity and only nine exhibited variation, seven of them with replicable results (Atkinson et al., 1986). In another study, including more than 200 genotypes from various origins and some progenies, six out of 12 tested enzymatic systems presented polymorphisms (Lanaud, 1986b). These six systems revealed the existence of 10 genes with 24 distinct alleles and total genetic independence (without linkage) and were used to investigate the origin, the genetic relations and the structure of various cacao populations (Lanaud, 1986a). More recently, Warren, 1994, used four enzymatic systems with 20 alleles to evaluate the diversity of nine cacao populations and demonstrated that the populations of Ecuador and Colombia presented more diversity than the four evaluated populations (P, PA, IMC, and NA) originally from Peru which is supposedly the region that includes the centre of diversity of cacao (Warren, 1994). The genetic variability and its partition between and within populations, arbitrarily divided, were also estimated by means of six enzymatic systems, encoded by eight loci for 86 clones of the USDA germplasm collection in Miami, Florida, USA (Ronning & Schnell, 1994). The genotypes originating from South America, considered Forasteros, contained all alleles of the systems investigated, with significantly different frequencies from the clones of Central America and the Caribbean, considered Criollos and Trinitarians. The greater part of the diversity was detected within the racial groups (instead of between groups). The use of isoenzymes presents a great potential for the characterization of germplasm and identification of cultivars, but the level of polymorphism shown is low for applications in genetic mapping. Comparisons between the estimation of genetic diversity by means of morphological, isoenzymatic, and RAPD data have been conducted in Trinidad and indicated a low correlation between these types of markers (Bekele et al., 1998). The use of molecular markers of the RFLP type in cacao was first proposed by Fritz et al., 1986. The first results, using only a genomic probe and two restriction enzymes in the analysis of four clones and two progenies (46 individuals), indicated the presence of polymorphism (Mirazon, 1988 and Mirazon et al., 1989). The markers of the RFLP type have been widely used in studies into genetic diversity using heterologous ribosomal gene probes (Figueira et al., 1992; Laurent et al., 1993a and Figueira et al., 1994); nuclear genomic and cDNA probes (Laurent et al., 1994a, b; N’Goran et al., 1994; Lerceteau et al., 1997a; Lerceteau et al., 1997b and N’Goran et al., 2000) and heterologous cytoplasmatic probes (Laurent et al., 1993b). RFLPs were also used for the construction of genomic maps (Lanaud et al., 1995; Crouzillat et al., 1996 and Risterucci et al., 2000a). Various studies of genetic diversity were realized using RAPD markers (Wilde et al., 1992; Figueira et al., 1992; Lerceteau et al., 1993; Russel et al., 1993; Figueira et al., 1994; N'Goran et al., 1994; de la Cruz et al., 1995; Lerceteau et al., 1997a; Lerceteau et al., 1997b; Whitkus et al., 1998 and Marita et al., 2000). This last marker type was also used in segregation studies (Ronning et al., 1995). The AFLP technique was used by Perry et al., 1998 for the identification of cacao genotypes, in comparison with RAPD. Queiroz et al., 1998, are using AFLP and RAPD markers to construct the genomic map of cacao aiming to identify regions associated with witches’ broom-resistance. Markers of the AFLP type were used by Risterucci et al., 2000a, for the saturation of a genomic map of cacao including 191 AFLP and 20 microsatellite loci, totalling 424 loci. At least 49 microsatellite loci have already been isolated, sequenced, and evaluated in CIRAD (Lanaud et al., 1999a). Markers of the ISSR type were used in cacao to examine the genetic diversity (Charters et al., 1999 and Charters & Wilkinson, 2000). Nuclear and chloroplast DNA: Cacao has a small genome, estimated by flow cytometry at about 0.43 picogrammes - pg or 0.415 x 109 base pairs- pb (Figueira et al., 1992), or even 0.38 pg or 0.388 x 109 pb (Lanaud et al., 1992), corresponding to nearly three times the genome size of the model plant Arabidopsis thaliana, considered the smallest genome known among the higher plants. The genome size of cacao was estimated at 0.201 x 109 pb by reassociation kinetics, with estimates of base compositions and relative extension of methylation (Couch et al., 1993). The small size of the cacao genome with little repetitive DNA suggests that gene cloning by chromosome movement from molecular markers flanking the region is possible. On the other hand, the size of the genome of the chloroplast of cacao was estimated at about 109 kpb by endonuclease restriction analysis (Yeoh et al., 1990), and at a 96 kpb limit by electron microscopy (Chung, 1988). There are no studies on the mitochondrial DNA of cacao. |
| Applications of molecular markers; Diversity evaluation, Diversity in cacao in south Bahia and Determination of the center of diversity. | Return To Table of Contents |
|---|
| Diversity evaluation: The traditional classification of Theobroma cacao assumes three horticultural races or main types or even racial groups: Criollo, Forastero and Trinitario. These terms have their origin in the classification used in the past century in Trinidad. The term Criollo represented pure native cacao and referred to the populations cultivated in Venezuela, which were originally known for having red or yellowish fruit when mature, with deep furrows and wrinkled, with a pointed apex and a fine husk. The seeds were large with white or light violet cotyledons, with supposed superior quality, but with less robust plants than the Forasteros. The Criollo came to be subdivided into those originating from Central America (with non-pigmented fruits) and from South America (with predominantly pigmented fruits) occurring in the north of Colombia and Venezuela (Cheesman, 1944). The Forastero group originally included all cacao trees considered non-Criollo, but a type of cacao came to be described with non-pigmented fruits, thick and hard husk and flatted and violet seeds. Forasteros can be subdivided into Upper Amazon (wild or semi-wild) and Low Amazon, characterized for having fruits in uniform format denominated Amelonados (green fruits of smooth surface without furrows, without basal constriction and without a pronounced apex). The Low Amazon Forasteros occur in the eastern Amazon and in the Guianas and was introduced in Bahia, Brazil, and in West Africa. They represent the principal group or type cultivated nowadays. This type of cacao may have been partially domesticated by pre-columbian forest dweller by the consumption of the pulp that covers the seeds (Barrau, 1979). Cacao trees with red fruits have not been found so far in the Amazon region, this being the only trait that permits a differentiation of the Amazon types from the others. Upper Amazon Forasteros are predominantly self-incompatible while Criollos and Lower Amazon Forasteros are predominantly self-compatible. Trinitarios are considered intermediate types between Criollo and Forastero or hybrid groups with traits that include the total variation of the species (Cheesman, 1944). Trinitarios were never found in the wild and it is presumed that they had their origin in a hybrid population (Venezuelan Criollos x Low Amazon Forasteros) of the estuary of the Orinoco river or alternately from crosses of Forastero introductions in Trinidad and Criollo plants from survivors of the destruction of the crop that had occurred in 1727 (Cheesman, 1944). Because Criollos and Forasteros are so heterogeneous the resultant hybrids cannot be distinguished from the original populations, making it impossible to define Trinitarios, except by geographic origin. The conventional classification of cacao into Criollos and Forasteros is based on distinct morphological, historical and commercial traits and these two horticultural races were considered as two subspecies in the last revision of the genus Theobroma (Cuatrecasas, 1964). This morphological classification principally considers shape and colour of fruit and seeds, which are the plant organs of commercial interest, and can therefore not be considered precise. The constant collection and description of various cacao populations (Lachenaud & Sallée, 1993; Almeida et al., 1987 and 1995) with traits that do not adapt to this classification, besides originally excluded traditional cacao types as, for example, ‘National’ from Ecuador make this classification yet more anachronistic and imperfect. The frequent movement of germplasm among the various producing regions since the beginning of cacao cultivation has drastically modified the cacao populations, making its classification difficult. Also, the recognition of really wild plants is difficult due to the strong association that cacao has with humans. Since it is a perennial and long-living plant, introduced and original populations and even their hybrids can co-exist in the same area for a long time (Bartley, s.d.). Another problem with the actual classification consists in the current myth that Criollos produce seeds of better quality for chocolate production, and would be more susceptible to diseases and pests. The chocolate industries differentiate cacao quality more in relation to the final use than the Criollo flavour quality itself. The quality standard is given by cacao of the Forastero type originating from West Africa (mainly from Ghana), while the little Criollo or Trinitarian type cacao mythically known as “fine” or of “quality” is used only in certain products with specific flavour or principally colour requirements, such as bitter chocolates with specification of origin, common on the European markets. The associations of chocolate manufacturers themselves assume that the characteristic chocolate flavour is generally stronger in cacao of the Amazon Forastero type than in Trinitarians with a predominance of Criollos (BCCCA, 1996). L.A.S. Dias (Chapter 3), however, defends the conventional classification for the use in improvement and assumes that Criollos effectively produce fine cacao. There is a tendency to demonstrate that this cacao classification is natural and that the genotypes can be categorized. A problem intrinsic to the attempt of categorizing genotypes lies in the identity of the analyzed genotype and its true origin as well as its a priori classification. For example, the fact that a genotype comes from Guatemala does not necessarily mean that one is dealing with a potential Criollo cacao; it could be an introduced Forastero. The initial attempts of categorizing cacao genotypes into these horticultural races by morphological descriptors (Engels, 1986) or isoenzymes (Lanaud, 1987a) failed, since there is a large overlapping between the groups, without the clear separation of the horticultural races. With the development of molecular markers it was believed that it would be possible to classify the genotypes definitively and unquestionably. Studies with molecular markers have generally demonstrated that Upper and Lower Amazon Forasteros, Trinitarios, and Criollos obviously present genetic differences. Various studies indicate that it is possible to separate Forasteros by isoenzyme analysis (Ronning & Schnell, 1994 and Warren, 1994) and molecular markers of the RFLP type using heterologous ribosome genes as probes (Laurent et al., 1993a), cytoplasmátic gene probes (Laurent et al., 1993b), or cDNA probes (Laurent et al., 1994a and 1994b and N'Goran et al., 1994 and 2000) and RAPD probes (N'Goran et al., 1994 and Marita, 1998). On the other hand, Figueira et al., 1994, Lerceteau et al., 1993 and Lerceteau et al., 1997a, concluded that it is not possible to distinguish the races clearly due to the occurrence of overlapping among the genotype clusters according to the conventional classification. The great heterogeneity within the races and the large number of existing hybrids limit the separation even more. The graphics of factorial analysis through the relative correspondence to the data on diversity for mitochondrial genes (Laurent et al., 1993b), probes of cDNA and RAPD (Laurent et al., 1994a; N'Goran et al., 1994 and Marita, 1998) demonstrate the difficulty of separating the races clearly, although the distribution of the various genotypes in the coordinates indicated congruent separation with the traditional classification in races. With this type of distribution, the unequivocal classification of a given unknown genotype is not possible (Figueira et al., 1994; Lerceteau et al., 1993 and Lerceteau et al., 1997a) and the various molecular markers types do not serve as the only classification criterion. Due to the difficulty of classification of the genotypes in the diverse horticultural races, Figueira et al., 1994, proposed to classify cacao in two large groups, considered as wild and domesticated. The wild would comprise the Upper Amazon Forasteros, while all the other types would be considered domesticated, taking the great human influence on cacao cultivation in the mesoamerican region where the Criollos and Trinitarios occur, and also the Lower Amazon Forasteros domesticated by the recent cultivation in the Amazon or even by natives in pre-Columbian times (Barrau, 1979) into consideration. This classification is compatible with the data generated based on the molecular markers. Among the 15 different identified units of the ribossomal gene (Laurent et al., 1993a), the genotypes considered domesticated of the Criollo and Trinitarian types only possess genes of the type A, B, C and D, while the Lower Amazon Forasteros have A, B, C, R, and M, and the 9 other wild types. The separation between wild and domesticated is also possible based on the factorial correspondence analyses derived from results of RFLP with cDNA (Laurent et al., 1994a and N'Goran et al., 1994) and RAPD probes (N'Goran et al., 1994). Based on the analysis of the genetic diversity of 270 genotypes of the CEPEC germplasm collection and using 38 primers to amplify 133 RAPD loci, Marita, 1998, demonstrated in a graph of principal coordinates (see item Germplasm characterization: - Figure 10.2) that the genotypes originating from Mexico and Central America tended to remain adjacent, close to the genotypes of the Caribbean region, in an area adjacent to the distribution of the genotypes considered as Low Amazon Forasteros. Therefore, in this principal coordinate analysis realized with RAPD data, the genotypes considered wild were clearly separated from the domesticated, concentrated in an adjacent area of the the graph. One notes, however, that the horticultural races can also be separated congruently by the conventional classification, but without a clear and definitive separation that would allow the unequivocal classification of a given unknown specific genotype. The classification of cacao genotypes into wild and domesticated presents a potential for use in cacao improvement. Pires et al., 1999a, proposed an improvement strategy through recurrent selection, using two populations, one derived from wild genotypes crossings, with favourable traits in terms of witches’ broom and black pod-resistance and high fat content, and the other involving crossings among domesticated genotypes (Trinitarios and Lower Amazon Forasteros) with larger fruits and seeds, high fruit productivity, and self-compatibility. The parents of the next selection cycle are selected from within the crossings between wild and between domesticated, while the crossings between wild and domesticated serve as base for the selection of hybrid clones and cultivars. Information on the genetic dissimilarity estimated by molecular markers would support the choice of the parents to be involved in the crossings for the formation of each population. Diversity in cacao plantations in the south of Bahia: Cacao was introduced in Bahia during the 18th century, from where it is believed to have later been taken to West Africa. Other introductions seem to have occurred in Bahia later, although cultivation was dominated by a homogeneous cultivar designated ‘Comum’. Based on 78 genotypes, Cascardo et al., 1993, estimated the genetic diversity and its geographic distribution in the cacao population in south Bahia. These genotypes included 30 cacao trees collected by CEPLAC near the probable introduction site of cacao in Canavieiras, on the banks of the rivers Salsa, Pardo, Cipó, and Jequitinhonha; two cacao trees from the region of Cairú, where there are records on the presence of cacao in 1780; five trees selected by CEPLAC in the valley of the river Mucuri farther south, close to the border with Espiríto Santo; 27 trees collected from farms that covered the greatest part of the cacao-producing region (Ilhéus, Itabuna, Itajuípe, Camacã, and Ubaitaba); 13 trees originated from local selections (SIC, SIAL, and EEG used in improvement programmes by CEPLAC); and an Amelonado cacao tree from Africa, used to establish its relation with the cacao of the Bahia. A total of 14 primers amplified 180 bands, 108 being polymorphic (60%). The mean genetic similarity among the 78 genotypes was 93.1%, varying from 86.2% to 99.2%, showing the low diversity found in the region. The cluster analysis brought forth three main groups, with the Amelonado included among them as an isolated group. The first group contained only genotypes from the region of the river Pardo/Jequitinhonha and from two farms in Ilhéus. The second group included genotypes from the traditional cacao region, the local selections, samples from the river Mucuri, and the samples from Cairú. The main component analysis showed that the genotypes from the river Pardo/Jequitinhonha region formed a distinct group, overlapping little with genotypes from the traditional region. The samples from Cairú grouped into the traditional group. These results suggest that the cacao population from the southern Bahia region probably stems from introductions to Ilhéus and not from Canavieiras. The genotypes from the river Mucuri valley as well as the selections from Espírito Santo are probably derived from the cacao population from Ilhéus. The Amelonado cacao trees from West Africa maintain little similarity with the populations from Bahia. This last result, however, should be considered with reservation since the Amelonado group was represented by only one cacao tree in this study. Determination of the centre of diversity: There are two hypotheses for the origin and the distribution of Theobroma cacao. Chapter 3, which deals specifically with the origin and evolution of T. cacao, presents three hypotheses and proposes a new scenario for the evolution process, supported by anthropologic, historic, paleogeologic, and biogeographic evidence and genetic population studies. Based on the occurrence of wild cacao that encompassed the entire morphological fruit diversity in the Upper Amazon region, Cheesman, 1944, proposed that the area between the rivers Caquetá, Napo, and Putumayo is the centre of diversity of the species. Cacao would accordingly have surpassed the natural barrier of the Andes mountains, probably with the help of humankind, and given origin to the populations of Central America and the north of South America (Colombia and Venezuela), later being domesticated by the Mayas or their ancestors. The physical separation by the Andes and the Isthmus of Panama would have caused the differentiation between the racial groups Criollo and Forastero, but would not have been sufficient to cause the separation into two species (Cheesman, 1944). The greater diversity of cacao originating from the Upper Amazon has been confirmed by evaluation studies into black pod and witches’ broom-resistance (Pires et al., 1999b); isoenzyme analysis (Lanaud, 1987a; Ronning & Schnell, 1994 and Warren, 1994); and of molecular markers of the RFLP type, using a ribosome gene as a probe (Laurent et al.,1993a and Figueira et al., 1994) or cDNA (Laurent et al., 1994a and 1994b and N'Goran et al., 1994) and RAPD probes (Figueira et al., 1994; N'Goran et al., 1994 and Whitkus et al., 1998). Lerceteau et al., 1997b, detected a greater allelic diversity for genotypes originating from Peru and Ecuador, but without presenting greater genetic diversity. The existence of wild cacao in Central America is questionable since even Cheesman, 1944, over 50 years ago, considered the occurrence of indigenous Criollos very unlikely, due to the long history of cultivation and domestication. Mora-Urpi, 1958 acknowledged that part of the Criollos cultivated in Mexico and Central America were extinct or crossed with Amazon introductions. Cuatrecasas, 1964, alternatively proposed that the natural distribution of the T. cacao species originally extended from the region of Guiana and Amazonia to the south of Mexico and eventually evolved into two forms separated by the Isthmus of Panama after the emergence of the Andes, which produced the two principal horticultural races, Criollos and Forasteros, under independent domestication. The differentiation between the two forms resulted in, according to Cuatrecasas, 1964, two subspecies: T. cacao ssp. cacao in Central America and Mexico and its cultivated forms would represent the horticultural group of the Criollos; and T. cacao ssp. sphaerocarpum in South America, whose cultivars would represent the Forastero group (Cuatrecasas, 1964). Evidence that sustains this hypothesis includes the discovery of wild T. cacao ssp. cacao forms in the Lacandon Rain Forest in Chiapas, Mexico (Cuatrecasas, 1964) and the recent discovery of cacao in “cenotes” in Yucatan, Mexico (Gómez-Pompa et al., 1990). Furthermore, Laurent et al., 1993b, evaluated 177 cacao genotypes using heterologous gene probes of the chloroplast and the mitochondria and detected greater variability in Criollo than in Forastero genotypes. The separation of genotypes in these two main groups by RFLP isoenzyme analysis (Ronning & Schnell, 1994 and Warren, 1994), using a ribosome gene as a probe (Laurent et al., 1993a) or cDNA probes (Laurent et al., 1994a and 1994b and N'Goran et al., 1994) and RAPD (N'Goran et al., 1994) gives additional evidence for the hypothesis of Cuatrecasas, 1964, in spite of the separation between Criollo and Forasteros not necessarily meaning differences of origin, since effects of foundation, selection, and genetic drift, caused by the isolation of the groups would allow the separation between the two populations. “Wild” cacao from the Yucatan cenotes was recently evaluated using 105 RAPD loci amplified with 24 primers and compared with the cultivars Criollo, Forastero, and wild cacao from South America (de la Cruz et al., 1995). It was demonstrated that cacaos from Yucatan are genetically distinct from all others as the cultivars from Criollos are more similar to wild cacao from South America. However, the occurrence of introduced exotic plants (for example, Musa sp., Citrus sp., and Cocos nucifera) in the cenotes, as described in the original study of Gómez-Pompa et al., 1990, does not exclude the possibility of cacao introduction from other regions, including that of Forasteros in the region, interfering in the results. In a subsequent study, Whitkus et al., 1998, evaluated the genetic diversity among individuals collected in Yucatan and other cacaos of Chiapas, Mexico, considered wild in relation to Criollos cultivated in Mexico in Chiapas and Tabasco, and genotypes of wild or not wild Forasteros, and Trinitarios. Fifty-seven relevant loci of RAPD amplified with 13 primers were used. The populations of Chiapas and Yucatan formed a significantly different group of cultivars, distinct from wild Forastero genotypes from South America. However, the genotypes considered as Criollos presented less affinity with the plants of Chiapas and Yucatan than with the cultivars and the wild Forastero genotypes diverging, therefore, from the hypothesis that the Criollos originate from the domestication of wild cacao from Central America and Mexico. On the other hand, Motamayor et al., 2000 analyzed cacao samples considered as ancestral Criollos by the authors collected in Venezuela, Colombia, Guatemala, Nicaragua, and Mexico (including the Lacandon Rain Forest and in archaeological sites). Twenty-five 25 cDNA probes for RFLP, revealing 66 alleles, and 16 microsatellite loci amplifying 150 alleles were used demonstrating that the ancestral Criollos from wild and cultivated genotypes of Central and South America were identical, with low genetic diversity and high homozygosity. These cacaos, so-called ancestral Criollos, were more similar to Colombian/Ecuadorian cacaos than they were to the Peruvian, indicating that the Criollo group originated from a few individuals from South America under human influence (Lanaud et al., 1999b). The “modern” Criollos actually cultivated or maintained in germplasm collections have introgression from Forasteros and show overlap with the genotypes considered Trinitarian, presenting high diversity. The proof or refutation of some of these hypotheses still need more evidence, with the real possibility of their never being demonstrated due to the constant exploitation of the areas involved. In practice, these studies indicate the real need of collection and preservation of the genetic resources from Central America and Mexico, independent of the proof of some of these hypotheses. The detectable morphological variability between the principal groups, indicating possible genetic differences, alone justifies the need for conservation. |
| Applications of molecular markers; Genomic mapping. | Return To Table of Contents |
|---|
| The construction of genomic maps permits the utilization of the strategy denominated marker-assisted improvement that favours the early selection of individuals containing genomic regions responsible for the desired traits. The identification of molecular markers associated with genes of interest of simple or complex inheritance is obtained by means of the compared segregation of the gene and the markers in progenies originating from crossings between contrasting parents for both. The construction of genomic maps involves the selection of the most appropriate population for mapping; the calculation of the pair recombination frequencies using this population; the establishment of linkage groups; the estimation of the distances in the map between marks; and the determination of the linear order of the markers to minimize the recombination events (Young, 1994). The choice of parents to develop genomic maps is critical in perennial tree plants, and the type of marker to be used must be considered. The parents must be important in the improvement programme, present a satisfactory level of polymorphism for the type of marker to be used, besides segregating for various qualitative and quantitative traits of importance (yield components and disease resistance, for example) and of product quality to maximize the acquisition of information. A maximum of genetic information can be obtained when using F2 populations and codominant markers while for dominant markers it is necessary to use progeny tests (F3 or F2RC1) to identify heterozygous plants (Staub et al., 1996). Populations derived from crossings between endogamous lines such as F2, backcrossings, or even endogamous recombinant lines are normally used for the genomic mapping of short-cycle plants. Homozygous populations for mapping can also be obtained through duplicated haploids, similar to the endogamous recombining lines, derived from anther culture of F1 or of inter-specific crossings (Kochert, 1994). The development of duplicated haploids was demonstrated for cacao (Lanaud, 1987b and 1988 and Lanaud et al., 1988) but this technique has not normally been used in improvement in spite of offering a great potential. In the case of absence of endogamous lines, F1 or backcrossing populations can be used for mapping, using single-dose polymorphic markers as dominant, segregating 1:1 for presence and absence of fragments (Ritter et al., 1990). It is expected that in allogamous plants, due to the high level of heterozygosity of various loci, a given segregant progeny (denominated F1) frequently presents segregation of the backcrossing test cross type (Carlson et al., 1991). The crossing of a heterozygote and a recessive homozygote (AB/ab x ab/ab) would obtain four progeny types (AB/ab, Ab/ab, aB/ab, ab/ab) in the frequency 1:1:1:1, that would produce distinct RAPD band patterns. This fact was demonstrated initially for pines (Douglas fir) and white spruce (Carlson et al., 1991), confirmed for pines (Pinus taeda), using markers of the RFLP type (Devey et al., 1991) and was used as base for the mapping strategy for both parents denominated pseudo test cross, proposed by Grattapaglia & Sederoff, 1994. In cacao, two strategies were used for the development of the two published genomic maps. The first used hybrid populations of parents, aiming to combine a genotype with greater heterozygosity than the other, evaluated by isoenzymatic analysis, to be able to use the configurations of “pseudo-backcross” that occur for various loci. This configuration presents the possibility of mapping codominant markers such as RFLP and isoenzymes, as much for the dominant ones as for RAPD and AFLP, but requires a greater number of individuals. The first linkage map for cacao was published by CIRAD (Lanaud et al., 1995), with 10 linkage groups, covering 759 cM and containing 193 loci, including 5 isoenzymes, 160 RFLPs, 101 of them cDNA, 55 genomic probes and 4 genes of known functions and 28 RAPD loci, in a ‘UPA 402 x UF 676’ progeny. The mean distance between markers was 3.9 cM and the physical distance per unit varied from 511 to 547 kpb/cM, based on the estimated size of the haploid genome of cacao (0.415 x 109 pb, Figueira et al., 1992; or 0.388 x 109 pb, Lanaud et al., 1992), the physical distance per unit varied from 511 to 547 kpb/cM, which favours gene cloning by the strategy of chromosome movement. This map was developed with a progeny of 100 individuals in the Ivory Coast, of which 55 individuals were in Bingerville and 45 in Davo. Morphological and agronomical traits, including number and percentage of fruits attacked by Phytophthora and by mirids, were evaluated in this progeny at the two sites for two and three years respectively (N’Goran, 1994). The parents presented significant differences at the two sites for the majority of the evaluated traits, the traits of ‘UF 676’ being superior. The morphological traits presented normal distribution in the population while the agronomical traits associated to Phytophthora and mirid-resistance presented deviations from normality. The quantitative trait loci, QTLs, were identified through a strategy of “pseudo-backcross” by using mapping methods by intervals and simple variance analysis (N'Goran, 1994). This linkage map was saturated with additional markers, including 15 more RFLP loci (3 cDNA and 10 genomic probes and 2 genes of known functions); 2 RAPD loci; 3 telomeric probes, 191 loci of AFLP and 20 of microsatellites, amounting to 424 markers, encompassing 885.4 cM, with a mean spacing of 2.1 cM between markers (Figure 10.1a and Figure 10.1b) and a population amplified to 181 individuals (Risterucci et al., 2000a). A total of 144 individuals of this population localized in Bingerville was used to map Phytophthora palmivora-resistance by trials of fruit loss in the field and trial in leaf discs (Lanaud et al., 1997). Various QTLs were identified; two regions explained 46% of the variation percentage for infected fruits while the data of the leaf disc trial were instable. Maps were developed for the identification of genomic regions associated with P. palmivora and P. megakarya-resistance for 12 progenies established in the Ivory Coast, Cameroon, Trinidad, and France. Besides the progeny ‘UPA 402 x UF 676’, the progenies ‘IMC 78 x Catongo’ with 128 individuals; ‘DR 1 x Catongo’ with 107 individuals; ‘S 52 x Catongo’ with 107 individuals (Clement et al., 2000); and two families ‘T 60/887 x IFC 2’ and ‘T 60/887 x IFC 5’ with 115 individuals in Zagne and 107 in Divo, in the Ivory Coast (Flament, 1998) were analysed for the identification of QTLs associated to P. palmivora-resistance and yield components. In Trinidad, genomic maps are being developed for the progenies of ‘IMC 57 x Catongo’ with 155 individuals (Motilal et al., 2000) and 174 individuals of the family ‘TSH 1077 x Catongo’ for resistance against P. palmivora. The evaluation of resistance to various isolates and species of Phytophthora by inoculation of leaf discs was conducted in Montpellier, France, with 154 individuals of the progeny [(‘SCA 6 x GS 36’) x IFC 1’)] (Paulin et al., 2000). The progenies ‘ICS 84 x UPA 134’ with 62 individuals; ‘SNK 10 x UPA 134’ with 78 individuals and ‘SNK 413 x IMC 67’ with 58 individuals were evaluated for resistance to P. megakarya (Flament, 1998). Another progeny involving ‘UPA 409 x Porcelana’ was also used for the construction of a partial genetic map (N’Goran, 1994). Among the 198 primers evaluated, 115 produced polymorphism, generating 186 contrasting bands between the parents, 77 of which were only present in ‘UPA 409’ and 109 in ‘Porcelana’. Among those present in ‘Porcelana’, only 28 were informative and were mapped in three linkage groups. Another linkage map was constructed by Nestlé researchers who used a population of 131 cacao trees derived from the backcrossing of ‘Catongo’ and the hybrid ‘Catongo x Pound 12’ and containing 138 markers of which 104 were RAPDs, 32 RFLPs and two morphological loci, comprising 1068 cM in 10 linkage groups. The mean distance between markers was 8.3 cM, varying from 0 to 36.9 cM with 1 cM equivalent to about 375 kpb, according to the estimates of the genome size. The segregations of the coloration by anthocyanin in flowers and new sprouts, of self-incompatibility, trunk diameter, height of the jorquette, number of days to the first flowering, and the number of ovules per ovary were analyzed (Crouzillat et al., 1996). ‘Catongo’ appeared as a highly homozygous cultivar (only 2% of heterozygosis estimated per 97 probes of RFLP and per 104 loci of RAPD). Four QTLs were detected for early flowering and two QTLs for trunk diameter, height of the jorquette, and number of ovules. In a complementary study (Phillips-Mora et al., 1995) QTLs were described for the number of fruits/tree/year, dry bean weight/tree/year, growth rate, leaf area, number of beans/fruit, field resistance to P. palmivora, and concentrations of caffeine and theobromine in the seeds. A genomic map was also constructed for the progeny ‘Catongo x Pound 12’ of 55 individuals, with 158 loci derived from ‘Pound 12’ and 4 loci of ‘Catongo’, representing a total of 772 cM and was used to identify QTLs associated with fruit yield (Crouzillat et al., 2000a). The total genetic variance for yield varied from 0 to 56% and among the 10 identified QTLs, two individually explained about 20% of the total variance for mean productivity of 15 years. This map was compared to the map of the backcrossing with ‘Catongo’, and 6 QTLs were identified for resistance to P. palmivora, estimated by the growth rate of lesions in inoculated fruits where one QTL was appeared in both populations and explained 48% of the phenotypic variance in the F1 population (Crouzillat et al., 2000b). The results published in relation to mapping represent the first initiatives of a genomic mapping of cacao and various other possibilities also exist. Specific populations for mapping have been developed to allow the identification of QTLs associated with important traits, many times not segregating in the fields of existing F1 hybrid cultivars. There is the example of the F2 population, developed in CEPEC from an F1 plant of ‘SCA 6 x ICS 1’, obtained to present segregation for witches’ broom resistance. Queiroz et al., 1998 are developing a genomic map of 82 individuals of this F2 progeny using AFLP and RAPD markers; and regions associated with witches’ broom-resistance were identified. |
| Applications of molecular markers; Genomic mapping - Considerations in cacao mapping. | Return To Table of Contents |
|---|
| The numerous segregating F1 populations that exist in the various improvement programmes in the form of hybrid trials have accumulated available data of yield and disease/pest incidence and could be used for genomic mapping and the identification of QTL-associated markers and/or qualitative traits of importance. These populations however present limitations regarding the restricted number of plants with normally 10 to 16 plants per plot, replicated 4 to 6 times, besides often not presenting the necessary contrast for the various traits of interest or even not segregating for the trait. The real identification and/or origin of these hybrids can also be doubtful, more easily identified by means of molecular markers. For the identification of QTLs, cacao has the advantage of the sexual propagation of hybrid plants, permitting the repetition of individuals and greater precision in the evaluations. Another great limitation that exists in cacao improvement and consequently in the use of markers for assisted selection is the lack of basic information on the diversity and inheritance in traits of interest. This lack of information on the existing diversity in germplasm collections limits the choice and construction of segregating populations from contrasting parents. The systematic evaluation of traits of agronomical interest in collections of germplasm has only recently been initiated. Pires et al., 1993 evaluated 681 accessions of the CEPEC germplasm collection in terms of yield components (fruit yield and seeds per plant, weight of fruits and seeds and loss by black pod) for three years. The incidence of black pod caused by diverse Phytophthora species was evaluated for four years under natural field conditions for 529 clones in the same collection (Pires et al., 1993). The genotypes from the upper Amazon, which is considered the centre of cacao diversity, presented greater frequency of factors of Phytophthora resistance while “domesticated” genotypes from Mexico, Central America, the Caribbean and those cultivated in Bahia, Brazil (SIC, SIAL, and EGG), presented greater susceptibility. The systematic evaluation of the same collection for incidence of witches’ broom in terms of number of vegetative brooms and flower cushions per plant over two years in six evaluations also identified a higher frequency of resistant genotypes among those “wild” types from the region of the upper Amazon, principally from Acre, Brazil (CSul and RB) and from Peru (SCA 6 and SCA 12) (Pires et al., 1999b). The traits associated with the industrial quality of cacao (content and hardness of the cocoa butter and flavour or chocolate flavour) present a great potential for manipulation by marker-assisted improvement for being complex traits and only possible to evaluate after the maturity of the plants with the fruit production and often needed in large quantities. Traditionally, the final quality of cacao has not been evaluated due to methodological and technological limitations. The final quality depends on industrial demands that are specific to each company, according to the desired traits of the final product. In spite of cocao butter being the most valuable product of the cacao tree, the crop is not considered oleaginous and the traditional improvement has ignored the potential of yield increase of this product (Pires et al., 1998), in contrast to what is happening with maize and soybean. The content of cocao butter affects the commercial value of the beans directly. There is an additional cost of grinding in the processing of beans with low butter content (Duncan & Veldsman, 1994). The diversity in content and productivity of cocao butter was recently evaluated (Pires et al., 1998). The genotypes originally collected in wild conditions in the upper Amazon exhibited a higher percentage of fat in the beans while ‘domesticated’ genotypes presented a tendency to lower values. However, when considering the cacao butter yield per tree, the tendency was the opposite, with the wild genotypes of the upper Amazon presenting low butter yield due to the significant and negative correlation between butter content and bean yield per tree. The mean values of butter content in progenies were similar to the mean of the parents (Pires et al., 1998). Likewise, the systematic evaluation of the composition of fatty acids, triacylglycerols, and hardness of the cacao butter was also realized with the intention of identifying contrasting parents for these traits, in order to conduct segregation and mapping studies (Figueira et al., 1999). Various populations of wild origin presented a composition of fatty acids and triacylglycerols and consequently of hardness of the cacao butter with a potential of utilization in an improvement programme aiming to reduce the butter unsaturation; increasing its quality (Figueira et al., 1999). There is also a selection potential of genotypes for a more unsaturated cacao butter yield, with application in the ice-cream industry. In terms of development of chocolate flavour, the inexistence of an objective and standardized evaluation method of the genotypic effect limited the progress in this area. The requirement of large seed quantities of a single genotype restricted, for several years, the analysis of the effect of the genotype in the development of the flavour. Until lately, due to the preponderance of the cultivation of hybrid cacao trees, it was not common to find large monoclonal cultivated areas that would permit the individual fermentation of various genotypes. The clone plantations established in Borneo, Malaysia, allowed the development of a standardized trial of fermentation in a small scale, where perforated plastic bags containing samples of less than 3 kg of fresh seeds of a given genotype are inserted into a fermentation mass of 50 kg containing seeds of genotypes mixtures (Clapperton et al., 1994a and 1994b). This trial allowed the evaluation of the flavour potential of various genotypes. The systematic evaluation of the fermentation of decreasing quantities of samples indicated that up to 750 g of fresh seeds (equivalent to 6 to 8 fruits) are sufficient for the evaluation of a given genotype, allowing the evaluation of individual trees (Clapperton et al., 1994a and 1994b). The flavour profile of the liquors of genotypes evaluated at BAL Plantations Sdn Bhd, Malaysia differed consistently in various replications indicating an important genetic contribution to the flavour (Clapperton et al., 1994a and 1994b). The flavour profile obtained from the various genotypes was maintained constant in Brazil even under distinct genotype growth conditions in relation to Malaysia (Figueira et al., 1997b). Direct comparisons of the flavour obtained from parents with the progeny demonstrated that the flavour traits are inheritable. These results indicate the possibility of initiating studies on the genetics of the chocolate flavour quality. However, there is also the need to standardize the sensorial analysis as one is dealing with a subjective evaluation, and the absence of well established standards for the various traits and of consensus between the large chocolate industries limit the progress with selection in this area (Clapperton et al., 1994b). Bucheli et al., 2000, described the initiative of Nestlé-France to identify QTLs associated with the appreciated flavour “Arriba” from Ecuador by means of the construction of a genomic map of a progeny with 171 plants derived from self-fecundation of a National cacao hybrid. A rapid method of marker-detection associated with specific genomic regions using segregating populations was proposed by Michelmore et al., 1991. The method denominated “Bulked Segregant Analysis” (BSA) consists in the comparison of two DNA sample groups (“bulks”) of a given population originating from simple crossing. Each group or bulk contains identical individuals for a certain trait used for the construction of the bulks. By means of marker analysis, the bulks must be contrasting for the specific genomic region associated with the trait used in the preparation of the same, but must have identical patterns for all the other regions that will appear heterozygous and monomorphic in both bulks. Summing up, the two bulks only differ genetically in one genomic region used as a criterion for the confection of the bulks. The possible polymorphisms present among the bulks and detected by molecular markers will probably be linked to the loci determining the trait used for the bulk separation. The use of this method permits the detection of markers localized up to 15 cM from the target locus in both directions. This method does not require any special type of genetic pool, such as pairs of Nearly Isogenic Lines (NIL), or an available linkage map. Using this method Michelmore et al., 1991, identified three RAPD markers among 900 amplified products linked to the resistance gene against downy mildew in lettuce (Dm 5/8) in an F2 population. The use of the population F2 with the heterozygosis tested in F3 configured the ideal situation, since homozygous dominant and heterozygous genotypes are not distinguishable at the F2 level due to the fact that RAPD markers are dominant. Using the F3 tested for resistance the resistant heterozygotes are eliminated, allowing the detection of markers linked in the phase of coupling or repulsion. The use of F1 populations originated from heterozygous parents in the BSA analysis restrict the RAPD/AFLP loci types that could generate some useful information on linked genes when compared to RFLP/microsatellites markers. In the following example, considering A the marker and B the allele for resistance, obviously not segregating loci in the population, linked or not, will not distinguish the bulks [AB/Ab x ab/ab; AB/Ab x ab/ab; aB/a x ab/ab] (Michelmore et al., 1991). Heterozygotes crossed with a dominant homozygote will not segregate RAPD/AFLP markers in even one case [AB/Ab x Ab/ab; AB/ab x Ab/Ab]; however a RFLP marker will only be able to detect polymorphisms among bulks if the heterozygote is the resistant (Barone et al., 1990). In the case where the heterozygote is crossed with the recessive homozygote, useful information will be obtained from the moment when the heterozygote is the resistant parent for both markers [AB/ab x ab/ab] (Barone et al., 1990). Crossings between heterozygotes are only useful for codominant markers of the RFLP type since for dominant markers, as in the example of RAPD, there will be no polymorphism among the bulks. Despite the dominant markers seeming to be disadvantageous they generate useful information when used in progenies normally found in improvement programmes of perennial plants. Due to the fact that the number of RAPD and AFLP markers are unlimited and are easier and quicker to be obtained, a sufficient number of markers can be evaluated to identify RAPD/AFLP markers linked to the gene of interest in the required form (resistant heterozygous x recessive homozygous). The frequent occurrence of the configuration of the testcross type in various loci in crossings of two heterozygous parents (described above) permits the detection of markers linked to the gene of interest. The use of the BSA method to detect markers linked to the resistance genes seems quite appropriate for the genetic improvement of cacao. In spite of its great potential BSA has been little used in cacao improvement. Cascardo et al., 1995 report its use to detect RAPD markers linked to the locus of susceptibility to sudden death of cacao caused by the fungus Verticillium dahliae. The resistance to this disease seems to be controlled by a gene with the dominant allele conferring the phenotype of susceptibility. The segregation for susceptibility was evaluated using a progeny of ‘Pound 7 (resistant, aa) x SIC 802’ (susceptible, Aa), artificially inoculated by immersion of the roots in spore solution. Two hundred and fifty primers were tested of which 65 were polymorphic (26%), but only 25 amplified bands in the susceptible parent. These primers were then tested using both parents and the two bulks, composed of 10 resistant and 10 susceptible plants. Only 5 primers presented the ideal configuration, which were then used to study the co-segregation with the susceptibility although only primer UBC 570, amplifying a fragment of 1350 base pairs, was apparently associated to the loci of susceptibility. This study, in spite of using simple genetic systems, shows the potential of utilization of the BSA together with the RAPD technique as useful tools in the improvement programme of cacao. The greatest difficulty faced in the further development of the identification of markers associated with traits of interest is blocked by the lack of basic genetic information on cacao. Another strategy tested was the attempt to identify molecular markers associated with witches’ broom-resistance based on the BSA analysis, using DNA bulks mixed with contrasting genotypes for resistance. The clones BE 5, 7, 10, CSUL 1, 4, 5, EET 376, 377, 390, RB 36, 38, 39, and SCA 6 were characterized as resistant in field conditions while clones GS 17, 29, SIAL 98, 105, SIC 802, SGU 50, 60, 71, UF 613, 650, and 677 presented a high susceptibility reaction in the field (Pires et al., 1999b). After amplification tests of all genotypes solutions of DNA mixture were prepared according to the class of reaction to witches’ broom (one resistant and the other susceptible), combining equal volumes of DNA solutions of identical concentration. These mixtures were then used in amplification reactions to select the primers that produce polymorphism between the two bulks. Of the 302 primers tested, 72 (23.8%) were selected for amplifying polymorphic fragments between the two mixtures, suggesting an association with the reaction of resistance or susceptibility. One expected that the polymorphism level would be very high as the two groups were contrasting for various other traits, besides resistance. The genotypes of the resistant group, besides representing tolerance reactions to witches’ broom, have other traits in common (small seeds, tolerance to Phytophthora, high fat content, etc) that are contrasting with the other group and could also be associated to the observed polymorphic fragments. Twenty-four primers were evaluated individually in RAPD reactions for the 20 genotypes, and eight of these (2.6% of the tested total) maintained the observed pattern in the group reactions and the polymorphic fragments remained unaltered. The presence of the same amplification patterns in the bulks and in the individual genotypes confirms the reliability of the amplification and its reproducibility. In general, the identification of polymorphic fragments was effective and stable, but the proof by means of studies of co-segregation of the presence/absence of the fragments with the reaction to witches’ broom would be necessary. Some susceptible or resistant genotypes do not present the typical polymorphic fragment observed for the other components of the group, probably due to the fact that the analysis included non-related, highly heterozygous genotypes that were not in linkage inbalance and exceptions were therefore expected. Primer OPB15 was tested several times with other genotypes (including with clones of the Huallaga and Ucayali series) and the amplification pattern remained similar to the visible polymorphisms. Similarly Marita, 1998 compared the frequency of RAPD loci of two subgroups containing 15% of the most resistant and 15% of the most susceptible ones. There was a significant difference in the frequency of 49 RAPD loci between the two subgroups that could be associated with witches’ broom-resistance and need to be tested in segregating populations. |
| Applications of molecular markers; germplasm characterization, Genotype identification and Pathogen diversity; Conclusion. | Return To Table of Contents |
|---|
| Germplasm characterization : The objective of ex situ germplasm collections used for the conservation of the genetic cacao resources is to ensure that a maximum of the genetic diversity is preserved and used in a rational manner. A maximum of genetic information must be obtained from the collections to allow its use for the improvement and conservation of the genetic resources. Molecular markers present diverse applications to efficiently obtain useful genetic information in diverse stages of the collection (Karp et al., 1997). The estimate of the genetic similarity between individuals of a given accession or between accessions of the collection permits information to be obtained on the identity of this accession, identify duplicates, and complement the characterization with morphological and isoenzymatic traits. The partition of the marker-estimated variability among individuals, accessions, populations, types, and geographic origin permits the identification of populations and/or geographic regions that present greater diversity or that are underrepresented in the collection, supporting strategies and the planning of new collections and acquisitions and the development of nuclear collections (core collections) that represent the diversity of the collection. Molecular markers have the potential to help in the identification of genes/alleles of interest in the collection. This strategy of characterization is being followed in the germplasm collection of Trinidad (Bekele et al., 1998). Marita (1998) evaluated the diversity and genetic relation of the sample composed of 270 genotypes of the Active Germplasm Bank (AGB) of CEPEC by RAPD markers with the objective of furnishing recommendations for the establishment of nuclear collections and to maximize the genetic diversity between genotypes with witches’ broom-resistance (Figure 10.2). This was the most comprehensive study on genotypes of a single collection, since genotypes had previously been provided by various collections. Thirty-eight primers produced 133 RAPD loci and the estimates of genetic diversity were used to generate a subgroup representing the maximum diversity that would serve as a nuclear collection (Marita et al., 2000). The inclusion of samples could be simulated and evaluated based on the availability of molecular data. The frequency of the RAPD loci associated with witches’ broom-resistant genotypes was used to indicate possible associations between the markers and the trait. Similar studies could be realized for other traits such as black pod-resistance and fat content evaluated earlier in this germplasm in order to accelerate the identification of associated markers (Pires et al., 1998). The cacao germplasm bank of ERJOH and CEPLAC in the State of Pará is result of the systematic collection of representative samples of the genetic variability of the populations existing in the Brazilian Amazon (Almeida et al., 1987 and 1995). The relations and genetic diversity between cacao genotypes contrasting in field resistance against witches’ broom (20 resistant; 10 susceptible), collected originally in various watersheds in the Amazon, besides two external susceptible controls (Bahia, Brazil, and Costa Rica) and two genotypes considered resistant (‘SCA 6’ and ‘SCA 12’ of Peru), were initially determined by RAPD (Silva et al., 1998). Of the 26 tested primers, 19 amplified 163 fragments of good resolution and reproducibility. The mean genetic similarity among all genotypes was 85.3%. All analyzed genotypes differed significantly from each other. The genotypes had a tendency to group according to their geographic origins; with genotypes collected in proximity presenting a greater similarity among themselves. The phenogram could be arbitrarily subdivided into two large groups. The first contained 11 resistant genotypes, including all genotypes collected originally in Acre (rivers Chandless, Acre, and Iaco) with one exception, besides two genotypes from the river Solimões and of the actual SCA 6 and SCA 12 clones (river Ucayali, Peru). This group presented the greatest genetic diversity. The second group contained 16 resistant and 7 susceptible genotypes, including principally “domesticated” genotypes from Belém and Alenquer, PA, from Bahia and from Costa Rica. The results demonstrated the genetic diversity between the various potential sources of resistance to witches’ broom. The structuring of the genotypes according to origin and geographic proximity suggests the potential of the use of this technique for an identification of genotypes and the establishment of core collections. Genotype identification: The various collections of cacao germplasm existing in the world embrace a large number of accessions with at least 16,300 clones identified by distinct acronyms in publications, representing close to 12,700 individual clones listed in the International Cacao Germplasm Database (ICGD) (Wadsworth et al., 1997). Various genotypes have more than one name or acronym of identification (synonymy) and some names refer to more than one genotype (homonymy) (Lockwood and End, 1993). The transference of results and recommendations between the various improvement programmes is hampered by the lack of agreement between the genotypes of the various germplasm collections. Few geneticists have sufficient experience to recognize individual genotypes and the identification of duplications and errors of identification only became possible and measurable with the advent of molecular markers. The correct identification of trees of progenies from controlled crossings is also necessary in programmes of genomic mapping. Figueira et al., 1997b demonstrated that the presence of errors in the identification of genotypes must be more common than expected in the diverse collections when they compared the identity of commonly used clones in improvement by means of RAPD analysis. Based on a sample of only 11 genotypes (Table 10.1; Figueira, 1998) at least two of their clones (ICS 6 and SIAL 93) or 18.2% did not show the same identity in two important collections (CEPLAC, Brazil, and BAL Plantations, Malaysia) in spite of bearing the same name (homonymy). Table 10.1: Degree of genetic similarity among the same accessions existing in the collections of CEPLAC, Bahia, Brazil, and BAL Plantations, Malaysia: Genotype/Similarity (%): - NA33/99.7; - UF168/99.7; - PA137/99.4; - ICS1/99.4; - CC11/99.4; - 10P/99.4; - CC10/99.2; - AMELONADO/97.6; - AMAZON2-1/94.6; - ICS6/93.3 and SIAL93/91.8. If the two ‘AMAZON 2-1’ accessions were also considered distinct at both sites based on the low similarity found, the error percentage would reach 27.3%. The results obtained by the RAPD analysis were corroborated by the aspect of the fruits (Figueira, 1998). Clone SIAL 93 from the Malaysian collection is derived from an accession of Mayaguez, Puerto Rico where a branch of the rootstock used (probably ‘IMC67 x PA 218’) might have grown and substituted the ‘SIAL 93’ crown. This accession with an identification error was detected in Mayaguez and was reintroduced in Bahia, Brazil, where it was renamed ‘CEPEC 533’ (B.G.D. Bartley, personal communication). The error in the identification of 'SIAL 93' was later confirmed with ISSR analysis (Charters & Wilkinson, 2000). Another demonstration of the applicability of molecular markers in the identification of genotypes refers to the study realized in Costa Rica to confirm the hybrid origin and paternity of the progeny resulting from the crossing of ‘Catongo x Pound 12’ (Phillips-Mora et al., 1995). By means of the RFLP analysis it was demonstrated that of the 120 trees that formed the trial, only 55 (45.8%) are true ‘Catongo x Pound 12’ hybrids, 29 (24.2%) are trees originated from self-pollination of ‘Pound 12’, three (2.5%) corresponded to ‘Catongo’ self-pollination, and the remaining 33 (27.5%) trees were of unknown origin. The low rate of confirmation of hybrid trees casts doubts on the results obtained in the various hybrid trials conducted in the country, besides indicating the fragility of the self-incompatibility system of cacao, considering that ‘Pound 12’ is self-incompatible. Similar results were observed in biclonal orchards of natural hybrid seed production installed in Togo, Ivory Coast, and Cameroon where self-fecundation rates of up to 97% were observed in self-compatible genotypes based on isoenzymatic markers (Lanaud, 1987a and Lanaud et al., 1987). The breaking up of the incompatibility is caused by the phenomenon denominated mentor pollen (see details on self-incompatibility in Chapter 6). Another application of genotyping with an apparent potential is the identification of the most probable parents for selections of witches’ broom-resistant cacao (VB selections) realized at plantation level in south Bahia. To carry out this heritability test the DNAs of plants of the VB selections and of the possible parents used in the hybrid generation programme of the period in question (in the 60s) were included in the RAPD analysis. The individuals were grouped based on their genetic similarity pattern (Corrêa et al., 2000). This technique allows the identification of the different resistance sources that contributed to the formation of the actual resistant materials. In the future it will allow orientation of the composition of materials on each farm with a view to amplifying its genetic base. The need for preservation of cacao genetic resources in active germplasm banks demands a great volume of resources and area due to the seed recalcitrance and the absence of efficient in vitro conservation and cryopreservation methods, besides exposing the collections to environmental and biotic risks (see also Chapter 5). However, the diversity of names and acronyms seems to substitute the genetic diversity established in the collections, and the occurrence of homonyms and synonyms diminishes the efficiency in the conservation of the genetic resources. Besides, administration problems in collections, errors of tagging, branch growth from the rootstock etc reinforce the problems of correct identification. The International Cocoa Genebank, Trinidad (ICGT) is being systematically characterized to optimize the conservation of genetic resources and facilitate the improvement. The genotypes are being characterized using morphological and agronomical traits, isoenzymes and RAPD with the objective of identifying duplications, errors of identification, and selecting clones with a potential for improvement. Christopher et al., 1999 described the use of markers of the RAPD type together with botanic descriptors to verify the identity of accessions. In the identification of 117 accessions, 23 RAPD loci were analysed and 35 accessions containing mixtures identified and of these thirteen did not have any tree with the original identity. Within the scope of the international project “Conservation and Utilization of Cacao Germplasm”, financed by the partnership CFC/ICCO/IPGRI (http://www.icco.org/projects/ germplasm. htm), Risterucci et al., 2000b, evaluated 148 individuals of 28 clones that were sampled in various national collections using 10 microsatellite loci and identified nearly 30% of the individuals with errors of identification. For the standardization of genotype identification through molecular markers the development of low cost and simple technology, stable systems with precision and reproducibility under the diverse conditions found in the countries that own germplasm and quarantine collections is also necessary. The molecular identifications should be deposited in data banks. Attempts in this sense were initiated by BCCCA, which conducted a test (“ringtest”) to evaluate the reproducibility of the RAPD technique for genotype identification at five laboratories localized in Brazil, Trinidad, France, England and the United States (Gilmour, 1995). The results demonstrated that the RAPD analysis was precise for the identification of duplicated genotypes when evaluated in the same laboratory but the reproducibility was low among laboratories. Alternatively, molecular markers of the ISSR type (Charters et al., 1996) are being evaluated at the University of Reading to characterize cacao genotypes, due to the greater stability and reproducibility (Charters & Wilkinson, 2000). The existence of stable molecular markers that could be used by the various programmes in the world would make the identification of synonymy and homonymy among the accessions easier. A recent initiative of the US Department of Agriculture (USDA) proposes the molecular characterization of all the important cacao germplasm collections of America using 15 microsatellite loci (Saunders et al., 2000). Diversity in pathogens: The fungus Crinipellis perniciosa (Stahel) Singer, etiological agent of witches’ broom, endemic in the Amazon region, was detected for the first time in south Bahia in May 1989, in the municipality of Uruçuca (Pereira et al., 1989). It was followed by the second observation six months later in the municipality of Camacã, 120 km south from Uruçuca (Figure 10.3). The area with infected plants in Camacã was greater however than that found in Uruçuca, suggesting that this introduction had occurred before the one in Uruçuca. The appearance of the disease at two distant and distinct sites, well in the middle of the cacao region suggests that more than one introduction of the pathogen had occurred, probably from the Amazon region and possibly by human intervention (Rocha et al., 1993 and Pereira et al., 1996). South Bahia comprises the largest contiguous cacao-producing region in the world. The predominance of the cacao cultivar ‘Comum’ (see Chapter 12 dealing with cacao cultivars), considered susceptible and presenting low genetic variability (Cascardo et al., 1993) contributed to a rapid expansion of the disease, causing progressive loss in the cacao yield. The most effective method of witches’ broom control, although long term, consists in the use of resistant genotypes. Clone SCA 6 has been to date the most used resistance source, in spite of its susceptible performance in the conditions of Ecuador, Colombia and Bolivia. Studies that were based on inoculations of the fungus onto ‘SCA 6’ progenies, on somatic incompatibility, morphological variability, and on biochemical reactions revealed genetic diversity among the isolates (Fonseca et al., 1985; Hedger et al., 1987; Bastos et al., 1988 and Wheeler & Mepsted, 1988). However these techniques only allow the evaluation of a small number of loci as opposed to the molecular RAPD markers that permit a rapid evaluation of a large number of loci of the fungus, being useful in the determination of the genetic diversity of pathogens. The use of these molecular markers is increasing in other pathosystems (Wolfe & McDermott, 1994). The determination of the genetic variability of the pathogen is an important aspect in the selection process of resistant genotypes that will be used in improvement programmes (see also Chapters 6, 7, and 8 that deal with the genetics of disease resistance). The great majority of the fungi do not fit into the classical model of population genetics. They have varied methods of colonization, dispersion, and survival and their populations can overlap with the passing of time (Anderson & Kohn, 1998). However, C. perniciosa multiplies uniquely by means of basidiospores which in turn have a mean lifespan of a few hours (Baker & Crowdy, 1943).The genetic analysis of isolates of the freshly collected fungus can reflect the genetic variability of the population at that moment with high confidence. In a pilot study realized by Cascardo & Figueira, 1995 in the cacao-producing region of south Bahia, infected branches of cacao (dry brooms) were collected during the year of 1994. These branches were placed within infected brooms to induce the reproductive structure of the fungus. Multisporic cultures were cultivated in liquid LB medium for the later extraction of mycelial DNA (Zola & Pukkila, 1986) and posterior analysis via RAPD. In this study, 156 loci RAPD were amplified using 10 primers. The genetic similarity was calculated using the DICE coefficient and a phenogram was constructed based on the UPGMA algorithm (Figure 10.4). The mean genetic similarity of the fungus isolates was 87%, varying from 60% to 100%. The isolates were grouped into two categories, according to the geographic origin and related to the supposed areas of independent introduction (Uruçuca and Camacã). An epidemiological monitoring of the disease done in the subsequent years after its discovery revealed that two foci concentrated radially around the sites of introduction of Uruçuca and Camacã in 1991 (Rocha et al., 1993). The disease was later disseminated from the initial sites; an overlapping of the two foci occurred in the region that encompasses the cities of Itabuna and Ilhéus (Pereira et al., 1996) (Figure 10.3). Comparing the geographic origin of isolates collected in these areas, a clear tendency of grouping of the same in relation to the presence or absence of determined polymorphic bands was observed, noticing also that some isolates of the intersection region presented bands in common. (Figure 10.5). These results suggest that two independent introductions may have occurred and that genetic recombination could be occurring in the intersection area. This hypothesis was confirmed since in other trials where the DNA of monosporic cultures originated from a single basidiocarp was analyzed, a high variability in the electrophoretic pattern of the RAPD bands was observed, suggesting genetic segregation. Multisporic cultures were therefore chosen to work with. It is known that in the PCR reaction the more frequent alleles are preferentially amplified in detriment of the less frequent ones. Consequently, a DNA bulk of a multisporic cultures would be more representative for the predominant alleles in that population (Michelmore et al., 1991). The other measure taken was to separate the cultures of the basidiocarps individually since in a same broom there may be basidiocarps from different origins. In a later study, when comparing these isolates together with other isolates from the Brazilian Amazon region (Andebrhan et al., 1999) it was confirmed that the infection foci were originated from two distinct introductions and the possible origin of the same was traced. In this study, specific polymorphic bands were also observed for a particular isolate group. Some isolates of the south were grouped with isolates of the north. This indicates that the attempt of eradication of the initial focus in the region of Uruçuca may have led to a reduction of the inoculum level of the same, contrary to the focus of Camacã that spread out quickly. The phenogram (Figure 10.4) shows that this tendency was also observed in the study realized by Cascardo & Figueira, 1995 where some isolates of the southern region grouped together with isolates from the north. These results demonstrate the potential of the use of molecular markers to evaluate the genetic diversity and pathogen gene flow. The studies, together with the work of identifying molecular markers of the pathogen associated with its virulence can contribute significantly to improvement programmes. This relation between the pathogen and its resistance response in cacao has already been demonstrated. For future perspectives, the RAPD analysis permits an evaluation of the gene flow and the genetic diversity of the pathogen throughout the years. Other cacao pathosystems were also studied using molecular markers. Nyassé et al., 2000, evaluated the genetic diversity of Phytophthora megakarya that infected cacao in Central and West Africa. A separation between two large groups, that of West and the other of Central Africa was found. In India, AFLP markers and ITS sequences were successfully used to differentiate various Phytophtora species found in intercropped cacao and coconut plantations (Chowdappa et al., 2000). Sequences of non-transcribed internal spacers (ITS) were also successfully used to identify different species that infect cacao Lee & Taylor, 1992, demonstrating another utilization of the PCR in studies of cacao pathosystems. Conclusion: The potential of using molecular markers in cacao was discussed in this chapter. These techniques appear to be very powerful with a wide range of applications. As cacao is a species of high genetic variability, the understanding of this variability will help a lot in improvement programmes of the crop. The study of the variability allows, from the establishment of nuclear collections, the planning of crossings even to help in the indication of regions of greater genetic variability for germplasm collection. Several linkage maps using molecular markers (RFLP, RAPD, AFLP, and microsatellite) together with other markers (isoenzyme and morphological markers) have already been drawn up. Some traits were already mapped and localized in these maps. It is important to point out the use of crossings of interest in mapping programmes. It would be important that these crossings segregated for various traits of interest. However, knowledge on the various agronomical traits of interest in the crop, such as genetic inheritance of the content and quality of cocoa butter, disease resistance (to witches’ broom, monoliasis, and black pod), as well as the various physiological yield components is restricted. Programmes of genetic improvement of cacao using molecular markers should therefore go hand in hand with evaluations of these traits. Knowledge on the variability of the pathogens that are harmful to the cacao crop is also very important. This type of information is a support in improvement programmes that aim at genetic resistance against diseases. The evaluation of the genetic variability of the pathogen also permits an epidemiologic control. Another use of molecular markers in pathogens is the identification of the chromosome regions responsible for the virulence of the same. This type of information can be very useful when linked to more in-depth studies of molecular biology, aiming at the cloning of genes intended for studies in genetic engineering. The genetic transformation of cacao (Figueira et al., 1997a) will allow the introduction of genes of interest in a very reduced space of time, being an option that is beginning to prove to be quite viable. |
| Chapter 11. Experimentation in breeding. L.A.S. Dias & M.D.V. Resende. | Return To Table of Contents |
|---|
| Contents: Introduction and Experimental principles; Choice of design, Simplicity, flexibility, and robustness, Multiplicity of objectives, Experimental designs in breeding; Statistical analysis strategies, General considerations and Analysis of mixed linear models (REML/BLUP); Repeated measures, Sample size for repeatability estimation and repeatability and Adequate number of measurements, Analysis of experiments with repeated measurements; Experimental accuracy; Final considerations. Summary: The experimental principles are reviewed after analyzing the peculiarities of research with perennial species. These peculiarities nearly always lead to the generation of unbalanced data and various possibilities of statistical analyses. The optimum plot size and number of replications are discussed under the focus of the statistical and genetic designs. For the latter, the plot size is function of the selection efficiency. Small plots of 1 to 6 plants are satisfactory in most cases. The number of replications is determined by dividing the number of individuals by progenies, which is necessary for the maximization of the selection accuracy by the chosen plot size. Numerical evaluations are presented for different situations aiming to optimize the selection efficiency. It is convenient that the designs used in research with perennials are as simple, robust, and flexible as possible. After reviewing the designs and arrangements it is emphasized that their choice should also consider the possibility of satisfying multiple research objectives. The mixed linear model analysis obtained in non-orthogonal designs is presented, which is highly adequate for the analysis of unbalanced data. Lastly, the focus turns to research with repeated measures, typical of perennial species. Designs of experimental analyses with repeated measures combining the uni- and multivariate approaches are presented. |
| Introduction and Experimental principles. | Return To Table of Contents |
|---|
| Introduction: Field research with perennial species such as cacao presents peculiarities that differ from research with annual species. The long cycle, the overlapping of generations, the extensive experimental area required, the high costs for implantation and maintenance of the experiments, the permanent exposition of the trees to pathogens, pests, and climatic adversity, and the hazards of fire and the dynamics of improvement programmes highlight the importance of well-designed experimental plans. This subject is emphasized in the present Chapter. There is, of course, no pretension of elaborating an encompassing approach on research used in cacao improvement. The intention of the Chapter is merely to outline recurrent themes in the area for which the solutions under way to date lack a scientific base. Experimental principles: There are three experimental principles (Fisher, 1970): replication, randomization, and the local control. Replication refers to the number of times the treatment appears in the experiment. It has the aim of estimating the experimental error and the increase of the precision of the estimates of the treatment means. In this last case, the greater the number of replications the smaller is the variance of the treatment means. The third and last purpose of the replication refers to the increase of the capacity of discrimination of the statistical tests such as the F and the other tests of means. As F is a ratio between variances (treatment variance/error variance) and the treatment variance weighed by r (number of replications) in the mathematical expectations of mean squares, it is deduced that the greater the r, the greater the F value. It is important to emphasize that in cacao improvement the concept of replication is frequently subverted. Some studies in germplasm sometimes deal with the clones, from the same accession as being replications. If, for example, each accession is represented by 10 clones, some studies assume that there are 10 replications and process the analysis as such. This is a great mistake that should be avoided. Germplasm banks in cacao are installed as clone collections without observing randomization and replication. Lastly the justified concern of installing new banks or of multiplying the old ones with at least two replications has arisen. Randomization consists in arranging the treatments randomly in the experiment so that all plots have the same chance of receiving a given treatment. It is therefore recommended to avoid systematic factors that would benefit some treatments in detriment of others. Its great benefit is to validate the estimates of experimental error and the treatment means and give them reliability. In terms of the genetic evaluation and the estimation of variance components, randomization is essential as a form of avoiding the correlation between genetic and environmental effects, a fact that would affect the entire basic model of estimation and prediction, which assumes independence between the referred effects. The local control is destined to control the environmental heterogeneity and has implications in restricting randomization. Blocks in the randomized block designs and lines and columns in the latin squares, for example, are local control strategies that allow to group homogeneous plots and randomize the treatments within them. The use of covariance is also a type of local control and will be commented on in the item Experimental accuracy. See Pearce, 1986 for a critical vision of these principles and the history of the development of experimental statistics. Optimum plot size and number of replications: The experimental accuracy can be increased by the choice of the most suitable design (see item Choice of design), of the adequate data analysis (see item Strategies of statistical analyses ), of the utilization of a greater number of replications and of the utilization of optimum plot size. According to Pearce, 1986, with the work of Smith, 1938, the notion that the variation in small areas is smaller than that observed in greater areas was disseminated. However, between the various methods used to increase the experimental accuracy, Anderson & Bancroft, 1952, recommend the increase of the size of the experiment, even with the reservation that this increase could result in greater environmental and material heterogeneity. The authors defend the thesis that when the number of blocks or of treatments increases, the degrees of freedom to estimate the experimental error would also increase. It is appropriate here to distinguish accuracy clearly from precision, two terms mistakenly used as synonyms. The accuracy of a given measure refers to the estimation of the true value of the quantity, while precision refers to the samples of values around the true mean of the measure, and this is not the true value if this mean is biased. Precision refers to the estimates with minimal or small variance and can be understood as the inverse of the variance under the Bayesian influence (Anderson & Bancroft, 1952). The researcher, for example, has the possibility of operating with unbiased estimates of low precision (high variance) or slightly biased but high precision estimates. Certainly, there is more than one method to determinate the optimum plot size. The most common is to determine it based on the precision or level of significance with which differences want to be detected between treatments. It is appropriate to mention here the exception that the inference that can be drawn on the difference between means (µ1 and µ2) of any two treatments results in two error types: error type I happens when the true null hypothesis (h0) is rejected, which assumes that there are no differences between the means of the two treatments (h0 : µ1 - µ2 = 0); and error type II when the wrong null hypothesis is accepted. To control error type I a certain level of significance is specified for the test of the null hypothesis. Against error type II the alternative hypothesis (ha) is specified for testing, where ha : µ1 - µ2 = ± 0. Returning to the subject, another method to determine the optimum plot size is the inspection of the point of inflection of the decreasing curve. This curve is formed by the values of the variation coefficients plotted in axis y, calculated for the various increasing plot sizes plotted in axis x. This method was presented by Federer, 1955. A third method consists in estimating the mean variances of the various plots, either from original or from transformed data. Transformation in this case depends on the function of data distribution. All these methods of determination of the optimum plot size use preliminary observation plots or uniformity trials by which the different plot sizes are simulated. A concern of field research with the optimum plot size in cacao improvement dates back to the 30s. Cheesman & Pound, 1932, analyzed uniformity trials in 4 x 4 latin square (see fundamentals of this design type in the item Simplicity, flexibility, and robustness), considering three simulated plot sizes (6, 12, and 18 trees ). The trials lasted for four years of successive harvests (1926/7-1929/0) in cacao trees with 18 years in the field. The yield component evaluated was the annually produced total number of fruits per plant. The results of these trials, in terms of number of fruits per year to detect significant differences between plots, at the level of 5% of probability of error type I, are in Table 11.1. These results revealed that to obtain a significant yield increase in plots of 12 or 18 plants, eight more fruits per tree would be necessary, which is not a small value. According to Cheesman & Pound, 1932, the mean number of fruits produced in the 18-tree plots was 23 in the four years, and the increase of eight fruits means a yield increase of about 35%. Therefore, a difference of 35% in the yield is subject to being detected among treatments at 5% of probability. These results led the authors to suggest plots of 12 cacao trees of uniform age. Others studies were developed to determine the optimum plot size in cacao, based on the method already mentioned of the maximum curvature of the variation coefficient. These studies, developed based on the analysis of the fruit yield, suggest the use of plots of 8 trees and 4 replications (Páez & Siller, 1963 and Esquivel & Soria, 1966), or even plots of 4 to 6 trees, as one is dealing with clonal or quite homogeneous genotypes (Esquivel & Soria, 1966). Plots with 9 trees and 5 replications were recommended by Pereira, 1972, as adequate for research with cacao in southern Bahia, Brazil. Plots composed of 12 up to a limit of 25 trees have also been recommended (Lotodé, 1971). This size was obtained based on the estimates of variances of the plot means using yield data of individual plants accumulated over two years. According to this study the increase in the plot size to over 20 to 25 trees results in little expressive reduction of the error variance. Plot squares with 25 trees have been recommended (Lockwood, 1981) in Ghana for situations where the trees show pronounced differences in vigour and growth habit and thus need internal borders. Under these circumstances, one works with actually measured plots of 9 trees. Based on earlier studies developed with cacao it can be concluded that the ideal plot size lies between 6 and 12 plants from the statistical point of view with a minimum of 5 replications. These sizes are well suited to the tests with more heterogeneous hybrid cultivars well and have been efficiently used in cacao genetic improvement programmes. Plots of 4 plants can be used for clones or even cultivars that are more homogeneous. The question of the experimental plot sizes in tree species has also been approached by various authors (Pearce, 1976; Cotterill & James, 1984; Pimentel Gomes, 1984b; Bonnot, 1995 and Resende, 1995). Using the standard methodology based on the coefficient of soil heterogeneity (Smith, 1938), which is greater when the soil heterogeneity is greater, Bonnot, 1995 recommended the following plot sizes: i) 1 plant when b is near 0 (b > 0.30) or; ii) at least 6 plants when b is near 1 (b > 0.70). With 0 > b > 1, it can be inferred that when 0.3 < b < 0.7, plot sizes between 1 and 6 would be recommended. Using the methodology of intra-class correlation coefficient [ratio among-plots variance/(among-plots + within-plots variance)], Pimentel Gomes, 1984b and Pimentel Gomes & Couto, 1985, arrived at the use of 1 plant per plot. Considering the experimental precision and the probability of detection of significant differences among treatment means, Cotterill & James, 1984, Loo-Dinks & Tauer, 1987 and Haapanen, 1992 also opted for the use of plots of 1 plant. The results and aspects discussed so far refer mostly to the adequate plot size for the detection of significant differences among treatments’ means and therefore also focus on statistics. It is however important to mention that the tests of comparisons of means are by no means relevant in the context of the genetic experimental design with perennial species. Given that the genetic effects are considered random by definition, there is no significance at all in the tests of comparisons of means (Duncan, Scheffé, Tukey, Bonferonni, Fisher, among others) whose methods are all derived on the supposition of fixed treatment effects, that is, fixed constants and not random variables. In the context of genetic selection, the plot size should be determined based on the selection accuracy (correlation between the true and the predicted genetic values) which is a measure of selection efficiency. This accuracy depends on the trait heritability, on the parentage between the individuals under evaluation, on the number of replications, and on the number of trees per plot. Thus, the number of individuals per treatment (progeny), which is function of the number of trees per plot (plot size) multiplied by the number of replications should be determined for each trait, based on its heritability. In the context of the prediction of genetic values, Resende, 1995, conducted a simulation study, evaluating the selection accuracy for different experimental conditions and levels of heritability: five different values (5; 10; 20; 40; and 80) of the relation (σ 2d / σ 2e - within-plots variance / among-plots variance); 14 levels of heritability (5% to 90%) and 30 different combinations of plot sizes (n) and number (b) of blocks. Plot sizes varied from 1 to 10 plants and number of blocks from 1 to 100. Fixing the total number (nb) of individuals, it is verified that plots with 1 individual and various replications always lead to a greater accuracy in relation to the plots with various individuals and less replications. This fact occurred for all the values of heritability and for the relationship σ 2d / σ 2e tested. On the other hand, plots of 1 plant may cause the problem of the loss of plots, this fact having little relevance for the prediction of genetic values, but being undesirable from the point of view of the statistical analysis. Therefore, the utilization of small plots but with more than one plant might be desirable. Another alternative suggested by Libby & Cockerham, 1980 refers to the use of non-contiguous plots with various plants, but with randomly distributed plants across the block. According to these authors, the design offers a comparable efficiency to that obtained with single plant plots. This strategy has been used in research with cacao: considering a mortality rate of 10% in the first two years, Cilas, 1995 recommends the utilization of 4 plants per treatment, randomized within each block. Summing up, plots of 1 to 6 plants should be used in research with perennial species. The only problem of the loss of plots formed by 1 plant, in terms of genetic evaluation, refers to the loss of representativity of some parents at certain levels of the fixed (block) effects that may bias the genetic comparisons if the BLUP procedure is not adopted. The determination of the number of replications is realized by dividing the total number of individuals per family, necessary for the maximization of the selection accuracy or the precise estimation of genetic parameters, by the chosen plot size. Sample size for parent selection: The number of individuals per half-sib progeny needed to obtain some values of accuracy in the prediction of genetic values of parents (Resende, 1995) are presented in Table 11.2, obtained based on the estimator of the accuracy. With a heritability of over 15%, 100 individuals per progeny result in accuracies above 90%. With heritabilities of 20%, about 80 individuals per progeny are necessary to attain 90% precision in the selection. On the other hand, with heritabilities below 35%, more than 100 individuals per progeny are necessary to attain 95% accuracy (Table 11.2). Sample size for selection of individuals: For the selection of individuals based on the multi-effect index or BLUP involving a single generation, the number of individuals necessary per half-sib family to obtain 90% and 95% of the maximal possible accuracy are presented in Table 11.3. It is verified that it is impossible to obtain an accuracy of 100% for the selection of individuals by combined selection. For heritability values (h2) varying from 5% to 90%, the maximum possible accuracy (with n →∞) varies from 53% to 95%. In general, h2 over 15% does not justify the use of more than 50 individuals per progeny (Table 11.3). Sample size for clone selection: The materials to be evaluated in clonal tests come from the base populations for selection (progeny tests) or the individual selection in areas of natural occurrence or commercial plantations (see Chapter 9). The first alternative is preferable since it allows a greater accuracy in this first stage of selection. Individual plants present a greater interaction genotypes x environments and therefore, the clonal tests should be installed in the greatest possible number of environments. Depending on the quantity of available material, the clonal tests could be realized in two stages: i) initial stage, aiming at the elimination of clones with a lower productive potential; ii) final stage or stage of recommendation of material for commercial plantations (see Chapter 9). The initial stage is recommended when a large quantity of genetic material for tests is available. It should be installed in linear plots and in a smaller number of sites. On the other hand, the final stage demands greater experimental rigor, often requiring the use of square plots with a greater number of plants. All these factors aim to minimize the effects of competition in the evaluations of the clones. Table 11.4 presents the adequate numbers of individuals per clone in clonal testing, in function of the broad-sense heritability at the plant level. It is verified that 100 plants per clone lead to accuracies over 90%, independently of the broad-sense heritability (Table 11.4). It is therefore unjustifiable to use more than 100 rametes per clone. However, knowing the trait heritability, a more adequate number can be chosen. For example, a heritability of 20% attains an accuracy of 95% in the selection, when using about 40 plants per clone, and an accuracy of 90% when using around 18 rametes per clone. Sample size for selection in diallels and factorials: The factorial (Paulin et al., 1994 and Paulin, 1995) and diallelic crossing designs (Cilas et al., 1995 and Dias & Kageyama, 1995) are commonly used in cacao improvement. For a 4 x 4 factorial or partial diallel with five parents, the values of accuracy and values of N = nb are presented in Table 11.5 to attain a certain accuracy in the selection by the additive genetic values associated to the various individuals of the experiment. These values were obtained based on the expressions of accuracy presented by Resende, 1999. It is verified that between 20 and 40 individuals per full-sib family are sufficient to attain over 90% of the maximal possible accuracy for heritabilities in the restricted-sense between 0.30 and 0.10 respectively. Sample size for estimation of variance components: Considering the progeny size generally adopted in the practice of improvement of some perennial species (6 plants per plot and 10 replications) - numbers that maximize the selection accuracy for the majority of the traits - the recommended sample sizes (T) (Resende, 2002) for the estimation of the heritability as well as the number of associated half-sib families are presented in Table 11.6. The T values presented (Table 11.6) reinforce the inference that when the total number of observations is under 1000, their use is not recommended for the estimation of genetic parameters. In this case, the prediction of genetic values should be realized using the mean values of the heritabilities obtained in literature. Harris, 1964 also concluded that at least 1000 observations are necessary for the estimation of h2, with a view to the construction of selection indices. For the precise estimation of the narrow-sense heritability based on half-sib progenies, approximate sample sizes of 8000, 4000, and 2600 observations for heritabilities of 0.10, 0.20 and 0.30 are recommended, respectively. For the estimation of the broad-sense heritability (h2a) one has that Var (h2a) = 8 h2a/N. This expression verifies that the estimation of h2a requires the smallest total number (N) of individuals for similar heritabilities and precisions. With n = 15 replications per clone it can be inferred that 17 clones are necessary to estimate h2a with reasonable precision. Such a precision is similar to the one obtained with N = 1000 individuals in half-sibs families structure, aiming at an estimation of h2 in the restricted sense. Thus, at least 250 seedlings (for example, 17 clones with 15 seedlings each) are necessary for an adequate estimation of the broad-sense heritability. |
| Choice of design, Simplicity flexibility and robustness, Multiplicity of objectives, Experimental designs in breeding. | Return To Table of Contents |
|---|
| Choice of design: Field research with tree species is frequently slow, tedious, and onerous. The long duration of the experiments exposes the trees to the action of pathogens, pests, climatic adversities such as drought and flood, and to damages such as by fire. The same experiment is commonly monitored and evaluated within and between years. In this period many trees and even entire plots are lost. Such facts can result in the generation of unbalanced data. These peculiarities distinguish research with perennial plants from that realized with annuals. For this reason, the simplicity, flexibility, and robustness of the experimental design to be used and the multiplicity of research goals are important aspects that should be observed. Simplicity, flexibility, and robustness: In virtue of the long duration of the field trials with perennials and to the dynamism of improvement programmes, it is not rare that changes of objectives occur that imply a revision of the original plans for the experiment. In addition, it is not rare that an experiment lasts longer than the career of the researcher. As a consequence, the ingression of new researchers results in new hypotheses that are normally tested on the trees of a pre-existing experiment. The scarcity of experimental areas and the high cost for the implantation and management of the experiments are common problems of research with perennials. It is therefore convenient that the designs used in this type of research are as simple, robust, and as flexible as possible: simple to be field-designed, analyzed and interpreted; robust to support unbalanced data caused by the loss of plots, treatments or blocks; and flexible to permit tests of new hypotheses. It is important to bear in mind that in research with perennials the plots consist of a single tree or few trees, as seen in the item Optimum plot size and number of replications. Thus, the variation due to the environmental effects is smaller and the variability between plots attributed more to the differences between plants. In any case the environmental effects tend to increase with the time of duration of the experiment. The choice of such designs should have the objective to reduce the variance of the experimental error. When this objective is attained, the power of the statistical tests to detect differences between treatments is amplified and the precision of the estimates of treatment means is improved. Consequently, the standard deviation and the confidence intervals associated with these mean estimates are reduced (Pearce, 1976; Swallow, 1981 and Martinez, 1989). Moreover, these designs should support a certain degree of unbalancing of the data, be simple, and offer flexibility in the change of research objectives with the same trees. The designs that best meet these demands are the following: Completely randomized: It is the design of greatest simplicity and robustness. Its utilization, however, is restricted to experiments involving a small number of treatments, subjected to more homogeneous environmental conditions. If these demands are met the reduction of the variance of the error is maximal and the unbalancing of the data does not cause problems. Randomized complete blocks: This design unites simplicity, robustness, and flexibility and is for this reason the most used in improvement of perennials. It is sufficiently robust to support the loss of one or more entire blocks or even of one or more treatments (Pimentel Gomes, 1990). The orthogonal result proportioned by the randomization of treatments and blocks makes this robustness possible. Additionally, the loss of one plot does not represent a serious problem in this type of design. However, it is not the most recommended when there are more than 20 treatments to test - a common situation in improvement where, frequently, one or more hundreds of progenies or cultivars are tested. However, even in such cases the use of randomized blocks is perfectly possible, distributing the treatments in to various experiments of 20 treatments each, all in randomized blocks. The trial of new treatments in an experiment of previously installed randomized blocks is also possible and depends basically on the number of blocks and the number of old and of new treatments (Martinez, 1989). Each new treatment should correspond to an old one, as in a latin square. The old treatments are arranged in the lines and the blocks in the columns. Incomplete blocks: Incomplete blocks, better known as lattices, can be of square, cubic and rectangular types. They are commonly used a lot in annual species, particularly in yield trials of cultivars. If there are, for example, 300 maize progenies to test, three 10 x 10 lattice squares can be installed. In the lattices, idealized for trials with a large number of treatments, the blocks are simple subdivisions of the replications. This way, each treatment appears once in the replication and each pair appears only once in the blocks. The blocks are randomized within the replications and the treatments randomized within the blocks. In the balanced lattices, despite all pairs of means being compared with equal precision, the number of replications increases in the proportion of the square root of the treatment number plus 1. In the unbalanced lattices, the situation is worse; the pairs of means are not compared with equal precision. These designs, which come to be a type of incomplete balanced blocks, are not robust in the extent to which the loss of a given treatment or a given block disturbs the balance and the symmetry of the same and complicates the analysis. Finally, the efficiency of the lattice is confirmed by the ratio between the variance of the error, in case the experiment had been installed in randomized blocks, and the variance of the effective error generated by the lattice. If this ratio is lower than 110% the use of the lattice will not have been effective (Cochran & Cox, 1957 and Pimentel Gomes, 1990). Randomized complete blocks with common treatments: The complexity and the absence of robustness and flexibility make the utilization of the lattices in research of perennials difficult. These, according to Pimentel Gomes, 1990, can be substituted advantageously by designs in randomized complete blocks with common treatments. To test a large number of cultivars in the order of hundreds, they can be distributed in groups installed in randomized complete blocks. Two or more standard cultivars are aggregated to each group. For general cases, there are g groups of experiments, each one with z regular treatments and c common treatments, amounting to v treatments, where v = gz + c. In this situation, each group is assessed in each experiment in r blocks of k plots, where k = z + c, with the following analysis scheme: Sources of variation / GL: Blocks / r-1; Treatments / z+c-1; Error / (r-1)(z+c-1) and Total / (z+c)(r-1). The joint analysis of all experiments obeys the following scheme: Sources of variation / GL: Experiments (E) / g-1; Blocks/E / g(r-1); (Adjusted) treatments / gz+c-1; Common treatments x and / (c-1)(g-1); Mean / g(r-1)(z+c-1) and Total error / (z+c)(gr-1). While the joint analysis involving all the experiments evaluated during various years of harvest obeys the scheme in the Table. The source of variation common treatments x E is not usual and tests the independence of the common treatments and, by extension, the regular treatments in the various trials. When not significant, it indicates that the common treatments behave similarly in all trials, suggesting the same for the regular ones. The situation is more complex when this interaction is significant. In this case, it is best to use the mean square of the source common treatments x E as source of error to test regular treatments of distinct groups, as suggested by Pimentel Gomes, 1990. The comparisons between pairs of treatment means are also given by Pimentel Gomes, 1990. This type of design is robust, once the loss of treatments does not represent an additional difficulty. The only recommendation is the use of at least two common treatments; the use of a single common treatment and its eventual loss would make the joint analysis unviable. The design is robust enough to admit distinct sizes of regular treatments if these are not very discrepant. There is a record of a single application of randomized complete blocks with common treatments realized by Almeida, 1991 involving two comparative trials of hybrids in cacao improvement. In fact, in the beginning the cacao improvement programme planned the installation of common controls in all comparative yield trials of hybrids (Vello et al., 1972). The importance of using such controls to make the results comparable was then emphasized. However, the analysis of these trials as designs in randomized blocks with some common treatments was never realized. It is hoped that the favourable properties of this design type lead to the generalization of its use. Latin squares: The latin square is a robust design type, to the extent that it introduces lines and columns as local control (Steel & Torrie, 1980). Each treatment appears a single time in each row or column. In spite of supporting unbalanced data, its principal limitation lies in the requirement that the number of replications must be equal to the number of treatments. This restricts its use to trials with four to eight treatments only and grants it little flexibility (Cochran & Cox, 1957). With few treatments and the introduction of rows and columns for a better control of the heterogeneity of soil and experimental material, the design provides few degrees of freedom for the error. Nevertheless, it is very appropriate for comparative yield trials involving few cultivars such as elite cultivars. Dias et al., 1998, for example, analyzed the performance of five cacao cultivars using the 5 x 5 latin square successfully. Similar to the randomized blocks, the latin square admits the overlapping of new treatments over the old if they are equal in number and independent from each other. Each new treatment should appear only once in each line, column and in each old treatment. The resulting design is called ‘greek latin square’ (Martinez, 1989). The term design has a double meaning (Pimentel Gomes, 1984a): it refers to the method of laying out the plots in the experiment or even to the method of organizing the treatments. Only the first has been approached so far. The second meaning will be outlined henceforth in the discussion of factorial and split plots, both typified as arrangements. Split plot: The split-plot design allows for the simultaneous test of more than one factor. One of them is allocated in the plots and the other, considered the principal or the most important, in the subplots. The plots and thereafter the subplots within the plots are randomized. There are two components of error: one associated with the plots (error a) and the other with the subplots (error b). The factor of the subplots is tested with greater precision in view of the greater number of degrees of freedom associated with error b (Pimentel Gomes, 1990). Split plots are classified in two types: in time and in space. In the improvement, for example, it is common to test cultivars and spacing in split plots in the space. On the other hand, the comparative yield trial of perennials involves, besides the cultivars, an additional factor, which is time. The latter is allocated in the subplots and constitutes the well-known split-plot in time (Steel & Torrie, 1980), widely used in cacao improvement (Pereira et al., 1987 and Dias et al., 1998). Factorial: These designs are characterized by containing all combinations of the factors under test. For example, with 10 cultivars trialed in three spacings there are 30 possible combinations. Theoretically they are more effective than those designs that work with a simple factor. However, the number of treatments increases a lot with the combination of the factors and affects the efficiency of the factorial. With a high number of combinations, the homogeneity of the experimental area can be affected and result in the need to use randomized blocks and/or in the adoption of confounding (Pimentel Gomes, 1990). On the other hand, as in research with perennials, the second factor of test is almost always time (year), the increase of the number of treatments by the combination of factors does not represent a major problem. Besides the demand of simplicity and robustness it is recommendable that the design has sufficient flexibility to give room to tests of new hypotheses and can thus meet more than one research objective. Multiplicity of objectives: Once again, field research with perennials demands highs investments for the implantation as much as for the maintenance of the experiments. Therefore the flexibility of the experimental plan required for this species type and which was the aim of the discussion in the previous item should be linked to attending multiple objectives. It is inadmissible that a single hypothesis is tested in an experiment whose duration lasts up to a decade and frequently outlasts it, as in the tests of cacao progenies. The breeder of perennials should pay attention to this aspect and use creativity to extract the maximum information from the experiment in order to support actions and future decision taking in the improvement programme. Thus, for example, based on the analysis of the results of five years (1986-90) of data involving the evaluation of five yield components in a 5 x 5 complete diallel installed in randomized complete blocks in CEPEC, Bahia, Brazil, the following important questions were addressed for the genetic improvement of cacao: i. - Is the inheritance of the yield components predominantly additive or dominant? ii - Is it possible to examine the classification of the racial groups using measures of multivariate genetic divergence? iii - Is it possible to predict and identify superior hybrids by using measures of multivariate genetic divergence of the parents of the hybrids? iv - Is this divergence stable in time? v - Are the multivariate methods of estimation of the genetic divergence equally consistent and concordant among each other? vi - Which is the minimum period of successive harvests to evaluate genotypes for yield components? Answers to these questions were obtained and published in diverse articles summarized by Dias, 1998. It is worth emphasizing that the analysis of a single experiment allowed answers to these six important and recurrent questions in cacao improvement. Another example of how it is possible to respond to multiple questions in cacao improvement based on the analysis of a single experiment occurred when three non-improved local cultivars (‘Maranhão’, ‘Pará’, and ‘Parazinho’) were tested against two improved cultivars (Commercial hybrid and ‘ICS 1 open pollination’), in a comparative yield trial. The trial was installed in a latin square and three yield components were evaluated during a decade (1984-93). The questions answered and published in various articles summarized by Dias, 1998, were the following: i. - Does ‘Commercial hybrid’ have a superior yield performance to the local cultivars? ii. - In what conditions are the improved cultivars temporally more stable than the local cultivars? iii. - Which is the minimum period of successive harvests to evaluate cultivars for yield components? iv. - Are the cultivars in test distinct from each other? Unfortunately the procedure of considering multiple research objectives is not common among breeders of perennials, especially of cacao. However, the increasing scarceness of funds for research seems to indicate that this is a good alternative to be followed. It is worth emphasizing that the implementation of this procedure requires a good dose of creativity from the breeder and the awareness that good research is based on good hypotheses. Designs for research in improvement: The adequate experimental design for the actions of improvement should obey the fundamental principles of research: replication, randomization, and local control (item Experimental principles). As local control the homogeneity within layers or blocks must be emphasized for which the designs in randomized blocks and lattices are principally recommended. The incomplete block designs (lattice, for example) are especially indicated in the situation of a large number of treatments and high environmental variability (quantified by the b of Smith, for example) in the experimental area. The relative efficiency of the experimental designs depends, above all, on the level of environmental spatial variation in the experimental area. Fu et al., 1998 used a spatial geo-statistical model which permits the specification of various levels of environmental variation and concluded that the incomplete block designs (lattice and alfa) are superior in a large number of situations in terms of statistical efficiency for the estimation of treatment means. Another class of designs that has been used more and more in the last years is that of augmented blocks of Federer (Federer, 1958 and 1998; Rios, 1997 and Wolfinger et al., 1997), which, by construction, are unbalanced. Analytical procedures (lattice, augmented blocks, intercalated control, and shifting means) for the evaluation of progenies in plant breeding programmes were compared based on the efficiency in the control of the experimental error and the precision of the estimates of genetic parameters (Rios, 1997). It was demonstrated that the utilization of augmented blocks is only viable in the initial programme phase and that these generate low precision estimates. The designs of augmented blocks and incomplete balanced and partially balanced blocks are non-orthogonal. In these cases, their use for selection leads to biased treatment means, even when the survival rate is 100% (Resende, 1999). This does not mean to say that these designs must not be used. However, when they have to be used, it is best to use them in association with BLUP, which adjusts the means to the BLUE estimates of the fixed or identifiable environmental effects. The designs in blocks are based on the supposition of a priori knowledge of the heterogeneity of the experimental area so that it would be possible to allocate all plots (treatments) in homogeneous blocks (Lotodé, 1971). In case the heterogeneity is not known a priori, the delimitation of the blocks becomes arbitrary, a fact that may imply a strong within block heterogeneity. In this context an alternative is the random allocation of plots of 1 plant in the experimental field and posterior control of the environmental heterogeneity, using methods such as the analysis of covariance. The method of Papadakis (Papadakis, 1984) can be used associating an independent covariable with the variable studied or the method of the regionalized or spatial variables – geo-statistical methods, according to Lecoustre & Reffye, 1986. The a posteriori adjustment to the environmental gradients in progeny tests presents a potential for a significant increase of the efficiency in the estimation of genetic parameters and in the selection. In this context the use of completely randomized plots of 1 plant (randomized complete design) has again become common (Lotodé & Lachenaud, 1988). In the geo-statistical models for spatial analysis the matrix of residual covariance is given by R = Σ = a non-diagonal matrix that considers the correlation between residuals, for example, auto-regressive lines and auto-regressive columns, to contemplate the spatial auto-correlation among the observations. These permit the study of the spatial variability of the soil in the experimental areas by means of procedures that permit the best criterion of environmental stratification (for mass selection or best definition of the fixed effects in the BLUP procedure). In this context, the spatial analysis is realized simultaneously (Cullis et al., 1998) with BLUP prediction. Knowledge and models of temporal series are very useful in this research area. Also, many techniques employed in the area of geology such as the construction of variograms and semi-variograms and the realization of kriging are used in the spatial analysis. The various designs are generally adequate for the estimation of genetic parameters and for the selection of parents based on the performance of their progenies. For these objectives the experimental design of randomized blocks with families of half-sibs and 1 tree per plot is especially recommended as the most appropriate, including the estimation of genetic correlations and study of the interaction genotypes x environments (Resende, 2002). |
| Statistical analysis strategies, General considerations, Analysis of mixed linear models (REM/BLUP). | Return To Table of Contents |
|---|
| General considerations: A first important aspect in the analysis of research data refers to the nature of the effects of the model which can be fixed or random. Designating the sample size as n and the population size as N it is possible to define these effects by the analysis of the ratio n/N. In case this ratio tends towards zero, the results of the analysis of the finite sample can be extrapolated to the infinite population and the effects of the model are called random. The effects will be fixed when this ratio tends towards 1. In this case, the sample is the population itself and the results generated by its analyses can not be extrapolated or generalized. Mixed effect models include both random and fixed effects. Note that the conception of effects is more related to the aspect of representativity of the population by the sample than to the aspect of the sample size. Thus, for example, random samples of a given population, despite being small in relation to size, characterize random effects. Attention must also be paid to the fact that many, if not most of the experiments in cacao improvement are of fixed effects. Therefore, all conclusions and implications based on them only serve for that set of treatments studied. The words of Fisher, 1960, are therefore still very topical and ought to be reflected on: “The researcher always assumes that it is possible to draw sound inferences on the research results; that it is possible to separate the causes from the consequences, the hypotheses from the observations; as a statistician would say, ‘the population from the sample from which it was extracted’, or as a strategist would say, ‘the general from the particular’”. A second important aspect is that the accumulation of data of successive years from the same plot is a typical characteristic of research with perennial species. The trait yield and its components, for example, is distributed in among and within years. The yield is measured in tonnes of dry beans per hectare per year, based on the accumulated data of fortnightly or monthly successive harvests. The question is to know what is the best way to deal with this trait. Pearce, 1976, believes that it is sufficient to work with the accumulated data from all the years. In fact if the total response (total yield, for example) is that which is required then this focus is adequate. Nevertheless, the previous focus sacrifices the possibility to analyze the interaction between the treatments and time period. The cacao farmer, the final beneficiary of the improved cultivar, is more interested in a cultivar’s stability year after year (Dias et al., 1998), since he harvests, processes and commercializes the dry cacao production every year. In other words, the optimum economical return of a given cacao plantation requires cultivars of low annual yield variability. It is possible to divide the time in periods, preferably equal ones such as crop or calendar years of harvests, to exploit this interaction. In this case time becomes a test factor and the experiment can be analyzed as split plot over time. Dias et al., 1998, worked with 10 years of yield component data obtained in a 5 x 5 Latin square and analyzed the total yield accumulated of the period and per year. Furthermore, split plot analyses were presented allocating the latin square in the plots and the years in the subplots, which seemed to be a good strategy. One has to be careful when delimiting the period. In some producing countries cacao exhibits the growth habit of biennial species such as coffee. In such cases it is convenient to define two-year-periods. Summing up, there are two types of possible analyses to be considered in handling data of experiments with perennial species. The first refers to the analysis of data of a single year or of a single harvest or of data accumulated over the years. In this first situation the analysis is similar to that processed for annual species. The second refers to the combined analysis of the data generated from various years or harvests. There are different possible approaches of use that are the object of discussion in the following items. A last and relevant aspect refers to the way the data is collected, which may be per plant or per plot. The focus is frequently the plot owing to the scarce and high labour cost. Strategies such as the prediction of genetic and genotypic values, presented and discussed in Chapter 6, are only possible with data collected from individual plants though. It is therefore necessary to review the manner in which field data is collected. On the other hand doubts persist if such data should be collected from the whole plot or only from the plot without borders. Data of total plots containing a smaller number of cacao trees seems to be more favourable (Lockwood & Martin, 1976 and Martin & Lockwood, 1979). There are instances where the elimination of the interference between the progenies by adopting plots with internal borders is not the most suitable strategy. In comparative trials of hybrid yields, for example, the competitive ability is a characteristic heterotic expression of the hybrids. Additionally, it is worth remembering that such hybrids are planted commercially in the form of a hybrid mixture and therefore, under open competition conditions; therefore the selection of the best hybrid progenies must take place in an environment of competition. The same reasoning is obviously not valid for the external border along the perimeter of the experiment, the latter considered to be essential. An additional difficulty in cacao field trials refers to the minimal evaluation period of the potential of a given cultivar in terms of yield and other traits. The coefficient of replicability has been applied successfully with the objective to determine such a period (Dias & Kageyama, 1998). This minimum period was determined to be two years of successive harvests for cacao grown in experiments in the South of Bahia, Brazil, (see also Chapter 6). Analysis of mixed linear models (REM/BLUP): In research with perennial crops the plants are exposed to risks for long time periods; a fact which, associated with the overlapping of generations, tends to generate unbalanced data. This, together with the need for a genetic evaluation at the individual level, is a plea for the use of the mixed model methodology (Resende et al., 1993 and Resende et al., 1996). The methodology of mixed models through the procedures BLUP (best linear unbiased prediction) and REML (restricted maximum likelihood) of prediction of genetic values and estimation of variance components respectively (Henderson, 1984), represents a flexible and indispensable tool for breeders of perennial plants. Such methodology allows for the estimation and prediction in situations of unbalanced data and non-orthogonal designs. As an example of the use of the mixed model methodology a research situation with half-sib progenies in the incomplete or lattice block design will be considered. Adjusting the block effect within the replications as a random effect one has the following statistical model at the individual level: y = Xß + Za + Tc + Sg + d, where: y, ß, a, c, g and d - data vectors of fixed effects (means of replications), of genetic additive values (random), of plot effects (random), of blocks within replications (random) and random errors, respectively. X, Z, T and S - incidence matrixes for ß, a, c, and g, respectively. The mixed model equations to predict genetic values by the BLUP procedure in an individual model are equivalent to:  where:   = intra-class correlation between individuals of the same block, `but from different plots.  = narrow-sense heritability.  = intra-class correlation due to the common environment of the plot. σ2F = σ2a + σ2c + σ2g+ σ2d = phenotypic variance. σ2a = additive genetic variance. A = matrix of additive genetic parentage among the individuals. The matrix of coefficients of the equations being a mixed model:  so the inverse is:  The variances of the prediction errors (PEV) (Henderson, 1975) of the genetic values when R (the matrix of the residues) = I σ2d, are given by PEV = Var (â - a) = M22 σ2d. The accuracies (correlation between predicted and true genetic values) associated with the precision of the genetic values are given by râa = [1-PEV / σ2 a]1/2. Equivalent estimators are: râa = (1-di λ1)1/2 and PEV = (1-r2âa) σ2a, where di is the i-th element of the diagonal of M22. The adjustment of the within-repetition effects of blocks as random is favourable due to the fact that additional fractions of the additive genetic variation are used in the prediction when such block effects are considered random. Proceeding in this form, one tends to obtain smaller variances in the prediction error and consequently greater precision and greater accuracies, which is highly favourable (Resende, 1999). Considering the replication effects as fixed provides invariant predictions of the fixed effects, thus removing bias from the genetic comparisons through the replications. This consideration is above all especially important when there is a non-random allocation of the genetic materials in the replications and in this case, the BLUP minimizes the bias caused by associations between effects of replications and the genetic level of the materials that develop in them. Invariant genetic evaluations of the replication effect, that is, comparable in an unbiased manner by the replications, are obtained depending on the consideration of the replication effect as fixed and also on the good association (capacity to estimate environmental differences between replications according to Weeks & Williams, 1964), between the replications. This is achieved by representing-the majority of the progenies at each level of the fixed effect. Complete blocks (blocks in the randomized block design and replications in the lattice design) can therefore generally be assumed as fixed effects and incomplete blocks (block within replications in the lattice design) as random effects (Resende, 1999). The estimates of the components of variance through REML may be obtained by algorithms that repeat in the mixed model equations as presented in Chapter 6. The approach of mixed models provides all information a breeder desires, such as the prediction of the genetic values, the estimation of variance components, the variance of the prediction error, the selective accuracy, and the genetic progress. Thus, it is hoped that this methodology will come to be used routinely in cacao improvement. |
| Replicated measurements, Sample size for replicability estimation, replicability and Adequate number of measurements. | Return To Table of Contents |
|---|
| Replicated measurements: Research with tree species generates multiple data over time. The genetic evaluation involving individuals of these species must therefore be based on models that consider the additional effect denominated ‘permanent environment’, as well as the parameter associated with the phenotypic correlation between replicated measurements of the same individual, denominated ‘replicability’. Aspects referring to repeated measurements in tree species based on Dias, 2000, Dias et al., 2001 and Resende, 2002 are presented in this item. Sample size for the estimation of replicability : The standard deviation of the estimate for replicability () can be calculated considering the estimator of the standard deviation of the simple linear correlation coefficient since the replicability is an intra-class correlation. Thus one has:  for the case of more than two measurements (m) and:  for the case of only two measurements. Considering the two measurements, the number (N) of individuals needed to obtain an estimate of r with a desired standard deviation σ, is given by N = (1-ρ2)2/σ2ρ + 1. The adequate values for N for some of the ρ e σρ are shown in Table 11.7. Considering a standard deviation of 0.05, it is verified that 100 to 400 individuals are sufficient to obtain precise replicability estimates that present parametric values between 0.70 and 0.10, respectively. Replicability and adequate number of measurements: The efficiency of use of more than one measurement in selection (Table 11.8) is given by the ratio between the selection accuracies based on various measurements and the selection accuracy based on only one measurement. For the individual or mass selection the efficiency is given by {m/[1+(m-1)ρ]}1/2. The efficiency in terms of genetic gain by unit of time (year for example) is given by the ratio between the accuracies with m measurements and with 1 measurement, multiplied by the ratio of the interval of generations (L) with 1 and with m measurements. Thus, the annual efficiency is given by [m/(1+(m-1)ρ)] 1/2 L1/Lm. The values of selective efficiency per selection cycle and annual values for the mass selection are presented in Table 11.8, considering L = (6 + m) and various values of accuracy and number of measurements. An interpretation of Table 11.8 should consider the different selection objectives: i) the CE (Cycle Efficiency) should be analyzed for the short-term improvement (maximization of the gain in the actual generation); ii) the AE (Annual Efficiency) should be considered for the long-term improvement by recurrent selection. For an estimated replicability of 0.75, it is verified that for short-term improvement, it is not worthwhile to carry out more than three measurements and for long-term improvement by mass selection is ideally based on only one measurement (Table 11.8). It is also verified (Table 11.8) that one or two measurements with a replicability of 0.50 equally contribute to the long-term improvement. The use of two and three measurements only becomes advantageous (AE greater than 1) for improvement in the long term when the replicability is equal to or below 0.45 and 0.35, respectively. The selective efficiencies per selection cycle (CE) for the different selection methods are presented in Table 11.9. Even with a fixed and known value of replicability the gain in efficiency with the use of various measurements depends on the selection method and the trait heritability. The efficiency of methods that use only the information of the individual is maximum and decreases rapidly by methods that emphasize more the information of family means. So the use of two measurements provides 10% more (in relation to one measurement) genetic gain for the mass selection modality and practically does not lead to a greater efficiency for parent selection. For the methods that use the information on the individual and the family the gain in efficiency is reduced to half (5%) or less (Table 11.9). In relation to the heritability, the efficiency diminishes with the increase of the heritability for the methods that use only the information of means. The efficiency increases with the increase of the heritability as much for the methods that use the information of means as for those that use the information of individuals (Table 11.9). In this last case, the greater h2 the greater is the weight attributed to the information of the individual and therefore the greater the efficiency of the use of replicated measurements. The results reveal the need to consider the selection method as much as the trait heritabilities in the recommendation of the adequate number of measurements for each individual and not only the replicability. |
| Analysis of experiments with replicated measurements. | Return To Table of Contents |
|---|
| Univariate approach: The evaluation of progenies or hybrids in the design of random blocks with various plants per plot is considered. The statistical model associated to the evaluation of ‘p’ progenies in ‘b’ blocks, with ‘n’ individuals per plot and ‘m’ measurements per individual is: Yijkl = µ + pi + bj + ml + pbij + pmil + bmjl + pbmijl + d*ijkl, where: µ = general mean, fixed, E(µ) = µ, and E(µ2) = µ2; pi = effect of the progeny i, random, and (pi)=0, and E (pi2)=σp2; bj = effect of the block j, fixed, and (bj)=bj, and E (bj2)=Vb; ml = effect of the measurement 1, fixed, and (ml)=0, and E (ml2)=Vm; pbij = effect of the interaction progenies x blocks, random, E(pbij)=0, and E(pbij)= σ pb2; pmil = effect of the interaction progenies x measurements, random, and pmil=0, and pmil2= σ pm2; bmjl = effect of the interaction blocks x measurements, fixed, and (bmjl)= bmjl, and E (bmjl2)=Vbm; pbmijl = effect of the interaction progenies x blocks x measurements, random, and (pbmijl)=0, and E (pbmijl2)= σ pbm2; d*ijkl = effect of the plant k within of the plot ij, in the measurement l, random, E (d*ijkl)=0, and E(d*2ijkl)= σ2d*; The effect d*ijkl can be decomposed in d*ijkl = dijk + tijkl, where: dijk = effect of the plant k within the plot ij, random, and (dijk)=0, and E (dijk2)= σ d2; tijkl = effect of the interaction plants x measurements within plots, random, and (tijkl)=0 and E(tijkl2)= σ t2. The scheme of the analysis of variance (Resende, 1999) at the level of individual plants with the respective mathematical expectations of mean squares of the various sources of variation is presented in the scheme of the analysis of variance table. Based on this scheme of analysis there are the following estimators for the components of variance:  Furthermore, there are the following estimators:  ^σ2p = estimates of variance among progenies, inflated by the interactions progenies x measurements (σ2pm), progenies x blocks (σ2pb), and progenies x blocks x measurements σ2pbm); ^σ2p* = ^σ2p -^σ2pm - ^σ2pb - ^σ2pbm = estimate of the variance among progenies, free of the interactions with measurements and with blocks. The estimator of the replicability or intra-class correlation between replicated measurements in the present case is given by:  It is important to mention that the very use of the model of replicability presupposes that the genetic correlation of the trait with itself is equal to 1 in another measurement. In this case, such a model admits that the interactions genotypes x measurements and genotypes x measurements x blocks may be due to the heterogeneity of variances only. When the genetic correlation of a given trait with itself in another measurement is smaller than 1, the use of the multivariate model (which assumes different measurements as different traits) is recommended instead of the model of replicability. In most cases, significance of the interaction blocks x measurements within a same experiment is not expected and therefore, the adjustment for this interaction would not be necessary. As the edaphic conditions are identical for the blocks from one measurement to another and the climatic conditions (precipitation for example) identical for all blocks in a measurement, it is not expected that certain blocks would be better in one measurement and other blocks the best in other measurements (which would characterize an interaction). Thus, disregarding the interaction B x M and bearing in mind that the components σpm2, σpbm2, and σt2 refer to the variance components of the temporary environment and, therefore, can be reunited in a single residue σt2, it is concluded that the reduced analysis model can be used efficiently. This reduced model is given by: YijKl = µ + pi + bj + eij + dijk + ml + tijkl, in que: µ = general mean, fixed, E(µ) = µ and E(µ2) = µ2; pi = effect of the progeny or family i, random, and pi = 0, and E(pi2 = σp2; bj = effect of block j, fixed, and (bj) = bj and E(bj2) = Vb; eij = effect of plot ij, random, and (eij) = 0 and (eij2) = σ(e2; dijK = effect of the individual k within plot ij, random, and dijk = 0 and E(dijk2) = σd2; ml = effect of the measurement l, fixed, and (ml) = 0 and E(ml2) = Vm; tijkl = effect of temporary environment, random, and (tijkl) = 0 and E(tijkl2) = σt2. The scheme of analysis of variance associated with the reduced model as well as the mathematical expectations of mean squares (E(QM)) are presented in the following for analysis at the level of plot means. (Table) There are the following estimators for the components of variance: σt2 = Q8 = variance of temporary environment; σd2 = Q7 - Q8/m = permanent variance within plot; σe2 = Q6 - Q7/n = permanent variance between plots; σp2 = Q5 - Q6/b = genetic variance between progenies; σF2 = σp2 + σe2 + σd2 + σt2 = phenotypic variance. One still has the following estimators: h2 = 4 σp2/ σF2 = heritability in the narrow-sense (considering half-sib progenies). Pb = (σp2 + σe2 + σd2)/ σF2 = (σa2 + σd+2 + σep2 + σe2)/σF2 = replicability at the level of individuals in the block. The analysis of replicated measurements according to the presented model, using the mixed model methodology (univariate model of replicability), is presented in the following. Mixed model y = Xb + Za + Wc + Tp + e, where: y, b, a, c, p, and e: data vectors of the (fixed) block effects, (random) genetic additive effects, (random) plot effects, (random) permanent effects, and of random errors, respectively. X, Z, W, and T: matrixes of incidence for b, a, c, and p, respectively. In this model, the fixed block effects (b-1 degrees of freedom), measurements (m-1 degrees of freedom), and interaction measurement x block [(b-1)(m-1) degrees of freedom], can be adjusted in a single effect (denominated combination block-measurement with mb elements or levels and mb-1 degrees of freedom), a statistically correct and computationally desirable and necessary procedure. The temporary effects of the interaction progenies x measurements and progenies x measurements x blocks are incorporated into vector e, together with the effect of temporary environment in the strict sense. In this model, the plot effects c refer to the permanent environment between plots. Mean and variance distributions and structures. y|b, V ~ N(Xb, V) a|A, σA2 ~ N(O, σa2) c| σc2, ~ N(O, I σc2) p| σp2, ~ N(O, I σp2) e| σe2, ~ N(O, I σe2) Cov (a, c`) = 0; Cov (a, p`) = 0; Cov (a, e`) = 0; Cov (p, c`) = 0; Cov (p, e`) = 0; Cov (c, e`) = 0; that is:  P = I σp2 V = ZA σa2Z` + WI σc2W` + TI σp2T` + I σe2. Mixed model equations.  where:  = individual heritability in the narrow sense in the block in a given measure;  = individual replicability in the block;  = coefficient of determination of the permanent effects within the plot;  = correlation due to/under the common environment of the plot. Estimators of the components of variance by REML via the algorithm EM. ^σ2e = [y'y - ^b' X'y - â' Z' y - ^c' W'y - ^p' T'y]/[N – r(x)] ^σ2a = [â'A-1 â + ^σ2e tr (A-1 C22)]/q ^σ2c = [^c'c + ^σ2e tr C33]/s ^σ2p = [^p'p + ^σ2e tr C44]/q in which: C22, C33 e C44 come from:  Multivariate approach: The main characteristic of the experiments of replicated measurements is that the same tree is measured repeatedly. In a comparative yield trial with perennial plants for example in which the yield is distributed over the year and during various years, it is possible to evaluate the increase of yield in time by means of measurements at regular time intervals. It must be assured that the plants in the plots equally receive the same management and are under the same influence of the crop years. Under these circumstances the plants in the plots compete in yield and are measured repeatedly to inform on the yield response to years. This type of experiment unites two principal advantages (Neter et al., 1990): i) it allows the causes of variation among plants to be excluded from the experimental error, improving the precision for treatment comparisons and ii) it makes research possible when the plants are present in a small number. In spite of being an elegant approach with interesting properties, the experiment of replicated measurements has rarely been used in plant improvement, very likely due to the complexity of its foundation, understanding and of the unconventional procedures of the analysis. It is worth remembering that this type of approach involves both the univariate (Danford et al., 1960) and multivariate variance analyses (Cole & Grizzle, 1966). Furthermore, this type of design has been mistakenly treated as an experiment in split plots in time in perennial species (Steel & Torrie, 1980), with the factor time allocated in the subplots. Time is however not an experimental factor whose levels may be randomly allocated to the subplots (Littell, 1989). The potential and the applicability of this type of experiment in genetic plant improvement programmes, especially for tree species, should therefore be investigated. This type of analysis establishes a compromise between the uni and multivariate approach (Littell, 1989), in order to validate the F tests for the effect of time and for the interactions involving this effect. Therefore when measurements are taken on the same plant at different times, a structure of covariance that can inflate the probabilities of the error type I for the F tests is created. This structure is calibrated by the condition H-F (Huynh & Feldt, 1970) when establishing that any orthogonal contrasts between the replicated measurements have the same variance and zero covariance. In these cases the approximated chi-square test is also applied to test the structure of covariance, also denominated Barlett’s test of sphericity. The usual F tests, processed in the univariate approach for the effect time and its interactions, are valid if the data do not violate the H-F condition and the qui-square is insignificant. In the cases of violation of the H-F condition the application of the multivariate analysis is justified as it implies no conditions on the structure of covariance of the replicated measurements. The statistic of Wilks is used (transformed into a value corresponding to F) for the significance tests of differences between the mean vectors for time and its interactions. The polynomial transformation is also frequently used as time represents a quantitative factor. When condition H-F is not satisfied, another interesting alternative is the adjustment G-G (Greenhouse & Geiger, 1959) in the value of the probability associated to the univariate F when the degrees of freedom of the numerator and denominator for the F test are adjusted. Probability values of significance for the condition H-F, adjustment G-G, and F multivariate are calculated using the SAS programme (SAS Institute, 1989). SAS, moreover, processes analysis of variance of replicated measurements with great robustness in diverse designs by means of the command “repeated” used in PROC ANOVA or GLM. The syntax of the analysis of variance of repeated measurements by SAS can be seen in Littell et al., 1991 and Dias, 2000 and has recently been applied in cacao (Dias et al., 2001). Again, measurements or successive evaluations of a given trait of the same plant or plot denote an experiment of repeated measurements and must be adequately analysed. In other words the correlation between multiple measurements of the same trait collected from the same plant is not necessarily equal to 1. Commonly, the measurements realized in closer intervals tend to be more correlated than those collected at longer intervals. This fact inflates the rates of error type I for the F tests involving time and their interactions. It is therefore necessary to test the structure of covariance of the repeated measurements and only then validate the univariate tests. As an example the analysis of cacau yield data measured in kg/ha/year of dry beans over 10 years (1984-93) of a trial installed in a 5 x 5 Latin square based on the evaluation of five cultivars is cited. The analysis in the traditional format - split plot in time - was presented by Dias et al., 1998, allocating the latin square in the plots and the years in the subplots, according to the scheme below for analysis at the plot level, considering ‘g’ evaluated genotypes in ‘a’ years and ‘b’ blocks and without diminishing lines and columns. See Table . Therefore, to know up to where these univariate tests for the effects of years and genotypes x years in that analysis are valid, the correlation structure of the repeated measurements needs to be investigated. The partial correlations computed within plots (Table 11.10) reveal that, in fact, the degree of correlation tends to diminish the more distant the measures were in time. For example, the correlation between year 1 and 2 is 0.88, 0.68 between year 1 and 3, and 0.79 between year 4 and 1, and so forth. Again, this is common in experiments of repeated measurements. The sphericity test of the matrix of covariance of the error, which tests condition H-F with the chi-squared approximated to orthogonal contrasts, produced a chi-squared value of 362.8 (P < 0.0001), casting doubts on the validity of the F test for cultivars x years. The multivariate test, on the other hand, does not require adherence to the H-F condition to test the effects of years and cultivars x years. This test produced values of P = 0.0001 and P = 0.0005 associated to the multivariate Fs for years and cultivars x years respectively as can be observed below; in contrast to P = 0.0001 in the case of the univariate Fs for both effects as can be seen in the Table. In the practice, the non-adherence to the H-F condition promotes the reduction of the P value associated with the univariate F for the interaction. Given that condition H-F was not met, the multivariate or the univariate adjusted tests are the most recommended. Although both tests reveal significance of the probabilities associated to the effects of years and cultivars x years, the P values for the univariate and multivariate tests and adjustment G-G were only concordant for the first effect (year), in this case P = 0.0001. The additional columns G-G and H-F represent adjusted probabilities for the correction of the unequal correlations between the pairs of repeated measurements. These are generated by multiplying the estimates of epsilon by the degrees of freedom of both the numerator and denominator; determining the new probabilities associated to the new F tests. The epsilon values commonly vary from 0 to 1 (Littell et al., 1991), and those found in the present analysis (0.22 for G-G and 0.30 for H-F) indicate that some adjustment is necessary for the correction of the probability levels, taking the correlations between the measurements into account. Although the univariate tests are basically more powerful than their multivariate counterparts, the consistent rejection (P < 0.0001) of the sphericity test of the correlation matrix between repeated measurements indicates that these should be interpreted with some caution. In this situation the realization of the multivariate and also of the adjusted univariate tests is justified. It must however be emphasized that all these procedures used revealed highly significant differences between the effects of years and cultivars x years. |
| Experimental accuracy; Final considerations. | Return To Table of Contents |
|---|
| Experimental accuracy: Besides the factors already commented on in the item Ideal plot size and number of replications, the analysis of covariance, although not common, is another strategy able to increase the experimental accuracy since it complements the local control. Its objective is the use of a given secondary variable as, for example, the stand or the yield of previous years, among others, to assist in the interpretation of a given variable of primary importance. Pearce & Thom, 1951, employed the analysis of covariance in cacao for the accumulated fruit yield of two previous years as a co-variable. Further details of the procedure of the covariance analysis can be seen in Pimentel Gomes, 1990. The co-variable, when measured in the same experiment, is denominated calibration (Martinez, 1989). It allows the comparison of plants’ behaviour in the presence of the treatments with its behaviour in their absence and works best in single plant plots (a mono-tree plot). In practice this means the temporary planting of a standard variety, the most uniform possible, to furnish data on the environmental variation. Lockwood, 1977, for example, used 6-monthly recorded data of the diameter of calibrator tree trunks as a co-variable in variance analyses. In this case, the calibrator trees - cacao trees of the group Amelonated - known to be homogeneous - were eliminated two and a half years after planting. Nevertheless, there are still problems and difficulties in the execution and the interpretation of the covariance analysis, which must be overcome to make it a routine technique. The choice of the calibrator trees, for example, is a difficult task. When put in competition with hybrid progenies, the calibrators can, if they are very uniform and homogeneous, be virtually annihilated by the great vigour of the hybrids. The accuracy of field experiments can be further improved by the adjustment of the plot data in relation to the data of the neighbouring plots (Pearce & Moore, 1976; Lockwood, 1980; Lotodé & Lachenaud, 1988; Paulin et al., 1993 and Lachenaud & Oliver, 1998), known as Papadakis’ method (Papadakis, 1984). For a given plot ‘K’ for example there will be eight possible neighbouring plots forming a square, as shown below, and seven possible combinations of neighbours in relation to X (A, B, C, A+B, A+C, B+C, and A+B+C):-
The residue of each plot is calculated as the deviation of the observation in relation to the treatment mean. The co-variables from A to A+B+C of each plot are calculated based on the mean residues of the respective neighbouring plots. The analysis of covariance is processed iteratively, in search of the best combination of neighbouring plots, until the stabilization of the estimates of the treatment effects. It is known that the greater the number of trees or plots used in the adjustment the better the results. Pearce & Moore, 1976, for example, verified that the adjustment involving eight neighbouring trees was better than the one involving only four, in view of the greater reduction of the standard deviations associated with the differences among treatment means. It is also known that the efficiency of the adjustment through neighbouring plots is greater when used in conjunction with the calibration (Lockwood, 1980 and 1981). Analysis of covariance as described earlier requires specific software and excessive time to process the data analysis. A simplified version of this type of analysis was proposed by Lotodé & Lachenaud, 1988, and practiced by Paulin et al., 1993, and by Lachenaud & Oliver, 1998, with the objective of improving the experimental precision of comparative trials of hybrid yield, conducted with single tree plots. In this case, the co-variable (X) in use was the mean of the data evaluated in the eight contiguous plants, as below: Where  Yi = value of the trait in the neighbouring plants; Yhi = mean value of the trait in hybrid i; n = number of contigous plants. The adjustment (Y) between the variable Y and the co-variable X in case of a significant correlation between both is given by the regression: Y' = Y – b(-x - -xg) where Y= raw data obtained from the hybrid; b= coefficient of regression; -x = mean of the covariable regarding the given hybrid; -xg = general mean of the covariable, and the adjusted means of the hybrids are used in the variance analyses. Final considerations: Research with perennial species, as with cacao, generally demands large areas, intense labour use and highs costs. Data are obtained in large quantities and research results generated in the medium and long term. These peculiarities differentiate research with perennial species, which demands care and redoubled attention on behalf of the researcher and his team. Data collection for example should be realized with the same team whenever possible and monitored closely by the researcher. There are particularities in this operation that can only be carried out successfully with a well-trained team accustomed to repetitive activities. The measurement equipment (scales, calipers, tapes, etc) should also be adequate and carefully checked before each operation. The recording of the field data should be in printed or electronic formats and deserves maximum attention. It is worth remembering that the selection of a given genotype that is done based on its performance per se is relative and therefore mistakes in collecting data lead to serious drawbacks for the improvement programme. Unbalanced data are common in research with cacao. However, they only present difficulties if they are not adequately dealt with. If, on the one hand, this type of dated requires the use of more sophisticated statistics, on the other hand it creates possibilities to combine diverse approaches to solve the same problem. For experimental data of perennials the possibilities of analyses are various and the one chosen should be that which unites simplicity, robustness, and exequibility. It is worth emphasizing that the occurrence of exceptional, unexpected, and the previously mentioned aberrations is possible on certain occasions. However in the interest of the best solution for the problem studied such data should, a priori, not be omitted from the analyses. Their origin should be further investigated to decide only then on their inclusion in the analyses. In the search for a generalization of the results of the research programme, experiments on farms or in rural companies should be combined with that traditionally conducted at experimental stations. Mult-isite and multi-temporal trials ensure the representativity of regional research. Additionally, long-term research should receive greater attention of statistics in the next years. Objectively, it should offer appropriate new analytical approaches to optimize the data treatment and make a maximum of information possible with respect to the problem under investigation. | ||||||||||||
| Chapter 12. The contributions of breeding. L.A.S. Dias. | Return To Table of Contents |
|---|
| Contents: Introduction; Breeding programmes; Trinidad’s programme, Pre-breeding and Use of recurrent selection; Traditional cultivars; Modern cultivars; Varietal differentiation; Breeding progress and Final Considerations. Summary: An account on the first initiatives of genetic improvement of cacao in the world and in Brazil is given. The improvement programme of Trinidad is re-visited and considered the most successful for having been developed in a permanent manner with a clear definition of the base population and of the objectives, for its long-term design and for making use of the advantages of cloning. Many clones of the series TSH and TSA produced by the programme and carriers of genes for witches’ broom resistance are presently recommended for the renewal of the disease-susceptible cacao plantations in Brazil. The work of locating genetic resources is emphasized and its continuity in more scientific terms defended. Pre-breeding is given special attention and should grow in importance in the next years. The traditional and modern cultivars that are part of the pool of variability are presented. Emphasis is placed on the great challenge that the real improvement is to produce dry beans of high quality in early cultivars with high productivity and temporal stability with resistance to a certain locally important disease or pest and with a low final cost. These objectives can be attained by the use of recurrent selection. A procedure for varietal differentiation that deserves breeders attention is presented now that the recommendation of clones is gaining global confidence. Lastly, the efficiency of the Brazilian cacao improvement programme is presented and discussed. |
| Introduction; Breeding programmes; Trinidad’s programme, Pre-breeding and Use of recurrent selection. | Return To Table of Contents |
|---|
| Introduction: The practice of cacao improvement in an organized and systematic form was implemented in Trinidad in the early 1930s after the bases for an efficient programme had already been established. On this occasion, first studies on fruiting, vegetative propagation, and the inheritance of pigmentation and yield components were realized. The presence of self-incompatibility in cacao of the upper Amazon was also detected and the first selection criteria for superior cacaos defined (Cheesman, 1931 and 1934; Pound, 1932a, 1932b, 1932c, 1933 and 1934). The major objective of the first steps of improvement was the selection of material resistant to Crinipellis perniciosa, the causal agent of the disease known as witches’ broom (henceforth designated WB). In those days this disease represented a great concern since it had decimated the cacao cultivation of Ecuador in the 20s and was already posing a serious problem for Trinidad. Thus, the pioneer of the genetic improvement of cacao - F. J. Pound - undertook two botanical expeditions to the upper Amazon region in Peru and Ecuador, in search of trees with absence of WB symptoms (Pound, 1938 and 1943), in other words, cacao that is presumably resistant under natural conditions. Two and a half thousand seedlings which were grouped in five populations were the result of the 1938 expedition, coded according to their provenance in: IMC (Iquitos Marañon Calabacillo), MO (Morona), NA (Nanay), PA (Parinari) and SCA (Scavina). The importance of a discrimination of the accessions according to provenance became very clear at the time but unaccountably was not considered thereafter. In the 1943 expedition also in Peru, Pound collected 32 clones that were later coded as P (Pound). In Ecuador, Pound collected 320 fruits from 80 WB symptom-free cacao trees found on farms. A great part of these collected genotypes is maintained to date in germplasm banks in the form of clones. It is worth noting that pioneer improvement studies were also realized by C. J. J. van Hall in Indonesia (van Hall, 1930) and A. F. Posnette in Ghana (Posnette, 1943 and 1945). Likewise, numerous botanical expeditions followed after Pound’s up to the present to save wild germplasm. Nevertheless, the importance of Pound’s collection exceeds the others, due to its impact on cacao improvement in Trinidad and in many other producing countries. The subject genetic improvement of plant species is dealt with very appropriately in the books of Allard, 1960 and Borém, 1997. Improvement programmes: The recommendation of a cultivar or hybrid for planting must meet three important criteria (Toxopeus, 1985): agronomical, commercial and local. From the agronomical point of view, the interest is in vigorous growth, earliness of dry bean yield and in high productivity. Commercially, an improved cultivar should produce uniform dry beans with a uniform and mean weight of 1.07 grams or 93 beans per 100 grams, or in other words, have a minimum individual weight of 1 gram (IS = 1). The mean fat content and husk of the beans should be 55% and 11%, respectively. Another important commercial criterion to be met is the fruit index (FI, see also Chapter 6 for a discussion on the main yield components). This index refers to the number of fruits needed to produce one kilogramme of dry beans and depends basically on the varietal type or on the local cultivated population. It is worth remembering that the activities of harvest and breaking up the cacao fruits makes up about 40% of the total production cost, and that a lower FI reflects a lower expense with these activities. This high harvest cost leads to the following economical paradox: high productivity of dry cacao in years of low prices can mean an even lower profit or even a loss for the cacao farmer. An improved cultivar should furthermore present tolerance or resistance to pests and diseases of regional importance and have broad local adaptability. The high dry bean yield and the resistance to locally important diseases, the former more than the latter, continue to be the two principal objectives of the improvement programmes. However, in the last decade, the lower production cost became a third objective to be pursued, because the prices of a tonne of the product on the external market are permanently repressed (see Chapter 1). The maintenance of large stocks in great quantities by the consumer countries and the high offer of the product on the world market impede the practice of more remunerative prices for the product. The reduction of production costs is obviously not achieved only by planting an early high-yielding and disease-resistant cultivar. Management practices and high technology in planting are two aspects that must be linked to the best cultivars (see Chapter 1). The fourth objective that is being incorporated in improvement programmes is the enhancement of the final product quality (greater bean weight, lower percentage of testa and greater fat content, among others; see also Chapters 6, 10, and 13). Lastly, the high production must be expressed in a stable form for years on end (high temporal stability); the cacao farmer is interested in obtaining a high yield on the same area every year, ensuring the profit stability financial return (Dias et al., 1998, 1999a, 1999b and Carvalho, 1999). So, the great challenge of actual improvement is to produce dry beans of high quality in early, high yielding cultivars that are temporaly stable, with resistance to a certain disease or pest of local importance and the lowest possible final cost. The point is to make the new production technology available for the producer so that he can offer more of a high quality product every year at a more competitive final price. The new cultivars must concomitantly offer all these attributes. In this context the choice of the most adequate cultivars for planting is a step of primary importance. How cultivars will be made available for the cacao farmers may be a function of the technological level of the rural propriety (Dias, 1993b). For farms that use a low technological level, cultivars of open pollination or the traditional commercial hybrid mixture will certainly be the most adequate. On the other hand, farms that practice a high cultivation technology will be able to exploit the advantages of the cloned cultivars. It should be observed that the technological level is independent of the size of the rural propriety. That means to say that even the small cacao farmers can make use of the advantages of clone cultivars if they are ready to use high technology in the crop. Trinidad’s programme: As an example of a successful and noteworthy improvement programme one can cite the programme implemented in Trinidad from 1930 to 1980, conducted by W. E. Freeman (Gonsalves, 1984; Kennedy et al., 1987, and Spence, 1994). The focus of the programme was always to increase the income of the cacao farmer. To attain this objective selected cacaos had to present: high productivity through heavy seeds, WB and Ceratocystis fimbriata resistance, earliness of production, and high adaptation (Freeman, 1969). Since large seeds and not their greater number are responsible for heavy fruits, Freeman, 1969, established that cacao candidates for selection should present a minimal fresh seed weight of 3.5 g, corresponding to a dry weight of 1.4 g (correction factor wet weight/dry weight of 40%). Measuring the seed weight however is rarely done. A minimal fruit weight was therefore adopted that would meet the minimum required fresh bean weight. Considering the maximum number of seeds per fruit as being 50, only cacaos with a mean fruit weight of 175 g (50 x 3.5 g) were selected. The clones IMC 67 and P 18 were intensively used in the crossings for Ceratocystis control, known for presenting resistance to this pathogen in the field and laboratory. For WB control the starting point was given by the use of the resistant clone SCA 6 in crossings with Trinitarios (ICS 1 and ICS 95) to increase the bean size. The reason is that the SCA 6 fruits did not even attain a fresh seed weight of 85 g. As described above, the programme deliberately involved only four parent clones (IMC 67, SCA 6, P 18 and ICS 1) in successive generations of inter-crossings and clone selection of superior cacao (Montserin et al., 1957 and Gonsalves, 1984), based on the fruit index (FI = 8 or 9, that is, 8 to 9 fruits per kg of dry beans) and resistance to the diseases Ceratocystis and WB (Kennedy et al., 1987). The greater weight in selection in the first generations was given to the fruit index while the resistance tests were given greater emphasis in subsequent generations. This programme resulted in the universally known clones TSH (Trinidad Selected Hybrids - hybrids of Amazon with Trinitarios) and TSA (Trinidad Selection Amazon - selections in the Amazon population). The TSH and TSA clones are early, vigorous, with pink seeds and WB and Ceratocystis resistant. Actually, these clones are being recommended for planting in south Bahia in substitution of the WB susceptible cacao (Ahnert, 1997). The relation of parentage of the clones TSA and TSH can be seen in Montserin et al., 1957 and Gonsalves, 1984. It is important to point out that the hybrids TSH 565 x SIAL 169 and TSA 656 x ICS 8 (see also Table 12.2) that involve the Trinidad clones are among the five best hybrids recommended for planting in the north of Espirito Santo after eight years of evaluation of the yield and its components (Pereira et al., 1987). Trinidad simultaneously demonstrated the importance of cloning in the improvement programme of a perennial species and the need that this programme be permanent and designed for the long term. It was also demonstrated that it is perfectly possible to initiate an improvement programme based on a deliberately restricted genetic base as long as the clones that represent the base population have a high mean and high genetic variation. For a complete revision on the genetic improvement of cacao practiced worldwide see Toxopeus, 1969, Cope, 1976, Kennedy et al., 1987 and Warren & Kennedy, 1991. The improvement practiced in Brazil can be studied in Vello et al., 1969, Dias, 1995 and Pereira et al., 1999. At this point, it is necessary to point out that the improvement programme of Trinidad was not only concerned about using the existing variability in the supposed centre of origin by collecting accessions in the upper Amazon but it also tried to exploit the existing variability on its local plantations. Around 50 thousand cacao trees were selected on these farms based on information of the producers and on visual selection. This number was later reduced to a thousand trees after negative mass selection, which eliminated the cacao trees with production of small fruits and thick husks. The thousand candidate cacao trees for selection were then evaluated for the fruit index and presence of rose coloured seed. It was then believed that the selection of cacao with this seed colour would result in a product with good flavour characteristics. The most rigorous selection, applying a differential of 1:500, finally resulted in 100 cacao trees that were cloned and installed in clone tests for a performance evaluation from 1937 on. These clones were numbered sequentially from 1 to 100 and coded as of the series ICS (Imperial College Selection) (Pound, 1932b). According to Cope, 1952, these clones presented a dry bean production potential of over 1000 kg/ha/year and a fruit index of 18. This result meant a substantial production gain since the productivity mean of Trinidad in the 30s corresponded to 500 kg/ha/year with a fruit index of 25-30. However, the characteristic chocolate flavour of the country was lost in the clones, demonstrating that the selection for cacao with rose-coloured seeds was inefficient to improve the product quality (Toxopeus, 1969). The genotypes collected by Pound in the upper Amazon and referred to in the Introduction were quarantined in Barbados and then transferred as clones to Trinidad (Kennedy et al., 1987). In search of germplasm to initiate an improvement programme in Ghana, A. F. Posnette visited Trinidad and had access to these clones. He then began a study of incompatibility reactions with these upper Amazons and took the seeds of crossings involving the clone selections IMC 47, 60, 76, NA 31, 32, 33, 34, PA 7, and 35 to Ghana. In Ghana, Posnette confirmed heterosis in these crossings (Posnette, 1943) that was later confirmed in Trinidad (Montserin et al., 1957). However, the detailed report on the phenomenon in the literature was presented by Russel, 1952. The heterosis observed was naturally due to the following complementation of traits: the adaptation of the local Trinitario clones of the ICS series to the high production and WB resistance of the upper Amazon clones, notably the Scavina (SCA) clones. The most outstanding among the hybrid progenies of greatest heterosis was hybrid SCA 6 x ICS 1. The discovery of heterosis unravelled programmes of successful hybridisation in various cacao-producing countries. The crossing of upper Amazon or Trinitario clones, frequently used as male parents, and local clones as female parents, aiming at the synthesis of heterotic hybrids is practiced to this day in a large part of these countries. The classification of the racial groups in upper and lower Amazon Forasteiros, Trinitarios and Criollos is described in detail in Chapter 3. In Brazil, the clone series SIC (Selection ICB), SIAL (Selection IAL), and EEG which belong to the racial group Forasteiro of the lower Amazon are frequently used up to today as female hybrid parents. Pre-breeding: One of the most important stages in the improvement programmes of plants is the screening. The reason for this is simple: frequently there is a large distance, in terms of performance, between the genetic sources maintained in situ or ex situ and those that are manipulated in the final stages of the programmes. This means that the introduction of non-improved genetic sources causes a reduction of the improved population means and consequently affects the progress with selection and the chronogram of the improvement programme. It is imperative that the wild material be elevated to performance levels that are not very different from those of improved material and this role is left to pre-breeding. Pre-breeding is nowadays considered in the international cooperation programme denominated as the CFC/ICCO/IPGRI project. The programme aimed to identify and make the accession collection (CFC project collection) formed by clones of broad variability and specific traits of interest available for breeders until 2003. The objective is the enrichment of the local core collection of each one of the 10 member countries of the cooperation agreement (see also Chapter 13). The accessions to be worked on belong to the ICGT of Trinidad. The breeders involved in the project proposed a programme to be carried out in three stages (Sounigo et al., 2000). In a first stage, a sample of 500 clones will be formed (sample A). Clones that belong to the different populations carrying genes for black pod and witches’ broom resistance and favourable agronomical traits (seed and fruit index and butter content) will be selected for this purpose. The broad sampling of the populations is justified considering that the classical racial groups of cacao (Forasteiros of the upper and lower Amazon, Central-American and South American Criollos and Trinitarios) present complementarity in the traits of interest. Forasteiros for example, are more robust and productive than the Criollos even though the latter produce cacao of superior quality (see Chapter 3). Finally, the broad sampling increases the chances that the collection will contain alleles that determine the adaptation to different environments - a trait of extreme importance for the referred collection of multinational nature. Incidentally, the adequate number of clones for the collection was estimated at 250. The knowledge was used that, with a sample taken at random corresponding to 10% of the original clone population (ICGT has a stock of about 2500 clones), 80% of the original alleles can be captured with 95% of probability. However, the use of an efficient sampling strategy will substantially reduce the adequate collection size and this is the expectation of the breeders involved in the project. The terminology of populations in cacao is not very clear. It can mean a group of trees sampled during a collection expedition such as the populations PA and NA of upper Amazon Forasteiros collected by Pound, for example. It can also mean clones collected from commercial plantations such as the SIC and SIAL selected in south Bahia by the now extinct institutes ICB and IAL. The second stage will begin with the application of multitrait selection indexes on the clones evaluated of the sample A in the first stage in order to create sub-sample B composed of 250 clones. The theory on multitrait selection indexes and their application in programmes of reciprocal recurrent selection emphasized in eucalyptus was presented by Baradat et al., 1994. The application of these indexes in the selection of cacao aiming particularly at the production of dry beans and resistance to black pod was presented by Cilas et al., 1994. The diversity of sub-sample B will be examined with DNA markers - in this case, RAPD markers. The Shannon-Weaver index will be applied for the quantification of the molecular diversity. Up to the present, only 20% of the ICGT germplasm pool was evaluated for morphological traits and isoenzymes and about 10% for RAPD markers. In the third and last stage, the breeders involved in the project proposed to form the collection with 100 clones based on diversity estimators (number of alleles per locus, percentage of polymorphic loci, and heterozygosity observed) previously generated by isoenzymatic data. The collection would have the following relative representation of the racial groups: 55 Forasteiro clones; 25 Refratários* (‘National’ cacao from Ecuador, ‘Trinitario’ from Trinidad, and upper Amazon ‘Forasteiro’ from Peru) and 20 Trinitarios. It is worth highlighting that sample A would be formed by 275 Forasteiro, 125 Refratários and 100 Trinitarian clones. Sample B would be composed of 138 Forasteiro, 62 Refratários and 50 Trinitarian clones. Finally, the collection of the project CFC will be characterized by codominant molecular markers such as RFLP and microsatellites that allow the application of the diversity estimators. From 2002/03 on the quarantine and the distribution of the clones of the collection to the member countries of the international consortium will begin. * Editor’s note: In CIRAD's researchers’ opinion, the classification of cacao racial or morpho-geographical groups is not very clear. In view of this they proposed a fourth hybrid group, named Refratário, which is derived from three different origins: Nacional from Ecuador, Trinitario from Trinidad and upper Amazon Forastero from Peru. Use of recurrent selection: Recurrent selection, henceforth designated RS, was originally proposed in maize (Comstock et al., 1949). It can be understood as the selective scheme through subsequent or recurrent cycles, involving the acquisition, evaluation, selection and recombination of individuals and/or progenies. Its application has the principal objective of obtaining continually improved populations for one or various traits by the gradual and continuous increase of the frequency of its favourable alleles. RS gained credibility among perennial plant breeders in the last decade of the 20th century after its application in tropical perennial species had been proposed (cacao, robusta coffee, latex/rubber, oil palm, and coconut) (Baudouin et al., 1997), perennial species of vegetative propagation (Pinto, 2000), eucalyptus (Namkoong et al., 1988; Resende & Higa, 1990; Baradat et al., 1994 and Resende, 1999) and robusta coffee (Leroy et al., 1993 and 1994). The application of RS is being implemented specifically in cacao in Malaysia (Lockwood & Pang, 1993), in the Ivory Coast (Paulin, 1994 and Paulin & Eskes, 1995), and in Brazil (Pires et al., 1999). The fundamental question is: why is RS important and adequate for an application in cacao improvement? In the first place, the hybrid seed programmes in cacao are traditional and relatively successful (see item Varietal differentiation). It can therefore be inferred that there is genetic diversity and a heterotic pattern among the racial groups in cacao. The presence of a broad diversity and of heterotic groups is condition sine qua non for the exploration of the specific combining ability (SCA) and application of RS programmes. In the second place, cacao improvement is rediscovering cloning in some producing countries such as Brazil. It opens up the possibilities of selection for SCA based on the total genotypic values, capitalizing on genetic additive value, dominance deviation and espistaticeffects. Resende, 1999, demonstrated that of the two possible selection units associated with selection for SCA - means of crossings and individual genotypic value - the second is the one that provides superior genetic gain. The selection by the individual value allows a greater intensity of selection and uses information of the individual and of the mean of its family in the form of selection index. Lastly, the cacao improvement, which so far had lacked base populations with broad diversity to begin an RS programme will soon be able to count upon the collection of the CFC project. This screened collection will enrich the local collections that contain genes for broad local adaptation. The scheme of recurrent reciprocal selection (RRS) (Figure 12.1) is presented in the following, adapted for cacao based on the references cited in this item, especially that of Resende & Higa, 1990. The objective is to improve the inter-population hybrid. Figure 12.1 illustrates the phases of the process involved in the first selection cycle: - Phase 1: Formation of base populations. The choice of the two populations must take into consideration the favourable traits, the degree of complementarity of these traits and the broad adaptation of each one of them. The choice will then fall on a population of the lower Amazon Forasteiros (LA) and another formed by the CFC project collection (CP). Population LA unites good adaptation, good production and self-compatibility. Morphological and molecular data of numerous clone accessions involving the lower Amazon are available in the germplasm banks of Brazil and of other countries, which can help in the formation of the LA population. Population CP encompasses all racial groups and will be a source of genes for resistance to witches’ broom and black pod, bean size, high fat content, and other genes that condition cacao quality. Each population must be composed of at least 100 superior clones and maintained in clone banks. - Phase 2: The 100 LA population clones will be used as female parents in crossing with CP and reciprocally, creating 200 inter-population hybrid families. - Phase 3: The 200 hybrid families must be evaluated in replicated multisite experiments (see Chapter 11 on research). Based on the results of these experiments, 30 parent clones of each population with the greatest combining ability with the reciprocal population will be selected. The superior cacao trees of the experiments within each trial should be cloned, evaluated in clone tests, and the best ones vegetatively propagated for the commercial plantations. The 70 non-selected clones should be uprooted from the clone banks of each population. - Phase 4: Intra-population recombination of the 30 superior parental clones and acquisition of inter-population hybrid families to exploit the selection for SCA. The five best parents that produced the best intra-population hybrids should be crossed with the five superior parents of the reciprocal population, totalling 25 full-sib families. These inter-population hybrid families will allow a maximization of the use of heterosis and should be evaluated in multisite experiments with replications. The superior hybrid families will be vegetatively propagated on the commercial plantations. - Phase 5: In the clone banks of each population seeds from crossings among the 30 selected parental clones will be collected aiming at the recombination of the first cycle. The recombined seeds of each population will bring forth two new populations, from among which new superior cacaos will be selected for cloning to initiate the second selection cycle. From this point on, all five phases described are repeated. A greater offer of superior clones is expected as soon as the programme advances in the selection cycles. |
| Traditional cultivares; Modern cultivars. | Return To Table of Contents |
|---|
| Traditional cultivars: The oldest cacao cultivars are derived from (Theobroma cacao ssp. cação) Criollo populations that were originally domesticated on the Yucatan plain by Mayas and possibly also by their ancestors around 3 thousand years ago. DNA marker analyses of genotypes of recently discovered cacao in the sacred caverns of the Mayas known as “cenotes” (Gómez-Pompa et al., 1990) seem to indicate that one is dealing with remainders of the cultivars these peoples had planted (de la Cruz et al., 1995). The DNA analyses further demonstrated that these primitive cultivars of the Mayas are distinct from those grown throughout Mesoamerica and that these, in their turn, present great similarity to the cultivars planted in the north of South America (Whitkus et al., 1998). These studies thus confirm the manipulation of cacao by pre-Colombian peoples and suggest that the cultivars planted in Mesoamerica are derived from those that exist in northern South America. There are records of intensive commerce among the Amerindians of Mesoamerica and those of northern South America. It can therefore be assumed that these merchants took diverse plant specimens with them for commerce, including cacao cultivars (see Chapter 3). More recently the introduction and the interchange of cultivars among the Americas were practiced by missionaries. Still, the precise description of the distribution of cultivars is even more uncertain and speculative than that of the cacao populations. It is known, however, that the Criollo cultivars dominated the world market until the mid 18th century. These present ample modification in a group of traits which constitutes the domestication syndrome, whose principal alterations are the demand for a shorter fermentation time (only two days) and the production of high chocolate quality. Chapter 3 presents a new hypothesis for the origin, dispersion, and domestication of cacao. From Latin America: A complete revision on the principal cultivars of cacao would go beyond the scope of this Chapter. The reader interested in the subject should consult Mora Urpi, 1958; Soria, 1963 and 1966; Vello & Garcia, 1971 and Wood & Lass, 1985 to become familiar with the traditional cultivars. Modern cultivars used in the producing countries are described in Toxopeus, 1969 and 1985). Only the cultivars that had an impact on the world’s cacao cultivation, especially the Brazilian, are brought into focus here. This is certainly not an easy task because ever since the domestication in Mesoamerica the intensive process of introduction of multiple cultivars and the free exchange of genes among them give rise to inter-varietal hybridizations and compromises the varietal purity. Similarly many traditional cultivars disappeared by being substituted by more productive cultivars and hybrids. Nevertheless, an approach that centres on the stock of traditional cultivars allows a vision of the variability of cacao exploited by humankind over the last three thousand years of cultivation. Cultivar ‘Porcelana’: This is a South American Criollo with small, nearly melon-shaped fruits with shallow furrows, a soft smooth husk and predominantly pink colouration, shiny and blended with white. The fruit apex has five furrows as in the Criollos. The seeds are white or light pink, round or flattened. The tree is small with small leaves and slow growth. As with the Criollos this cultivar is classified as of extra fine quality on the international market. The principal cultivation region of ‘Porcelana’ lies on the south-western banks of the Maracaibo lake, Venezuela, in the sources of the rivers Catatumbo and Escalante (Soria, 1966). Cultivar ‘Matina’: This cultivar is planted on the Atlantic coast of Costa Rica, belongs to the Forasteiro population, and is of unknown origin. It seems to have been introduced from Brazil since it presents great similarity with the cultivar Comum (Soria, 1966) or with ‘Maranhão’ (Mora Urpi, 1958) of that country. It has medium-sized melon-shaped fruits, a thick green husk and purple seeds of small to intermediate size. It presents a good production potential and seems to be self-compatible and susceptible to the diseases black pod and sudden death. Cultivar ‘National’: Supposedly from the eastern flank of the Andes, in the Amazon forest of Ecuador, this cultivar has large melon-shaped, nearly oval fruits, with a slight narrowing at the base, a thick green wrinkled husk, and deep furrows. The seeds are intermediate to large and of purplish or reddish colour, although white seeds are commonly found among representatives of this cultivar. The trees are vigorous, tall and have large leaves. Based on this group of traits the cultivar was classified as belonging to the group of the Amazon Forasteiros (Soria, 1966). This same author points out that there are few plantations where National cacao can be found in a pure state, due mainly to its susceptibility to witches’ broom (WB). Also according to Soria, the majority of the cacao actually classified as from the Arriba region comes from ‘National’ and Trinitario hybrid plantations. The ‘National’ gave birth to cacao cultivation in Ecuador in the 18th century, beginning with few fruits dispersed on either side of the Equatorian Andes. The high quality of the National cacao gave the cacao produced in Ecuador an outstanding position until 1920-30 when the cacao cultivation of the country was decimated by WB. Actually, based on its morpho-anatomical and agronomical traits of seeds and flavour this cultivar was reclassified as belonging to the Criollo group (Enriquez, 1993). From Brazil: The traditional cacao cultivars of Brazil belong to the racial group Amazon Forasteiro, identified as ‘Comum’, ‘Pará’, and ‘Maranhão’ (Vello & Garcia, 1971). Nevertheless, Soria, 1963, judges that cacaos classified as pertaining to the Comum cultivar are really a mixture of forms or types that include ‘Pará’, ‘Maranhão’ and intermediate types, due to natural hybridizations. In fact, ‘Pará’ and ‘Maranhão’ present well-defined forms known as Parazinho, Maranhão Liso, and Maranhão Rugoso as a result of mutations, segregations, and even documented or undocumented introductions (Vello & Garcia, 1971). It is thought that ‘Comum’ had been the first cacao cultivar that was introduced in Bahia in 1746, brought from the state of Pará by Luiz Frederico Warneaux. The seeds, which were given to Antônio Dias Ribeiro, were planted on the fazenda Cubículo, on the right-hand bank of the river Pardo, in the municipality of Canavieiras. Seeds of the cultivar Maranhão were also brought from the state of Pará in 1874/76, this time by Steiger. According to this hypothesis, the cultivar Pará had its origin in Bahia, based on the natural hybridization that occurred between ‘Maranhão’ and ‘Comum’ (Bondar, 1920 and Miranda, 1947). Later Bondar, 1958, affirmed that the second cultivar introduced in Bahia would have been ‘Pará’ and that ‘Maranhão’ was result of the hybridization of Pará with Comum. It is more reasonable, however, to consider that ‘Pará’ and ‘Maranhão’ are really distinct cultivars as they do not present the typical intermediate hybrid traits which are common to this or that parent. The introduction would have occurred independently and at different times. The origin of the cultivar Pará would be from the state of the same name, given its great similarity in terms of rusticity, high yield and of “calabacillo” fruits such as wild cacaos found from that state to Venezuela and part of the Guianas. ‘Maranhão’ in turn would have come from the region of the river Marañon in the upper Amazon owing to its melon-shaped fruits, similar to those of the wild cacao existing in that region (Soria, 1963). Biometric data of these cultivars are shown in Table 12.1 and a short description, based on Soria, 1963 and Vello & Garcia, 1971, is presented in the following: Cultivar ‘Comum’: Its fruits are of average size, intermediate between ‘Maranhão’ and ‘Pará’, melon-shaped, with a smooth or slightly rough husk, thin and of green-whitish colour, with 10 pronounced furrows and a slight basal constriction. Its purple seeds are small, however larger than those of the cultivar Pará and with a dry weight above 1 gram. The productive potential of the cacao ‘Comum’ is good especially because it presents tolerance to black pod disease (see Table 12.1) and among the traditional cultivars it is the most planted. Curiously there is a great similarity between ‘Comum’ from Brazil and ‘Matina’ from Costa Rica, as mentioned before, indicating a possible common origin. Cultivar ‘Pará’: The small-sized fruits have a round shape, a fine smooth husk with superficial furrows and light green or whitish green colour. The seeds are purple and small, although this cultivar presents good yield (Table 12.1). Cultivar ‘Maranhão’: With its large fruits, larger than those of ‘Pará’ and ‘Comum’, this cultivar presents a high productive potential, particularly ‘Maranhão Rugoso’, due to the high production of fruits with heavy seeds (Table 12.1). Its fruits have a basal constriction, a thick smooth or wrinkled husk, but are suave to the touch. As with ‘Pará’, ‘Maranhão’ is rare in isolated plantings but found in mixed stands together with the cultivar Comum. The negative feature of this cultivar is its susceptibility to black pod (Table 12.1). The forms of ‘Maranhão’ and of ‘Pará’, which have a large number of traits in common, do not need to be described separately. Only the excessively small fruit size produced by the form Parazinho should be emphasized (Table 12.1). Modern cultivars: The cacao genotypes currently cultivated by the producing countries consist mostly of traditional cultivars and non-conventional hybrids, that is, hybrids synthesized by means of the crossing of heterozygotic clone parents, and to a lesser degree also by clones (Figure 12.2). In both cases the producers receive a mixture of biclonal hybrids or of clones for planting. This strategy guarantees the dry bean yield in case the mixture contained some self-incompatible and/or inter-incompatible hybrids or clones. Under these circumstances one can deduce the difficulty of classifying the cacao genotypes used in today’s stands. Although not really adequate and due to the absence of an appropriate terminology, the genotypes designed for planting will henceforth indistinctly be denominated ‘cultivar’. The short summary of the modern cultivars presented in the following is mainly based on Toxopeus, 1985. The hybrids TSH and TSA, referred to before in the item Improvement programmes, are used for planting as seed and also clone cultivars in Trinidad. Costa Rica produces and distributes seeds of the hybrid mixtures for planting in all the Caribbean and Central America. CATIE, in Turrialba, synthesizes such a mixture with over 35 bi-clonal hybrids whose parents are the upper Amazon clones such as SCA, P, and IMC in crossings with the local clones UF and EET of Ecuador, principally. Producing countries on the West African coast such as Ghana, Nigeria, Ivory Coast and Camaroon plant bi clonal hybrids that are basically derived from the crossing of local Amelonado selections with upper Amazon clones. It is worth remembering that the Amelonados of West Africa were derived from the lower Amazon clones introduced from Brazil. Southeast Asia, comprising Malaysia, Indonesia and Papua New Guinea represents the novelty in cacao improvement in the world. Similar to the other cacao-producing countries this region has been exploring the hybridization, principally after the introduction of melon-shaped fruits from Ghana, of Trinitarios and of upper Amazons. However, the cloning was intelligently incorporated into the hybridization programme so that genetic gains could be included at every stage of the process. Thanks to the permanent character of the research realized in that region, to the well-defined funding structure, the well designed crop expansion policies, and the exploration of cloning on a commercial scale, the region could assume a position among the world’s five greatest producers in the 90s, with Malaysia highlighted. Brazil also uses the hybrid mixture for planting, of which hybrids basically include Trinitarian and upper Amazon clones as male parents in crossings with clones of the series SIC, SIAL and EEG, selected in the local Amazon Forasteiro populations. With the arrival and advance of WB, hybrid SCA 6 x ICS 1, renamed Theobahia, came to be recommended as the resistant “cultivar” for seed planting from 1998 onwards. To classify a hybrid of heterozygotic clone parents as a “cultivar” is, in the least, imprudent. When one is dealing with a non-conventional hybrid this material segregates as much for WB resistance as for self-incompatibility genes, both inherited from the parental clone SCA 6. Consequently, plants with varied degrees of resistance and of self-incompatibility can be found in commercial plantations within this hybrid. For clone plantings, clones of the series TSH and TSA (TSH 516, TSH 565, TSH 774, TSH 1188, TSA 654, TSA 656, and TSA 792) are being recommended, developed by ICTA of Trinidad, as well as the clones EET 397 from Ecuador and CEPEC 42 from Brazil. This planting material does not of course express immunity towards the WB pathogen. On the contrary, this material is a carrier of genes for horizontal resistance and their planting in large-scale aims at a reduction of the inoculum source. The importation of WB resistant clones for planting in substitution to the cultivated susceptible cacao touches on two important aspects of the Brazilian cacao improvement programme. The first is that when the disease arrived the programme did not have and does not have to date locally developed resistant genotypes. The second relevant aspect is also that this importation of resistant clones should have a character of emergency only. The continuation of this emergency with the importation of this material will be a witness to the national programme’s failure to solve the question. It is worth emphasizing that, although there are nine clones recommended for planting in south Bahia, together they present a very narrow genetic base since they are basically derived from only four parental clones, as seen in the item ‘Improvement programmes’. Around a thousand clones selected by producers and technicians in the cacao plantations of the south Bahia region, codified as series WB, are in the phase of test and evaluation. This is really the most promising local action of improvement with chances of success if well managed. It is worth calling to mind once more that the seeds of the hybrid mixtures have been distributed and planted since the 70s. And, as mentioned, the constituent hybrids of these mixtures, in a minimum number of five, involve crossings with upper Amazon clones that are acknowledged bearers of WB resistance genes. As one is dealing with hybrids among non endogamic clones there is a strong segregation in this material for different traits, including those for resistance to diseases. Finally, it is estimated that the total cultivated area with such mixtures is 250 thousand hectares (Dias et al., 1998). In consequence, a great quantity of genetic variation for WB resistance and other traits must be available for exploitation in the cacao plantations of south Bahia. |
| Varietal differentiation. | Return To Table of Contents |
|---|
| Cacao has been cultivated in the region of the Low Rio Doce in the State of Espirito Santo since 1917 with seeds from the State of Bahia, Brazil. These cacaos are from the racial group Amazon Forasteiro, represented by the cultivars Comum, Pará, Maranhão, and the forms Parazinho, Maranhão Liso, and Maranhão Rugoso. The expansion of the cultivation in the following decades required genetic material of high yield potential. It was then decided to produce and distribute biclonal hybrids involving local selections (SIC, SIAL and EEG) and selections introduced from other producing countries such as SCA, UF, IMC, DR, ICS, among others. From 1966 seeds of the hybrid mixture produced in Bahia began to be distributed for planting to the cacao farmers of the state and in the north of Espirito Santo (Vello et al., 1969). Cultivars and forms of cacao have been differentiated principally according to the productivity, fruit traits such as form, weight and size, and the weight and number of seeds (Vello et al., 1969), based on univariate analyses (Vello & Garcia, 1971; see Table 12.1). However, the application of univariate tests for differentiation of such material has a limitation, since it is not possible to identify each cultivar or form by the assessment of a single trait (Vello & Garcia, 1971). In similar cases the utilization of multivariate tests that consider the different measurement scales of the traits and the existing correlations among these traits can be advantageous. The potential of multivariate analysis in the examination of the classification of the racial groups and for the prediction of the hybrid combinations’ performance in cacao has already been demonstrated (Dias, 1994 and Dias & Kageyama, 1997a and 1997b). In cacao (Dias et al., 1997) as well as in other perennial cultivations such as the example of robusta coffee (Fonseca, 1999), multivariate analyses have been useful in the evaluation of the phenetic divergence between clones. In the following, the procedure for an analysis of cultivar differentiation by the multivariate technique will be applied as proposed by Dias et al., 1996a. The data analysed refer to a comparative yield trial involving five cultivars: the commercial hybrid mixture recommended for planting in 1982, and four other cultivars: three local (‘Maranhão’, ‘Pará’ and ‘Parazinho’) and an introduced cultivars (‘ICS 1’ open pollination) that were evaluated during ten years (1984-93). The trial was installed in February 1982 at the Experimental station “Filogônio Peixoto”, Linhares, ES, in a 5 x 5 latin square design with plots of 196 plants (Dias et al., 1996a, 1996b, 1998, 1999a and 1999b). The yield components studied were the number of healthy fruits per plant (NFSP), the weight of fresh seeds per fruit (PSUF) in g/fruit and weight of fresh seeds (PSUH) in kg/ha. For differentiation of the cacao cultivars the criterion of varietal distinction proposed by Weatherup, 1994 was adopted. This criterion foresees the application of the generalized Mahalanobis, distance designated D2 (Mahalanobis, 1936) since this statistic considers the differences among two or more cultivars when evaluated by various traits, measured in different scales and with different degrees of correlation between them. The Mahalanobis distance between pairs of cultivars was calculated based on the analyses of variance and covariance applied to the traits i and j evaluated in the group of cultivars, based on the means of the years. The sums of squares (SQ) and of products (SP) referring to g cultivars and a years were obtained according to the scheme of analysis of variance and covariance below:  The standard deviation s within years and cultivars (residual standard deviation) for the traits i and j was given by:  - and permitted the calculation of the coefficient of residual correlation (rij) for the traits i and j within years and cultivars:  Thus, considering two cultivars X and Y evaluated by p traits one has the following expression for D2:  - where ~d is the vector of differences among the means of pairs of cultivars X and Y for all p traits, standardized by their respective standard deviations within years and cultivars and ~d is its transpose as in the expression below:  R-1 is the inverse matrix p x p of the residual coefficients of correlation of years and cultivars. The test of significance of D2 for declaration of distinction was conducted with the generalization of the statistic T2 of Hotelling, 1931, as below:  - where F corresponds to the F test with p and (ag-a-g-p+2) degrees of freedom, evaluated at the level of 1% probability. Cultivar ICS 1, a Trinitario, appeared as the most differentiated from the others, above all in relation to the local cultivars Maranhão, Pará, and Parazinho (Table 12.2). A study of multivariate genetic divergence among local cacao selections from Bahia (‘CEPEC 1’, ‘SIAL 169’, and ‘SIC 19’) and selections introduced from Costa Rica (‘CC 10’) and from Trinidad (‘ICS 1’) demonstrated that the latter is the cultivar of greatest relative distance (Dias & Kageyama, 1997a). Dias & Kageyama, 1995 also related a similar fact when they applied univariate analyses to the same data. The commercial hybrid mixture appeared distinct from ‘ICS 1’ and equally distinct and equidistant from ‘Pará’ and ‘Parazinho’ (also in Table 12.2). This result was coherent once the multidimensional analysis by D2 indicated that ‘Maranhão’ is at the same distance from ‘Pará’ and ‘Parazinho’ without however, distinguishing ‘Pará’ from ‘Parazinho’. The form derived from ‘Pará’, the so-called Parazinho, was therefore not confirmed by the multivariate analysis. The commercial hybrid mixture was at the same time close to ‘Maranhão’, indicating that this improved material has a high proportion of genes from this local cultivar, more than genes from ‘Pará’ and ‘Parazinho’. It is necessary to remember that the cacao hybrid has a local clone of known performance as female parent. Besides, the hybrid is distributed for planting to the cacao farmers in the form of hybrid combinations, in more than five combinations. Thus, the parent clones contribute with unequal gene proportions to the formation of the commercial hybrid mixture, to which the local clones contribute with the greater proportion. It is worth remembering that local clones such as SIC and SIAL used as hybrid parents were selected in the cacao populations formed by the cultivars ‘Maranhão’, ‘Pará’, ‘Parazinho’ and ‘Comum’. ‘Maranhão’ has a superior performance in relation to the other local cultivars (Dias et al., 1998) and the selection of the clone series SIC and SIAL may therefore have preferentially occurred in this latter cultivar, more than in the others. The fact explains the greater proximity of the hybrid mixture to ‘Maranhão’ and its consequent distance from the cultivars ‘Pará’ and ‘Parazinho’. The procedure for varietal differentiation aiming at the protection of the clone cultivar should gain relevance now that clone stands (see Chapter 9) are already a reality in the cacao region of south Bahia. |
| Breeding progress and Final considerations. | Return To Table of Contents |
|---|
| Progress with improvement: Cacao has been cultivated since the 18th century in the State of the Bahia, Brazil, where the environmental conditions favoured the development of the crop. Cacao farmers selected the best trees on their own plantations and from these they collected the best seeds for new stands. As much in Bahia as in Espirito Santo State where cultivation began in 1917 with seeds from Bahia, the traditional cultivars ‘Comum’, ‘Pará’, and ‘Maranhão’ and, highly probably the spontaneous hybrid types among them were planted. With the expansion of the cultivation prompted by PROCACAO in the 70s and in virtue of the high prices of the product on the internal and external markets, the demand for high yielding genotypes resulted in the development of biclonal hybrids involving the crossing of the locally selected clones SIC, SIAL and EEG with introduced clones SCA, UF, P, IMC, DR, among others. From 1966 on, seeds of the hybrid mixture produced in the orchards of Bahia were distributed to the cacao farmers of the State and to those in the north of Espirito Santo. However, the majority of the producers were not satisfied with the performance of these mixtures, compared to the traditional local cultivars. The hybrids constituting the mixture presented great competitive capacity, a vigorous trunk, but a low dry bean yield so that that many cacao farmers decided to continue planting their traditional cultivars. One of the reasons alleged for this inferior performance of the hybrid mixture was the heterozygosity of the parent clones, leading to a strong segregation in the progeny hybrid and a high proportion of unproductive trees in the stand. Studies showed that in 1 hectare cultivated with hybrids only 33% of the cacao trees produced effectively (Lainez, 1991). The cacao research programme of Linhares, in the State of the Espirito Santo, was the only one to face the question rationally, avoiding emotional discussions about the relative value of the hybrids and of the hybrid mixture when comparing their performance with that of the local cultivars (Pereira et al., 1987 and Dias et al., 1998). One of the most direct forms of measuring the efficiency of a genetic improvement programme is by a comparison of the performance of the products it brings forth with the previously existing products. In other words, it is necessary to compare the performance of the improved genotypes in terms of the traits for which they had been selected with the performance of the traditional cultivars. Note how the superiority of five hybrids selected among 81 others in Linhares is reproduced in Table 12.3. The selection was practiced after eight years of evaluation of four trials in complete randomized blocks with four replications and plots varying from 12 to 16 trees. These hybrids resulted basically from the crossing of clone selections of local origin EEG, SIC and SIAL with the clones SCA, IMC, PA, P, TSA, TSH, CC, UF, ICS, and DR introduced from diverse countries. The selected hybrids, if compared to the cultivar ‘Comum’ and the hybrid among the local selections (EEG 48 x EEG 64) produced about 20% more fruits with a 46% greater fresh bean yield. The greater production was more a result of the greater weight of beans per fruit (23%) than of the greater fruit yield (20%) (Table 12.3). This increase in bean weight is of fundamental importance for Linhares, because this municipality lies in a region that is subjected to an annual drought period, which can last up to five months and contributes to a drastic reduction of the expression of this trait. Another comparative trial of improved genotypes between traditional cultivars allowed the simulation of a condition much closer to that of the cacao farmer (Dias et al., 1998). The study provided interesting results, disproving beliefs assumed as true in cacao improvement. It evaluated the unimproved traditional cultivars ‘Maranhão’, ‘Pará’ and ‘Parazinho’ for performance and temporal stability against improved genotypes represented by the open pollinated ‘ICS 1’ and the hybrid mixture - the same distributed for planting to the cacao farmers. The design used was the 5 x 5 latin square with plots of 196 plants, totalling nearly 1 hectare of trial area (980 cacao trees) with each one of the five genotypes. The number of fruits per tree, the fresh seed weight per fruit and also per hectare were evaluated over 10 years (1984-93). The mean fresh production per hectare will be analyzed in detail in the following (Table 12.4). In relation to yield performance, the improved cultivars were superior to the traditional cultivars (Table 12.4). Even in the climatically most unfavourable year for production (1987/8) the hybrid mixture was superior to the other cultivars. In the meantime, the great productive superiority of the mixture and of the open pollinated ‘ICS 1’ was demonstrated in the climatically favourable years, such as 1991/2 and 1992/3, particularly in the latter year. The responsiveness of this improved material is another aspect that deserves mentioning. It is also worth pointing out the reasonable average productivity obtained with the traditional cultivars, especially in 1992/3. In respect to temporal stability (Table 12.5) of the bi-segmented linear regression, calculated by the method of Cruz et al., 1989, the same hybrid mixture revealed a good adaptability in unfavourable years, with the highest productivity in those years and in the total years, without however being responsive to better environmental conditions, opposite to the open pollinated ‘ICS 1’. The traditional local cultivars showed little productivity during the unfavourable years and little responsive to improved environmental conditions (Dias et al., 1998). ‘Pará’ and ‘Parazinho’ curiously had a very similar performance in the unfavourable as well as the favourable years, corroborating the thesis that these two cultivars are only one (Dias et al., 1996a). The association of high productivity and high stability in the improved materials suggests that the improvement practiced for increase in cultivar yield has also contributed to increase the temporal production stability. It furthermore suggests that these two traits cannot have an independent genetic control. The temporal stability is important for the cacao farmer who is interested in income stability over many years. In terms of improvement its importance is in allowing the recommendation of genotypes which are production-stable over the years (Dias et al., 1998 and Carvalho, 1999). Despite all the controversy about the agronomical value of the hybrid, heterosis played and will still play an important role in the improvement of cacao (Soria, 1961 and 1964; Morera & Mora, 1991; Dias, 1993a; Paulin et al., 1993; Dias, 1995; Dias & Kageyama, 1995 and 1997a; Dias et al., 1998 and 1999a and 1999b). In these articles the yield of the hybrids commonly surpasses 3000 kg/ha/year. The difficulty is that yield values such as these are frequently limited to experimental station conditions (Hunter, 1990). In the majority of the producing countries the production mean is around 500 kg/ha/year (see Chapter 1). There are many reasons for this difference in performance: the great heterogeneity of the plantations, inadequate management, the use of unimproved seeds, the advanced age of the plantations as the economical life of cacao trees is about 40 years, or the relatively small area actually cultivated with hybrids in relation to the total. There are no statistics that register the mean Brazilian productivity obtained with cacao over the estimated 250 thousand hectares planted with hybrids, considering a total planted area of about 650 thousand hectares. It is often possible to find all these factors acting together in a given plantation. Comparing the hybrid performance specifically under experimental conditions and in commercial stands, other limiting factors could be alleged, as has often been done: the occupation of marginal soils of low fertility, the presence of genetic incompatibility and the narrow genetic base. Particularly in Brazil, where the productivity mean had once attained 750 kg/ha/year in the 70s it is possible to admit that much of the low productivity of the commercial hybrids could be due to the distribution of planting material that did not correspond to that recommended based on yield trials (Dias et al., 1998). It is important to remember that in the 70s and 80s, the demand for hybrid seeds was greater than the production capacity of the seed orchards, and that there was a hurry to meet the demand of the cacao farmers (Alvim, 1973). Dias et al., 1998 justify this allegation based on the analysis of the comparative production trial conducted in Linhares, ES, as presented above. The seeds of the cultivars ‘Maranhão’, ‘Pará’, ‘Parazinho’ and the open pollinated ICS 1 used in the experiment were harvested randomly from representative trees of local plantations. In turn, the seeds of the hybrid mixture were obtained at the extension office of CEPLAC by taking random fruits from among those earmarked for distribution to the cacao farmers. As a result, the analysis of the performance and stability of production revealed the superiority in production of the improved cultivars (Hybrid mixture and open pollinated ‘ICS 1’). As one is dealing with the same hybrid mixture distributed to the cacao farmers that had been superior to the other cultivars under experimental conditions, why is it then that in conditions of commercial cultivation it appeared mediocre in production and became a target of producers’ criticism? The reason could not be problems of self-incompatibility or else the hybrid mixture would also be mediocre in production in the experiment, which was not the case. The occupation of marginal soils is also not a reason. Linhares offers adverse edaphic and climatic conditions for cacao and the soil where the experiment had been planted is chemically poor and physically sandy, so that after 16 field years the experiment was completely destroyed by drought. It was in this environment where soil and climate are equally unfavourable that the hybrid mixture appeared superior to the four other cultivars. Finally, the way the hybrids are synthesized suggests that the genetic base is not the problem. So the explanation seems to lie in the fact that hybrids and hybrids were distributed, that is: in the years in which the demand for seeds was compatible with the offer of the seed production orchards the distributed hybrid was the one that had been validated in productivity trials and therefore proven to be superior. But, in situations of seed scarceness, other genotypes were distributed when facing an excessive demand. This procedure condemned the hybrid mixture as untrustworthy. Final considerations: The easy and quick success obtained with the complementation of traits provided by hybridization between heterozygotic, presumably divergent clones was the reason that this was the only strategy of improvement adopted by the majority of the cacao-producing countries. In Brazil, this strategy persisted for nearly three decades. The hybrid mixture synthesized by means of a single selection cycle was distributed to the Brazilian producers. The cycle was initiated by the acquisition of hybrid progenies from among clones selected intuitively by the breeder. These progenies were then evaluated in replicated experiments and after some years of evaluation (a never well-defined period) the best were selected, applying selection at independent levels. Finally, the selected progenies were recombined in clone orchards and the hybrid seeds obtained were distributed to the cacao farmers for planting. New selection cycles were reinitiated, but now involving new clones in the hybridization. This rather unscientific strategy only acknowledged the advantages of applying the scheme of recurrent selection cycles in the beginning of the 90s, after the initiatives in this direction developed by Malaysia and Ivory Coast. Another frustrating aspect of the Brazilian cacao improvement was the complete ignorance of commercial cloning as a strategy of fixing genotypes at any stage of the programme, capitalizing on greater genetic gains and accelerating the improvement programme. It was the arrival of WB in the south Bahia region in 1989, the success achieved with cloning in Malaysia and Indonesia and the publication of papers (Dias, 1993b) that point out the advantages of the technique and the adoption of forest improvement as new paradigm for the improvement of cacao (Dias, 1993a) that definitively made the Brazilian programme incorporate cloning. A proof is that technicians of Veracel Celulose S.A. were asked to transfer know-how from eucalyptus cloning to cacao. Intriguing indeed is the fact that cloning had not been commercially used in the past when there was no WB in the region and therefore in a time when the technique would have had all possibilities of success. Now that the disease affects the entire regional cacao cultivation the technique is proclaimed as the solution for the problem. Henceforth it is expected that commercial cloning will be used with skill and intelligence so that it is not discredited in the producer’s eyes by misuse. It is still causing concern that the clones of the series TSH, TSA and EET imported from Trinidad and Ecuador are being recommended for commercial planting in south Bahia. It is necessary to remember that cloning in itself is last resort strategy. As a strategy of high value it must be linked to a population and hybrid improvement programme (see Chapters 9 and 13) in order to optimize it. The Brazilian cacao cultivation offers all the necessary conditions for the development of local clones: presence of great genetic variability in the regional plantations, existence of exhaustively tested elite hybrids and, above all, the acceptation of clone stands by the cacao farmers. Thus the question is to outline a long-term improvement programme and to work and work!. |
| Chapter 13. New approaches in breeding. L.A.S. Dias. | Return To Table of Contents |
|---|
| Contents: Introduction; Cacao breeding; Re-thinking breeding; Hybrid breeding, Hybridization strategy; Population improvement, Natural population model; Crop-pathogen coexistence, Breeding for total quality and Witches’ broom x shade; Asexual breeding, Micropropagation; Biotechnology and breeding; Final considerations and prospects. Summary: The genetic improvement of cacao already has enough accumulated knowledge to permit foreseeing its application on more scientific bases. The predictive techniques provided by quantitative genetics and biotechnology therefore need to be more intensely used, since one simple conventional selection cycle in cacao can last over a decade. In this Chapter, the information on the most recent research results in cacao improvement is presented and discussed. Fundamentally, the concomitant and harmonious practice of sexual improvement (for hybrid and population) and asexual improvement (cloning) with biotechnology are proposed, aiming at promoting significant advances in the production of cacao. Forest improvement is furthermore proposed as the new paradigm for the genetic cacao improvement. The present state of the art, the new concepts and the tendencies and future prospects for cacao improvement are also presented and discussed. Finally, some short, medium, and long-term actions are prescribed with which cacao improvement can advantageously face the scenario of low prices and of coexistence with the disease witches’ broom. |
| Introduction; Cacao breeding; Re-thinking breeding. | Return To Table of Contents |
|---|
| Introduction: The species Theobroma cacao L. is the only one among the 22 species of the genus Theobroma that is commercially exploited on a large scale. From its fermented, dried, roasted, and ground beans the cacao liquor (mass of cacao in non-alcoholic and semi liquid form) is extracted, and after pressing the cocoa butter and the powder are extracted - all of which are raw material for manufacturing chocolate. Around 70% of the world production comes from only five countries: Ivory Coast, Ghana, Indonesia, Brazil and Malaysia (see details in Chapter 1). In the first half of the 90s the mean world production was about 2.3 million tonnes/year in relation to a demand that was slightly higher. Today a tendency of decline is observed, especially when compared to the production of the second half of the 80s. The price on the international market dropped from US$ 2,500.00 in the second half of the 80s to US$ 1,000.00 per tonne in the first half of the 90s, and attained the all-time low of the last 27 years (about US$ 630.00) in February 2000. The Ivory Coast, Malaysia, Indonesia, and Brazil were the countries that contributed most to this disastrous price drop by the strong increase verified in their production. Furthermore, the combination of low prices, incidence of witches’ broom disease, loss of the producers’ capital, and ageing of the plantations provoked the drop in production of Brazilian cacao from 365 thousand tonnes obtained in the harvest 86/87 to 133 thousand in 99/00. Historically, the Brazilian cacao farmer sometimes acts as the villain and sometimes as the hero of the Atlantic rainforest (Muniz, 1994). As the villain he always occupied native forest areas as much at times of high prices, aiming at a production increase, as in moments of price drops to create additional profit through wood extraction and sale. The Brazilian cacao cultivation is now going through this last situation. The loss of capital due to the low product prices on the internal and external markets and to the profit loss through witches’ broom make the cacao farmer destroy the remaining Atlantic rainforest. Of the 2.3 million hectares of forest that existed in south Bahia only 55 years ago there are now only 130 thousand hectares left (Silva & Rezende, 2000). Only the stabilization of the product prices at reasonable levels will ensure the integrity and survival of the remainder. In this situation, the cacao farmer behaves as a hero, joining the cacao activity with the preservation of the Atlantic rainforest. Within this panorama, science can and must search for rational solutions in order to fulfill its social role. These solutions must pass through the capacity of genetic improvement in offering cultivars of high yield and the means to express their superiority. Bearing in mind that the species takes a long time to complete one generation of conventional improvement since one selection cycle can last 10 to 12 years, one word becomes essential. This keyword, in the holistic sense, is prediction - the capacity to foresee or estimate facts of interest related to cacao improvement (see also Chapter 6). Naturally there are numerous approaches by which the prediction can benefit improvement. For example, prediction can be applied: in the formation of the base population, the selection of cacaos with superior yield, to detect superior genotypes in the germplasm, to access the genetic diversity in natural populations, to access the genetic diversity in elite germplasm aiming at the synthesis of superior hybrids, for the recommendation of hybrids and clones for planting in a given region, in accessing the minimum harvest period for an evaluation of the genetic potential of a given genotype for yield and disease-resistance, in the evaluation of the temporal and geographical stability of a given cultivar recommended for planting, and in many other important stages of the improvement. Still with respect to the duration of the selection cycle in cacao it must be underscored that this will only be long (10 to 12 years) in the initial stage of the improvement programme; with the subsequent recurrent cycles it will become as fast and dynamic as in annual species (Ramalho, M.A.P., personal communication). It is worth emphasizing that not only a greater genetic gain per cycle should be pursued in cacao improvement but also the cost reduction by gain and the preservation of the broad genetic base, as much as possible. Besides, the improvement should aim at the sustainable production of cacao without advancing on the forest remnants. This means that the emphasis in cacao improvement should be the rehabilitation and substitution of decadent cacao plantations by improved cultivars and never the expansion of the planted area. This Chapter consists of an updated and amplified version of the article of Dias, 1998. The most recent research results in cacao and their implications for the genetic improvement are presented and discussed. The Chapter further aims at drawing up pathways and prospects for cacao cultivation, in particular that of Brazil. Cacao improvement: Historical records show the evolution of the improvement programmes of cacao implanted by the principal world research centres (Toxopeus, 1969; Cope, 1976; Kennedy et al., 1987; Warren & Kennedy, 1991) and of Brazil (Vello et al., 1969; Dias, 1995). Actually, there is no long-term programme going on in any tropical country of the western hemisphere dedicated to cacao improvement, according to Hunter 1990. This author states that the strategy frequently adopted in these programmes has been of the “quick and dirty” approach by which the exploration of heterosis occurs without a solid genetic base. The inefficiency of the improvement programmes in the producing countries was also attributed to the pressure for fast results and the ignorance of the inheritance mode of the traits of economical importance (Lockwood & Pang, 1994). These records demonstrate that consistent, broad and permanent cacao improvement programmes were and still are rarities in the world. For many specialists the establishment of an international cacao research centre is essential. This centre would act internationally coordinating and structuring a research programme in the long-term, raising funds for such research, promoting better training of scientists in the area and providing them with cutting-edge equipment to conduct their studies (Gotsch, 1997). Certainly the recent creation of INGENIC in the 90s - an international forum focussing on genetic improvement - improved the communication among scientists of the area and the coordination of research efforts. Also according to Gotsch (1997), for other specialists the success of long-term research depends more on a close collaboration among the different countries, especially regarding the interchange of cacao germplasm. Following this tendency, an international project was created in December 1997 for the use and conservation of cacao germplasm with the financial support of the institutes CFC/ICCO/IPGRI, under the coordination of A. B. Eskes, CIRAD, France. With an initial duration of five years (1998-2002), the project represents an internationally coordinated effort for selection and evaluation of cacao cultivars that are resistant to the principal pests and diseases, aiming at a sustainable low-cost production (Eskes et al., 1998). It unites research institutes of 10 countries (Brazil, Cameroon, Ivory Coast, Ecuador, Ghana, Malaysia, Nigeria, Papua New Guinea, Trinidad, and Venezuela) and has a budget of 10 million dollars. An initiative of the University of Reading, England with the support of the BCCCA in 1998 culminated in the most complete international data bank on cacao germplasm, the ICGD. Presented in CD-ROM, the data bank unites agronomical and morphological information of 13455 clones and their 13925 synonymies. It is distributed for free and can be requested by e-mail (icgd@reading.ac.uk). Each clone is presented with its synonymies and homonymies and contains additional information on its origin and history, pedigree, besides several relevant traits such as disease-resistances, among others. Another important data bank is the Bibliography of Cacao and Chocolate (Gerritsma, 2000; http://www.agro.wau.nl/agro/research/pps-pr05.htm), the result of efforts of the University of Wageningen together with five Dutch chocolate companies. This bank compiles references in agronomy, botany, cultivar protection, improvement, processing, manufacturing, biochemistry, consumption, and economy of cacao and chocolate globally. All in all there are 10.2 thousand references covering the entire 20th century and there are 1067 references dealing specifically with improvement (W. Gerritsma, personal communication). These computerized data banks make the information available instantaneously, democratize the knowledge, and intensify the interchange among researchers globally. On the one hand, the researchers gain since they are stimulated to compete and train themselves. The expectation is that these scientists can interact intensely and this seems to be already happening. On the other hand, the cacao farmers that can start counting on the collaboration of international groups of scientists who are aware of the questions linked to cacao cultivation and improvement gain as well. In general, there are few breeders and cacao research institutions in the world. Nevertheless, the number of articles published in improvement of the species has grown significantly in the last two decades (Figure 13.1) and it was not only the quantity of articles that increased; their quality also improved a lot. In these last decades one could observe a worldwide tendency of publishing in renowned journals with an editorial board and a good level of impact. This healthy fact represents an important change of attitude of cacao breeders. In Brazil, the situation of scarcity of breeders and institutions is not different. Only one institution, (CEPLAC), determines the path to be followed by cacao improvement. Still, in this as in many other fields of science, the diversity and the critical thinking are essential. Only the interaction of different ideas, programmes, breeders and institutions will ensure the success of cacao improvement. Re-thinking breeding: It is necessary to rethink the improvement of cacao (Dias, 1993a). It is known that the improvement of perennial species, including that of cacao, was always benefited by the strategies developed and used in the improvement of annual species. Evidence of this process of “anualization” of perennials are implied in the objectives outlined by Enriquez, 1985 and Kennedy et al., 1987, for cacao improvement in the 90s. However, many of the strategies of the annual species are not adequate or even viable for the use in perennials. Thus, it is necessary that “specific pathways” are followed, altering the paradigm of cacao improvement. Following the option of “annualizing” cacao would involve opening new research lines that emphasize: the reduction of the plant height by the introduction of mendelian control genes of this trait; the manipulation of the insertion angle of the branches so that the light absorption is maximized; the selection of genotypes with long-living leaves; the use of hormones inducing early flowering to make the synthesis of endogamous lines viable; the selection of genotypes of high metabolic efficiency that would be responsive to marginal soils of low fertility and the implementation of plantings in high density under some type of mechanization, among many other possibilities. On the other hand, the option of searching a specific pathway for the improvement of the species involves the better understanding of aspects linked to the origin of cacao: - tree of the tropical rain forest, dwelling in the understory under intense shade (see Chapter 4 on ecology of natural populations). In this case, forest improvement becomes the new paradigm to be followed. Thus, efforts should be concentrated, among other aspects, on the study of the interaction of the populations of the species with light. The possibility of such an interaction was first mentioned in the elegant work of Purseglove, 1968, and recently treated by Hadley & Yapp, 1993. Preliminary data show that cacao, when cultivated under full sunlight, presents signs of photo-inhibition stress, which can reduce the production in the absence of adequate irrigation and fertilization (Serrano & Biehl, 1999). It is necessary to concentrate efforts on the selection of genotypes which express genetic variation in response to the different light qualities, making differentiated improvement lines for a continuum, from little light (shade) to full sunlight possible, and also on the population improvement in order to ensure synthesis of the best hybrids. Under these circumstances, the maximization of genetic gains should not be pursued, thus avoiding the excessive narrowing of the base genetic. The challenge is to combine reasonable genetic gains with balanced cultivation thus minimizing the incidence of pests and diseases (Dias, 1993a). These two approaches discussed can be combined without losing sight of the base referential and, eventually, with the objective of exploring the possibilities thoroughly. |
| Hybrid breeding, Hybridization strategy; Population improvement, Natural population model. | Return To Table of Contents |
|---|
| Hybrid breeding: The history of the improvement of cacao can be divided into before and after the development of hybrid cacao (Dias, 1993a). Superior hybrids have significantly contributed to lift the cacao productivity, in particular that of Brazil (see also Chapter 12). In Brazil cacao hybrids have made possible the renewal of decadent areas with early, high yielding material that presents some disease-resistance. The type of hybrid selected for production in commercial scale is similar to the inter-varietal maize hybrid; a non-conventional, synthesized hybrid of non-endogamic clone parents. It is worth emphasizing that the importance of the non-additive over the additive genetic effects for yield components of cacao such as the number of fruits and the wet seed weight per plant was already demonstrated (Dias & Kageyama, 1995 and Resende & Dias, 2000; see also Chapter 6). The phenomenon “heterosis” for production, yield components, vigour, disease-resistance and earliness has been demonstrated based on the comparison of the hybrids with their self-fertilized parents (Russel, 1952; Atanda & Toxopeus, 1971; Atanda, 1972 & 1973 and Dias & Kageyama, 1995). Frequently, this is due to the actual structure of the cacao hybridation programme. Although the parents of the hybrids can be propagated clonally, the hybrids resulting from the crossings among these clones are planted using seeds, so that the comparisons are realized exclusively among seed-propagated genotypes. In consequence, the hybrids are installed in the competition trials together with their self-fertilized parents, all reproduced by seeds. Accordingly the term heterosis, as used in cacao, refers to the difference between the hybrid mean F1 and the mean or the greatest value of the parents S1. (Dias & Kageyama, 1997a) thus warn about the use of the concept of heterosis in cacao. Although the genes of the clones and of the self-fertilized parental clones S1 are the same, both heterosis and effects of combining ability can appear distorted if endogamic depression occurs. Dias & Kageyama, 1995, for example, found values of mean heterosis for yield components that are approximately half of those reported by Atanda, 1973 and Atanda & Toxopeus, 1971 and justify their results based on the use of self-compatible clone parents of the generation S1, whose relative homozygosis makes them less depressive. Beyond this, the way they are planned, seed production orchards require many years to be substituted when new hybrids are validated in the trials of competition. In spite of these difficulties, the need of using hybrid seeds of improved cultivars in the planting has been reinforced. The improved material has presented superior performance for traits such as weight of fresh seeds per hectare and per fruit when compared to the unimproved material (Dias et al., 1998, & 1999a). In the case of perennial crops, the benefits from the use of improved hybrid cultivars are durable and, particularly for cacao, such benefits are obtained without additional cost for the producer since the hybrid seeds are donated to the Brazilian cacao farmers by CEPEC. Hybridization strategy: The exploration of the genetic variation in the improvement programmes of cacao has been quite limited, even though the species presents expressive variability. The majority of these programmes conducted in the producing countries use practically the same germplasm. This could not of course be different since such programmes follow, basically, the same strategy of clone selection and hybridization among superior clones. Frequently, clone pairs are inter-crossed for the synthesis of hybrids. The evaluation and selection of the best hybrids is practiced in progeny tests. Subsequently, the superior hybrids are reconstituted in various orchards of commercial seed production, usually biclonal orchards. In the Ivory Coast for example, after the evaluation of 434 hybrids, 12 were selected and distributed to the producers for planting from 1975 on, in the form of seeds produced in biclonal orchards (Paulin et al., 1993). Besides prolongating the selective cycle, this strategy requires area, time and an excessive amount of work. An alternative can be the recombination of the hybrids selected in the actual progeny tests. For this purpose it would be enough to eliminate the hybrids with inferior performance, making the recombination among the best plants of the best hybrids possible, such as a synthetic variety that could be destined for commercial planting. With this, the effect of permanent environment would be capitalized on, maximizing the genetic gain. This simply means that the progeny test is transformed into a seed production orchard by seedlings, to use the terminology of forest improvement. One should bear in mind that the ample pollen pool and open pollination guarantee the equilibrium of the allelic frequencies and assure that the performance of the synthetic variety is the closest possible to the set of superior hybrids from which it is derived. To avoid contamination with alien pollen these tests should be adequately isolated or pollinated manually. In virtue of the peculiar requirements of cacao regarding shading, this alternative could only be applied in regions where the ecological conditions would allow it. It is important that such an alternative guarantees material for the next cycle in a programme of recurrent selection. Programmes of recurrent selection in cacao improvement are being conducted since 1990 in the Ivory Coast (Paulin & Eskes, 1995) and in Malaysia (Lockwood & Pang, 1993) and were also proposed for Brazil (Pires et al., 1999) and are already ongoing. The parents used in the formation of the base populations are chosen based on their genetic values per se and on the relative genetic diversity. The schemes of crossings generally used in the first cycles by these programmes are the diallels and the factorials, while the partial diallels are used in the subsequent cycles. The use of these designs allows the estimation of genetic parameters to guide the following actions. The designs also foresees the practice of selection between and within progenies and the use of progeny and clone tests in any cycles. In the Ivory Coast, the programme of recurrent selection arose from the partnership between CIRAD/France and IDEFOR, the local research institute. In the formation of the base populations crossings among parents of the groups lower Amazon and Trinitarios were used, the latter for being carriers of genes for high cacao quality. In Malaysia, the programme is conducted by the private initiative (Bal Plantations, Sabah) and uses parents of the upper Amazon group. At the moment, the programme of Malaysia seems paralyzed; the Bal Plantations were sold and the buyer company does not cultivate cacao any more (A. Figueira, personal communication). The Malaysian chocolate industry collapsed and the focus is now on the production of oil palm. From a cacao production of 240 thousand tonnes per annum, the country now produces only 40 thousand (Chok, 1999). The Ivory Coast programme actually receives support from the international project CFC/ICCO/IPGRI having been invigorated from 1997 on and the first pollinations of the second cycle are being done (A.B. Eskes, personal communication). As has been shown, cacao hybrids are of the non-conventional type (obtained from non-endogamic clone parents) with the genetic structure of double hybrids. If each alogamous S0 plant originates from the fusion of two gametes, the resultant hybrid S0 x S0 is similar to the double hybrid. Therefore, there is sufficient genetic variability in this material for a viable selection within hybrids. This way, the best cacao within the best hybrids would be maintained for recombination in the progeny tests. On the other hand, superior hybrids could be vegetatively propagated with the objective of producing improved commercial material. The strategy of cloning superior cacao hybrids is the one that provides the greatest genetic gain, as demonstrated in Chapter 6. These two strategies have been discussed (Dias, 1995). In another alternative, the selection simulated in the progeny test would already provide information for the installation of the clone orchards in the first years of evaluation. At the time of the definitive selection, the clones installed in the orchards would already be in production, and the superiority of those selected would be confirmed in the progeny tests. This strategy would shorten the selective cycle. In spite of all these strategies, what should be emphasized is that only by the improvement of the parent populations it will be possible to obtain the best hybrids. There are cases in which the crossings of cacao involving clones of the same racial group have expressed comparable heterotic vigour to that obtained with crossings between clones of different racial groups (Vello et al., 1972). Therefore, there is a certain degree of divergence among these clones, produced by the segregation phenomenon and/or restrictions to the gene flow during their life histories. Thus, the estimation of the multivariate phenetic distance among clones of the same racial group is justified (Dias et al., 1997) and can provide an important vision of the divergence in cacao. Also within this context, the hybridization in cacao is characterized by randomness, developing through trial and error. Still, Dias & Kageyama (1997a) quantified the degree of association between the multivariate genetic distance of a set of cultivars, evaluated during five consecutive years, and the mean and heterotic performance of their hybrids. Thus, it is possible to use this strategy to identify heterotic groups of clones, even before doing the crossings (see also Chapter 6). At the same time this divergence presents a relative stability degree over the years and can be calculated using data of a single climatically favourable year for the cultivation (Dias & Kageyama, 1997b). It was also demonstrated that the statistics of multivariate distances that use the matrixes of variances and residual covariances, as in the examples of the D2 distance of Mahalanobis and of the Euclidean distance obtained based on canonic variables, are the most robust and adequate for the prediction of heterotic hybrids (Dias & Kageyama, 1998a). In all these studies, the data used to calculate the genetic distance were obtained from a diallel analyzed in detail in terms of mean genetic components (Dias & Kageyama, 1995). Large fluctuations in the productivity of cacao have been observed in the producing countries in many regions and throughout the years of cultivation, making the recommendation of cultivars by the breeders difficult and reducing the return cacao farmers’ return. Analyses of adaptability and phenotypic stability were conducted in search of the most stable cultivars for given regions (Paulin et al., 1993 and Pinto et al., 1993). Lamentably, the geographic stability has been given greater attention by the breeders. However, data on temporal stability are more important for the cacao producer who is interested in income stability, enabling him to explore a more stable cultivar for years at the same local (Dias et al., 1998, 1999b and Carvalho, 1999). This information will be useful in the future when the heterogeneity of the years will be taken into consideration in the cultivar recommendation for a given region. Dias et al. (1998 and 1999b) compared three local unimproved cultivars - ‘Maranhão’, ‘Pará’, and ‘Parazinho’ - with two improved cultivars - open pollinated ‘ICS 1’ and commercial hybrid mixture. These two last cultivars were more adapted and stable during the ten years of production. The hybrid mixture and ‘ICS 1’ were also identified as the cultivars with the best performance (Dias et al., 1998and 1999a), suggesting that the genetic improvement practiced for productivity increase also contributes to a greater cultivar stability. This fact reinforces the importance of using improved cultivars for planting and the need to realize selection for temporally stable cultivars. While the local cultivars presented low productivity means during the whole 10 years, the hybrid mixture presented general stability, with the lowest relative interaction with the years (2.9%). The interest of cacao improvement research is in the early evaluation of the genetic potential of a given cultivar. This evaluation is based on the record of successive harvests throughout some years. However, literature lacks information on the duration of the minimum adequate harvest period to evaluate the genetic potential of a given genotype with accuracy. Analyses of correlation between the annual yields and the accumulated yield of the period show that the evaluation of cacao should be carried out over a minimum period of two years of consecutive harvests after the fifth year of planting (Soria & Esquivel, 1967) or even after the eighth year of planting (Dias et al., 1996b). The first situation occurred and seems to be more adequate for regions with favourable climate conditions and the second for regions under adverse climates, particularly those subjected to water stress. The alternative to the analysis of correlation to identify the minimum harvest period is the application of the coefficient of repeatability, as proposed by Dias & Kageyama, 1998b, making the selection in the field quicker, earlier and more efficient. This strategy certainly represents a reduction in time, work and costs. As mentioned before, a marked irregularity of yield and other traits can be observed in the biclonal cacao hybrid, a result of the heterozygosity in the parent clones of the hybrids. Lainez, 1991 for example showed that only 33% of the progenies of the hybrid Catongo x Pound 12 presented a high yield. Although the complete homozygosity promoted by the process of successive self-fertilizations is difficult, a certain degree of endogamy should be applied to the hybrid parents, principally to the self-incompatible parents. Hybrids of partially endogamic parents should be field-tested by comparison with the actual non-conventional hybrids (Dias, 1995). In many cases, cacao has expressed reduced endogamic depression (Carletto et al., 1977; Bartley, 1969 and 1970), which will certainly make viable the production of more uniform hybrids, with fixation of genes for disease-resistance and exploration of greater variability in the inter-population hybrids. It is worth remembering that the production of dihaploids in cacao is already feasible (Lanaud, 1987a and 1987b) and this can accelerate the acquisition of endogamic lines. For some unknown reason, this promising research line seems to have been abandoned, once later studies with dihaploids in cacao were no longer published. Population improvement: The discovery of hybrid vigour for earliness of production, disease resistance and fruit yield (Russel, 1952) in cacao obtained from crossings among divergent populations has given rise to large-scale hybridization programmes in some cacao-producing countries. However, many of these programmes are subjected to stagnation once population improvement is not practiced. As is known, population and hybrid improvement are complementary and not conflicting programmes (Paterniani, 1969). With continuous and gradually improved populations it is possible to produce better hybrids than the previous ones. The way they are, the actual programmes tend to stagnate since the probability to obtain a hybrid genotype that would be superior to the first by means of successive samplings in the same population is quite remote. Nevertheless, the strategy of population improvement requires the assimilation of a new concept. In the new concept, the tree becomes less important than the population. Natural population model: When focussing on the population instead of the tree, the cacao breeder has to return to the natural population model of the species. Knowledge about the form of occurrence, the population density and the magnitude and distribution of the genetic variation in these populations will be important for a biologically more stable cultivation. The natural cacao populations of Amazonia, for example, show expressive intra-population variability, occur in low density and, frequently, in association with rubber (Hevea Braziliensis), tonka bean or ‘cumaru’ (Coumarouna odorata), crabwood (Carapa guianensis), Brazil nut (Bertholletia excelsa), pau d'arco (Tabebuia avellanedae) and mahogany (Swietenia mahogani), besides various palm species (Almeida & Almeida, 1987; see also Chapter 4 for discussion on ecology of natural populations). Consequently, if the natural population model was followed, cacao should not be grown in monocultivation only but in association with other crops under moderate or low planting density, lower than that currently used and involve the combination of different superior genotypes. This proposition essentially involves the maintenance of the genetic diversity found in the natural populations of the species. For each intercrop, the ideal cacao stand in association with other crops will be a research target. Mixed stands are nothing new and have been investigated in the Philippines, Papua New Guinea, and Malaysia, involving the combination coconut-cacao (Leach et al., 1971). Recent experiences with an agroforest system in Rondônia have disseminated the combination rubber-cacao at a planting density of 404 trees/ha for both species (Almeida et al., 1995). In Bahia, it is reckoned that seven thousand hectares are cultivated under the cacao-rubber system, with a productivity mean of 600 kg/ha/year for both beans and dry rubber; however, very high productivity means of 1625 and 1200 kg/ha/year of dry rubber and dry beans, respectively, were verified in a selected propriety (Sena Gomes et al., 1996). A broad revision on agroforestry and multi-cropping systems can be found in Alvim, 1989. In support of the thesis that the use of the diversity is the recommendation of the mixture of hybrid cultivars, involving at least five hybrids, is that this is practiced in Trinidad and Brazil. However, the basic objective in these cases is to minimize problems with self-incompatibility presented by cacao, rather than to avoid the genetic uniformity. Only multi-cropping at a low density can make the coexistence of the plantation with the pathogens possible. The precious systems are only some examples of already tested multi-crops. Nevertheless, there are still some other combinations to be tested, such as cacao-coconut and cacao-oil palm. Despite cacao being purposely included in this proposal of intercrops, because of the familiarity of the producers in south Bahia with the plant, other agribusinesses should also be encouraged, aiming at the salvation of the regional economy. The cultivation of coconut, for example, occupies only 73 thousand ha in the region and generates an annual return of 200 million Reais for the State. Oil palm occupies 40 thousand ha, of which 30 thousand are with semi-spontaneous oil palm trees of the Atlantic rain forest, and generates an annual return of 14 million Reais with the production of 10 thousand tonnes per year of oil. It is worth noting that palm oil is the second most consumed in all the world. In view of the enormous potential represented by the rational exploration of the agribusiness in south Bahia, the Secretary of Agriculture of this State, in partnership with diverse State organs, banks and private companies released a development programme for the cultivation of oil palm and coconut. The programme allows for the allocation of resources for the financing and the re-uptake of the production, for the associated agribusiness and for the training of technicians involved with these agribusinesses. |
| Crop-pathogen coexistence, breeding for total quality, Witches’ broom x shade. | Return To Table of Contents |
|---|
| Crop-pathogen coexistence: The new cultivation strategy proposed in this Chapter, even if providing only reasonable productivity, can ensure the survival of the cacao crop in regions subjected to an intense incidence of pathogens, such as the principal cacao-producing region of Brazil, in the case south Bahia. This region was recently devastated by the fungus Crinipellis perniciosa, the causal agent of witches’ broom (henceforth denoted WB), the most destructive cacao disease. The proposed approach - the biologically more stable multi-crop with low cacao density - represents the possibility for a coexistence of the crop with the pathogen. This strategy impairs the dissemination of the fungus spores, usually effected by the wind. The diversification of agriculture, which generates additional return for small and intermediate cacao farmers, is another advantage provided by the multi-crop. The causal agent of WB that seems to have co-evolved with cacao in the Amazon basin, causes overdevelopment of cacao meristematic tissues. By physiological disorder at the level of the infection site, it leads to hypertrophy and to the uncontrolled growth of shoots, which assume the aspect of brooms. The fungus infects fruits and beans, making them commercially worthless. The production loss in infected regions can vary from 30 to 90%, especially where no cultural control is realized. Indirect production loss is also caused by the debilitation of the pathogen infected tree. The disease was first described in Suriname, in 1895, decimating the local cacao cultivation and spreading rapidly across the entire Amazon basin by means of the wind (see Chapters 7 and 8 for an in-depth discussion on WB). One century later, or more precisely in 1989, the disease was dispersed throughout the cacao cultivation in Bahia causing the ruin of the regional economy (Pereira et al., 1989). This dissemination over such a long distance must be accredited to humankind, whether purposely or not. The extensive continuous area cultivated with cacao (over 650 thousand hectares), to a large extent with susceptible traditional cultivars, and the regularity in the rain distribution favouring the growth flushes of the trees over the year were the appropriate conditions found by the fungus for its rapid dissemination. Strong social, economical and environmental changes are occurring in the cacao region of Bahia with the appearance of WB, according to a detailed study elaborated by Trevisan, 1999. A risk analysis of a WB-control project in the Brazilian Amazon using simulation indicated that productivity levels must be higher than 600 kg/ha/year to obtain an economical return, considering the stage of the technology and the price paid to the producer at the time of the study (Mendes & Almeida, 1999). Half of the WB-control involves the phytosanitary pruning and the use of resistant or tolerant material (IWBP, 1993). The systematic removal of the fungus-infected parts (brooms, fruits and flower cushions) by burning or covering of these parts with leaves, impedes the formation of basidiomata and constitutes the phytosanitary pruning. This practice has great efficiency in the disease control and is realized in Amazonia during the dry season and repeated at the beginning of the rainy period. In the cacao region of Bahia, where the greatest peaks of the disease occur in September and October, four removals per year are recommended: February, May, August, and November, according to the strategy of maintaining the infection at controllable levels (Light et al., 1997). Undoubtedly the best strategy for a long-term disease control would be the planting of resistant or tolerant cultivars. Nevertheless, cultivars with resistance or tolerance are still rare or inexistent. Besides, the cacao farmer does not have sufficient capital available for a replanting with such cultivars and even less to wait for the new planting to begin production. Side-grafting on unproductive adult cacao tree thus exerts an important role since it avoids replanting and does not interrupt the production. In Malaysia, the flowering of the graft on adult cacao occurs nine months after grafting, while harvesting begins one year after the pruning back of the rootstock (Yow & Lim, 1994). In Trinidad, the hybrids ‘SCA 6 x ICS 1’, ‘SCA 6 x ICS 6’, and ‘SCA 6 x ICS 60’ were mentioned by Purseglove, 1968 as having high WB-resistance and high yield. Hybrid ‘SCA 6 x ICS 6’ was particularly recommended for commercial planting in Trinidad in the 60s. From 1960 on, the emphasis shifted to the improvement of the clones TSH and TSA which showed field resistance (see Chapters 6 and 12). In respect to the low witches’ broom incidence exhibited by ‘SCA 6 x Trinitario’ hybrid progenies, this tolerance to the disease seems to decline with the advancing cacao tree age (Bartley, 1970 and Fonseca et al., 1999), when an increase of the intense and progressive infection and mortality levels of adult cacao occurs (Machado et al., 1992 and Fonseca et al., 1999). The occurrence of Ceratocystis wilt (Ceratocystis fimbriata) causing mortality of the SCA 6 hybrids should not be disregarded either. On the other hand, the clones ‘SCA 6’ and ‘SCA12’ present differential performance of tolerance among sites and among years within sites (Purdy & Schmidt, 1996). Moreover, the majority of the ‘SCA 6’ progenies present small seeds, a trait considered undesirable by the chocolate industry. Other difficulties that restrict the success in the improvement for WB-resistance are the existence of distinct pathogenic races of the fungus C. perniciosa (Laker, 1990), requiring the conduction of local improvement programmes for resistance and the differential response of cacao clones to the infection by these races, as observed by Wheeler & Mepsted, 1982. The important point that emerges in this discussion is that the breeder must recognize which are the principal pathogens that occur in his production zone and, from this starting point, practice the improvement for resistance to the specific pathogens. The selection for resistant genotypes however, is a complex task. For this reason, it is better to access resistance indirectly. In this case, resistance would be evaluated as protection against production loss, considering that the resistant genotypes would be those with the smallest loss. In this context, the following hybrids have presented superior performance in the Amazon region where the WB pathogen exerts a strong selection pressure: ‘SCA 6 x ICS 1’, ‘SCA 6 x ICS 6’, ‘SCA 6 x CAS 2’, ‘SCA 6 x CAS 3’, ‘SCA 6 x MOCORONGO’, ‘SCA 6 x MOCORONGO 1’, ‘SCA 6 x BE 9’, ‘SCA 6 x BE 10’, ‘SCA 6 x PA 150’, ‘PA 121 x MA 13’, ‘POUND 7 x SIC 644’, and ‘POUND 7 x SIAL 644’ (Machado et al., 1992). As can be observed, the majority of these superior hybrids have a Trinitario or upper Amazon parent clone crossed with Scavina (‘SCA 6’ or ‘SCA 12’) clones. The three hybrids listed last result from the crossing between non-Scavina upper Amazon clones and lower Amazon clones, combining the broad adaptability of the last with the WB-tolerance of the first. The importance of the wild populations denominated ELP, which were recently collected in French Guiana and appear totally free from the disease symptoms should also be considered for the WB-resistance improvement programme (Lachenaud et al., 1997). It is not possible to conclude for sure that this or that genotype is resistant and favourable for planting without rigorous experimentation in this respect. It is necessary to know how these resistant genotypes perform in different production environments in terms of their yield and quality of the final product. In this way the hybrid SCA 6 x ICS 1 for example, renamed and re-released in Brazil as the “cultivar” Theobahia, presented intense mortality by Ceratocystis wilt also known as ’mal-de-machete’ when cultivated in the south Bahia cacao region in a high planting density. CEPEC now considers removing it from the list of cultivars recommended for planting in the region; but suggests, however, the recommendation of two other seed “cultivars”: Theobahia 2 and Theobahia 3 (‘SCA 6 x ICS 6’ and ‘SCA 6 x ICS 8’, respectively), supposedly resistant to Ceratocystis wilt. Advancing in the discussion, when the programme chooses direct selection for resistant genotypes it will be necessary to know the minimum period of time sufficient to evaluate the resistance potential of the cultivar accurately. For Rondônia, this period was evaluated in three years of successive annual count records of the number of vegetative and reproductive brooms from the fifth year of planting on (Carneiro et al., 1997 and Carneiro et al., 2000). It is worth remembering that the WB-resistance can also be evaluated in seedlings. However, the resistance expressed in seedlings is not always directly related to resistance in adult plants, hence the importance of accessing it in adult cacao also. In Brazil, CEPLAC initiated its selection programme of WB-resistant genotypes in 1992. For a large-scale evaluation of genotypes under greenhouse conditions, the semi-automated system with basidiospore inoculation proposed by Frias & Purdy, 1995 is being used. Another selection programme of field-resistant cacao is also being conducted on hundreds of cacao farms. In the search for solutions to re-invigorate the cacao economy of Brazil the fund of support for cacao research FUNDECAU was created. In this topic, prediction, as defined in the beginning of this Chapter, also assumes an important role. The prediction of the chaotic scenario that would occur by the arrival and dissemination of a given pathogen of great destructive power is a task for genetic improvement. Unfortunately this role was not exerted by cacao improvement with respect to the entrance of WB in the southern region of Bahia. The measures of exclusion adopted by CEPLAC, aiming to impair the entrance of infected cacao propagules in south Bahia and coming from Amazonia were not sufficient to block the arrival of the disease. The pathogen, once introduced, found an extensive production area cultivated with susceptible genotypes (around 400 thousand hectares) and is at present disseminated over more than 90% of this area. Unhappily, the lesson of coffee improvement in respect to rust, when disease-resistant Brazilian genotypes began to be tested in 1950 in Portugal even before the rust settled in Brazil in 1970, was not assimilated by the cacao improvement. The advance of moniliasis, a disease of great destructive power caused by the fungus Moniliophthora roreri is also worrying. This disease is disseminated across Central America and the Peruvian, Colombian, and Equatorian Amazonia, already quite close to the Brazilian Amazonia and M. roreri spores have a viability of over two or three months. Improvement has to anticipate the arrival of this pathogen in the national territory too if it does not want to be succumb again. As encouragement, it is worth remembering the recent history of the successful upturn of the Ecuadorian cacao cultivation, whose cacao is classified as of fine flavour (aroma and taste). Ecuador suffered the impact of the arrival of WB and of moniliasis in the 20s. Being the market leader in the first and second decades of the 20th century, producing 50 thousand tonnes in 1915, the country noted a strong decline, recording a production of 10 thousand tonnes in the early 30s (Wood & Lass, 1985). The Ecuadorian harvest in 96/97, however, was close to 100 thousand tonnes. This story is an irrefutable proof that with science and planning it is possible to overcome even pathogens of great virulence such as WB. Breeding for total quality: Cacao constitutes an agricultural commodity produced exclusively by tropical countries in development and consumed, principally, by developed countries. Cacao production must therefore simultaneously meet the producers’ and the chocolate industry’s demands. Various components are used to evaluate it: number of fruits per plant or plot, weight of fresh seeds per plant or plot and per fruit, and weight of dry beans per plant or plot and per fruit (see also Chapter 6). Besides these yield components there are two others of great importance: the fruit index (FI), representing the number of fruits needed to generate 1 kg of fermented and dry beans, and the seed index (SI), represented by the weight of 100 fermented and dry beans. A lower FI value means a cost reduction for the producer in harvest, transport and breaking up of fruits. A higher SI is important for the industry inasmuch as large beans have a lower husk percentage and are more easily manipulated at grinding. For the Brazilian cacao cultivation conditions an ideal SI of over 100 grams and an FI under 40, with beans presenting between 10 and 12% husk and over 55% fat are assumed. Tests with 490 accessions of the CEPEC germplasm bank (Pires et al., 1998) indicated a broad variability of the fat content (45.4% in the accession ‘CC 57’ to 60.3% in ‘NA 312’) and negative correlation, although near zero, between this trait and the dry bean yield per plant. Thus, high yield genotypes with high fat content can be selected. The possibilities of progress with selection for hardness of the cocoa butter and composition of fatty acids and triglycerids were also demonstrated in these same accessions (Figueira et al., 1999). Cocoa butter is solid at ambient temperature, but melts at human body temperature (36.5oC) representing a fundamental property for chocolate manufacture. Its quality is determined by the hardness (melting resistance), which is intimately linked to the composition of fatty acids. At the moment, there is consensus among cacao breeders that the search for maximum yield is not enough. It is necessary to associate it to the factors pest and diseases resistance, good local and temporal adaptability and quality improvement. All these criteria together compose what can be called total quality and will be inevitably accompanied by the price dynamics. Therefore, the emphasis in the genetic improvement of cacao now seems to be concentrated on maximizing the production of beans of high quality. According to Clapperton, 1994, the genotypes used for planting and the adequate post-harvest processing determine the development of the cacao flavour of a given region and the current studies permit their manipulation and improvement. Efforts in this direction are being made by Brazil as well. Figueira et al., 1997 also demonstrated the genetic contribution to the flavour when observing similar flavour profiles in the same genotypes planted in Brazil and also in Malaysia, subjected to the same conditions of post-harvest processing. A revision on the principles, the practice and the factors that affect the fermentation of cacao is found in Dias, 1992. It is necessary to create a proper market niche, based on the differential product quality. The cacao ‘National’ of Ecuador for example, accounts for half of the global production of fine cacao (Crouzillat et al., 1998) and the prices offered on the market for this type of product have a surcharge (premium) varying from 10 to 40%, reaching up to 100% above the price offered for ordinary cacao (Wood, 1978). A successful recent experience in this sense in south Bahia occurred in the company Unacau. The planting of selected genetic material and the careful post-harvest processing guaranteed quality to the product of the company whose cacao was labeled “unacau cacao”. It is also necessary to pay attention to the modus operandi of the production that minimizes the damage to the environment, which would topically be designated as an ecologically correct production system. Actually, European consumers are aware of this question which, in the close future, will define the success or failure of exportations of the cacao-producing countries. Chadler, a chocolate manufacturer and Brazilian subsidiary of Barry-Callebaut is selecting cacao farmers in Bahia who are willing to produce organic cacao. To be admitted the producer’s farm will be inspected by a company specialized to emit certificates similar to those of the ISO. To produce organic cacao, agro-toxics and chemical fertilizers must not have been used on the farm for the three previous years. It is predicted that the price paid for organic cacao exceeds by up to 80% the price of the non-organic cacao. In summary, it is up to the breeder to interpret the diverse market expectations and convert these expectations into one or more products that may meet them. Witches’ broom x Shading: A combination of strategies to combat WB seems possible and desirable. After the expedition to the Amazon region from 1942-43, Pound, 1943 suggested resistant cacao cultivars in environmental conditions that could maximize the resistance expression. During the expedition, Pound observed that healthy or slightly fungus-infected cacao was found under full sunlight. On the other hand, observations made by Allen & Lass, 1983 in wild populations of Ecuador were in the opposite sense, that is, the majority of healthy or slightly infected trees were found under shade. Shading therefore seems to play an important role in the expression of resistance, although this is not completely elucidated. The planting of resistant or tolerant cultivars under moderate shading is actually recommended. Although there is still little experimental information, this point deserves attention. In theory, the permanent shading would contribute to reduce the infection levels since shaded cacao would reduce the peaks of leaf flushing, flowering and fruit setting and all these parts are susceptible to infection by the pathogen. This theory gives the measure of the dependence of the pathogen for meristematic tissues. Simultaneously, the shading would contribute to reduce the production of fungus spores, by maintaining the temperature and the humidity at the canopy level stable. The level of shading that would benefit cacao and reduce the pathogen incidence is however not known. Along the same line of combining strategies to combat WB, the biological control by means of the production and large-scale use of the fungus Trichoderma stromaticum, which is antagonistic to Crinipellis perniciosa, is already possible. Despite still depending on definitive tests, the biological fungicide, whose market name is Tricovab, is being recommended at a quantity of 2 kg/ha, in four annual applications during the rainy months from May to August, after the removal of the brooms. The application of this product reduces about 90% of the spore production in the brooms removed by the pruning. However, phytosanitary pruning is so far the most effective method of WB control. In isolated plantations, where this practice is realized with frequency, the production loss is maintained between 3 and 5% (Andebrhan et al., 1998). |
| Asexual breeding, Micropropagation. | Return To Table of Contents |
|---|
| Asexual breeding: The production of hybrid cacaos obtained from partial or completely endogamic parents should be implemented. Still, this problem cannot only be considered in relation to the parents. If the objective is to maintain the uniformity for the yield and other traits of the hybrids S0 x S0, the most promising alternative would be the cloning of the superior hybrids, as a form of fixing these genotypes and eliminate irregularities (Dias, 1995). In this context, Cheesman, 1929 already defended the use of vegetative propagation as a means to obtain uniformity in the cultivar. It is important to remember that vegetative propagation as a strategy of perpetuating superior genetic combinations is comparatively cheaper than the production of lines aiming at the synthesis of hybrids. The strategies of vegetative propagation can furthermore accelerate the improvement of perennial crops that present long juvenile periods, as it occurs with cacao, to shorten the selection cycles (see Chapter 9). When vegetative propagation was developed in the 30s (Pyke, 1933), the cacao committee of Trinidad and Tobago soon distributed a mixture of rooted clone cuttings to the cacao farmers in the early 40s, formed by at least four clones, being 30% of them self-compatible. Later, many of the areas planted with clones were replanted with hybrids multiplied by seeds (Purseglove, 1968). In the late 50s, Brazil also used vegetative propagation for new plantings and for the renewal of decadent plantations (Yungtay, 1958). Recently, multi-clonal stands were suggested for cacao farms of high technological level (Dias, 1993b). This author believes that the technological level of the cacao farm should be the criterion to opt for the planting of clones or of hybrids by seeds. Clones should be planted on highly technological farms. In the opposite situation, seed-propagated hybrids will be recommended. Inspired by the very successful models of silvicultural cloning, the suggestion foresees that multi-clonal stands will be composed of monoclonal blocks, in areas varying from five to ten hectares. To facilitate management the similarity in growth, development, and production of the clones will be taken into consideration for the formation of these blocks. The genetic incompatibility (see Chapters 4 and 6) should also receive special attention. Cheesman, 1938 warned of the danger of the stands in monoclonal blocks without exhaustive genetic compatibility tests. Also, the planting of pollinator clones should not be discarded to avoid problems of incompatibility. In the previous paragraph it was shown that in the 40s Trinidad distributed four clones with at least one of them being self-compatible. Considering the Brazilian reality and the evolution of the improvement programme, even if the ideal number of clones for recommendation is not available, something around 15 to 20 clones tested for each region seem adequate to ensure variability and avert problems of incompatibility. In the recommendation of clones for commercial plantations, the multivariate methodology of quantification of the genetic diversity by the D2 statistic of Mahalanobis applied by Dias & Kageyma, 1997a can also be used so that only clones with high performance and great relative divergence are recommended for each region. The D2 statistic was also used to process analyses of cultivar differentiation (Dias et al., 1996a). These authors demonstrated that ‘Para’, ‘Maranhão’, ‘ICS 1’, and the commercial hybrid mixture are distinct cultivars. However, also based on the D2 statistic, ‘Pará’ and ‘Parazinho’ were considered as one same cultivar (see Chapter 12). This strategy can be interesting for the protection of clone cultivars and for the interchange and transference of germplasm among institutions. It is worth underlining that errors of identification of clones planted concomitantly in Brazil and Malaysia were observed by Figueira, 1998, using RAPD analysis. Cacao cultivation represents a rare case among the cultivated perennial species inasmuch as the sexual reproduction had greater commercial importance than the vegetative propagation. Some of the reasons for the failure of the asexual propagation (Dias, 1993b) at the time were the following: i) demand for propagators of high cost; ii) need for trained labour; iii) very low and irregular growth of the clones, and iv) high cost of production and distribution of rooted clone cuttings. Theoretically, while 1 hectare can produce up to 1 million seeds per annum with 90 to 100% germination, the same area produces merely 50 thousand cuttings a year with only 50 to 60% viability. Nevertheless, the principal reason for the failure of cloning was the inferior performance of the clones compared to the seed hybrids. The clone selection practiced in Brazil was inefficient. Although this selection had only been practiced after two or more years of records of successive monthly harvests, no subsequent clone test was conducted. Thus, the phenotypic selection practiced was inefficient, since the dry bean yield of cacao is a complex trait that is formed by some yield components of quantitative inheritance and of low heritability (Soria, 1978; see Chapter 6). About a thousand buds of orthotropic branches per cacao tree can be obtained annually by conventional vegetative propagation (Glicenstein et al., 1990), maintaining the cacao trunk curved and forcing it to grow numerous suckers. Despite being promising in terms of the cost/benefit relationship, the referred methodology was not put into practice by the species’ improvement programmes as should have been expected. Nevertheless, it is an identical strategy to the one used by wild cacaos to spread out across the tropical forest (Almeida & Almeida, 1987), which demonstrates that many of the present phyto-technical difficulties can be solved by the simple observation of the model of natural populations. Conventional vegetative propagation has received special attention of CEPEC. With the support of CNPq, of other research institutes, and of FUNDECAU a logistic structure was created in Itabuna for a large-scale production of cacao clone seedlings from cuttings. The prevision is that around 7.2 million seedlings will be handed out annually for free to the cacao farmers when the full capacity of the nursery is reached. It is also planned that around 2.5 millions of cuttings will be distributed annually in the form of grafts (A.B. Pereira, personal communication; see also Chapter 9 on clone breeding). This nursery received the name “biofábrica” (bio-factory). Nurseries with a budget of nearly 1.5 million dollars, with a production capacity of 15 millions seedlings a year, and fully automated are common among companies of the forest sector. However, they continue to be nurseries, with the principal function of producing clone or seminal seedlings. The simple manipulation of this data of production capacity of clone seedlings produced by the nursery of CEPLAC would provide a prospect of the real situation of the cacao cultivation in Bahia. Assuming the conventional planting density of 1111 trees per hectare and a maximal annual production of 10 million clone seedlings (7.5 million seedlings plus 2.5 million grafts), 48 years would be needed to substitute 2/3 (433 thousand hectares) of the actually cultivated area with cacao. The importance of combining alternative technologies to make the process of renewal of the cacao cultivation feasible can thus be understood. Micropropagation, in this context, represents an important and supplementary alternative to conventional clone multiplication of genotypes and will be dealt with in the following item. Micropropagation: The introduction of the micropropagation technique in cacao occurred based on the in vitro propagation of axillary buds (Flynn et al., 1990). The procedure, according to the authors, would allow the production of thousands of plants arising from a simple bud within a few months, which was not confirmed in practice (A. Figueira, personal communication). The culture can also count on protocols for the regeneration of cacao via somatic embryogenesis, induced based on nuclear tissues and flower buds (Figueira & Janick, 1993; Lopez-Baez et al., 1993; Sondhal et al., 1993; Alemano et al., 1996 and Guiltinan et al., 1998). A lot of the pioneer work in somatic embryogenesis of cacao should be accredited to the Brazilian researchers A. Figueira and M. Sondhal, and a broad revision on the subject can be found in Figueira & Janick, 1994. The conversion of somatic embryos in plants (Figueira & Janick, 1993 and Lopez-Baez et al., 1993) was later improved (Lopez-Baez et al., 1999). These last authors studied the recovery rate of seedlings, the performance ex vitro, and the genetic field stability of the plants derived from the somatic embryos. The regenerated seedlings with at least two leaves were transferred to a greenhouse and later acclimatized. Of the 385 seedlings transferred, 220 (57%) survived. All the regenerated trees preserved the diploid of the original and presented an apparently normal development in field conditions; after 15 months they formed the jorquette and between 15 and 24 months they initiated flowering and fruit setting. Actually, the best results for conversion of somatic embryos into plants (A.O.V. Figueira and J.B. Teixeira, personal communication) are being obtained with the protocol developed by Guiltinan et al., 1998. Still, the frequency of regenerants is low and the work quite tedious. In conclusion, micropropagation now rendered possible, though yet to be optimized, will have an enormous importance if applied as routine technique, making the large-scale multiplication of selected genotypes possible. A recent project in Brazil, integrating efforts of CENARGEN/EMBRAPA and CEPEC/CEPLAC, proposes putting into practice supplementary alternatives to clone multiplication of WB-tolerant genotypes by means of tissue culture. The preliminary tests began in 1997, initially using the protocol of Lopez et al. (1993) and later that of Guiltinan et al. (1998), using flower cushions of five cacao trees. The latter worked better, obtaining a high index of embryogenic calli. Still, the conversion rate in plants continues to be very low and efforts are being made to increase it. In March 1999 the definitive work was initiated by collecting flower buds of eight elite CEPEC genotypes. A new collection of flower buttons was realized at the end of 1999. It is expected that the clones produced in CENARGEN will be in field tests of CEPEC by 2000 (J.B. Teixeira, personal communication). Another ongoing project in Brazil also works with similar objectives (A. Figueira, personal communication). What is basically proposed throughout this entire Chapter is the harmonious union of the sexual improvement with the asexual and with biotechnology. The sexual improvement should support the asexual programmes which together will ensure continuous genetic gains. Biotechnology will furnish the tools to monitor and guarantee such gains. In this sense, the geometric shape that best represents this harmonious union is the equilateral triangle (Figure 13.2). |
| Biotechnology and breeding. | Return To Table of Contents |
|---|
| The introduction of biotechnological techniques can change the present panorama of cacao improvement. Biotechnology can accelerate the generation of basic knowledge on biology and genetics of cacao. Consequently, improvement can become more scientific and precise, with time reduction for the development of high-yielding cultivars. Another benefit is the amplification of the genetic base of the cacao improvement programmes, reducing the crop’s vulnerability. Biotechnology can provide a quantitatively greater and qualitatively better evaluation of the germplasm available for the breeder (Dias, 1995). If this is confirmed, then a great effort should be made in the area of genomic mapping of cacao, in order to make the early selection for traits of interest feasible. The genomic mapping and its application to the genetic improvement programmes of agricultural and forest species are dealt with in-depth in Tanksley, 1993, Ferreira & Grattapaglia, 1996, Milach, 1998, and Chapter 10. For a complete approach to the state of affairs in quantitative molecular genetics considering the analysis of complex phenotypes the reader should address Paterson, 1998. The first genetic linkage map of cacao was developed by CIRAD, France (Lanaud et al., 1995), containing 193 RFLP, RAPD, and isoenzymatic markers. This map presents a total length of 759 cM, with a mean distance between two markers of 3.9 cM and was constructed using a population of 100 individuals derived from the crossing of the clones ‘UPA 402’ (Forasteiro of the upper Amazon) and ‘UF 676’ (Trinitario). On the other hand, CATIE in Costa Rica presented a preliminary linkage map of low density, with 18 RAPD markers linked in seven groups and constructed using 38 cacaos derived from the backcrossing [‘Catongo x (‘Catongo x Pound 12’)] (Osei et al., 1995). Later, the map of Osei was amplified (Phillips-Mora et al., 1995 and Crouzillat et al., 1996). Phillips-Mora et al., 1995, mapped 81 RAPD, 12 RFLP and two morphological markers by the analysis of 137 cacao trees. Its length was 887 cM, covering 10 linkage groups with a mean distance between markers of 10.5 cM. Finally, Crouzillat et al., 1996, amplified the markers to a total of 132, analyzing 131 trees of the same backcrossing. The map extended to 1068 cM, with mean distance between markers of 8.3 cM. To see advances in the genomic mapping the reader should refer to Chapter 10. Nevertheless, the strategies adopted to construct these two maps need to be examined in detail. The map of Osei used the conventional strategy of crossing among relatively contrasting parent clones to create the mapping population. In this case, ‘Catongo’ is a Forasteiro cultivar of the low Amazon, moderately susceptible to Phytophthora palmivora, with white seeds and self-compatible. ‘Pound 12’ is a Forasteiro of the upper Amazon, self-incompatible, moderately resistant to P. palmivora and has purple seeds. Besides, ‘Catongo’ was selected as a spontaneous mutant of white seeds in Bahia over 50 years ago, and ‘Pound 12’ was collected at the same time in the headwaters of the river Amazon. Considering that the map of Osei analysed one-year-old DNA of seedlings derived from backcrossing, around 15 years would be required to initiate the mapping, as described in Fritz et al., 1995. Even considering that the experiment mentioned had not originally been planned for mapping, the strategy of backcrossing would require at least four years to initiate the mapping if used in cacao. In contrast, mapping involving the crossing of two heterozygotic parents, such as done by Lanaud, would require only one year. Consequently, this strategy is more favourable as it demands less time and cost. Due to the fact that cacao improvement still aims at the parent clone, any strategy that permits the construction of linkage maps for the individual in particular deserves attention. Obviously, such maps should be constructed for elite clones. The mapping strategy used by Lanaud and previously denominated “pseudo-testcross” by Grattapaglia & Sederoff, 1994 therefore seems to be adequate for cacao. The pseudo-testcross sets up a false crossing test, once the true test is done to confirm the genotype of the individual F1. By this strategy, the crossing between genetically divergent individuals, possibly from distinct geographic origins, generates a 1:1 segregation of polymorphic markers in its progenies as in a true crossing test, since many markers will be heterozygotic in one parent and null in the other. Naturally the result is a linkage map for each tree, as for an elite cacao clone, for example. (See Chapter 10 dealing with genomic mapping). The versatility of this type of mapping allows dominant as well as codominant markers to be used. In this case the RAPD technique can be used for countries in development as it is more accessible. In these countries funds for research are scarce and financing is of vital importance (Dias, 1995). Also from the technical point of view, the RAPD markers have shown to be adequate for the study of the cacao genome since the majority of their loci segregate in a mendelian manner (Ronning et al., 1995). For one or few genes (oligogenes) controlling a trait such as resistance to a given disease, molecular markers linked to them can be identified quickly by the methodology of bulk segregant analysis (BSA) (Michelmore et al., 1991). The viability of the application of BSA aiming at the early selection of RAPD markers linked to genes of susceptibility to Verticillium dahliae in cacao was demonstrated by (Cascardo et al., 1995). The resistance to V. dahliae presented monogenic inheritance, conditioned by a recessive gene and the susceptibility by the dominant allele. For traits of low heritability controlled by polygenes, the identification of linked markers demands a more complex QTL analysis. In classic improvement, the selection is primarily based on the phenotype. In this context, the individual selection is practiced for traits of high heritability and the selection at family level used for traits of low heritability, aiming at increases in the accuracy of the selection process. However, molecular markers such as RFLP, RAPD, among others, when linked to mendelian genes or QTLs can make indirect selection based on the phenotype of such markers possible. This strategy was designated MAS. Markers that are genetically linked to traits of interest are chosen and trees with the particular combination of these markers are selected. Treated in this way, MAS seems to be a simple and easily workable strategy. The reality is not quite like that. Keeping the more recent studies in mind, one can verify that there are still great difficulties to be overcome before MAS will be effective. The presence of interactions of the QTLs by environments (localities, years and times of planting) (Melo, 2000) and by genetic background, generations and age (Grattapaglia, 1994 and Campinhos, 1996) limit the application of the strategy to very particular situations. It is worth emphasizing that the interaction QTLs by annual productions was also detected in cacao (Clement et al., 2000). Besides, to detect association between the marker and the trait, high precision is needed in the field experiments, which is not always achieved. Among other critical points of this strategy is the requirement for large samples, with reference to the mapping population and the number of markers (Beavis, 1998). As if these difficulties were not enough, the studies with simulation up to the moment demonstrate that the economic viability of MAS, in comparison to traditional phenotypic selection, is only concretized for traits with low heritability (Knapp, 1998; Bernardo, 1999). In spite of the limitations presented by MAS, the expectation is that the advance of genomic genetics causes a deep impact on the classic genetics, especially in relation to the understanding of the inheritance of quantitative traits. Until recently, for example, it was believed that such traits were controlled by great number of genes of small and equitative effects. The era of the QTLs, however, revealed that the quantitative traits can be explained by the great number of loci, although of unequal effects. It is furthermore not rare that few QTLs of large effects are verified which explain a great part of the genetic variation of the trait (Tanksley & McCouch, 1997). If this represents a general rule, the marker-assisted backcrossing will become the most promising MAS strategy. Besides, already in 1995, the analysis of a survey on DNA marker applications in the cultivar development - a survey answered by 55 breeders of diverse crops and from diverse countries - supported the introgression of oligogenes using MAS as the actual and future application of greatest consensus among these scientists (Lee, 1995). The cloning of the first QTL was recently announced (Frary et al., 2000), which was revealed to be a regulatory gene. Certainly, the study of this and other QTLs will open new prospects for the understanding of the functioning of the genes and of the genetic architecture of complex traits. Focussing on the resistance to black pod, a disease caused by Phytophthora palmivora and of presumed polygenic inheritance, Risterucci et al., 2000 detected eight resistance-linked QTLs, all localized in chromosome 5. The mapping population was synthesized based on the crossing of clone FIC 1 (susceptible) and hybrid ‘SCA 6’ (resistant) with an unknown clone by using AFLP (191) and microsatellite (11) markers. With the same objective, Fritz et al., 1995 and Phillips-Mora et al., 1995 also processed a QTL analysis, using RAPD and RFLP markers in the backcross ‘Catongo’ x (‘Catongo x Pound 12’). They then detected three QTLs, two of them in chromosome 5, involved in the resistance expression. Together, the three QTLs explained 43% of the phenotypic variance of the resistance. The observation of these results suggests that chromosome 5 plays an important role in the resistance to <>P. palmivora. This same backcross population was used for the localization of QTLs for early flowering, trunk diameter, height of the jorquette and number of ovules, using the genetic linkage map of Crouzillat et al., 1996. Analysis of variance and interval mapping revealed between 2 and 4 QTLs (P <0.01) involved in the expression of these traits, whose combined effects explained between 11.2% and 25.8% of the phenotypic variance. Actually AFLP and RAPD markers are being utilized in the construction of a genomic map of cacao aiming to identify regions associated with resistance to witches’ broom (Queiroz et al., 1998). The mapping population is composed of 82 individuals of the F2 generation derived from the hybrid SCA 6 x ICS 1. In the specific case of witches’ broom, heritability seems to be high, which in a way reduces the economical viability of the MAS application. Nevertheless, the expectation is that MAS realizes all its potential and increments the selection efficiency, above all for traits of low heritability. In forest improvement as well as in that of perennial species in which the tree has great individual value and a long cycle, the use of MAS has demonstrated some applicability. Markers linked to wood density and rust resistance in Pinus for example were successfully identified (Nelson et al., 1993). Another example refers-to the identification of a RAPD marker denominated AT09/900, at the Federal University of Viçosa, co-segregating in Eucalyptus grandis progenies with the gene that conditions the rust caused by Puccinia psidii (Junghans, 2000). This last research opens the immediate possibility of studies into function and sequencing of the gene, using it as model and in the future practice MAS for early resistance selection. Cacao, for having saturated genetic linkage maps (Lanaud et al., 1995 and Crouzillat et al., 1996) and a small genome (only twice the genome of Arabidopsis thaliana) and with little repeated DNA is, theoretically, enabled to localize mendelian genes and QTLs. For the first map, the physical distance by unity of genetic distance was estimated at 511 kpb/cM, making it favourable for the cloning of genes. Lanaud et al., 1995, announced that larger populations for mapping are being evaluated for the identification of QTLs for traits of disease resistance and bean quality. For diverse reasons, MAS will be particularly appropriate in the reduction of the number of backcross generations needed to introduce a simple resistance gene to some important disease into cultivated germplasm, as for example, the gene or oligogenes controlling WB and to evaluate accessions in germplasm collections. Another great contribution of biotechnology to improvement will be in the field of parent selection. Transgressive segregants or maximal heterosis for agronomical traits are expected when genetically divergent parents are used in the crossing. The molecular markers help in the discrimination of the parent, even before the realization of the crossings and this enormous potential is discussed in detail in Chapter 6. |
| Final considerations and prospects. | Return To Table of Contents |
|---|
| Cacao improvement has been practiced in a systematic manner since 1930. This is a relatively short period for a species with a secular productive cycle such as cacao. Still, this statement cannot and must not justify the long and frequent periods of stagnation observed in the majority of the cacao improvement programmes implemented around the world. The species presents expressive genetic variability and the practice of simple improvement methods can result in significant production gains. Since T. cacao is a species of relatively recent domestication, the simple use of mass selection for example can make expressive gains possible. The actual knowledge of the genetic and physiological mechanisms of cacao obviously favours the application of more refined improvement methods, such as recurrent selection (see Chapter 12). The arrival of WB in south Bahia - the principal cacao region of Brazil - caused a socio-economical and ecological disaster of never recorded dimensions. The fact, according to Light et al., 1997, will enter the history of world agriculture as one of the most dramatic examples of the impact of a plant disease on an extensive cultivation area. The disease, besides its capacity for devastation, brought with it the clear perception that the coexistence with the pathogen and the administration of the cacao farms on more enterprising bases should be two of the new trend-setting goals of the Brazilian cacao cultivation. For the post-WB cacao farmer the period of amateurism in farm administration has come to an end; the survival of the activity will depend on good management. For research, the message left by the disease was no less severe. The advent of WB represents the opportunity for cacao improvement to show competence. It is necessary to make it more dynamic, so that it could offer solutions to the producer and guarantees of survival for cacao cultivation. The question now is: how can cacao improvement face this new scenario advantageously? Even though the complete response would be difficult to work out, it is possible, in the actual state of the art of the improvement, to prescribe actions in the short, intermediate and long-terms. In the short term, measures of integrated control such as phytosanitary pruning, reduction of the of the plant height with individualization of the crown and the chemical and biological control and the recommendation of tolerant and productive cultivars most be offered. This is being done in south Bahia. Light et al. (1997) and Ahnert, 1997 report that resistant clones imported from Trinidad, specified for replanting or substitution of the crown of susceptible cacaos are already in use in Bahia (see Chapters 9 and 12). Care must be taken, however, not to distribute only one tolerant clone genotype, but a mixture, aiming at meeting the multi-clone stands discussed before. Not to lose sight of the varietal diversity is a supplementary strategy to genetic resistance in terms of conferring protection to the crop. Historically, the epidemics emerged in situations of absolute genetic homogeneity of the host. With a simple host genotype, the pathogen can easily adapt itself physiologically and genetically to it, quickly overcoming its defences and promoting infection. In situation of greater diversity of the host genotypes, the pathogen would require more time for adaptation. Concomitantly, efforts should be made in the formation of a base population to start a scheme of recurrent selection. All tools of quantitative genetics and biotechnology (Chapters 6 and 10) should be used to guide the formation of this population. Cloning via grafting and cuttings should comprise the diverse stages of the programme, be it to accelerate the selection cycles or be it to make the distribution of superior genotypes to the cacao farmers possible. In spite of all the effort reported before, the most important short-term control measure has not been adopted yet. This is the selection and subsequent commercial cloning of the best individuals within the best hybrids, as proposed by Dias, 1995. It should be remembered that CEPEC, for over 3 decades, has an extensive net of experimentation with hybrids that brought forth and still bring forth a great volume of data for yield and its components. These data are sufficient to make the selection of productive and witches’ broom-resistant cacao possible, not counting the fact that such genotypes present broad adaptation to the local conditions, since they are derived from crossings of introduced parent clones with local maternal clones. From this point of view, there is no doubt that the lack of adaptation of the clones imported from Trinidad will be one more obstacle in the attempt of unravelling the problems of the cacao cultivation in Bahia. In the medium term, micropropagation should realize all its potential making a large scale propagation of genotypes possible; selected as much at the level of plantation as those that come from controlled crossings, derived from the advanced generations of the recurrent selection programme. However, it is necessary to point out that the improvement of eucalyptus practiced in Brazil and considered the most advanced of the world, owes a great part of the its success to the use of clones propagated by mini and micro-cuttings, multiplied by conventional techniques. This does not mean to say that in the case of cacao one should only invest in conventional techniques of vegetative propagation. On the contrary, promising technologies such as micropropagation must be researched and optimized, however without interfering with research into other already acknowledged propagation techniques. By the way, the success of cloning in the revitalization of eucalyptus, devastated by canker in the 1970s must not be the justification for its use on a large scale in cacao cultivation. Cloning in itself is not a strategy of disease control; on the contrary, it narrows the genetic base of the cultivation even more and consequently amplifies its genetic vulnerability. Therefore, caution is advised. Nevertheless, cloning reveals to be the most agile strategy to perpetuate resistant genotypes in situations of extreme pathogen pressure, as is the case of Crinipellis in south Bahia. Note that cloning will be fully successful if the strategy is linked to a robust genetic improvement programme. In the long-term, derivations should be promoted in the improved populations of the most advanced selection cycles, aiming to meet the different market objectives (see Chapters 6 and 12). WB brought a new awareness to the community of Brazilian cacao producers. The crisis began with the appearance of WB and, aggravated by the drop in the remuneration of the producer, triggered an epidemic and boiling process of discussion. Regional newspapers, magazines and other means of communication report that the community is rethinking the cultivation and all its complications. The ABC for example organized a discussion list via Internet (ABCACAO-L) that today has become a debate forum on the past, the present and the future of Brazilian cacao cultivation. There are over 100 participants, among them cacao farmers, researchers, technicians and other segments involved in the cacao economy exchanging information and experiences. The statement of the cacao farmer Jamacy Costa de Souza, from the county of Wenceslau Guimarães, Bahia, in the list of discussion on cacao does not leave any doubt as to this new creation of awareness: “I am not afraid of WB and cacao cultivation has been profitable for me. I am optimistic about cacao and believe that we will once thank the crisis for changing some structural aspects of the region. We will get out of this crisis by trying to optimize the use of the resources, diversifying the crop, adopting the technological innovations more effectively and - though it may seem contradictory - with a greater ecological conscience ”. |
| Sign in |
