Potential for Joint Management and Multiple Use of Nyungwe Forest, Rwanda

Jean Bernard Ndayambaje

This study documents the local knowledge about uses and importance of trees to the Kaguru people in the Morogoro Region. Tanzania. The study was carried out as a first step towards the improved use and management of trees within the region. Quantitative ethnobotany, using the informant consensus method, was used to estimate use-values for 47 tree species found within a village and on adjacent farmlands. All species were found to have uses and 27 species are within the top 3 in at least one of 11 use-categories for meeting livelihood needs. Thus, a diversity of tree species plays an important role in people's livelihood. The most important tree species identified were the indigenous multipurpose trees Acacia tortilis (Forssk.) Hayne, Lannea schweinfurthii (Engl.) Engl, and Azanza garckeana (F. Hoffm.) Exell & Hillc. For more specific uses, the most important trees were exotic species such as Eucalyptus tereticornis Sm. for construction. Carica papaya L. for fruits and Azadirachta indica A. Juss. for medicine. The firewood and construction use-categories had the highest sum of all tree rankings. Relatively few of the 47 trees species were ranked as useful for fodder and only 3 ranked as very useful. Other tree species identified in neighbouring areas but absent in Majawanga, could be introduced.

Conservation of indigenous tree species is crucial for restoration of ecosystems and provision of livelihood support functions among rural communities. However, most tree planting initiatives have promoted exotic species, ignoring native species which populations have for long depended on. This study assessed the indigenous tree species conservation and Use Values in the Kyoga plains and the highland ranges Agro Ecological Zones of Uganda. A semi-structured questionnaire was used to capture primary data from a total of 422 household heads. Data were coded and entered in SPSS statistical package (Version 16). Descriptive statistics were generated to establish the distribution of indigenous trees on farms and tree management options employed by farmers. Chi-square test at 5% probability level was used to assess whether there were any difference in indigenous tree establishment, propagation and management options in the four study districts. The Use Value index technique was used to rank and prioritize 20 indigenous tree species. Most of the local communities in Eastern highland ranges prefer planting and retaining indigenous tree species in crop fields, while deliberate tree planting in the vicinity of family compounds (homestead) was predominant in the Kyoga plains. Indigenous tree species mode of establishment, propagation and management options significantly differed across the four study sites (P<0.05). The overall tree species Use Value Indices generally correspond to the number of uses of a particular tree species and vary between 2.4 and 0.2. Our study findings show rather a narrow range between the highest and least use value, implying that communities generally valued a particular tree species for at least more than one preferred use. Local communities prefer protecting naturally growing indigenous tree seedlings than planting. They facilitate tree regeneration when the benefits of their investment are guaranteed and the economic value such trees can provide to the household. Therefore, while evaluating indigenous tree species for relative importance and use, the end objectives of such evaluation must be critically examined in the light of the interest of the different tree resource user groups.

Conservation of indigenous tree species is crucial for restoration of ecosystems and provision of livelihood support functions among rural communities. However, most tree planting initiatives have promoted exotic species, ignoring native species which populations have for long depended on. This study assessed the indigenous tree species conservation and Use Values in the Kyoga plains and the highland ranges Agro Ecological Zones of Uganda. A semi-structured questionnaire was used to capture primary data from a total of 422 household heads. Data were coded and entered in SPSS statistical package (Version 16). Descriptive statistics were generated to establish the distribution of indigenous trees on farms and tree management options employed by farmers. Chi-square test at 5% probability level was used to assess whether there was any difference in indigenous tree establishment, propagation and management options in the four study districts. The Use Value index technique was used to rank and prioritize 20 indigenous tree species. Most of the local communities in Eastern highland ranges prefer planting and retaining indigenous tree species in crop fields, while deliberate tree planting in the vicinity of family compounds (homestead) was predominant in the Kyoga plains. Indigenous tree species mode of establishment, propagation and management options significantly differed across the four study sites (P<0.05). The overall tree species Use Value Indices generally correspond to the number of uses of a particular tree species and vary between 2.4 and 0.2. Our study findings show rather a narrow range between the highest and least use value, implying that communities generally valued a particular tree species for at least more than one preferred use. Local communities prefer protecting naturally growing indigenous tree seedlings than planting. They facilitate tree regeneration when the benefits of their investment are guaranteed and the economic value such trees can provide to the household. Therefore, while evaluating indigenous tree species for relative importance and use, the end objectives of such evaluation must be critically examined in light of the interest of different tree resource user groups.

POTENTIAL MULTIPLE FOR JOINT MANAGEMENT AND USE OF NYUNGWE FOREST, RWANDA BY Jean Damascene NDA YAMBAJE Thesis presented in partial fulfilment of the requirements for the degree of Master of Science in Forestry Sciences at the University of Stellenbosch, South Africa. Supervisor: Dr. Isla M. GRUNDY March 2002 University of Stellenbosch .' '1·, ..-- ~ ~ Stellenbosch University http://scholar.sun.ac.za DECLARATION I, the undersigned, hereby declare that the work contained in this thesis is my own original work and has not previously in its entirety or in part been submitted at any university for a degree. Signature Date Stellenbosch University II http://scholar.sun.ac.za ABSTRACT Nyungwe Forest in the South West of Rwanda is surrounded by a dense human population (350 people/km') which depends on subsistence agriculture and livestock. Previous reports on the Forest have shown that its survival is threatened by illegal use of its resources. Participatory rural appraisal was conducted in five Communes next to the Forest to study the community's perceptions, uses and interests with respect to access and management of the Forest. Respondents perceived the Forest as a source of 18 categories of products of which only 14 were said to be of value for the respondents'livelihoods. Land, timber and minerals (gold) were the priority resources preferred by more than 55% of the respondents. Other important resources included fodder/pastures, wooden goods, fuelwood and poles. Fourteen different species were most preferred for timber, 16 for poles, 45 for medicinal materials and six for wooden goods. However, the majority of these species were preferred for multiple uses as timber, poles and medicines. A forest resource assessment was carried out to identify the stock of the preferred tree species. Not all woody resources mentioned as being preferred were available, with some tree species not found at all in the Forest. In order to identify tree species that can be used sustainably, different criteria including abundance (density), dominance, diameter size-class distribution and regeneration, were used in combination. Of the 12 tree species recorded in the Forest and most preferred for timber, only three species were present in sufficient abundance and sizes to allow sustainable utilisation. Of the 12 pole species identified, six were not vulnerable to exploitation, and of the preferred medicinal tree species only five were considered to fit into this category. Low resource availability and the need to address the interests of adjacent communities necessitated management options which enable access to some resources and benefits the communities as well as conservation measures to protect the biodiversity. A range of joint forest management options is discussed in order to assess the feasibility of a collaborative approach in the management of Nyungwe Forest. Some recommendations are made with respect to access to resources, the use of substitutes and areas for future research. Stellenbosch University III http://scholar.sun.ac.za OPSOMMING Nyungwe Bos in die suidweste van Rwanda word omring deur 'n digte bevolking van 350 mense/km- wat van bestaanslandbou en lewende hawe afhanklik is. Vorige verslae het aangedui dat die voortbestaan van die Bos bedreig word deur die onwettige gebruik van sy hulpbronne. 'n Evaluering is in vyf gemeenskappe (Communes) langs die Bos gedoen om die gemeenskap se persepsies oor, gebruike van en belangstelling in die toegang tot die Bos, sowel as die bestuur daarvan te bestudeer. Die mense van hierdie plattelandse gebied is by die evaluering betrek. Respondente sien die Bos as 'n bron van produkte wat in 18 kategorieë ingedeel kan word, maar waarvan slegs 14 kategorieë van waarde is vir die respondente om 'n bestaan te kan maak. Meer as 55% van die respondente het grond, hout en minerale (goud) as die belangrikste hulpbronne aangedui. Ander belangrike hulpbronne sluit in veevoer of weivelde, houtartikels, brandhout en pale. Die respondente het 14 verskillende houtspesies verkies as timmerhout, 16 vir pale, 45 vir medisinale gebruike, en ses vir houtartikels. Die meerderheid van hierdie spesies is egter verkies vir veelvuldige gebruike soos timmerhout, pale en medisyne. 'n Evaluering van die hulpbronne in die bos is uitgevoer om te bepaal hoeveel bome van die gunstelingspesies daar in die bos is. Nie al die houthulpbronne wat deur die respondente verkies is, was beskikbaar nie, en sommige boomspesies is glad nie in die Bos gevind nie. Ten einde boomspesies te identifiseer wat vir lewensmiddele gebuik kan word, is 'n kombinasie van verskillende kriteria gebruik, insluitende die hoeveelheid bome (digtheid), dominansie, die verspreiding van die verskillende klasse deursneegroottes, en regenerasie. Van die 12 boomspesies in die Bos wat na aanleiding van hierdie evaluering opgeteken is en wat voorheen as gunstelingspesies vir timmerhout aangedui is, is daar slegs drie wat volop en groot genoeg is om vir lewensmiddele gebruik te word. Van die 12 spesies wat vir pale geïdentifiseer is, is ses teen oorontginning bestand, en van die gunsteling medisinale boomspesies val slegs vyf in laasgenoemde kategorie. Lae hulpbronbeskikbaarheid en die behoefte om die aangrensende gemeenskappe se belange aan te spreek, het bestuuropsies genoodsaak wat toegang tot sommige hulpbronne moontlik maak en wat die gemeenskappe bevoordeel, sowel as bewaringsmaatreëls ten einde die biodiversiteit te beskerm. 'n Verskeidenheid gesamentlike bosbestuuropsies is bespreek ten einde die uitvoerbaarheid van 'n benadering van samewerking in die bestuur van die Nyungwe Bos te ondersoek. Enkele aanbevelings is gemaak wat betref toegang tot hulpbronne, die gebruik van plaasvervangers, en studiegebiede vir verdere navorsing. Stellenbosch University IV http://scholar.sun.ac.za ACKNOWLEDGEMENTS I wish to express my sincere gratitude to the following people and organisations: • The Rwanda Government, through the Ministry of Education and the Rwanda Embassy in South Africa, for covering my scholarship. • The Rwanda Agricultural Research Institute, my employer, for supporting my application for a scholarship and for giving me permission to undertake my postgraduate studies. • Dr. Isla M. Grundy, my supervisor, for her guidance, comments and constructive ideas and interests shown throughout the study. • The Nyungwe Forest Conservation Project and its staff, for the provision of information on Nyungwe Forest and for access to their mini-library. • The leaders in the Communes surveyed, for facilitating household interviews and group discussions. • Mrs. A. Sadie, a lecturer at the University of Stellenbosch, for her assistance in data analysis and interpretation. • Prof. D. Ward, a lecturer at the same university, for the provision of ideas for site selection and sampling design. • Mr. J. P. Mwitwa, a Ph.D. student, for encouragement, assistance III data organisation and analysis. • Miss C. Riginos, for checking and correcting grammatical errors and for useful comments on the presentation of the results. • To all my fellow students in Community Forestry, for the exchange of ideas in data analysis and editing of the initial draft. • The Rwandan students in Western Cape, particularly those in the University of Stellenbosch, for their encouragement during my studies. • To Mr. D. Niyibaho, who took me around the Forest and showed patience during my fieldwork. • Finally but not least, I wish to thank my wife and sons who supported my absence during the course of my studies. Stellenbosch University v http://scholar.sun.ac.za DEDICATION This thesis is dedicated to my dear wife, Mrs Assumpta, to my sons Jean Yves, Jean Hubert, Ghislain and Joel Clint, and to my sister, Donatilla. Stellenbosch University VI http://scholar.sun.ac.za TABLE OF CONTENTS DECLARA.TI ON 1 11 ABSTRA.CT OPSOMMING III ACKNOWLEDGEMENTS IV DEDICATION V TABLE OF CONTENTS VI LIST OF TABLES X LIST 0 F FIGURES XI LIST OF APPENDICES 1 INTRODUCTION XIII 1 1.1 RWANDA: 1.2 POPULATION .........................................................................................•.......... 2 1.3 STATUS OF FORESTRY IN RWANDA 3 OVERVIEW 1 1.3.1 Forest plantations 3 1.3.2 Montane forests and savanna woodlands .4 1.4 FOREST PRODUCTS 6 1.4.1 Wood products 6 1.4.2 Non-wood products 7 1.4.2.1 Medicinal plants 7 1.4.2.2 Honey 8 1.4.2.3 Wildfruit andforestry seeds 8 1.4.2.4 Fodder plants and grasses 9 1.4.2.5 Wildjlowers 9 1.4.2.6 Essential oil plants 10 VII http://scholar.sun.ac.za Stellenbosch University 2 1.4.2.7 Mushrooms 10 1.4.2.8 Game meat 10 1.4.2.9 Ecotourism 10 1.4.2.10 Other goods 11 1.5 FOREST POLICY AND LEGISLATION ll 1.6 PROBLEM STATEMENT 13 1.7 JUSTIFICATION OF THE STUDY 14 1.8 AIM AND OBJECTIVES OF THE STUDY 18 1.8.1 Overall aim of the study 18 1.8.2 Objectives of the study 18 1.9 POTENTIAL BENEFITS 19 1.10 THESIS STRUCTURE 19 STUDY AREA AND METHODS 2.1 STUDY AREA ..••••.......•••••••......•...•..........••.......••••.•.......•.................•.•.•.....••••••.. 2.1.1 Description ofNyungwe Forest.. 21 21 21 2.1.1.1 Size and location ofNyungwe Forest 21 2.1.1.2 Geology and drainage 22 2.1.1.3 Flora 23 2.1.1.4 Fauna 24 2.1.1.5 Management history and conservation interests 25 2.1.1.6 Adjacent communities 28 2.2 RESEARCH METHODS 29 2.2.1 Selection of the study sites 30 2.2.2 Participatory Rural appraisal 30 2.2.2.1 Selection of households 30 2.2.2.2 Selection of groups and key informants 32 2.2.2.3 Data collection 32 2.2.2.4 Data processing and analysis 33 2.2.3 Forest resource assessment 34 2.2.3.1 Sampling design and plot sizes 34 2.2.3.2 Data Collection 36 2.2.3.3 Data analysis 38 Stellenbosch University VIII http://scholar.sun.ac.za 3 RESUL TS 3.1 PARTICIPATORY 42 RURAL APPRAISAL •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 42 3.1.1 Socio-demographic characteristics 42 3.1.2 Farm size and land ownership .44 3.1.3 Livelihood activities and sources of income .45 3.1.3.1 Crop farming 47 3.1.3.2 Other plant species grown 48 3.1.3.3 Livestock farming 49 3.1.3.4 Tree crops 51 3.1.4 Soil fertility status and management 54 3.1.5 Respondents' perception ofNyungwe Forest resources 54 3.1.6 Resource preferences and ranking 56 3.1.7 Tree species preferences and uses by the respondents in the study area 59 3.1.7.1 Timber tree species 59 3.1.7.2 Building poles 60 3.1.7.3 Medicinal plants 61 3.1.8 Uses of wild foods 63 3.1.9 Respondents' knowledge of wildlife species 64 3.1.10 Current resource access and use 66 3.1.11 Respondents' perception of conflicts and conflict resolution 66 3.1.12 Respondents' perception of past and present forest management.. 68 3.1.13 Respondents' views of appropriate institutions for forest management 3.2 FOREST RESOURCE ASSESSMENT 68 71 3.2.1 Identification of tree species 71 3.2.2 Tree species abundance and density analysis for the study area 72 3.2.3 Diameter size-class distribution of trees among and within the study sites 76 3.2.4 Presence of preferred tree species in the study area 77 3.2.5 Abundance of the most preferred tree species in the sample area 80 3.2.6 Dominance of the most preferred tree species 82 Stellenbosch University IX http://scholar.sun.ac.za 3.2.7 3.2.8 4 Size-class distribution and regeneration of most abundant preferred tree species 83 Quality of harvest able tree species for timber and poles 86 DISCUSSION 4.1 RESOURCE 88 PREFERENCES, USES AND AVAILABILITY IN NYUNGWE FOREST ••••••••••••••.••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••.•••••••••••••••• 88 4.1.1 Wood products 88 4.1.1.1 Timber 88 4.1.1.2 Building poles 90 4.1.1.3 Fue/wood and charcoal 90 4.1.1.4 Wooden items 91 4.1.1.5 Bean stakes 92 4.1.2 Non-wood products 93 4.1.2.1 Land 93 4.1.2.2 Gold mining 94 4.1.2.3 Wildlife and ecotourism 95 4.1.2.4 Medicinal plants 96 4.1.2.5 Fodder plants andpastures 98 4.1.2.6 Beekeeping. 99 4.1.2. 7 Wildfoods 100 4.2 CHARACTERISTICS 4.3 POTENTIAL OF THE PREFERRED TREE SPECIES •••••••••••••••••••••••••••••••I0l SUPPLY AND VULNERABILITY OF THE PREFERRED TREE SPECIES TO EXPLOITATION 4.4 5 POTENTIAL 104 FOR JOINT MANAGEMENT OF NYUNGWE FOREST ••••••••••••••••••107 CONCLUSION AND RECOMMENDATIONS 114 REFERENCES 119 APPENDICES 130 x http://scholar.sun.ac.za Stellenbosch University LIST OF TABLES Table 1.1. Distribution of forest plantations in Rwanda by tree species and ownership in 1990 4 Table 1.2. Protected forests' cover change in Rwanda between 1960 and 1999 6 Table 2.1. Latin and scientific names of some primates found in Nyungwe Forest. Table 3.1. Education level of the respondents in the study area 25 .44 Table 3.2. Important food and cash crops grown by the respondents in the study area 47 Table 3.3. Estimates of crop harvests by the respondents in the study area .48 Table 3.4. Numbers and distribution oflivestock types in the study area .49 Table 3.5. Causes of soil fertility decline listed by the respondents in the study area 54 Table 3.6. List of important resources for community's livelihoods in the study area 55 Table 3.7. Assessment of the availability of resources in Nyungwe Forest by groups in the study area Table 3.8. Ranking of resource categories in the study area 56 57 Table 3.9. Tree species most preferred for timber by the survey respondents around Nyungwe Forest 60 Table 3.10. Tree species whose stems are preferred for building poles in the study area Table 3.11. Plant species used for medicinal purposes in the survey area 61 62 Table 3.12. Wild food resources found in Nyungwe Forest, as mentioned by the respondents in the study area 63 Table 3.13. Respondents' knowledge of locally extinct wild animals in Nyungwe Forest. Table 3.14. Nature of conflicts listed by the respondents in the study area 66 67 Table 3.15. Importance of conflicts over resource use and access, as rated by the respondents in the study area Table 3.16. Proposed institutions for management by the respondents 67 68 Stellenbosch University XI http://scholar.sun.ac.za Table 3.17. Individual responsibility of the respondents for sustainable forest management in the study area 70 Table 3.18. Categories of trees used during the analysis 71 Table 3.19. Analysis of variance of the densities of trees among tree categories 72 Table 3.20. Occurrence of the 10 most dominant tree species, by site and tree category 73 Table 3.21. Diameter size-class distribution oflarge trees among the study sites 76 Table 3.22. Kolmogorov - Smimov tests comparing distributions of trees of different sizes in pairs of sites 76 Table 3.23. Chi-square analysis to investigate the distribution of species among the study sites and within diameter size classes 77 Table 3.24a. List of preferred timber and pole species and their occurrence categories in the study area 79 Table 3.24b. List of preferred medicinal species and their occurrence categories in the study area 79 Table 3.25a. Relative densities of the preferred timber tree species 80 Table 3.25b. Relative densities of the preferred pole tree species 81 Table 3.25c. Relative densities of the preferred medicinal tree species 81 Table 4.1 a. Classification of the preferred timber and pole species into 'vulnerable' and 'non-vulnerable' categories l05 Table 4.1 b. Classification of the preferred medicinal tree species into 'vulnerable' and 'non-vulnerable' categories l05 LIST OF FIGURES Figure 1.1. Geographical location of Rwanda Figure 2.1. Nyungwe Forest, showing altitudes inside the Forest. 2 22 Figure 2.2. Communes neighbouring Nyungwe Forest, showing the survey sample distribution 31 Figure 2.3. Subdivision and size of the sample units used for forest resource assessment, N yungwe Forest. 35 Figure 2.4. Different levels of stem quality used in the resource assessment, Nyungwe Forest. Figure 3.1. Age classes of the respondents in the studyarea 38 .43 Stellenbosch University XII http://scholar.sun.ac.za Figure 3.2. Family size of the respondents in the study area .43 Figure 3.3. Farm sizes of the respondents in the study area .44 Figure 3.4. Processes ofland acquisition by the respondents in the study area 45 Figure 3.5. Priority enterprises of the respondents in the study area 46 Figure 3.6. Sources of fodder during the dry season in the study area 50 Figure 3.7. Plantation areas owned by the respondents in the study area 51 Figure 3.8. Tree species planted by the respondents in the study area 51 Figure 3.9. Annual income from the sale of charcoal by the respondents in the study area 52 Figure 3.10. Annual income from the sale oflumber by the respondents in the study area 53 Figure 3.11. Annual income from the sale of building poles by the respondents in the study. area 53 Figure 3.12a. Stem densities per hectare of the 10 most dominant tree species represented by small trees 74 Figure 3.12b. Stem densities per hectare of the 10 most dominant tree species represented by medium trees 74 Figure 3.12c. Stem densities per hectare of the 10 most dominant tree species represented by large trees Figure 3.13. 74 Comparison of stem densities per hectare of the 10 most dominant species between tree categories, by study site Figure 3.14. Occurrence categories of preferred tree species in the sample area 75 78 Figure 3.15. Size-class distribution of the most abundant and dominant trees preferred for timber, poles and medicines 84 Figure 3.16. Relative abundance of seedlings and saplings of the abundant and dominant tree species across the five sites 85 Figure 3.17. Proportion of trees with stem suitable for timber. 87 Figure 3.18. Proportion of trees with stem suitable for building poles 87 XIII http://scholar.sun.ac.za Stellenbosch University LIST OF APPENDICES Appendix 1. Questionnaire for household interviews Appendix 2. List of on-farm medicinal plants mentioned by the respondents in the survey area. Appendix 3. Palatable tree and grass species mentioned by 5 or more respondents in the survey area Appendix 4. List of preferred timber tree species by the respondents in the study area. Appendix 5. List of preferred medicinal plants found in Nyungwe Forest, as mentioned by the respondents. Appendix 6. List of scientific and vernacular names of tree species recorded in all study sites. Appendix 7. Relative frequencies of trees recorded within the large tree category. Appendix 8. Relative frequencies of trees recorded within the medium tree category Appendix 9. Relative frequencies of small trees recorded within the small tree category. Appendix 10. Basal areas of tree species within the large tree category in the sample area. Stellenbosch University http://scholar.sun.ac.za 1 1.1 INTRODUCTION RWANDA: OVERVIEW Rwanda is a small (26 338 km') land-locked country in Central Africa between latitude 1°05' and 2°50' South and longitude 28°50' and 30°25' East. It borders Burundi in the South, Uganda in the North, Tanzania in the East and the Democratic Republic of Congo in the West (Figure 1.1). It is hilly with altitudes less than 1 500 m in the eastern plateau but rising to between 1 500 and 2000 m in the central plateau area and higher in the West and North. The following geographic regions are distinguished, moving from West to East (MINETERRE, 2000): • The north-south ridge of the watershed between the Congo and the Nile, with peaks rising to between 2 400 mand 3 000 m. The remnants of Rwanda's natural forests are found on this ridge; • The highlands that are an extension of those in Uganda and form the centre of the country between the Congo-Nile Crete in the West and the eastern zone, and decrease in altitude from West to East. Within the altitudinal range of 1500 mand 2000 m, the centre of the country is characterised by a mosaic of hills with rounded tops, separated by large swamps which cover almost half of the country's area. A ridge of five volcanoes in the far North-West dominates the Congo Nile Crete. From the West and moving eastwards are found the peaks of Kalisimbi (4507 m), Bushokoro (3 711 m), Sabyinyo (3 634 m), Gahinga (3 474 m) and Muhabura (4 127 m). • A large eastern zone of hills and swampy lakes, located from Kigali to the border with the Republic of Tanzania, occurs at altitudes between 1 000 and 1 500 m. A large section of this zone lies within Akagera National Park. Rwanda has a sub-tropical climate with an average annual temperature of 18°C. The average rainfall is 1250 mm, falling in two seasons separated by two dry seasons. 2 Stellenbosch University http://scholar.sun.ac.za UGANDA .~ BURUNDI '_Rusumo ._ 0 ... 0 • "''''' 30mi Figure 1.1. Geographical location of Rwanda 1.2 POPULATION The population of Rwanda was estimated to be 7.733 million in 2000 with a natural increase rate of 2.1 % and a density of 320 inhabitants per km2 (ADB, 2001). The rural population density of Rwanda is one of the highest in Africa (MINAGRI, 1998). In 1991, 94% of the population lived in rural areas and the remaining 6% in urban areas. A socio-demographic survey by MINECOFIN (1998) found that 91.1% of the active population is in the agricultural sector, 1.7% in industry and 7.2% in the services sector. At national level, the average size of the household is 5 persons and the farm size less than one hectare. The current demographic features lead to mixed farming where diversified cropping and livestock farming are practised. The farming techniques are exclusively manual (MINAGRI, 1998). About 70% of 1.3 million ha of land is cultivated, the majority of it on family farms. The latter are found on all types of land, including land of marginal quality for agricultural production (MINAGRI, 1998). A large number of these farms are not profitable and households live on the threshold of rural poverty. In 1997, the GDP per capita was US$ 240 (MINECOFIN, 1998), which was less than half of the most Stellenbosch University 3 http://scholar.sun.ac.za recent estimate for low income Sub Saharan Africa of US$ 490 (World Bank., 1998). The main constraints to production include a strong pressure on natural resources, the natural land infertility and the lack of use of agricultural inputs (MINAGRI, 1998). Land shortage is particularly prevalent in high altitude zones (MINAGRI, 1997). 1.3 ST ATUS OF FORESTRY IN RWANDA 1.3.1 Forest plantations Forest plantations have been established in Rwanda since 1920 with the introduction of fast growing exotic tree species, mainly Eucalyptus spp. (Eucalyptus tereticornis and E. saligna), for the purpose of protecting high mountainous areas from erosion and for the supply of fuelwood. been established. By 1967, about 20 000 ha of forest plantations had Wide scale plantation forestry started in 1967 with the commencement of the first donor-funded Forest Project, Projet Pilote Forestier (PPF). From 1967 to 1975, PPF established 5500 ha in Kibuye Prefecture. The year 1975 was a turning point for plantations with the start of a tree planting campaign and large scale development organisations (Habiyambere, projects, abroad. 1997). supported financially and technically by several The total area planted during that year was 25 500 ha In 1990, 247 500 ha of forest plantations were created (Habiyambere, 1997). Between 1993 and 1995, all silvicultural activities stopped due to war and genocide in Rwanda. In 1995, these activities started again, including establishment and rehabilitation of damaged plantations. The total forest plantations were estimated at 247500 ha in 1991 (Habiyambere, 1997) and at 256300 ha in 1997 (MINITERRE, 2000). Apart from Eucalyptus spp., other new exotic tree species in plantations Grevillea robusta, Pinus patula, Cupressus lusitanica, Callitris include spp., Acacia melanoxylon and Casuarina spp. These are established largely on government lands, as only a small proportion of plantation land is privately owned (FAO, 1999; Mihigo, 1999). A few local tree species were planted, particularly in buffer zones around indigenous forest reserves (Habiyambere, 1997). Beside plantation forestry, trees on farmer's fields are an integral part of the production systems. tree species with agroforestry potentialities Both local and exotic are mixed or inter-planted with food Stellenbosch University 4 http://scholar.sun.ac.za crops. They are normally established around the homesteads, along erosion control ditches, scattered on farms or used to demarcate farms. In 1989, trees on farms covered an estimated area of 87 200 ha (FAO, 1999). Effective inventory data for forest plantations are reported to be unavailable (FAO, 1999). Based on ownership, forest plantations in Rwanda are divided into governmental, communal and private categories. The area of forest plantations belonging to the second category was the largest in 1990 (Table 1.1). Table 1.1. Distribution ownership in 1990. of forest plantations in Rwanda by tree species and 70 Total Area (ha) 161 010 4360 5 32630 13 8 8720 10 21 550 9 6940 7 - - Il 220 5 3 2970 3 - - 4800 2 - - - - 4360 5 4360 2 Casuarina spp. 1 230 2 1 980 2 - - 3210 1 Others - - - - 8720 10 8720 3 Total 61200 100 99 100 100 87200 100 247500 100 Species Eucalyptus spp. Government Area % (ha) 30600 50 Communal % Area (ha) 69370 70 Private Area (ha) 61 040 Pinus patuia 18360 30 9910 10 Cupressus lusitanica 4900 8 7930 Acacia menaloxylon 4280 7 Callitris spp. 1 830 Grevillea robusta % % 65 .. Source: Mihigo (1999), MINAGRI (1997) and Murererehe (1999) The plantation Habiyambere forest cover declined following the 1994 cnsis In Rwanda. (1997) reported that the large displacement of people and cutting of forest plantations for fuel wood and shelter needs resulted in a loss of 15 000 ha and extensive damage to 25 000 ha. 1.3.2 Montane forests and savanna woodlands Rwanda's protected forests consist of closed forests that are generally submontane or montane, concentrated in the west of the country, and in the mountains bordering Kivu Lake (FAO, 1999). deciduous The submontane forests are mainly broad-leaved, forests and are characterised by species such as Syderoxylon moist adolfi- friederici and Entandrophragma excelsum. These are largely secondary growth, with Stellenbosch University http://scholar.sun.ac.za 5 most of the original forest having been cleared and only small patches of primary forest remaining in inaccessible areas. Above 2 000 m, the predominant forest type is broadleaved moist evergreen forest, with characteristic species including Syzygium parvifolium, Xymalos monospora and Psychotria spp. Stands of Arundinaria alpina are found in the mountains of the Zaire-Nile ridge above 2 300 m to 2 400 m in almost pure condition or mixed with Podocarpus latifolius and Polyscias fulva (Bosshard, 1971). Arundinaria stands are also found in the Birunga National Park bordering Uganda and the Democratic Republic of Congo. Major montane forests in Rwanda include three natural forests, namely Nyungwe, Mukura and Gishwati. These forests belong to the afromontane forests of the Albertine Rift and are characterised by high biodiversity (MINITERRE, 2000). The Volcano National Park, in the northern mountainous border of Rwanda and the Democratic Republic of Congo and Uganda, is a remnant montane forest set aside as a forest reserve since 1933 (Spinage, 1972). The east of the country has extensive areas of savanna woodland, with Acacia seyal and Combretum binderanum found on deep soils and Pappea ugandensis on lateritic soils (Hecq and Froment, 1961). Open sclerophyllous forests of Croton dichogamus and Euphorbia dawei cover part of Akagera National Park (FAO, 1999). Savanna woodland and gallery forests make up about 163 ha and are home to a variety of plant species used as traditional medicines, foods and for other human needs (MINITERRE, 2000). Pressure for agricultural land, settlement and forest products has led to a reduction in size and to an irreversible loss of biodiversity of protected forests (MINAGRI, 1998). Table 1.2 shows the reduction in size of these areas between 1960 and 1999. Nyungwe Forest has undergone a significant rate of deforestation. lts current area is 89 150 ha (Gapusi, 1999), much bigger than that of the second largest existing natural forest, Mukura (1600 ha). remaining The Gorilla sanctuary, Birunga National Park, has a area of 12 760 ha from an initial cover of 30 000 ha in 1924 (Twarabamenye, 1999). There is no precise estimate of the cover of Gishwati Forest and savanna woodlands. Parts of Gishwati Forest and Akagera National Park have been declassified for human settlement (Mihigo, 1999; Williams and Ntayombya, 1999). 6 http://scholar.sun.ac.za Stellenbosch University Table 1.2. Protected forests' cover change in Rwanda between 1960 and 1999 (MINAGRI, 1999). Cover Forest cover (ha) Protected forest 1960 1970 1980 1999 1996 1990 change, % 89150 21.8 1600 1600 46.7 14000 12760 12760 62.5 241 000 220000 90000 66.3 Nyungwe 114025 108800 97000 97000 94500 Gishwati 28000 28000 23000 8800 3800 Mukura 3000 3000 2000 2000 Birunga 34000 16000 15000 Akagera 267000 267000 267000 Other* 150000 150000 90000 50000 20000 Total 596025 572 800 494000 412800 352660 86.7 193 510 * Gallery and savanna woodlands. 1.4 1.4.1 FOREST PRODUCTS Wood products The major products from exotic and indigenous tree species include timber, poles, crafts and fuelwood. Quantitative data are disparate and only imprecise estimates exist for these products. Eucalyptus poles are used chiefly for construction purposes. The annual consumption of construction wood in the country's capital, Kigali, is about 15 000 to 20 000 m3 per year, while it averages 40 000 m3 in rural areas (Habiyambere, are Eucalyptus 1999). Timber is entirely produced by pitsawyers. Preferred species spp. (45%), Cupressus spp., and Grevillea robusta. consumption of timber is estimated at 36 000 rrr' (MINECOFIN, from a standing volume of about 72 000 m3 (Habiyambere, The annual 1998), processed 1999). Eucalyptus represents 45% of the market for timber; Cypress and Grevillea both 23% and Pinus 4.5%. The remaining 4.5% of this market is made up by imported wood from neighbouring countries Entandrophragma and consists of high value tree species such as excelsum. MINECOFIN (1998) reported that forestry supplies 96.2% of the energy consumed in the country, of which 90.8% is firewood and 5.4% charcoal. In 1997, fuelwood consumption averaged 7.1 million nr', 5% of which was used by small industries. For charcoal, which is traditionally produced and mainly consumed in urban areas, the Stellenbosch University 7 http://scholar.sun.ac.za annual production amounted to 46 000 tonnes, most of which (80%) was sold in the country's capital, Kigali (Habiyambere, 1999). In 1996, the charcoal deficit was approximately 4.5 million rrr' per year (FAO, 1996). The result is the use of crop residues and other dead materials by rural people in place of fuelwood and charcoal (Niang and Styger, 1990). In the absence of alternative sources of energy, the forestry sector will continue to be the major source of energy supply. However, the sustainability criteria of woodlands and protected forests will not be met due to encroachment through fuel wood gathering and charcoal for sale in towns (Habiyambere, 1999). 1.4.2 Non-wood products Non-wood forests products (NWFP) are still considered as secondary forest products in Rwanda (Murekezi, 1999). For many years, forests were considered as source of wood products, namely fuelwood, timber, building poles and woodcraft. Today, the interest in NWFP is gaining popularity as responses to national and international concern over the conservation of biodiversity and the multiple roles ofNWFP in local livelihoods (FAO, 1992). In Rwanda, NWFP are found in managed forest plantations, in indigenous forests and woodlands. Commonly used NWFP include medicinal plants for traditional human and veterinary medicine, wild ornamental plants, honey, wild fruit, tree seeds, essential oil plants, mushrooms, game meat, and ecotourism (Murekezi, 1999). 1.4.2.1 Medicinal plants Medicinal plants used in traditional human and veterinary medicine are increasingly in demand for research interests and traditional medicinal purposes. Mbarubukeye (1992) found that 80 000 patients For instance, have visited the Centre for Traditional Medicine of Bare in south-eastern Rwanda, since its creation in 1978 (Kayinamura, 1986). Additionally, the Pharmaceutical Centre (CURPHAMETRA) within the Rwanda Institute of Scientific and Technological Research (IRST), based in Butare Province in the South of the country, trains traditional healers' groups in collaboration with medical doctors in order to enhance the value of medicinal herbs. Stellenbosch University 8 http://scholar.sun.ac.za In veterinary medicine, 536 healers were identified in seven provinces of Rwanda (Mbarubukeye, 1992). Woody plants whose parts are used in traditional human and veterinary medicine are found in natural forests, in savanna woodlands or in forest plantations. Mbarubukeye and Niang (1996) outlined a list of 59 tree species used in traditional human medicine and 39 tree species for veterinary medicine. Despite their value, these species are disappearing through due to population pressure deforestation and burning (Habiyambere, 1999). 1.4.2.2 Honey Forests and woodlands play an important role in beekeeping. Habiyambere (1999) reported a list of 27 plant species associated with honey production in Nyungwe Natural Forest and 14 plant species outside this Forest. The potential for honey production in natural forests and forest plantations is not fully exploited (Nyilimana, 1997). Because harvesting methods are still traditional and include the use of fire, forests are at high risk of burning. At National level, honey production declined between 1990 and 1994 due to the death of beekeepers during the genocide and following massive destruction of hives and other bee-keeping materials and equipment (Murekezi, 1999). In 1989, honey production reached as high as 441 365 kg (Murekezi, 1999). Since 1994, the production has been well below average. Available data on honey production in 1998 at the Rwanda Association for the Promotion of Integrated Development (ARDI) indicated a production level of 23 000 kg (ARDI, 1998). 1.4.2.3 Wild/ruit and/orestry seeds According to Murekezi (1999), wild fruits are present in the forests but they receive little interest from the people, who prefer to consume fruits from agroforestry orchards. Wild fruits are generally eaten by children and by adults only during the periods of food deficit. available. Quantitative information on wild fruit production is not In Nyungwe Forest, woody plants producing edible wild fruits were 9 http://scholar.sun.ac.za Stellenbosch University identified as Clerodendron (Murekezi, 1999). fuscum, Parinari excelsa, and Myrianthus holstii In savanna woodlands, edible fruits are produced by Rumex maderensis, R. usambarensis, Cajanus cajan, Carica edulis, Rhus vulgaris, Solanum taitente and other Solanum spp. (Dessouter, 1991). Local collection of tree seeds in natural stands and plantations is carried out by the Tree Seed Centre (TSC) within the Rwanda Agricultural Research Institute (ISAR). The seeds are supplied for tree planting and afforestation country. activities all over the Before 1994, the annual delivery of tree seeds by TSC was nine tonnes, purchased by both tree planting projects and private growers (CGF, 1993). 1.4.2.4 Fodder plants and grasses Natural forests and savannas are valuable sources of animal feeds. However grazing is not allowed within state's forests. In the case of Nyungwe Forest, silvopastoral practices were permitted in new forest plantations in the buffer zone. About 20 grass species were grazed by livestock on the ground layer of these plantations (Gasana, 1988). The pressure of pastoralists on Gishwati Natural Forest resulted in the degazettement of 1200 ha for grazing between 1983 and 1993 (Murekezi, 1999). The Akagera National Park in the eastern region of the country contains palatable grass species, the most important of which are Themeda triandra and Panicum maximum (Mvukiyurnwami, 1987). 1.4.2.5 Wildflowers Many wild plant species are invaluable sources of horticultural material. Fischer and Hinkel (1992) produced a list of some species producing spectacular individual blooms or mass displays of colour and interesting forms. Unfortunately, the export market of cut flowers is not developed. Netherlands and Belgium. Potential markets have been identified in the A local co-operative, Highland Flower Association, exports about 1500 kg of flowers per week; but exports are limited by the number of flights to Kanombe International Airport (Kigali) where only one airline makes two trips per week (Murekezi, 1999). 10 http://scholar.sun.ac.za Stellenbosch University 1.4.2.6 Essential oilplants Many local plant species are used to produce medicines and perfumes in Rwanda. The Rwanda Institute for Scientific Research and Technology (IRST) extracts various oils which are used for pharmaceutical purposes. Some of the plant species producing valuable oils are Plantago lanceolata, Calendula officinalis, Datura stramonium, Eucalyptus globulus, E. smithii, Capsicum frutescens, Neorautanenia mitis, Thymus vulgaris, Pentas longiflora, Pelarganium graveolens, P. raduta, Cymbopogon afronardus, C. winterianus, C. citratus and C. jlexanomus (MICOMART, 1998). 1.4.2.7 Mushrooms Wild mushrooms occur in most of Rwanda's forests and are associated with Termitomyces (Hayes, 1991). Artificial production of these mushrooms is not yet possible due to difficulties in collecting the micro-organisms (Murekezi, 1999). Thus the potential productivity of wild mushrooms is not sufficiently enhanced. women and children collect and eat the mushrooms. mushrooms Only However, the demand for increases in towns as well as in rural areas. At the national level, artificial production of mushrooms in 1991 was about six tonnes (Hayes, 1991). The potential for mushroom production is high and the export market is available either in Europe or in Africa (Murekezi, 1999). 1.4.2.8 Game meat In the past in Rwanda, game meat was obtained from buffalo, antelopes, zebra, guinea fowl, and duikers among others, through hunting. Today, illegal hunting for bushrneat is still underway in protected forests. As a result, a few mammals and birds are present in both Akagera National Park (Ndayambaje, 1999) and in Nyungwe Forests (Gapusi, 1999) as a result of human disturbances. 1.4.2.9 Ecotourism Rwanda's forests and savannas are diverse. The country has a number of endemic plant and animal species with high ecotourism potential. Akagera National Park, Stellenbosch University 11 http://scholar.sun.ac.za Nyungwe Natural Forests and Birunga National Park are important wild lands that, when their maximum exploitation for ecotourism is attained, will contribute a large share towards the availability of foreign exchange in the national economy (Habiyambere, 1999). For instance in 1989, the turnover from ecotourism was about US$ 1 078000 of which US$ 475 500 accrued from Birunga National Park (ORTPN, 1989). This income takes account of direct revenues from entry fees and is considered to be low in relation to the ecotourism potential of the country (Murekezi, 1999). Forest reserves are the major focus of the ecotourism industry (MINAGRl. 1998). The challenge, however, is to satisfy the interests of the visitors and at the same time avoid social conflict and pressure on the environment. The best way is to ensure that both the country and the communities share the benefits from conservation of ecotourism resources. 1.4.2.10 Other goods Various other goods are manufactured stands. from plants obtained mainly from natural These are used to make various crafts and curios. Mbarubukeye and Niang (1996) illustrated a number of plant species whose parts are used to manufacture various objects, including baskets, mats, beehives and calabashes. 1.5 FOREST POLICY AND LEGISLATION Rwanda lacks a formal forest policy and legislation. The first piece of legislation providing for protected areas dates back to 1925 when the country, together with Burundi and the Democratic Republic of Congo, were under Belgian colonial rule (Spinage, 1972). Official reserves were created by the Belgian authorities in 1933 (Vedder and Weber, 1984). Since Rwanda achieved its independence in 1962, Law Ordinances have provided for establishment of managed areas. The Law Ordinance of 18 June 1973 established rules applicable to national parks, hunting reserves, and strict nature reserves. December This law has been slightly modified by the Decree Law of 18, 1973, establishing by presidential decree special reserves, national parks and hunting reserves. 12 http://scholar.sun.ac.za Stellenbosch University In 1988, Rwanda reshaped its forestry legislation, with the adoption of Law No. 47/1988 (JORR, 1989). This law was based on an old decree of 18 December 1930 that provided, among other things, for the creation of forest reserves and the protection of forest species. Dating from the colonial mandate, it applied to felling carried out by, or for, European settlers, leaving the Rwandan population to use the forest in conformity with their customary rights. The rapid population growth resulted in severe deforestation that the old legislation was unable to control (FAO, 1989). In response, the revised forest legislation defined a legislative framework for action and intervention in forestry. community It highlighted aspects of forest ownership and private forest land), forest conservation (public, and protection, forest establishment and tending, silviculture and management, harvesting, funding of forest programmes, national and communal forest planning, as well as forest law enforcement. Generally, the State reserves the right of ownership and control of all forest management operations (MINAGRI, 1984; JORR, 1989). In 1984, because of the conservation importance of natural forests, the Rwandan Government adopted an Action Plan aimed at maintaining essential ecological processes and at preserving genetic diversity of flora and fauna as well as satisfying national demand for direct and indirect benefits from forests (MINAGRI, 1984). Major programmes within the plan focussed on: • The establishment of integral natural reserves of 59 340 ha, of which 15 000 ha are in Birunga National Park, 5 000 ha in Gishwati and 39 840 ha in Nyungwe forests; • The demarcation of 7 500 ha of forests for restricted use; • The establishment of buffer zones between the different forests and the adjacent communities. In 1987, Rwanda adopted a ten-year National Forest Plan (1987 - 1997). This plan defined objectives, perspectives and a plan of activities in the forest sector. It has been a very useful document but its implementation has sometimes been constrained by disparate objectives and means, very wide interpretation of its content and lack of monitoring and evaluation for a regular review (AGRIFOR CONSULT, 1997). 13 http://scholar.sun.ac.za Stellenbosch University 1.6 PROBLEM STATEMENT Protected areas in Rwanda have undergone a tremendous rate of regression since 1960. Gapusi (1999) reported a regression rate of21.8% ofNyungwe Forest between 1960 and 1998. The study by Muderevu (1974) estimated the annual rate of regression of Nyungwe Forest at about 750 ha. He concluded that the Forest would disappear completely within a period of 84 to 134 years if adequate strategies are not adopted and implemented. in Nyungwe Many authors have indicated that the loss of biodiversity Forest results from illegal use (Gapusi 1999; MINAGRI, 1998; MINITERRE, 2000). Protected forests in Rwanda are under acute threat by adjacent (MINITERRE, 2000). communities The underlying causes are reported to be high population densities, overwhelming poverty and dependence on land and other natural resources (MINAGRI, 1998). Therefore, these factors result in great land and resource degradation, and consequent pressure on forest reserves. Both at national and local levels, the capacity to manage forest reserves with respect to human and financial resources is inadequate (WCS and ORTPN, 1995). As a result of these factors, the legal status, mandate and management of these forests have suffered substantially. Various strategies have been proposed by many authors and consultants (IUCN, 1987; IUCN and WWF, 1985; Mackinon and Mackinon, 1986; Perlez, 1988; Wilson, 1991) on how to improve the conservation, protection and management of forest reserves in Rwanda. However, resource degradation through encroachment by local people has been increasing over the years and culminated during the 1994 crisis in Rwanda. Various studies from Rwanda have emphasised the importance of improving the relationship between local communities and the management of the forest reserves (Bahigiki and Vedder, 1987; Gapusi, 1999; IUCN, 1979; MINITERRE, Renner, 1991; Renner, 1992; Vedder and Weber, 1984; Weber, 1992). 2000; Constant pressures on protected forests, however, have resulted in incremental loss of resources from the reserves. The survival of Rwanda's forest reserves depends on mechanisms that recognise the need for access and use by local people, while simultaneously co-ordinating local 14 http://scholar.sun.ac.za Stellenbosch University development and conservation of forest reserves. Following gazettement of all natural forests as forest reserves in 1984, these local communities received limited access to some resources (Renner, 1991). At present, there is a tendency by relevant state departments to seek ways of accommodating the needs and interests of local communities, with increased focus on maintaining a balance between conservation and utilisation of forest resources in protected areas (MINETERRE, 2000). Therefore sustainable utilisation of forest resources in protected areas, including Nyungwe forest, is a mechanism for involving communities in conservation and management of these areas. 1.7 JUSTIFICATION OF THE STUDY Prior to the establishment of Nyungwe Forest as a Forest Reserve, the local people around the Forest had access to wood and NWFP. From 1933, they were barred from removing the forest products from the Forest (Mbonyintwari, 1989). Consequently, conflict between the Forest Department and the local people began. For example, during the establishment of the buffer zone between the local communities and the Reserve, the land that was previously exploited by local people was taken back by the authorities, instilling disenchantment in the people (Renner, 1992). On gazettement ofNyungwe Forest as a Forest Reserve, local households utilising the buffer zone for settlement, agriculture and livestock farming were settled outside this zone or compensated for having to leave the Forest. After the people left, conservation measures were instituted involving the planting of exotic tree species. This, it was believed, would serve to mark the boundaries of the Nyungwe Forest as stated in the 1984 Action Plan for the conservation and management of natural forests on the Congo Nile Crete, where Nyungwe Forest is located (MINAGRI, 1984). With respect to conservation of Nyungwe Forest, the Rwandan Government, in collaboration with funding agencies, adopted strategic actions aimed at reducing the dependence of local communities on the Forest. Access to some forest resources was also granted only in the buffer zone, including fuelwood gathering, taungya agroforestry and beekeeping (Barbier, 1992). Some of these rights to local people were established following socio-economic studies and recommendations by various Stellenbosch University 15 http://scholar.sun.ac.za consultants including Bahigiki and Vedder (1987), Barbier (1992), MINAGRI (1992), Renner (1991), Stebier et al. (1984), and Weber (1992). Socio-economic studies show the importance of the Forest to local communities in terms of wood and NWFP (Bahigiki and Vedder, 1987; Renner, 1992). On the other hand, Nyungwe Forest is also an important water catchment's invaluable role in soil protection (Budowski, 1976; Sorg, 1978). area and plays an Despite all these advantages and perceptions, Nyungwe Forest is threatened by illegal activities that impact negatively on its size and biodiversity (MINAGRI, 1998). Human disturbances and unlawful activities are the major causes of regression and losses of plant and animal species (Gapusi, 1999; Muderevu, 1974). At present, Nyungwe Forest is managed by the Nyungwe Conservation Project (PCFN) in partnership with the Rwanda Office of Tourism and National Parks (ORTPN) with limited financial and human resources (Barakabuye, 2001). Activities in Nyungwe Forest and in the buffer zone are limited to forest protection and public awareness programmes, ecological studies and monitoring of animal populations (WCS and ORTPN, 1995). Tourism activities are not fully enhanced. Against this background, opportunities for income-generating activities to local people are too few or lacking. Few local people are employed as forests guards for the protection of the Forest. The revenue from tourism in Nyungwe Forest is too little and has no impact on the livelihoods of the local communities (Weber, 1992). The major focus of the management of the Forest is to contain all illegal activities coming from the neighbouring communities, The Government of Rwanda, represented by the Ministry of Land and Environment, has vested interests in the conservation of Nyungwe Forest, with its rare and endangered flora and fauna and potential for sustainable ecotourism, for which only a small area of the original size is available (MINAGRI, poverty, local peoples' 1997). Faced with rural strategies include access to land through deforestation, agricultural intensification, migration, and off-farm income activities in rural areas (Habimana, 1982; Weber, 1989). Stellenbosch University 16 http://scholar.sun.ac.za It may be futile to aim at conserving the biological diversity in Nyungwe Forest without understanding the problems underlying its conservation. Addressing the needs of the neighbouring communities can partly alleviate threats to conservation. In addition to tangible benefits, intangible benefits that local people derive from the Forest near to them should be considered. zone are rich in resources. Nyungwe Natural Forest and its buffer It is important to identify activities of interest to local people that could be allowed so as to utilise the resource sustainably conservation purposes. Due to dependence on the Forest, communities organised at the local level to use the resources. and for could be This approach could result in the local people to utilise the resources in a sustainable way, leading to their conservation. The adoption of such an approach would tend to answer questions related to community involvement in forest management and to the benefits this community may obtain in return. In Nyungwe Forest, resource exploitation is not practised, but there is a need to determine whether the local communities could use the resource sustainably if legal rights for resource exploitation are given to them. With all its boundaries surrounded by poor households, the long-term future of the Nyungwe Forest depends on a strong partnership including local communities. enhancing biodiversity among various stakeholders, Addressing the conflict over the Nyungwe Forest, and improving availability of resources units require a management option which protects the Forest from the damaging use and supervises the use of the Forest. Joint Forest Management (JFM) or Collaborative Management is generally accepted as a viable arrangement that benefits the forest and the major partners, namely the government and the local communities (UFRIC, 2000). JFM emerged from India and encompasses management schemes based on the sharing of responsibilities, benefits and obligation from the forest between local people and usually the state (Wollenberg, 1999). The communities receive benefits from such involvement, which are clearly spelt out at the start of the programme (Nhira and Matose, 1996). JFM is most applicable to public/state land, or to collective property where resources , are managed under local management rules. For its effectiveness, there must exist a decision-making authority, recognised and respected at the local level. In the case of Stellenbosch University 17 http://scholar.sun.ac.za conflict, the presence of a mediator capable of leading negotiations, is important (Dubois, 1998). Monitoring of the community level forest management enables the provision of checks and balances community-level management. as well as the identification It also contributes to the building confidence, as has been the case for the community-managed Pacific (Fisher, 1999). of successful of trust and forests in Asia and the Thus, this study takes the first step in assessing ways of managing Nyungwe Forest in a sustainable manner with the participation of local communities who lost access to forest resources. Approaches to achieving this stem from the understanding of the priority needs and interests of local communities in the Nyungwe Forest. Resources of interest to local people are scattered throughout the forest at different intensities. Hall and Bawa (1993) and Peters (1996) have proposed methods for collecting information on the abundance and distribution of NWFP from tropical forests. In some countries like Uganda, some work has been done on the density and distribution of different plant species being harvested in Multiple Use Zones of montane forest areas (Bitariho, 1999; ITFC-EMP, 1999; Kamatenesi, 1997; Muhwezi, 1997; Ogwal, 1998). In Nyungwe Forest, many plant species have been harvested illegally and are still in high demand for different uses (Gapusi, 1999). In the case of Nyungwe Forest, no work has been done on resource exploitation and impact. This study aims at providing baseline data on forest resources of importance to local communities. Further, it evaluates the Forest in terms of the numbers and sizes of trees and of what species may be harvested. The assessment of regeneration potential of valuable plant species is also emphasised, since it provides valuable data for resource managers (ITFC-EMP, 2001). The general approach of the study involves focussing on the preferred forest resources through the use of participatory rural appraisal techniques. The potential of the Nyungwe Forest to provide these resources has been assessed by focused research on the density, size-class distribution and regeneration potential of preferred tree species for timber, poles and medicinal materials. In line with the formulation of policies that involve the management and conservation of natural resources, it is worthwhile to assess the possibilities of recognising local people rights and responsibilities as well as partnerships in decision-making. The Stellenbosch University 18 http://scholar.sun.ac.za government may pass certain user rights and responsibility to the local communities as a strategy to overcome the conflict between the community needs and the conservation of the Nyungwe Forest. Therefore, it is important to provide guidelines that may lead the government and non-governmental partners to develop and implement decisions over the management of Nyungwe Forest with the dual focus of improving the local community livelihoods and enhancing proper management of Nyungwe community Forest. Legislative gain in sustainable and policy framework utilisation would take account and conservation of biodiversity of ill Nyungwe Forest. 1.8 AIM AND OBJECTIVES 1.8.1 OF THE STUDY Overall aim of the study The overall aim of this study was to contribute to improved livelihoods of rural communities by involving them in sustainable utilisation of forest reserves. Focus was placed on assessing the potential supply of wood and non-wood products for adjacent community needs in Rwanda's forests. Thus, the backbone of this study was to assess the current status of resource use and management ofNyungwe Forest and to determine the potential for community involvement in management and sustainable utilisation of wood and NWFP from the Forest. 1.8.2 Objectives of the study In view of the above aim, the study of the potential for joint management and multiple use ofNyungwe Forest has the following objectives: 1. To identify local people's problems, needs, preferences and interests in access to and use of resources in Nyungwe Forest. 11. To identify the potential ofNyungwe Forest to provide the most preferred timber, pole and medicinal species by local communities. Stellenbosch University 19 http://scholar.sun.ac.za 111. To inventory the forest resources in order to ascertain the abundance, the sizeclass distribution and the regeneration potential of the preferred tree species in the periphery ofNyungwe Forest. . IV. To determine the tree species that can be used sustainably. v. To recommend community a management involvement option that enables access to resources and in management and utilisation of Nyungwe Forest resources. 1.9 Potential benefits The potential benefits of this study relate to the participation of local communities in management and conservation of Nyungwe Forest, particularly deriving increased benefits from the sustainable utilisation and conservation of the Forest. Within this framework, local communities are expected to work for the protection of Nyungwe Forest, thereby avoiding illegal use of the Forest. The study helps to assess the potential for sustainable utilisation ofthe Forest and provides recommendations which decision-makers may use in devolving some power and responsibility for collaborative management of state forests. 1.10 THESIS STRUCTURE The thesis is structured in the following manner: • Chapter Two describes how the research was conducted. It covers the methods used both for the Rapid Rural Appraisal exercise and the N yungwe Forest resource assessment. It is also designed to give an account of the methods used in the analysis of the households survey and forest inventory data. • Chapter Three covers the findings of the study. The results of the participatory assessment cover household surveys, group and key informant interviews. Also, within this chapter, forest inventory results are provided, with emphasis on preferred tree resources identified during the surveys. Stellenbosch University 20 http://scholar.sun.ac.za • In Chapter Four, the discussion of the various results is presented, based on the findings of the study and supported by relevant literature where appropriate. • Chapter Five summanses recommendations this study's findings and provides suggestions for the way forward. and Areas of future research related to community utilisation and sustainability of Nyungwe Forest are proposed. Stellenbosch University 21 http://scholar.sun.ac.za 2 STUDY AREA AND METHODS 2.1 STUDY AREA 2.1.1 Description of Nyungwe Forest This section provides information about Nyungwe Forest by reporting its biophysical characteristics and biological diversity. An overview of the Forest is given, relating to conservation interests and the characteristics of the adjacent population. 2.1.1.1 Size and location of Nyungwe Forest Nyungwe Forest is located in the South West of Rwanda, near the southern shore of Kivu Lake (Figure 2.1). Nyungwe Forest covers approximately 90 000 ha (Gapusi, 1999). It is one of the largest blocks of lower montane forest in Africa (Weber, 1989). The majority of its area stretches within Gikongoro and Cyangugu Provinces, while a very small proportion of the Forest is found in Kibuye Province (Sorg, 1978). It also extends to Kibira National Park in Burundi (Budowski, 1978). Nyungwe Forest is found at latitude 2017'S - 2049'S and longitude 29°05' - 29°29' E (Sorg, 1978). It occurs within the western part of the tectonic Rift along the Zaire-Nile Divide (Weber, 1989). The altitudes within the Forest range from 1 600 m to more than 2 700 m (Budowski, 1976). The mean annual rainfall is 1744 mm, typical of an African rainforest (Sun et al., 1996). The temperatures in Nyungwe Forest are generally cool, with an average minimum temperature of 10.90e and an average maximum temperature of 19.60e (Sun et al., 1996). Temperatures rarely fall below o'c in high altitude zones of the Forest (Budowski, 1976). 22 Stellenbosch University http://scholar.sun.ac.za N + newafrics.com Il'n ac.1e 1:44,5. 8 Legend Information Cantra @ Mountains Campsite /'v? Main Road /'V Motorabie Track Foot Path ... . ' .. .. - UND International Boundary /'V River Lake .. Reserve . ::~:~ ~:~ : BUR BpS , Boundary/Extent Source: Newafiica (2000). Figure 2.1. 2.1.1.2 Nyungwe Forest, showing altitudes inside the Forest. Geology and drainage The area of Nyungwe Forest is mainly composed of very old Precambrian rocks (Storz, 1983). Storz (1983) differentiates three geological parts of the Nyungwe Forest. In the eastern part of the Forest, the substratum is derived from granites and Stellenbosch University 23 http://scholar.sun.ac.za granitic rock. The homogenous character of these rocks creates narrow steep-sided valleys dominated by round hillcrests. Sorg (1978) reported that soils derived from such rocks are usually poor. In the western part of the Forest schist, quartzite and dolerite are the main parent rocks: the soils derived from these rocks are fertile (Sorg, 1978). These rocks result in a topography characterised by very steep slopes and very narrow valleys. Soils in the area of the Forest near Gisakura are basic, with contents of Silica, Iron and Magnesium. The whole area of Nyungwe Forest is a vital water catchment containing sources of several perennial rivers that drain the Congo Basin and the Nile Basin (Barabwiliza, 1992). For the Congo Basin, five rivers found in Nyungwe Forest flow to Kivu Lake: Koko, Ruhwa, Rubyiro, Kamiranzovu and Kilimbi. The Nile Basin, on the other hand, receives water from the Nyabarongo, Nshili, Akanyaru, Rubyiro, Rukarara, Mbirurume and Mwogo Rivers. 2.1.1.3 Flora Nyungwe Forest contains about 1 400 plant species, of which 400 are identified tree species, hundreds of Orchidaceae and over 200 mosses and lichens (MINAGRI, 1997). Its flora contains 42 endemic tree species, various medicinal plants and valuable hardwoods (Habiyambere, 1999). On the basis of altitude ranges, four vegetation strata are identified and recognised by dominant tree species, as described by Sorg (1978). Between 1 600 m and 2000 m, dominant tree species reach 35 to 40 m in height. These include Parinari excelsa, Newtonia buchananii, Symphonia globulifera, Entandrophragma excelsum and Albizzia gummifera. Orchidaceae and ferns are abundant. This vegetation type covers 17% of the total area. At the altitude range of 2 000 - 2 300 m, dominant canopy tree species are Entandrophragma excelsum, Parinari excelsa, Prunus africana, laurifolia and Chrysophyllum gorungosanum. Ocotea usambarensis, Ficalhoa The shrub stratum is less dense and consists mainly of ferns. This vegetation represents 21 % of the Forest. Podocarpus latifolius dominates at altitudes between 2 300 m and 2 500 m. This valuable tree species reaches 15 to 20 m in height. At this altitude mosses and lichens also occur, but the grass layer is poor and discontinuous. This type of vegetation occupies the largest area, estimated at 43%. At very high altitudes between 2500 mand 2700 m, Stellenbosch University 24 http://scholar.sun.ac.za the structure of the vegetation is dominated by shrub and herb growth forms. Tree species found at these altitudes include Phillipia benguellensis, Agauria salicifolia, Faurea saligna and Hagenia abyssinica. Due to high humidity, mosses, lichens and epiphytes are present. This vegetation is distributed over 19% of the total Forest area. Kamiranzovu wetland within Nyungwe Forest contains typical wetland vegetation types over an area of more than 3 500 ha in the northwest part of the Forest at an altitude of 1 860 m (Bouxin, 1974). The other large swamp of the Forest is Uwasenkoko. Nyungwe Forest has relatively high species diversity. Studies of species dynamics found 62 plant species in an area of 12 500 m2, which is higher than the 45 plant species per 10 000 m2 found in South American montane forests (Gapusi, 1999). At an international scale, 13 species of endemic Orchidaceae are recognised (MINAGRI, 1998). These include Disa eminii, D. robusta, Catyrium crassicaule, kassneriana, horsfalii, Cynorkis Disperis anthoceros, Calanthe sylvatica, Stolzia cupuligha, Eulophia Polystachya virginia, P. vulcanica, Chamaeangis sarchopylla, Draphananthe burtii and Cyrtochis arcuata. Gapusi and Mugunga (1997) have listed a number of tree species in Nyungwe Forest that are rare or in danger of extinction. The major species are Habenaria macrantha, Disa robusta, Dispais anthoceros, Cynorkis kassneriana, Calanthes sylvatica, Polystachia virginia, Chamiseangis sarcophylla, Cyrthorchis arcuata, Aningeria altissima and Conyza volkensii. 2.1.1.4 Fauna A broad array of fauna is present in Nyungwe Forest, made up of approximately 265 bird species, 25 to 30 large mammals, many other small mammals, reptiles and insects (MINAGRI, 1998), including 126 species of Lepidoptera (Habiyambere, 1999). Mammals are dominated by primates that are unique to the area; a description of the most common species is given by Storz (1983). Table 2.1. provides a list of Latin and English names of some primate species found in the Forest. 25 http://scholar.sun.ac.za Stellenbosch University Table 2.1. Latin and scientific names of some primates found in Nyungwe Forest (Storz, 1983). Latin names Pan troglodytes schweinfurthi Papio anubis Colobus angolensis adolfi-friederici Cercopithecus lhoesti Cercopithecus mittis doggetti Cercopithecus mona denti Cercocebus albigena johnstoni Perodicticus potto Galago spp. Cercopithecus ascanius schmidti Cercopithecus aethiops English names Eastern chimpanzee Olive baboon Angolan black-and-white colobus L'Hoest monkey, mountain monkey Sliver monkey Mona monkey Grey-checked mangabey Potto gibbon Bush baby Red-tail monkey Grivet monkey A number of animal species is reported to be rare or endangered. include the forest hog Hylochoerus meinertzhageni, The most rare ones the blue duiker Cephalophus monticola, the bushbuck Tragelophus scriptus, the small-scaled tree pangolin Manis tricuspis, the giant pangolin M gigontea, the spotted hyena Crouta crocuta, and the leopard Panthera pardus. MINAGRI (1998) reported characterising afromontane forests that are protected by CITES. 36 animal species Examples of these animals include the lturi Forest chameleon Chamaeleo adolfi frederici, the flap- necked chameleon C. dilepis idjwensis, the Ruwenzori three-homed chameleon C. johnstonii, Rhampholeon boulengeri and the mountain monkey Cercopithecus lhoesti. 2.1.1.5 Management history and conservation interests Nyungwe Forest was gazetted as a Forest Reserve under Legal Notice No. 83 bis Agris of December iz", 1933 (PCFN, 1989). prevent the total disappearance of the Forest. The Colonial authority aimed to In 1934, the status was modified slightly by the Legal Notice No. 33/Agri of May 24th, 1934 to include a special exploitation regime of the Forest (Mbonyintwari, 1989). Within the framework of management and conservation of natural forests on the Congo Nile Crest, the 1984 Action Plan aimed at conserving Nyungwe Natural Forest qualitatively and quantitatively (Renner, 1991). responsible for its management and conservation. The Forest Department was To facilitate this, four management zones of Nyungwe Forest were delineated as stated in the 1984 Action Plan. In Stellenbosch University 26 http://scholar.sun.ac.za partnership with the Swiss Government, the World Bank, and the Central Fund for Economic Co-operation, management the Rwandan of the Forest. Various Government developmental promoted activities conservation and integrated rural development with conservation. 'The Projet Pilote Forestier (PPF)' was the first project that operated in Nyungwe Forest and the buffer zone since 1970 (MINAGRI, 1984). This project established exotic plantations with the objective of marking the boundaries of Nyungwe Forest and keeping it from encroachment by neighbouring communities (Grison and Braesco, 1984). The project initiated a series of other activities consisting of rural afforestation so as to secure fuelwood and pole supplies to local communities adjacent Communes. in At Rangiro (Cyangugu) and Gisovu (Kibuye), a large number of pit sawyers who encroached on the Forest were used as project workers. In 1982, the 'Projet Crête Zaire Nil' (CZN) was opened in the North-East Nyungwe. This project integrated forestry, agriculture and social activities. of It also created a buffer zone and managed pastures (MINAGRI, 1984; PCFN, 1989). Both CZN and PPF did not cover the whole buffer strip between Nyungwe Forest and the local people. The southeast and southwest borders of Nyungwe Forest lacked adequate protection. Thus forest degradation continued, particularly due to illegal logging and gold mining (PCFN, 1989). Since 1982, based on geographical zones, Nyungwe Forest has been managed by four projects named "Unité de Gestion de la Zone" and referred to as UGZ 1 to UGZ 4 respectively (UGZ1, 1987; Weber, 1989), under the financial assistance of many bilateral and multilateral organisations. In the North-West, Nyungwe was managed by UGZI under financial support of the Swiss Co-operation since 1986. Two projects operated in the northeast: UGZ2, funded by European Development Fund (EDF) since 1982, and Mudasomwa Project, assisted by Belgium. The latter project concentrated its action in Mudasomwa Commune and conducted afforestation activities in rural areas. The southeast of Nyungwe Forest has been managed by UGZ3 since 1989, using funds provided by the Rwandan Government and the World Bank. Since 1987, the southwest of the Forest was under the responsibility of UGZ4, which was funded by the French Government (MINAGRI, 1991). The various UGZ realised different Stellenbosch University 27 http://scholar.sun.ac.za activities, including the establishment of 8 300 ha of forest plantations. In 1991, the various projects established about 10 000 ha of forest plantations of predominantly exotic tree species (MINAGRI, 1991). Other important roles played by the UGZ include (MINAGRI, 1991): • The construction and maintenance of infrastructures: offices, staff accommodation houses, roads and trucks, forest roads, schools, water adduction, dispensaries, etc. • Timber harvesting from thinning and marketing of timber and timber products; • Inventory and research in Nyungwe Forest in collaboration with research institutions; • Protection of the Forest from encroachment through regular patrols inside and outside the Forest; • Establishment of forest management plans to guide interventions in the buffer zones; • Public awareness and education programmes on the protection and conservation ofNyungwe Forest. Due to its high biological diversity, PCFN in conjunction with ORTPN aimed at preserving Nyungwe Forest with increased focus on ecotourism, research and public awareness of conservation (ORTPN, 1991). PCFN was funded by USAID, African Wildlife Fund, Digit Fund and Wildlife Conservation International (WCI). Funds were used to improve technical and management capabilities, to increase the level of on-going biological, physical and socio-economic studies, to raise public awareness on the importance of biodiversity conservation and to provide various forms of administrative support to the project (PCFN, 1989). Activities of the various projects operating in Nyungwe Forest were planned to be co-ordinated by a Conservation Centre that had not become operational before 1994. Offices and staff houses were established at Kitabi (Mudasomwa Commune) but they were damaged during the genocide in 1994 (Gapusi, 1999). Stellenbosch University 28 http://scholar.sun.ac.za 2.1.1.6 Adjacent communities On an administrative basis, the population of the Republic of Rwanda is organised into Prefectures and the latter into Communes. Depending on its size, each Prefecture comprises a number of Communes, which in turn are subdivided into administrative Sectors. A Sector is further broken up into administrative Cells, each being made up of at least 50 households. At each level, the leaders are elected locally. The population living within the boundaries of Nyungwe Forest was organised and unevenly distributed over 11 Communes (today, two or more Communes have been fused together to form a District). This population was estimated at 500 000 people in 1997 (MINAGRI, 1997). MINAGRI (1984) distinguished two types of communities living around the Nyungwe Forest. The eastern ridge comprised less populated Communes within the agricultural region of Congo Nile Crest, characterised by poor soils on steep topography, relatively higher farm size than at national level, low population growth, relatively low quality of life and food insecurity. The other characteristic of this region was the high decrease in farm size that made agricultural intensification by use of agricultural inputs difficult. Land in the border strip of Nyungwe Forest was of particular interest due to its high organic matter content. MINAGRl (1997) defined the western ridge of Nyungwe Forest as falling within the Communes of the "Impala and Kivu Lake Borders" agricultural farming conditions were favourable. region, where The average farm size was very small (about 0.39 ha per household) but sufficiently productive. A large area of Nyungwe forest was cleared through expansion of arable land in this locality. Besides agriculture, the population as a whole derives a livelihood from activities in the Forest and in the buffer zone. In the buffer zone, livestock farming is not well developed and is dominated by sheep, goat, and pig rather than cattle (MINAGRI, 1984). Within the forest of Nyungwe, activities are illegal and consist of gold mining, logging, hunting, bee-keeping, agriculture and collection of medicinal herbs (MINAGRl, 1984; PCFN, 1989, Gapusi, 1999). 29 http://scholar.sun.ac.za Stellenbosch University 2.2 RESEARCH METHODS This chapter describes the methods employed in the collection and analysis of quantitative and qualitative data in order to achieve the study objectives outlined in section 1.8.2: The methods used for the study have two components: • Participatory rural appraisal (PRA) techniques were used to determine the problems, needs, priorities and interests of local communities in Nyungwe Forest. Interviews were conducted in the surrounding areas of the Forest. Interviews within individual households were in the form of open-ended questions. Further, data were obtained through in-depth discussion on general issues with small groups and key informants. • Forest resource assessment was carried out as the focal point of the study in order to find out how well stocked Nyungwe Forest is to guarantee a sustainable supply of resources that are mostly preferred by local communities. In addition, it aimed at collecting information on which to base recommendations for resource uses. Because of time and financial constraints, it was not possible to cover more than a small part of forest area adjacent to communities, and so the survey focussed only on the highly preferred plant species identified during the PRA exercise. The survey was carried out to achieve the following specific objectives: - 1. To estimate the abundance (presence or absence) of the forest resources that are believed to be important to the livelihoods of local communities; 11. To determine the abundance of each species present and the size-class distribution of individuals at different diameter classes; 111. To assess the regeneration potential of useful plant species to local communities; IV. To recommend the resources for which the communities could have access. 30 http://scholar.sun.ac.za Stellenbosch University For the purpose of this study, basic information on species types, density, frequencies, diameter-class distribution, species dominance and regeneration were collected. Some of these parameters determine sustainable levels of resource use as described by Peters (1994). 2.2.1 Selection of the study sites The study area was stratified into two sections, based on agro-ecological zones defined by Niang and Styger (1990) using biophysical characteristics and crop types. The Bufundu and Bushiru highlands in the eastern and northern part of the forest comprise eight ex-Communes of similar population densities in the range of 200 to 250 persons per krrr'. The second area, the Impala Ferralsols Region, stretches over three ex-Communes (Karengera, Kagano, Busozo) in the western ridge of the Forest with a population density of250 to 350 inhabitants per krrr'. A further subdivision of the region was based on past project management zones. Within each zone, one Commune was selected to represent the project area. Other bases for selection of sample sites included the interface of the Communes with the Reserve, the degree of deforestation at the forest edge and altitude ranges within the Forest. The latter criteria were used because of their potential impact on the availability and diversity of resources. Based on geographical contiguity with the Forest, a total of five Communes were selected for this study. The five Communes were Kagano and Busozo at the western ridge of the Forest, Nshili in the south-east, Mudasomwa and Musebeya in the eastern side of the Forest (Figure 2.2). 2.2.2 2.2.2.1 Participatory Rural appraisal Selection of households The households of each Commune are not equally close to the Forest. Administrative Sectors extending into Nyungwe Forest were identified and further stratification aimed at selecting a variety of households at "Cell" level, using the criteria of the closest location to the Forest. At most, two Sectors next to the Forest were considered for each Commune and the respondents randomly selected in two Cells. 31 Stellenbosch University http://scholar.sun.ac.za Source: Figure 2.2. Communes neighbouring Nyungwe Forest, showing the survey sample distribution. Within each Cell, the number of households interviewed was 10, representing a fifth of 50 households per Cell. Therefore, a total number of 200 households spread over five Communes in different Sectors and Cells in the vicinity of the Forest were interviewed. A sample of this size is much larger than the sample size of 70 households used by Wambugu (1999) in Embu district of Kenya and recommended by Palm et al. (1993) as appropriate for making inferences about the larger population. Due to heterogeneity and the large adjacent population, the sample size of200 households was thought to represent the conditions around Nyungwe Forest. Stellenbosch University 32 http://scholar.sun.ac.za 2.2.2.2 Selection of groups and key informants In the communities neighbouring Nyungwe homogenous groups using the forest resources. Forest, there are no identified Groups of five to 10 people living near the Nyungwe Forest were randomly identified with the aid of the head of each Cell or members of the local administrative committee at Cell levels, since they are living with and leading the people living close to the Forest. Individuals known to be the heaviest users of the Forest and forest products were selected for group discussion. Various key informants were approached to identify the current challenges and opportunities facing managers of the Forest. These respondents were important stakeholders in the future conservation and management of Nyungwe Forest. They included the administrative authority of each Commune being surveyed and staff of a number of relevant bodies, including PCFN, ORTPN, the Department of Environment in the Ministry of Land and Environment, various NGOs and Co-operatives, the Regional Agricultural Services, tea factories and private enterprises. 2.2.2.3 Data collection Household and group interviews provided information on resource availability within Nyungwe Forest as well as species preferences and their uses, land tenure and landuse, socio-economics and general information on agriculture and livestock keeping, and perceptions of institutions that might shoulder the responsibility for sustainable utilisation and conservation of Nyungwe Forest resources (Appendix 1). Interviews with groups and key informants focussed on resource preferences and ranking. The discussions also focused on utilisation and conservation strategies of the Forest with active participation and increased benefits to local communities. Information that relates to resource abundance and distribution were explored through subjective assessment based on the interviewee's knowledge of a particular resource. The abundance of a given resource was measured on a scale of 1 to 5, corresponding to the DAFOR (Dominant = 5, Abundant = 4, Frequent = 3, Occasional = 2 and Rare = 1) scale (Smith et al. (1985). From a list of available resources in the Forest, three important resources to communities were prioritised and appraisals were made to determine the uses of the resources. In order to identify habitats of the Forest that Stellenbosch University 33 http://scholar.sun.ac.za might reeerve intensive use by the local communities, information on resource abundance and distribution were sought in relation to landscape, i.e. whether the resource occurs on hilltops, on slopes, in the bottomlands or in the swamps. Nyungwe Forest resources were categorised based on household use of the resources. Subjective quantity estimates of marketed products were investigated, focussing on the season of resource harvest and market location. Harvesting of some parts of plant resources such as roots and bark has an impact on the availability of the forest resources and their response to disturbance (Peters, 1996; Wild and Mutebi, 1996). Consequently, respondents were asked to identify parts of plants used for different purposes. For wildlife resources, emphasis was put on the game used to provide various game products to local communities. Methods and time of hunting were also emphasised as they were thought to influence the wildlife population dynamics. Land tenure, land-use, and socio-economic status of local communities were obtained through semi-structured interviews with groups and individuals and were collected by means of an open-ended questionnaire and discussions. Data on socio-economic status concerned the personal economic and financial situations in terms of the level of education, the enterprises undertaken by the household and the sources of income. Information on land tenure and land use aimed to understand the process of land acquisition and the management of soil fertility. These were thought to influence the dependence of people on forest resources. The questionnaire also included questions on institutional arrangements that are believed by households to enable the conservation of the resources for the benefit of rural communities. Questions focused on mechanisms at local level that enabled access to resources by local communities. Finding strategies to coordinate local needs and the conservation of Nyungwe Forest and linking the management of the Forest to community involvement, sustainable utilisation, and governance was emphasised. 2.2.2.4 Data processing and analysis Prior to analysis, the interview data were coded and organised into categories following the procedure described by Babbie and Halley (1998), and Neuman (2000). Numerical codes were assigned to semi-structured responses on the background Stellenbosch University 34 http://scholar.sun.ac.za information of the respondents. Responses were all recorded on paper and, depending on the frequency of the answers, broad categories were created. The survey data were entered and organised into Statistical Package for Social Sciences (SPSS) (Babbie and Halley, 1998) data format. The reeode procedure was used to transform all pre-existing variables into different variables either by discarding the "don't know" and "no answer" responses or by dividing a pre-existing variable into different categories and coding each category differently. The latter procedure has been applied to questions that related to family size, level of education, farm size, quantities of crops at harvest, and others. In some instances, categorical data were transformed into continuous values without contorting the respondents' responses. As the interview was flexible and allowed detailed and anticipated information, some responses from some local people could not be coded and therefore the information was used to support the discussion of the results. For the survey data, limited quantitative analysis was performed, focusing mainly on frequencies and associated histograms so as to provide visual illustrations and to support descriptive results. 2.2.3 Forest resource assessment 2.2.3.1 Sampling design and plot sizes The size of plots depends on the characteristics of the community being sampled. Peters (1996) recommended plot sizes of 16 to 100 m2 for shrubs and understorey vegetation. For the study of small trees and shrubs, Stanley and de Olieveira Castro (1959) recommended the use of small plots of 4 x 4m or 2 x 2m, always located in the same corners of the larger plots. However, small plots produce density estimates with a large error term. In contrast, Kruger et al. (1997) recommended using plot sizes that are scaled to the canopy height of a particular forest. In their study in South African forests, they used 20 m by 20 m plots, randomly located in a forest with a canopy height of 20 m. For this study, plots were established with the aid of a prismatic compass, tape measure, machete, and ranging poles. They were placed 1 km inside Nyungwe Forest 35 Stellenbosch University http://scholar.sun.ac.za from each side of the forest edge community being surveyed. In each part of the forest area, two plots each measuring 2500 m2 (i.e. 50 m x 50 m) in size (Sample unit A) with the corresponding nested areas (Sample units Band C) were established (Figure 2.3). The survey procedure was based on the method used by Fedlmeier (1998) for the study of the development of secondary forest in Costa Rica. Over sample unit A, trees greater or equal to 10 cm diameter at breast height (OBH) were recorded. Sample unit A: 50 x 50 m Sample unit B: 707 m2 ample unit C: 12 x 4 m2 Figure 2.3. Subdivision and size of the sample units used for forest resource assessment, Nyungwe Forest. Sample unit B was measured off from sample unit A for appropriate assessment of abundance of small trees greater than 1.30 m height and DBH of 2.5-9.9 cm. A circular shape of 15 m radius (i.e. 707 m2) was selected. Seedlings and saplings were recorded in twelve 2 m x 2 m nested sub-units (Sample unit C) within sample unit B. The sub-units were delineated following the diagonals of Sample unit A and placed at 3 m intervals. Individuals ofO.10 m to 0.30 m height and DBH ofless or equal to 2.5 cm were recorded. Peters (1996) scaled small trees of this size to tree seedlings or saplings. Therefore, regeneration assessments were done on an area of 48 m2, lying in the range of plot sizes recommended by Peters (1996). For the study, a total number of 10 sample units with 10 corresponding nested subunits and 60 sub-subunits were used. 36 http://scholar.sun.ac.za Stellenbosch University 2.2.3.2 Data Collection Quantitative and qualitative information was recorded on tree species found in the sample plots. The aim was to gather information on the availability and abundance of the preferred tree species in relation to other tree species present. Preferred tree species were defined as those that were identified during the participatory rural appraisal as mostly being sought after by local communities for use as timber, poles and medicinal materials. They were also referred to those species that were frequently used in the past by the local people before the Forest was gazetted a forest reserve. The following parameters were recorded during the forest assessment: tree species composition (local names), stem densities, diameter at breast height (DBH), and stem quality of the most preferred timber and pole species. Plant identification and species composition Within each sample unit, tree species were identified in the field by a team of five local persons including two forest rangers who provided vernacular names. Local people were included in the survey team because they were knowledgeable of the local names of the tree species. Scientific names were identified by making reference to the work on Rwanda Flora by Troupin (1985, 1983, 1978), Troupin and Bridson (1982) and Troupin and Donneaux (1982). Individual plants of every species were also recorded. Density, frequency and diameter size-class distribution Density, or the count of individuals of species within the plot (Davis and Johnson, 1987; Kent and Coker, 1992), is the ecological parameter of interest to ethnobotanists (Peters, 1996). Hall and Bawa (1993) and Peters (1994) reported that density and size-class distribution are indicators of the impact of exploiting a given population. Density describes not only the degree to which the site is being utilised, but also the intensity of competition between trees. In this study, density measures that combined tree sizes and the number of individuals were used. Individual trees were counted in each of the sample plots and recorded in 37 http://scholar.sun.ac.za Stellenbosch University the appropriate diameter class. Frequency as a measure of species abundance (Kent and Coker, 1992) was used to provide the proportion of individuals of the most preferred tree species in relation the number of individuals of all other tree species present in the sample plots. Tree species abundance was quantified by tallying the number of stems of each species in different DBH size classes. Large canopy tree species or tree species greater than 10cm DBH were grouped into 10 cm diameter class. Species dominance Species dominance was used as a measure of the contribution of each individual species to the forest standing biomass (Owiunji, 1997). The assessment of standing biomass was measured by the basal area, measured at breast height, of individuals of a particular species. Species dominance was used as an index to ecological dominance and an indirect measure of biomass (Cunningham, 2000; Peters, 1994). Assessment of the regeneration of the most preferred tree species Regeneration assessment was used for quantifying the density of seedlings and saplings in the forest area adjacent to communities surveyed around Nyungwe Forest. The aim was to assess whether or not the canopy tree species that were most preferred were capable of producing seedlings and saplings in the Forest. Seedlings and saplings were recorded for their numbers. Quality of the most preferred timber and pole species A quick visual assessment of the straightness of the tree bole in each plot was made. Stem quality was evaluated on a six-point scale in terms of their suitability for utilisation as timber or poles. The quality of the stem was matched with the quality of the stem provided on Figure 2.4 below. Therefore, points 1 to 6 were used to mean stems of excellent, very good, good, fair, poor, and very poor quality stems. 38 Stellenbosch University http://scholar.sun.ac.za 1 = excellent; 2 = very good; 3 = good; 4 = fair; 5 = poor; 6 = very poor. Figure 2.4. Different levels ofstem quality used in the resource assessment, Nyungwe Forest. 2.2.3.3 Data analysis Tree species identification and composition A list showing the scientific and vernacular names of the tree species was made by compiling all tree species found in all forest areas next to the Communes being surveyed. Tree species composition was assessed for large trees (trees greater than 10 cm DBH), for medium trees (trees of 2.5 - 9.9 cm OBH) and for the seedlings and saplings (individuals ofO.lO - 1.30 m height and < 2.5 cm DBH) at the different study sites. Density and diameter size-class distribution Analysis of size-class distribution was limited only to the tree species most preferred by local communities. The selection of these species was based on the results of the household survey and on the relative frequency in the species composition data obtained from forest inventory. Diameters of the most preferred tree species were compiled into size-class histograms showing the distribution of the stem densities into different OBH size classes. Species distribution curves were generated to reflect the 39 http://scholar.sun.ac.za Stellenbosch University tree species demographic pattern by plotting DBH size classes versus the stem densities per hectare (ITFC-EMP, 1999). The resulting histograms reflected the current status of the Forest in relation to the most dominant canopy tree species. A sample provides an estimate of the population parameter of interest, as it is impossible to measure the entire population (Mead et al., 1993). Confidence limits define a range of values around the sample mean, which is expected to include the true population mean with a certain degree of confidence (Snedecor and Cochran, 1967). In this study, stem densities were calculated as a measure of abundance of the most preferred tree species for timber, poles and medicinal materials. The calculated confidence interval of 95% for the mean stem densities and the lower limit confidence interval were used as conservative estimates of the density of the resource. In contrast, stem densities at the upper limit of the confidence interval were considered abundant enough for utilisation. These confidence intervals were then used to determine the level at which the preferred tree species were abundant enough for utilisation. In their research on stand density management in the United States, Drew and Flewelling (1979) quoted by Davis and Johnson (1987) found that the stem density of 0.55 stems per hectare is the level at which total growth per hectare level out and competition-induced mortality of trees begins. As the density of 1.0 is an empirically fitted maximum in nature (Davis and Johnson, 1987), preferred tree species with a stem density of 1.0 stems per hectare and above were considered to have a high enough stem density to recommend them for utilisation. The guiding concept for the harvesting of timber, pole and medicinal materials was to reduce the density to a level such that the density of the preferred tree species will grow back to or slightly above the 0.55 stems per hectare at the next scheduled harvesting ofthese products. 40 http://scholar.sun.ac.za Stellenbosch University Species dominance Species dominance was measured by the basal area (G), i.e. cross-sectional area of an individual of a particular species measured at breast height (Philip, 1983). The following formula was used for the calculation of the basal area: G = 7t x ((DBH)2/4), where: G= basal area (m/ha); 7t = 3.14; DBH = diameter at breast height (cm). The contribution of each individual species to the forest biomass was measured by the relative dominance (D), which is defined as the cross-sectional area for each species divided by the basal area for all species. D = GsfGtob where; D = relative dominance or relative basal area (m2/ha); Gs = basal area for a given species (m2/ha); Gtot = basal area for all species (m2/ha); The calculated proportion was used to compare the contribution of each species to the total standing biomass. With a total index of 1.0 for all the species in the sample forest area, a tree species with a relative basal area of 0.01 and more was subjectively said to have a 'high contribution' to the forest standing biomass and that with less than 0.01 to have a 'low contribution'. Species of the latter category were immediately not classified with harvestable tree species. Regeneration characteristics The most abundant and dominant preferred tree species were compared in terms of corresponding available seedlings and saplings. Seedling and sapling densities of these species were compared by means of histograms showing the relative abundance of trees across the study sites. abundance of seedlings/saplings The relationship between tree species and relative was considered to be a measure of species Stellenbosch University 41 http://scholar.sun.ac.za regeneration and thus was used to predict the ability of the preferred tree species to produce significant resources. Quality of the boles of the most preferred timber and pole species Data on the quality of the stems were presented as percentages. The six levels of the quality of the stems were aggregated into three categories namely good, fair and acceptable. For the most abundant and dominant preferred tree species, histograms were constructed to provide a visual display of the percentage of individuals with good, fair and acceptable stem quality. Stellenbosch University 42 http://scholar.sun.ac.za 3 RESULTS 3.1 P ARTICIP ATORY RURAL APPRAISAL This section presents the information collected during interviews with individual households, small groups and key informants. The results of the participatory rural appraisal (PRA) assessments include the of the availability socio-economic of resources status of the in Nyungwe respondents, the Forest, their relative abundance estimates and the prioritisation of important resources to community livelihoods. a list of key institutions The section also presents that were acknowledged by the interviewees to be appropriate for better management of the utilisation of Nyungwe Forest. Also presented here are the experiences of individual key informants relating to Nyungwe Forest management problems and the strategies adopted to reduce the conflicts between local people and forest managers. In many cases, the total percentages of responses exceeds 100% because the respondents mentioned more than one item for examples the types of crops farmed, the resources available in Nyungwe Forest, the uses of resources, the types of conflicts and the strategies to avoid them, among others. 3.1.1 Socio-demographic characteristics In the study area, interviews were conducted in 200 individual households of which 161 (80.5%) were represented by men and 39 (19.5%) by women. Thus, most of the information was collected from men who were the heads of the households. Information on marital status indicated that 85.5% were married, 7.5% widows, 6.0% single, and 1.0% divorced. respondents. Figure 3.1 illustrates the age distribution of the 43 Stellenbosch University http://scholar.sun.ac.za Figure 3.1. Age classes of the respondents in the study area, n = 200 The majority (43.5%) of the households contained 5 - 7 members (Figure 3.2). This coincides with the national statistics, which indicate an average family size of five persons (MINECOFIN, 1998). Large households (JO or more members) were reported by 36% of the respondents. One of the reasons for large family size appears to be the hosting of the children orphaned as a consequence of the 1994 genocide. 2 - 4 merTtlers 5 - 7 merri>ers Figure 3.2. Family size of the respondents in the study area, n = 200 Table 3.1 shows the level of education of the respondents and indicates that the majority of the respondents had little or no primary school education. This was particularly the case in the Kagano and Mudasomwa Communes, where people had 44 Stellenbosch University http://scholar.sun.ac.za abandoned school education m order to engage in tea harvesting for income generation. Table 3.1. Education level of the respondents in the study area, n = 200 Educationlevel No school Some primaryschool Primary schoolcompleted Some secondaryschool Secondaryschoolcornpleted Total % 34.5 39.5 19.0 6.0 1.0 100.0 Other reasons for school abandonment, such as the lack of school fees, were expressed by 6% of the respondents. A very low percentage (1%) of the respondents had completed their secondary education, and none of the respondents had reached university level. This implies that people educated to secondary and university levels were living outside their native area, probably due to better employment and income generating opportunities in other parts of the country. As a result, less educated people were encountered in all the households interviewed. 3.1.2 Farm size and land ownership Land is a major form of capital for the production of crops, livestock and tree resources. Figure 3.3 shows the farm sizes of sample households in the study area. >2ha 11% 34% 1-2 ha 33% 0.5 - 1 ha 22% Figure 3.3. Farm sizes ofthe respondents in the study area, n = 200 Stellenbosch University 45 http://scholar.sun.ac.za Many respondents (35%) owned less than 0.5 ha of farmland. The majority (41%) obtained their land through state allocation (Figure 3.4). 50 40 <Il C 30 ... 20 0 ""0 C 0 C<Il 0 ....... 0 'cJ. 10 o Allocation by state Inheritance Purchase Purchase and inheritence Rent Illegal occupation Figure 3.4. Processes of land acquisition by the respondents in the study area, n = 200. Another 36% obtained their land through inheritance. Some respondents (15%) had purchased farms from neighbours, while a few respondents (6.0%) indicated that their land was both purchased and inherited. Where land was acquired through donation by the state, as in Musebeya Commune, the average farm size was 2 ha. Areas where land was largely inherited, for instance in Nshili Commune, the farm size was less than 1.0 ha. Previous research by MINAGRI (2000) found that the average farm size in Gikongoro Province, which includes three of the five Communes surveyed (Mudasomwa, Musebeya, Nshili), was 1.26 ha, comparatively higher than the 1.0 ha owned by 79% of all households in Rwanda. 3.1.3 Livelihood activities and sources of income Not a single respondent indicated only one livelihood activity. Most people combined food production and livestock farming with other activities such as beekeeping, charcoal making, pitsawying, handcraft, employment, commerce and mining. As a Stellenbosch University 46 http://scholar.sun.ac.za result, agriculture was a source of livelihoods to 99.5% of the respondents, and livestock keeping to 74.0%. Tree-related livelihood activities included pitsawying (9%) and charcoal making (4.5%) (Figure 3.5). Only a few respondents ranked the practices of beekeeping (one respondent) and mining (three respondents) in Nyungwe Forest as important activities for people's livelihoods. Beekeeping Mining Commerce Charcoal making Paid employment Handcraft Pitsawying Labour ••• Livestock keeping •••••••••••••• _ Food production o 20 40 80 60 100 % of responses Figure 3.5. Priority enterprises of the respondents in the study area, n = 200. Agriculture was a source of income to 75% of the respondents, while 40% of the respondents obtained cash from livestock keeping. Income earnings from tree products were reported by 10% of the households. About 30% of the respondents had off-farm income, and only 6% had paid employment. respondents A small percentage (2.5%) were engaged as casual labourers on neighbouring of farms, particularly in tea harvesting where tea factories were operational. Other sources of cash income included the sale of craft works, the sale of fruit and gold mining. Respondents also indicated that credit systems in the study area were not developed and, as a result, financial resources for many households (78%) were self-generated, with little or no access to bank credits. A few respondents (8%) had Stellenbosch University 47 http://scholar.sun.ac.za established a mechanism of a 'rotating fund' among group members. This system enabled each group member to be awarded a lump sum of money from the monthly contributions of all other members. The respondents highlighted the importance of this system in enabling members to cover high costs, including the expenditure for house building, the purchase of livestock and the payment of school fees. Despite the advantages of such schemes however, most (69%) of the respondents registered with any group or association. were not Other respondents (20%) were members of farmers' associations. 3.1.3.1 Cropfarming All the respondents managed their farms predominantly for the production of food crops. They grew traditional food crops, including sweet potatoes, Irish potatoes, bananas, maize, wheat, peas, and cassava, with sweet potatoes and maize being the most prominent in the study area (Table 3.2). The principal objective of crop farming was to generate enough food for domestic consumption and to meet other basic needs through income accrued from the sales of excess production on local markets. grown for food and cash income were predominantly Crops Irish potatoes, bananas and wheat. A few respondents grew other cash crops such as tea (7.5%, n = 15) and trees (4.5%, n = 9). Table 3.2. Important food and cash crops grown by the respondents in the study area. Crops Irish potatoes Maize Wheat Peas Beans Cassava Sweet potatoes Banana Vegetables Percentage Food crops 17.0 34.7 20.5 23.5 37.0 19.0 45.5 6.0 11.5 of respondents Food and cash crops 51.0 19.5 22.0 14.0 7.0 13.5 16.0 22.5 6.0 Generally, crop harvests were small, with the majority of the households harvesting less than five bags (one bag is approx. 100 kg) even in areas where the crops were most suited to soils and climate (Table 3.3). Irish potatoes, maize and wheat were 48 http://scholar.sun.ac.za Stellenbosch University common in Musebeya, Mudasomwa and Nshili. Roots and tubers (cassava and sweet potatoes) were preferred crops in Busozo, and banana production was specific to the agro-climatic conditions of Kagano Commune in the North-West of the study area. Factors that were said to influence the quantity at harvest included the non-use of improved inputs and the destruction of crops by wild animals. Table 3.3. Estimates of crop harvests by the respondents in the study area. Results are presented for important crops mentioned by more than 20% of the households. No. bags (1 bag = 100 kg) <1 1-2 3-5 6-8 9-10 > 10 No estimate Total 3.1.3.2 Irish potato 3.8 25.0 31.1 4.5 18.9 16.7 100 Maize Percentage of respondents Cassava Wheat Bean Peas 26.2 43.0 14.0 2.8 11.2 1.9 0.9 33.3 37.0 23.5 1.2 2.5 62.5 29.5 6.8 65.8 25.0 7.9 - - - - 1.1 1.3 100 100 2.5 100 100 11.6 40.6 24.6 5.8 5.8 1.4 10.1 Sweet potato 3.9 16.4 18.8 11.7 3.1 4.7 41.4 100 100 Banana 10.7 89.3 100 Other plant species grown Apart from crops, trees, shrubs and grasses were also farmed by many of the respondents. The presence of fruit, medicinal and fodder plant species is widespread in the survey area. Exotic fruit plants on farms included avocado (Persea americana), Japanese plum (Cyphomandra betacea), passion fruit (Pasiflora edulis), citrus (Citrus lemon), guayava (Psidium guayava), mangoes (Mangifera (Anona reticulata) and pawpaw (Carica papaya). 48.5% had planted avocado. indica), custard apple Of the households interviewed, Only 21.5% and 19.5% of the respondents had plum trees and passion fruit on their farms, respectively. The remaining fruit plant species were planted by fewer than five households. The respondents mentioned the presence of 20 medicinal plants traditionally used to treat humans and livestock on their farms (Appendix 2). These plant species were indigenous, and they were planted or conserved on farms. The most common were Iboza riparia (33.0%), Plecthranthus (13%). barbatus (20.5%) and Vernonia amygdalina 49 http://scholar.sun.ac.za Stellenbosch University Different fodder species were included as components of the farming system with the objective of controlling erosion. At the same time, they were used to feed domestic livestock during periods of fodder shortage. Many respondents (59%) established Setaria spp. on erosion control ridges or on terraces. Other fodder species mentioned were Panicum hochstetteri, Cynodon aethipicus, Tripsacum laxum, Digitaria hackelii and Triumfetta cordifolia. The elephant grass Pennisetum purpureum was present on 18.5% of farms and its leafy biomass and young plants were cut to feed livestock during the dry season of July to August. Horticultural crops grown in the study area were largely cabbages and amaranths. the 200 households interviewed, 39.5% produced cabbages, particularly Of in the swamps. Amaranths were present in the home-gardens of 24.5% of the respondents. The prevalence of these vegetables was attributed to an existing local market as well as to the need for food supplements within the households. Other vegetables and fruits of limited interest included carrots, pineapples, onions and courgettes. 3.1.3.3 Livestock farming Within the study area, livestock farming was complementary to food production. Different livestock types were present, but they were unevenly distributed among the respondents (Table 3.4). Table 3.4. Numbers and distribution of livestock types in the study area. No. of livestock None 1 2-4 5-7 8 - 10 > 10 Total Cattle 63.5 9.5 21.0 4.0 2.0 100.0 Goats 44.5 5.0 24.0 18.5 4.0 4.0 100.0 % of res20ndents Pigs Shee~ 59.5 78.0 6.5 12.5 18.0 7.5 11.0 2.0 5.0 Chicken 89.0 1.0 4.0 3.5 1.5 100.0 100.0 100.0 Rabbits 93.5 2.0 1.5 1.5 1.5 100.0 Goats (56.5%) and sheep (40.5%) were the most commonly kept livestock among the respondents, while only 36.5% kept cattle. Pigs were important in the study area and were kept by 22% of the people interviewed. Chickens and rabbits were also kept. Some households kept all the types of livestock, while others had none. The purpose of livestock farming was predominantly for manure production, except in the case of Stellenbosch University 50 http://scholar.sun.ac.za rabbits and chickens, which provided about 30% of the respondents with protein. Income was also an important aspect of livestock keeping, especially in the case of sheep and goats. Many households obtained an income from sales of goats (70.8%), sheep (56.3%) and pigs (44.4%). About 22% of the households obtained cash from cattle through the sale of unproductive cows, calves or milk. Of the households interviewed, 60 to 80% used manure from cattle, goats, sheep and pigs on their farms owing to the low soil fertility caused by erosion and continued cultivation. For better productivity, when affordable, manure or compost combined with inorganic fertilisers including NPK and lime, were used. Grazing practices by the respondents are shown on Figure 3.6. consisted predominantly Grazing practices of free range grazing (67.7%) and zero grazing (32.3%). During the dry season, fodder plants were collected everywhere, from wetlands and swamps, as well as from farms (terraces and erosion control ridges). Croplands Household residues Woodlots and forests Along roads Fallow farms •••••• Erosion control ridges •••••••••••••• Wetlands ••••••••••••••• o 5 10 15 20 % of responses Figure 3.6. Sources of fodder during the dry season in the study area. 25 51 Stellenbosch University http://scholar.sun.ac.za 3.1.3.4 Tree crops Apart from fruits and medicinal plants present on farms, 66.3% of the households owned plantations of different sizes (Figure 3.7), usually as small woodlots of less than 0.5 ha. The farm sizes of the respondents showed a strong positive relationship with tree crop area (pearson's correlation r = 0.35, p<O.OOl, n = 200). The planting of Eucalyptus spp. was common, alone or in combination with other tree species, namely Pinus patuia and Grevillea robusta (Figure 3.8). 1.0·2.0 ha 4.0% 0.5 • 1.0 ha 13.1% None 33.7% < 0.5 ha 49.2% Figure 3.7. Plantation areas owned by the respondents in the study area. 60 ..... Cl) c Il) "0 C 0 0Cl) 50 40 30 .... Il) 4-< 0 ~ 20 10 0 2 3 4 5 6 7 Planted trees 1 = Eucalyptus spp.; 2 = None; 3 = Eucalyptus spp. and Pinus patula; 4 = Cupressus lusitanica; 5 = P. patula and C. lusitanica, 6 = Arundinaria alpina, Cypress and Hagenia abyssinica 7 = Eucalyptus spp. and Grevillea robusta. Figure 3.8. Tree species planted by the respondents in the study area. Stellenbosch University 52 http://scholar.sun.ac.za The majority of people (54.4%) indicated that they planted trees as their main source of building poles. Those who planted Pinus patula, Cupressus lusitanica and Grevillea robusta (10.3%) aimed to produce timber by pitsawying, as they recognised an increasing market demand for this product. Trees were converted to fuelwood by 28.7% of the respondents and to charcoal by 9%. Charcoal was a common product for sale. Others said that money was obtained from the sales of poles. On average, income from tree products for many households did not exceed 10000 Rwandan Francsl (RWF) per year (Figures 3.9 through 3.11). 40 <Zl ..... s:: 11) '0 s:: 30 0 0... <Zl 11) I-< ~ 0 20 'cf- Annual income, RWF Figure 3.9. Annual income from the sale of charcoal by the respondents in the study area 1 At the time of writing, 1US$ was equivalent to 473 RWF Stellenbosch University 53 http://scholar.sun.ac.za 30 Vl ..... s:: <l) "'0 s:: 0 0.. Vl 20 .... <l) 4...; 0 ~ 10 Annual income, RWP Figure 3.10. Annual income from the sale of lumber by the respondents in the study area. 40 Vl 1:: <l) "'0 s:: 30 0 0.. CIl <l) .... ~ 0 20 ~ Annual income, R WF Figure 3.11. Annual income from the sale of building poles by the respondents in the study area. Stellenbosch University 54 http://scholar.sun.ac.za 3.1.4 Soil fertility status and management Soil fertility in the study area was a key concern to many households. When asked to rate the fertility status of their farms, 55.5% of the respondents said it was moderate, while others considered the fertility of their farms as very poor (18%) and poor (16.5%). Many people (85%) reported that soil fertility decline in the study area was a major impediment to the productivity of crops. Many causes of soil fertility decline were cited, but inadequate use of manure, fertiliser and compost was a crosscutting problem among all the Communes surveyed (Table 3.5). Table 3.5. Causes of soil fertility decline listed by the respondents area. Causes of soil fertility decline Lack of use of manure, fertiliser and compost 'Senility' of land Continued cultivation, no fallow Erosion Soil acidity Lack of use of terraces Lack of use of improved crop varieties Small farm size Inadequate agricultural practices Overgrazing No. responses 131 64 33 20 9 7 6 4 3 1 in the study Percent of responses (%) 65.5 32.0 16.5 10.0 4.5 3.5 3.0 2.0 1.5 0.5 Farmers' strategies to improve the fertility status of their farms were varied, but the majority (53.5%) were using manure, fertiliser, compost or a combination of these. The respondents with land shortage problems said that they were renting land from the neighbours (22%) or that they served as casual labourers on neighbours' fields (15%). The remaining respondents had adopted various solutions to cope with low fertility, including the practice of natural fallow, the use of terraces to prevent soil erosion, the registration as farmer's association members and the use of improved seeds. 3.1.5 Respondents' perception of Nyungwe Forest resources The 200 respondents listed 18 categories of resources present in Nyungwe Forest. Of the 18 resource categories, only 14 were said to be important for the respondents' Stellenbosch University 55 http://scholar.sun.ac.za livelihoods. Table 3.6 shows these resources, their abundance estimates and their uses by the respondents. The most important products from the Forest, according to more than 50% of the interviewees, were land, timber and minable gold. Nyungwe Forest was seen as a source of fodder and pastures but by few respondents (16%). Appendix 3 shows a list of palatable tree and grass species mentioned by 5 or more respondents. The most mentioned were Panicum hochstetteri, Triumfetta cordifolia, Mariscus thomaiophyllus and Digitaria hackelii. Table 3.6. List of important resources for community's livelihoods in the study area, n = 200. Resources % respondents Abundance estimate Land 68.5 D (49) Timber 58.5 D (45) Gold mine 56.5 F (72) Fodder and pastures 16.0 F (8) Wooden and woven item 13.5 D (21) Fuelwood 11.0 D (18) Poles 10.5 D (20) Employment 9.0 A (5) Honey (beekeeping) 7.5 A (6) Charcoal 3.5 NM Medicinal plants 3.5 A (5) Rain and humidity 3.5 NM Ecotourism 2.5 NM Game-meat 1.0 F (5) A = Abundant; D = Dominant; F = frequent; NM = not mentioned S = subsistence; C = commercial; SIC = subsistence and commercial ( ) indicates the percentage of responses for each resource. The perception of resources Uses SIC (75.4) C (62.9) C (89.7) SIC (59.4) SIC (48.3) S (78.3) S(74.1) S (90) C (71.4) C (87.5) C (62.5) NM C (100) S (77.8) in the Forest by individual interviews compared favourably with that of the group participants in all the study sites. On the whole, 21 resources were listed and their abundances estimated (Table 3.7). Timber, poles, fuelwood and beekeeping were the major abundant resources common to all study survey sites. However, land availability was a controversial resource on which the participants had a long discussion before a consensus was reached. Land availability was perceived differently in terms of its size and soil fertility. For some participants, the land area was large but not productive enough as the soil fertility declined after one to two years of continued use. Others classified land as abundant because there existed fertile wetlands in the Forest. A similar debate occurred again for the participants' perceptions of indigenous timber trees. Some claimed that they did not know what was in the Forest, as they were not allowed to go there. Others said that timber trees were abundant because the Forest was protected, was regenerating and no pitsawying was carried out, as it was in the past. Stellenbosch University 56 http://scholar.sun.ac.za Table 3.7. Assessment of the availability of resources in Nyungwe Forest by groups in the study area. Resources Gold Timber Land Medicinal plants Charcoal Ecotourism Fuelwood Wildlife Rain and humidity Goods Poles Employment Beekeeping Pastures Bean stakes Water Vegetables Mushrooms Wild honey Fruit Bushrneat Musebeya F 0 D 0 D F 0 0 0 0 0 D D 0 D 0 R R R R F Nshili F D F 0 NM NM F D D 0 0 F D 0 NM D R R NM NM NM Kagano A D D A F NM 0 A D F D 0 0 NM NM D R R NM R F Mudasomwa F D 0 F D NM D A 0 F D A A A NM D R R R R 0 Busozo F F A 0 NM NM D A D NM A A D F NM D 0 0 0 R F D = Dommant; A = Abundant; F = Frequent; 0 = Occasional; R = Rare; NM = Not mentioned From household interviews, only a few respondents indicated the availability of building poles (10.5%, n = 21), fuelwood (11%, n = 22), medicinal plants (3.5%, n=7) and charcoal (11%, n = 22) in the Forest. Ecotourism potential was a resource mentioned by a very few people, only 2.5% of the respondents, and only 9% identified the potential of the Forest to provide local employment. The Forest's ecological functions, such as its influence on local climate, were cited as resources by 3.5% of the respondents. 3.1.6 Resource preferences and ranking This exercise enabled the participants to list the resources that they thought would be of value to their livelihoods if they were given access to use them. In general, however, as some resources were mentioned and listed, the participants from some groups reacted by commenting that their preferences could not be met, as the Forest was not theirs but for the State, the Rwanda Office of Tourism and National Parks (ORTPN) and the foreign tourists. Stellenbosch University 57 http://scholar.sun.ac.za Preferred resources were ranked using both discussion and the pairwise ranking of importance of resources to communities. Table 3.8 shows the ranked order of the resources in order of their importance in relation to each other. The most important resources listed by the participants were land, timber and gold, but the ranking appeared to be dependent on the agro-ecological characteristics of the population. zone and the socio-economic Land preference ranked high in areas where farm sizes were small, as in N shili and Kagano, but not in Musebeya where people had an average farm size of 2.0 ha. Whereas timber was the first priority in Mudasomwa, gold mining was important for the community in Busozo. These preferences partly explain the problems encountered by the local authorities and the conservation services with regard to the illegal use of the Forest. Deforestation for arable land was a severe problem in Busozo, particularly in the Bweyeye Sector. Table 3.8. Ranking of resource categories in the study area Rank order 1 2 3 4 5 6 7 8 9 10 11 12 l3 14 Musebeya Nshili Kagano Mudasomwa Busozo Land Timber Fuelwood Pastures Employment Poles Ecotourism Charcoal Beekeeping Rain Gold Medicinal plants Goods Bushmeat Land Timber Gold Bamboos Pastures Poles Fuelwood Employment Beekeeping Charcoal Bushmeat Land Gold Timber Poles Employment Fuelwood Goods Beekeeping Pastures Medicinal plants Bushmeat Timber Gold Land Charcoal Fuelwood Beekeeping Medicinal plants Bushmeat Goods Gold Land Timber Pastures Employment Rain Poles Bushmeat In Nshili and Kagano, available information suggested that both the buffer zone and Nyungwe Forest were used to produce food crops. Food production and hunting in the Forest were mostly practised by about 1 000 indigenous people, Batwa, who lived inside the Forest. Unlawful users of the Forest were given fines or were jailed. At Kagano, the illegal use of the Forest and the buffer zone was condemned by a fine ranging from 10 000 RWF to 20 000 RWF. Strategies to reduce the rate of deforestation were many and included the management of the largest wetlands in the environs of the Forest, namely Kamiranzovu and Uwasenkoko. Other suggested ways of reducing the conflict from land problems were the delimitation and management of Stellenbosch University 58 http://scholar.sun.ac.za pasture lands, the intensification of agriculture involving the use of erosion control methods and fertilisers, and the establishment of projects that could provide local employment. Timber was the first priority needed in Mudasomwa, second in Musebeya and Nshili and third in Kagano. Although the participants denied the harvesting of timber from the Forest, The Nyungwe Forest Conservation Project (PCFN) and the local authorities recognised illegal logging of commercially important timber tree species in all the Communes being surveyed. They indicated that illegal logging was taking place throughout the year, with intense activity during the dry season. Timber was sold locally, outside the area or in neighbouring countries such as in the Republic of Burundi that borders with Nshili and Busozo Communes in the South of the Forest. The local authorities and the conservation services were much concerned about charcoal making and fuelwood gathering in Nyungwe Forest and in the buffer zone. Proposed solutions for reducing the conflict over illegal timber exploitation, charcoal making and fuelwood gathering included among others: • Licensed harvesting and pitsawying of mature or dead trees in specified areas of the Forest; • Harvesting timber trees in the buffer zone; • Privatisation of the management of Nyungwe Forest; • Distribution of seedlings communities lil of fast growing preferred tree species to local order to enable them to establish their own sources of tree products; • Revival of forestry projects. Gold mining was the highest priority lil Busozo, followed by Kagano and Mudasomwa. At Busozo, mining was an alternative form of livelihood, and the group participants indicated that it was the only source of income for the majority of households, whereas people in Kagano and Mudasomwa were getting income from tea factories. For instance, the Tea Factory at Gisakura in Kagano Commune employed annually approximately averaged 40 million RWF. 5 000 casual labourers for which the salaries 59 http://scholar.sun.ac.za Stellenbosch University The Rwanda Office of Tourism and National Parks suspended gold mining because the activity led to the deviation and pollution of rivers and loss of biodiversity of the Forest. Mining in the valleys in the vicinity of the Forest resulted in the cutting of trees, erosion, hunting, agriculture and the sale of beer in the Forest. According to the conservation services, mining rights in bottomlands and valleys could only be granted to organised groups and the activity guided by a memorandum of agreement. Except for the high rankings given to land, timber and gold resources, preferences for other resources ranked variably, with exception of bushmeat which appeared to be of less interest to all five communities. Bamboo, Arundinaria alpina, bamboo ranked among the most important resources at Nshili compared to other study sites. This indicated the extensive use of bamboo collected without permit from the Forest for house construction and for basketry material. Groups in Musebeya, Kagano and Mudasomwa plants as important resources. Communes identified medicinal Current use of the resource was permitted by the Nyungwe Conservation Project (PCFN) and had shown little impact to the Forest. The practice of beekeeping was also considered important, particularly where beekeepers' associations were formed. PCFN restricted the activity to the buffer zone and in open areas inside Nyungwe Forest. Beekeeping, however, was reported to result in runaway fires in all the study sites. Conservationists estimated that the possibility of runaway fires could be addressed through the formation of beekeeper's associations. Beekeeping was practised during the period of April to August when tree species favoured by bees were in flowers. On the local market, honey was sold at 500 RWF to 800 RWF per kilogram. Honey was consumed as food, as an ingredient of beer, and wax was used for the local production of candles. 3.1.7 Tree species preferences and uses by the respondents in the study area 3.1.7.1 Timber tree species Hardwood tree species from Nyungwe Forest were preferred for their sizes, durability and wood colour for different uses. Table 3.9 shows the most commonly used tree species. Other preferred timber species are presented in Appendix 4. 60 http://scholar.sun.ac.za Stellenbosch University Table 3.9. Tree species most preferred for timber by the survey respondents around Nyungwe Forest, n = 200 Only tree species mentioned by 10 or more respondents are presented. Tree species Percentage of responses 64.0 56.0 38.0 15.5 13.5 12.0 11.5 10.0 10.0 8.0 7.5 6.0 5.0 5.0 Podocarpus latifolius Entandrophragma excelsum Podocarpus falcatus Symphonia globulifera Parinari excelsa Faurea saligna Carapa grandiflora Dasylepis racemosa Syzygium parvifolium Hagenia abyssinica Ocotea usambarensis Newtonia buchananii Polyscias fulva Strombosia scheffleri A range of timber species was mentioned by the respondents as being used for the production of wooden items. Generally, goods were obtained from the following key tree species: Polyscias fulva (a musical instrument, locally known as 'inanga'), Ocotea usambarensis (household utensils, beer boat), Markhamia lutea (household utensils), Carapa grandiflora (household utensils), Dasylepsis racemosa (beer boats) and Asparagus falcatus (beer boats). Other wooden and woven goods were obtained from non-timber forest species and included bamboo Arundinaria alpina, different climbers (Smilax anceps, Urera cameroonensis, U hypselodendron) and shrubs (Triumfetta cordifolia). 3.1. 7.2 Building poles The preference of tree species for poles in the study area was based on the quality of the stems. Major characteristics of suitable trees for poles included stem straightness, durability and a suitable diameter of at least 10 em. Tree species preferred for quality building poles are given in Table 3.10. Poles were used for house building, for fencing of livestock and the home compound. Bamboo was also used for building particularly in Nshili Commune next to the forest area where it occurs. 61 http://scholar.sun.ac.za Stellenbosch University Table 3.10. Tree species whose stems are preferred study area, n = 200. for building poles in the Only tree species mentioned by 10 or more respondents are presented. % of responses Tree species 36.0 Macaranga neomildbraediana 28.5 Cassipourea ruwenzoriensis 26.0 Strombosia scheffleri 13.5 Syzygium parvifolium 12.5 Carapa grandiflora 12.5 Casearia runssorica 6.5 Olea hochstetteri 6.5 Arundinaria alpina 6.0 Parinari excelsa 5.5 Alchornea hirtella 5.5 Podocarpus latifolius 5.5 Oxyanthus speciosus 5.0 Faurea saligna 5.0 Cassipourea gummiflua 5.0 Mil/etia spp. 5.0 Symphonia globulifera 5.0 All tree species Only 9.5% of the respondents indicated that Nyungwe Forest was a source of building poles in the past, where quality-building poles were abundant. Even now, of the 25 respondents who mentioned building poles as one of the most important resources for their livelihoods, 80% (n = 20) categorised the availability of building poles in the Forest as very abundant. The majority of the respondents (52.1 %) said that poles occurred dominantly everywhere in the Forest. Most people (74.1 %) estimated that access to Nyungwe Forest would satisfy their subsistence needs for building poles. 3.1. 7.3 Medicinal plants Respondents listed 45 medicinal tree species found in Nyungwe Forest (Appendix 5; Table 3.11). Products from these plants particularly the bark, were traditionally used by medical practitioners to treat people or livestock. Bark material was used from 56.5% (26 out of 45 species) of the medicinal plants mentioned by the respondents. For some species, both bark and leaves could be used as sources of medicine (Table 3.11). Trees are the major life form category of the medicinal plants, with the exception of Asclepiadaceae scandens which are climbers. spp., Begonia meyeri-johannis and Sericostachys Stellenbosch University 62 http://scholar.sun.ac.za Most people reported that medicinal plants were used locally to treat intestinal parasites in people and livestock. were mentioned There is no commercial trade in medicines. Men as the major harvesters of medicinal plants by 67.4% of the respondents, while 30.2% said that both men and women harvested medicinal plants. Table 3.11. Plant species used for medicinal purposes in the survey area, n = 200. Onl~ Elant sEecies mentioned b~ 3 or more resEondents are Qresented. Part used Local name Plant sEecies Bark Umushwati Carapa grandiflora Bark, roots Umwumba Prunus africana Bark, leaves Syzygium parvifolium Umugote Bark Zanthoxyllum gilletii Umuturirwa Bark Croton megalocarpus Umunege Dichrostachys cinerea Umunkamba Leaves Fagaropsis angolensis Umugomera Bark Ocotea usambarensis Bark Umutake Rapanea melanophloeios Bark, leaves Uruneke Ocotea michelsonii Umuganza Bark Mitragyna rubrostipulosa Leaves, bark Umuzibaziba Vernonia amygdalina Umuravumba Leaves Asclepiadaceae div. spp. Indarama Tubers Canthium oligocarpum Bark, leaves Umushabarara Cyathula polycephala Igifashi Leaves Maesa lanceolata Leaves, roots Umuhanga Begonia meyeri-johannis Leaves Irebe Xanthoxyllum usambarensis Intareyirungu Bark Symphonia globulifera Bark, leaves Umushishi Pentas decora Leaves, bark Isagara Brillantaisia cicatricosa Icyunga Bark Faurea saligna Umutiti Bark Myrica kandtiana Isubyo Wood Tabernamontana johnstonii Umuronzi Leaves, bark Canthium g/abriflorum Imvuvu Bark No. resEondents 51 27 22 16 16 10 9 8 8 8 7 7 6 6 6 5 5 5 3 3 3 3 3 3 3 The respondents varied widely in their estimates of the amount of medicinal plants that could be collected from the Forest. From a total of 129 respondents who knew about medicinal plants, 44.2% (n while 36.4% (n = = 57) estimated the amount collected to be small 47) said moderate quantities were harvested. Forty-four percent of the respondents said that medicinal plants were found everywhere in the Forest, while 26% (n = 35) located many medicinal plants in the bottomlands. The remaining respondents found medicinal plants in several different locations, including valleys, disturbed areas, and hilltops. Medicinal plants such as Vernonia amygdalina, Iboza riparia, Mitragyna rubrostipulosa, and Lantana trifolia were also found outside Nyungwe Forest in disturbed areas and on fallow farms. Stellenbosch University 63 http://scholar.sun.ac.za 3.1.8 Uses of wild foods Wild foods collected from the Forest (Table 3.12) are used either for subsistence or as a source of modest income from their sales on the local market. They are seasonally obtained from the Forest, most frequently during the wet seasons. Table 3.12. Wild food resources found in Nyungwe Forest, as mentioned by the respondents in the study area, n = 200. Food Percentage of responses 88.5 14.5 32.0 66.5 11.5 Fruit Game Mushrooms Vegetables Wild honey The most important edible fruits to the respondents are Myrianthus holstii (88.5%, n=l72) and Rubus rigidus (17%, n = 34). Other wild fruits mentioned by the respondents are Cyphomandra betacea (5.5%, n = 11), Physalis peruviana (5.0%, n = 10), Passiflora edulis (7.5%, n = 15) and Citrus lemon (0.5%, n = 1). These were all reported to be available in Nyungwe Forest from seeds dispersed from plants originally established on farms. Edible wild mushrooms identified by their local names included ibizinu, ibihumyo, imegeri, utwoba, intyabire and ubushikiri. According to the respondents, these mushrooms are commonly growing in association with termite mounds or with dead trees. Mushrooms provided seasonal income from sale in local markets. Prices for one mushroom ranged from 10 RWF to 50 RWF, depending on the location of the market. Mushroom gathering was practised mostly by women and children. Wild vegetables from Nyungwe Forest are another seasonally available resource that was used as a dietary supplement or as a source of small income. whose leaves are used as food include different Amaranthus nigrum. Annual herbs spp. and Solanum Local names of some Amaranthus spp. were bwizabwishyamba, umuhuruza and umushokorankoko. inderama, The vegetables were found in various locations in the Forest. Most respondents mentioned their occurrence in swamps (39.8%, n = 53), and in disturbed areas, including forest areas that had been cleared for agriculture and 64 http://scholar.sun.ac.za Stellenbosch University burned areas (28.6%, n = 38). The wild vegetables were mostly gathered for subsistence use, but a local market for Solanum nigrum did exist. For this vegetable, out of 75 respondents, 41.5% suggested a market price of 5 to 10 RWF per handful. As for other NWFP from Nyungwe Forest, the major collectors of wild vegetables were children and women who collected small amounts at a time. Honey is a wild food that can be found in the Forest. A few respondents (11.5%, n = 23) mentioned its availability from ground or tree-nesting bees known locally as inkura. Wild honey is collected from hollow tree trunks or branches as well as underground. Wild honey is used solely for household consumption. Honey production was listed as an important activity in the Forest, and associated with the availability of plant species favoured by bees and hills inside the Forest that provided the best sites for placing hives. Forest plants that were important sources of nectar and pollen to bees were listed as Sericostachys scandens, Brillantaisia cicatricosa, B. nitens, Mimulopsis excellens, Syzygium parvifolium and Vireetaria major. The areas preferred for hive placement were those with high densities of flowering trees, climbers and herbs. Bushmeat was of value to adjacent communities. Animal species were hunted for subsistence. Other wild foods, consisting of two tuberous species, were mentioned by only a few respondents. Discorea spp. (2%, n = These were Afromomum angustifolium (2%, n = 4) and 4), reportedly gathered by poor people and the indigenous people Batwa, who were living inside the Forest. The low numbers of users suggests that the two wild foods were not commonly of value in all the survey sites. 3.1.9 Respondents' knowledge of wildlife species The respondents mentioned 32 wildlife species and estimated their abundance in the Forest. Their presence was determined by the noise of wildlife from the Forest or observations inside or outside the Forest. Commonly cited wildlife included crop raiding animals such as bushpigs (82%, n = 164), chimpanzees (48%, n = 97), grivet monkeys (25.5%, n = 51), black and white colobus monkeys (44.5%, n = 86) and olive baboons (73%, n = 146). Stellenbosch University 65 http://scholar.sun.ac.za Wild animals caught for domestic meat were listed by many respondents. Respondents perceived most of the crop raiding animals to be abundant because they were easily seen in groups on farms or heard from the Forest. According to many respondents (25% to 50%), dominant wild animals consisted of monkeys, baboons and colobus. A few wild animals were estimated to be rare in the Forest: elephant (Loxodonta africana) (92%, n = 23), buffalo (Syncerus caffer) (80%, n = 4), civet (Viverra civetta) (54.5%, n = 6) and bushbuck (Tragelaphus scriptus) (55.6%, n = 15). Generally, edible animals were said to be few. Animals hunted for subsistence included the black-fronted duiker Cephalophus nigrifrons (98.2%, n = 167), the yellow-backed duiker C sylvicultor (99%, n = 102), the bushpig Potamochoerus poreus (100%, n = 127), the African elephant Loxodonta africana (36.5%, n = 35), the African buffalo Syncerus caffer (98.7%, n = 75), and the bushbuck Tragelaphus scriptus (98.9%, n = 86). Many other small animals, including the Gambian rats, rabbits, a bird species identified by its local name igikoyi, and a small mammal known locally as inkezi, were valuable sources of animal proteins. Of the 200 respondents, only 17 said that monkeys were hunted from the Forest either for food (76.5%) or for reducing the population of crop raiding animals (23.5%). Hunting was practised in all seasons. Of 183 people who answered, 71% estimated that hunting was done throughout the year, although a few argued that hunting is frequent during the rainy season because animal footprints are visible on the ground. Commonly cited hunting methods included dogs (22.3%, n (29.3%, n = 55), and snares (23.4%, n = = 42), dogs and spears 44). According to the people interviewed, two factors, namely the reproduction rate and the 'no hunting rule', contributed to the change in animal population size of wild animals in the Forest. Compared to the past, an increase in the population of monkeys was observed by 40 % of the people interviewed, baboons by 24.5%, bushpigs by 8%, colobuses by 6.5% and chimpanzees by 4.5%. Chimpanzee's populations particular were thought to be increasing because they are not eaten. in Where the number of wild animals was perceived to have decreased, this decrease was attributed to hunting. Respondents perceived a reduction in numbers of black-fronted duiker (C nigrifrons), yellow-backed duiker (C sylvicultor), and porcupine Atherurus africanus. Stellenbosch University 66 http://scholar.sun.ac.za The hunting for meat, skins, teeth, horns and medicines caused the local extinction of some wild animals as recognised by the respondents (Table 3.13). Table 3.13. Respondents' Forest, n = 200. knowledge of locally extinct wild animals in Nyungwe English names Elephant Buffalo Leopard African lion 3.1.10 Current Latin names Loxodonta africana Syncerus caffer Panthera pardus Panthera leon % of respondents 60.5 30.5 8.5 6.0 resource access and use The majority of the respondents (97.5%) said they had no access to resources in Nyungwe Forest. A few individuals (3%) admitted access to cut bamboo (Arundinaria alpina) from the Forest. In Nshili Commune in the south-eastern zone of the Forest, more than five persons said they were aware of the regulations but that, due to poverty and lack of alternatives, they kept on harvesting bamboo from the Forest. Others perceived that they benefited from the ecological function of the Nyungwe Forest in their areas, even though they were not permitted to access land and other resources in the Forest. In the buffer zone, however, the respondents collected dead branches for use as fuelwood (31.5%) and grasses (23%) for fodder. Agriculture in the bottomlands in the environs of the Forest was practised by 11.5 % of the interviewees. Others used the buffer zone as a settlement site because they argued that they had not been compensated for leaving the forest plantations. 3.1.11 Respondents' perception of conflicts and conflict resolution In the study area, the respondents listed four types of conflicts arising from their cohabitation with the Forest (Table 3.14), the major conflicts being between local people and either the conservation services or the government. These conflicts occurred when the Nyungwe Forest Conservation Project (PCFN) and the extension services evicted the local people who were encroaching respondents also accused the Rwanda on the Forest. Office of Tourism and National Some Parks (ORTPN) of employing outsiders as forest guides and rangers. Conflicts between the local people and the government were attributed to lack of government decisions Stellenbosch University 67 http://scholar.sun.ac.za permitting the use of resources in the Forest by neighbouring communities. The respondents adjacent to the Forest were mostly concerned about their crops that were raided by wild animals. While illegal users of the Forest were severely sanctioned, respondents claimed that in contrast they were not compensated for crops damaged by wild animals. Table 3.14. Nature of conflicts listed by the respondents 200. in the study area, n = % of respondents Conflict Local people vs. government Local people vs. wildlife Local people vs. conservation Local people vs. militia Local people vs. extension service Within local people 45.5 24.5 48.0 20.5 2.5 0.5 Twenty percent of the survey respondents reported that, during the last two weeks, armed bandits coming from the Forest had threatened their lives and stolen their livestock. This was an uncommon problem, and respondents proposed the use of force for its resolution. Apparently, tensions or competition among local villagers over Nyungwe resources did not exist within the community surveyed (Table 3.15). Table 3.15. Importance of conflicts over resource use and access, as rated by tbe respondents in the study area. Conflict Not important n 107 50 97 41 5 Local people vs. government 6.5 Local people vs. wildlife Local people vs. ORTPN* 21.6 Local people vs. militia 2.4 Local people vs. agricultural 60.0 extension services Within local people 1 100 * Rwanda Office of Tounsm and National Parks. Importance of conflict Important Very important 9.7 6.0 23.7 12.2 - 12.7 8.0 23.7 7.3 40.0 - - Most important 71.0 86.0 30.9 78.0 - Only 89 respondents suggested strategies for the resolution of local - government conflict. Of these, approximately 24% suggested consultation between the local people, the government authorities and the conservation services in order to establish resources that could be used. Other commonly mentioned strategies included creation of job opportunities in the area, the demarcation of a use-zone where resource use Stellenbosch University 68 http://scholar.sun.ac.za would be permitted, the use of fines, fencing of the Forest to avoid crop destruction by wild animals and compensation for damaged crops by wildlife. 3.1.12 Respondents' perception of past and present forest management Of the households interviewed during the survey, 85.7% said that the management of Nyungwe Forest was under the responsibility of the state. When asked what role the state played in forest management, 62.5% of the respondents said that the state was involved in forest protection activities. Twenty percent thought that the management of the Forest by the state had encompassed the allocation ofland to local people in the Forest adjacent areas. Only 11.7% believed that the Forest belonged to everyone, as access was open to all. Harvesting rights were said to be exclusive to the state and the colonial authority, and the provision of harvesting licenses was recognised by 16% of the respondents. At the local level, agricultural extension services and the local administration were reported to enforce regulations and to sensitise local people on the importance of conserving Nyungwe Forest. 3.1.13 Respondents' views of appropriate institutions for forest management Community organisations (Wild and Mutebi, 1996). are important for collaborative management of forests Survey respondents were asked to name institutions in which they had trust and confidence (Table 3.16) and to give reasons for their answers. Table 3.16. Proposed institutions for management by the respondents, n Institutions District Cell ORTPN Local protection committee Sector MINAGRI* Local administrative committee Developmental Project Government Local farmers' associations and co-operatives Total * Ministry of Agriculture, Livestock and Forestry. % of respondents 24.0 16.5 12.5 10.5 10.0 7.5 6.5 6.0 3.5 3.0 100.0 = 200 Stellenbosch University 69 http://scholar.sun.ac.za Twenty-four percent were of the view that the District Administration was the most appropriate institution that could enable community use of resources and manage the Forest sustainably because of the power and authority that the District holds over all development issues in the region. Many argued that local concerns were best known at District level. Other reasons for the choice of District as the appropriate organisation included: • The capability of the District to provide employment to local people; • The ability to mobilise enough resources for management and protection; • The part of the Forest near to the respondents lies within the District geographical zone; • The leadership at District level was elected by local people; • The District had a vested interest in the conservation of Nyungwe Forest, more than any other organisation. Similar institutions Communes surveyed. were identified during interviews with groups in the five Generally, institutions in which local people had confidence were local and concerned with local administration and development (Box 3.1). Box 3.1. Proposed institutions for management of Nyungwe Forest utilisation District Administration. The District based administration has been introduced in Rwanda very recently within the framework of good governance and decentralisation of administrative authority. A District is geographically based and comprises one or more adjacent exadministrative Communes. The leaders of the Districts were elected locally in order to manage the affairs of the Districts and to coordinate activities in the administrative Sectors. Administrative Sector. This is a lower level of the District Administration. It is based on the administration of at least 250 households within the same geographical area. The leaders at this level consist of an executive committee of nine people elected on their merits by local people. Each member has specific responsibilities in the sector, including health, education, communication, and security, among others. The sector Coordinator is the head of the Sector and he is assisted by a secretary. Administrative Cell. This is a further subdivision of the Sector and refers to an administrative area occupied by at least 50 households. The structure of the Cell is similar to that of the Sector, with the community leadership elected locally and its representation found in women, youth and men. Nyungwe Forest Conservation Project (PCFN). A conservation project managed by ORTPN and funded by USAID and WCS. It was established in 1988. It deals with conservation, environmental education and research on fauna and flora in Nyungwe Forest. Stellenbosch University 70 http://scholar.sun.ac.za The interviewees perceived key roles that these institutions should play with regard to resource use and access in Nyungwe Forest. Below is a compilation of the roles that came out from household and group interviews: • To participate in protection and conservation ofNyungwe Forest; • To employ local people in pitsawying mature trees and in enrichment activities on degraded and burned areas in Nyungwe Forest; • To negotiate access to resources from the highest authorities and to participate in the organisation of associations for resource utilisation (for example, associations of beekeepers, gold-miners, herbalists, and basketry makers); • To identify and delimit use areas in the part of the Forest adjacent to communities, including land for agricultural development; • To raise awareness on the importance of conserving Nyungwe Forest; • To provide technical and financial assistance to organised resource users; • To provide incentives for protection; • To resolve conflicts over resource use and access. Individual responsibility for forest management was assessed by asking the respondents what they could do to sustain Nyungwe Forest resources. Answers were many and diverse. The responses in Table 3.17 show that the respondents have an interest in having access to resources in the Forest. Table 3.17. Individual responsibility of the respondents for sustainable forest management in the study area, n = 200. Activity Participation in protection and management Respect of laws, regulations and guidelines Reporting offenders Non-destructive use of the Forest Harvesting mature trees and enrichment planting Participation in awareness raising Use of demarcated use zone Selective harvesting of mature trees Rotational harvesting and enrichment planting Enforcement of patrol % of respondents 29.0 21.0 21.0 19.5 11.5 9.5 8.5 4.0 2.5 1.0 The most common suggestion for community involvement (29%, n = 200) in forest protection and management was through participation in related activities, followed Stellenbosch University 71 http://scholar.sun.ac.za by the respect of laws, regulations and guidelines (21%, n = 200) and the report of offenders to authorities (21%, n = 200). Only 19.5% of the respondents said that they would commit themselves to what they pretended to be non-destructive activities in the Forest, including agriculture and mining in the swamps, gathering of fuelwood and other non-wood products. 3.2 FOREST RESOURCE ASSESSMENT Species composition and tree abundance were analysed for distribution among and within the forest areas adjacent to the communities surveyed in Musebeya, Nshili, Kagano, Mudasomwa and Busozo Communes. Abundances of tree species in different diameter sizes are presented for the three categories as indicated in Table 3.18. Table 3.18. Categories of trees used during the analysis. Categories Small trees Medium trees Large trees Height (m) 0.30 - 1.30 > 1.30 DBH (cm) < 2.5 2.5 - 9.9 2::10.0 The subdivision into small, medium and large trees corresponds to the definition of seedlings and saplings, poles and trees by Mwima (2000) and Owiunji (1997), respectively. The small trees were recorded for the assessment of the regeneration of tree species represented by large and medium trees. Data on tree and species abundance are presented on the basis of diameter size-class distributions among and within the study sites. Further, the assessment of abundance, dominance, size-class distribution and regeneration is presented for the most preferred tree species identified during participatory rural appraisal (PRA). 3.2.1 Identification of tree species The local and scientific names of all tree species are given in Appendix 6. Some local names were applied at "generic level," and a single local name could correspond to two or more species, as reported by Cunningham (1996). local name "ingongo" corresponded For example, the single to three Cassipourea species, Cassipourea ruwenzoriensis, C. ndando, and C. congoensis; "umunkamba" was a local collective 72 http://scholar.sun.ac.za Stellenbosch University name applied to various species including Ormocarpum tricocarpum, Dichrostachyus cinerea, and an on-farm medicinal plant, Clematis simensis. The matching of local names to scientific names also required attention because a local name given to a single plant species sometimes differed from one area to another. For instance, local people in Nshili and Busozo Communes used the local name "rwamba" to describe Prunus africana whereas others used "umwumba" to describe the same species. Despite this shortfall, local knowledge of tree species represented a practical, less time consuming and cost-effective method for identifying resource of interest to local communities in the absence of an operational Herbarium and qualified ethnobotanists in Rwanda. 3.2.2 Tree species abundance and density analysis for the study area For the whole sample area, there was an average of 439 large trees per hectare (standard deviation = 52 stemslha), represented by 54 species. The mean density of medium trees per hectare was 1 407 (standard deviation = 542 sterns/ha), also represented by 54 tree species. Small trees averaged 5 146 individuals per hectare (standard deviation = 3776 sterns/ha) but they were represented by only 34 tree species. The large standard deviations indicate that the stem densities differed greatly among the study sites. The results of the analysis of variance indicated significant differences in stem densities among the three tree categories (Table 3.19). Table 3.19. Analysis of variance of the densities of trees among tree categories Tree category Small trees Medium trees Large trees F-value p-value * Stem density/ha 5146A 1407AB 439 B 6.97* 0.02 indicates significance at p < 0.05. The Tukey's test of the comparison of the means showed that the densities of small trees did not differ significantly from the densities of medium trees but differed significantly from the densities of large trees. The analysis of the whole sample indicated that there were no significant differences in the densities of individuals between the study sites (d.f = 4, F=0.59, p>0.5) with a minimum difference of 8380 stems per hectare. significant Stellenbosch University 73 http://scholar.sun.ac.za A separate Chi-square analysis of the whole sample showed that the distribution of trees among the five study sites was not even (X2 = 1880.31, d.f. = 8, p<O.OO1). 2 However, tree species were evenly distributed among these sites (X = 11.55, d.f. = 8, p = 0.17). The difference in the distributions may reflect the heterogeneity in the study area, as the plots were located in areas with different altitudes, soils and different degrees of disturbance. Some species had many representative individuals, contributing significantly to the total density of trees recorded on each study site. The contribution of each tree species to the total density of individuals recorded in each study site and at different tree categories is presented in Appendices 7 through 9. Table 3.20 summarises the occurrence of the 10 most dominant species by site and tree categories. Table 3.20. Occurrence category. of the 10 most dominant tree species, by site and tree Tree species Strombosia schejjleri Carapa grandiflora Cleistanthus polystachyus Myrianthus holstii Casearia runssorica* Symphonia globulifera Cassipourea ruwenzoriens is * Parinari excelsa* Macaranga neomildbraediana* Chrysophyllum gorungosanum .. 1 = Musebeya; 2 = Nshili, 3 = Kagano, 4 * Species that were not among the 10 dominant as large and medium trees. Tree categories Large trees Medium trees Small trees 1,2,3,4,5 1,3,4,5 2,3,4,5 1,2,3,4,5 1,3,4,5 1,2,3,4,5 2,3,5 2,3 2,3,5 1,2,3,4,5 1,2,3,4,5 1,2,5 1,2,3,4,5 1,2,3,5 1,2,3,4,5 1,2,3,4,5 1,2,3,4,5 1,2,3,4,5 3,5 1,2,3,4,5 1,2,3,4,5 1,3,4 1,2,3,4,5 3 3,5 1,2,3,4,5 1,2,3,5 1,3,4,5 1,2,3,4,5 1,2,4,5 = Mudasomwa; 5 = Busozo . most dominant species in the small tree category but were Seven tree species were represented in the large trees and the small trees in all the five sites. Only seedlings and saplings of Symphonia globulifera were recorded in all the five study sites. The marked differences in the presence of seedlings and saplings of tree species in all the study sites could be attributed to environmental conditions, to shade intolerance of some species, and to disturbances in the periphery of Nyungwe Forest. These differences are also reflected in the stem densities of the species (Table 3.20 above). Figure 3.12a to 3.12c provide a comparison of the stem densities of the 10 most dominant tree species recorded in the three tree categories. Figure 3.13 indicates that the densities of small and medium trees were generally higher than those of large trees in all the study sites. 74 Stellenbosch University http://scholar.sun.ac.za 2500 2000 os .é .... 1500 8.. .g '"I: 1000 E 500 -8 b3 0 2 5 4 3 6 8 7 10 9 Species Figure 3.l2a. Stem densities per hectare of the 10 most dominant tree species represented by small trees 350 l ... 300 0. ;.-, 250 <1) .~ 200 ~'" 150 ~ 100 ei) 50 o 2 3 4 5 6 7 9 8 10 Species Figure 3.l2b. Stem densities per hectare of the 10 most dominant tree species represented by medium trees. 150 ii 120 Co .~ '"" 90 "0 ë ~ en 60 30 0 2 3 4 5 6 7 8 9 10 Species • Musebeya • Nshili D Kagano • Mudasomwa • Busozo 1 = Carapa grandiflora; 2 = Casearia runssorica; 3 = Cassipourea ruwenzoriensis; 4 = Chrysophyllum gorungosanum; 5 = Cleistanthus polystachyus, 6 = Macaranga neomildbraediana; 7 = Myrianthus holstii; 8 = Parinari excelsa; 9 = Strombosia scheflleri; 10 = Symphonia globulifera. Figure 3.12c. Stem densities per heetare of tbe 10 most dominant tree species represented by large trees. 75 Stellenbosch University http://scholar.sun.ac.za Ste 2=N;hili ~1=MEeh:ya 12345 6 7 8 9 ID 5 6 7 8 9 ID I 2 3 4 5 6 7 8 9 10 ~ ~3=Kag;m I 234 12345 6 7 8 9 ID ~ ill) 1<Ul .!!Il) ~(fi) ~<U) jm O~~~~~~~~~-.~~ 123 456 7 8 9 ID 1 = Carapa grandiflora; 2 = Casearia runssorica; 3 = Cassipourea ruwenzoriensis; 4 = Chrysophyllum gorungosanum; 5 = Cleistamhus polystachyus, 6 = Macaranga neomildbraediana; 7 = Myrianthus holstii; 8 = Parinari e:xcelsa;9 = Strombosia scheffleri; 10 = Symphonia globulifera Figure 3.13. Comparison of stem densities per hectare ofthe 10 most dominant species between tree categories, by study site. 76 http://scholar.sun.ac.za Stellenbosch University 3.2.3 Diameter size-class distribution of trees among and within the study sites Table 3.21 shows the densities of trees by size-class categories in all the five study sites. Table 3.21. Diameter size-class distribution of large trees among the study sites. 1 . included . The analysis only species WhiiChifi registered a frequency greater or equa to rve DBH size classes (cm) 10.0- 19.9 20.0 - 29.9 30.0- 39.9 40.0 -49.9 > 50 Sites Stem densities per ha s 46 38 84 Musebeya 19 184 88 64 30 Nshili 10 60 62 60 56 32 62 Kagano 10 142 90 32 20 Mudasomwa 154 58 13 88 34 42 36 Busozo 12 194 80 s = number of species With a frequency greater than five. Kolmogorov - Smimov tests were applied to the data to test the differences in the distributions of stem densities between pairs of sites as these tests are more appropriate than the chi-square test when the number of cases is small and when the expected frequencies at each category are also small (Bryman and Cramer, 1997; Zar, 1999). The results of the analysis are given in Table 3. 22 below. Table 3.22. Kolmogorov - Smirnov tests comparing distributions different sizes in pairs of sites Sites Musebeya vs. N shili Musebeya vs. Kagano Musebeya vs. Mudasomwa Musebeya vs. Busozo Nshili vs. Kagano Nshili vs. Mudasomwa Nshili vs. Busozo Kagano vs. Mudasomwa Kagano vs. Busozo Mudasomwa vs. Busozo Most extremes differences Absolute Positive Negative -0.60 0.60 0.20 -0.20 0.20 0.20 -0.40 0.40 0.00 0.40 -0.40 0.20 -0.20 0.40 0.40 -0.40 0.40 0.40 -0.40 0.40 0.40 0.20 -0.20 0.20 -0.40 0.40 0.20 -0.20 0.40 0.40 of trees of K-SZ 0.95* 0.32 0.63* 0.63* 0.63* 0.63* 0.63* 0.32 0.63* 0.63* 2-tailed p 0.33 1.00 0.82 0.82 0.82 0.82 0.82 1.00 0.82 0.82 K-S Z: Kolmogorov - Smrrnov Z * indicates significance at p-value = 0.05. The tests indicated that there was no difference in the distribution of stem densities between Musebeya and Kagano, and between Kagano and Mudasomwa. The observed differences in the distributions of trees of many pairs of sites illustrated differences in the two distributions. This could result from some species being 77 http://scholar.sun.ac.za Stellenbosch University represented by more trees than expected at some diameter size classes, to some species that do not grow up to larger sizes or to the effect of selective logging of suitable species and sizes for timber or poles. The analysis of distribution indicated that some tree species of different sizes were evenly distributed among two or more sites. Table 3.23 compares the distribution of the species among the study sites. Carapa grandiflora, Casearia runssorica, Maesa lanceolata, and Cleistanthus polystachyus were found in all the five study sites, but they were unevenly represented among the diameter size classes. Among the tree species that occurred in all the study sites, the size-class distribution was even for Myrianthus holstii and Cassipourea ruwenzoriensis. Table 3.23. Chi-square analysis to investigate the distribution of species among the study sites and within diameter size classes. The differences in degrees of freedom were due to species that were not common to all the study sites and sneci Wit . h no representative . III . dilVIiduals in some diameter size-c . Iasses. n to species ? Occurrence Tree species DBH size n d.f. Xclasses 1,2,3,4,5 A,B,C,D,E Carapa grandiflora 126 16 32.53** 1,2,3,4,5 Casearia runssorica A,B,C,D,E 62 16 29.43* 1,2,3,4,5 A,B,C,D,E Cassipourea ruwenzoriensis 45 12 8.67 A,B,C,D,E 1,2,4,5 Chrysophyllum gorungosanum 28 12 9.99 Cleistanthus polystachyus 20.78** A,B,C,D,E 1,3,5 68 8 Cremaspora trifolia 1,3,4 21 1.40 A,B 2 Diospyros gabonensis 17 0.62 A,B,C,D 3,4 3 Grewia mildbraedii A,C,D,E 1,4 Il 3 0.24 A,B,C,D 1,5 Maesa lanceolata 20 20.00*** 3 Myrianthus holstii A,B,C,D 1,2,3,4,5 68 12 11.32 Newtonia buchananii 18 1 0.07 A,E 3,4 Oxyanthus troupinii A,B,C 1,4 19 2 0.28 Parinari excelsa 28 14.16 A,B,C,D,E 2,3,4 8 Strombosia scheffleri 1,2,3,4,5 217 16 52.12*** A,B,C,D,E Symphonia globulifera A,B,C,D,E 1,2,4,5 44 12 13.04 * Significant at p<0.05, ** significant at p<O.OI, *** significant at p<O.OOI 1 = Musebeya, 2 = Nshili; 3 = Kagano; 4 = Mudasomwa; 5 = Busozo. A = 10.0 - 19.9 cm; B = 20.0 - 29.9 cm; C = 30 - 39.9 cm; 0 = 40.0 - 49.9 cm; E = ~ 50 cm 3.2.4 Presence of preferred tree species in the study area During the PRA, the respondents identified different tree species for different uses. A minimum frequency of 10responses was used as a guiding criterion for narrowing down the list of preferred tree species. The forest resource assessment established that not all woody resources mentioned as being used by the respondents were found in the Forest. Only 12 (i.e. 85.7%) of the 14 most preferred timber species were found as large trees and 12 (i.e. 75%) of the 16 preferred for pole species were found as 78 Stellenbosch University http://scholar.sun.ac.za medium trees. Some of the most preferred timber tree species that were not found included Faurea saligna and Ocotea usambarensis. Among the tree species preferred as building poles, Oxyanthus speciosus, Faurea saligna and Milletia sp. were not encountered. Of the 45 most preferred medicinal trees, only 17 species were recorded. Some medicinal plants mentioned during PRA surveys were not found at all during the forest resource assessment, including Prunus africana and Ocotea usambarensis. Of the six tree species preferred for wooden goods, only three were found in the sample area. These were Po/yscias fulva, Dasylepsis racemosa and Carapa grandiflora. Tables 3.24a and 3.24b show the most preferred tree species identified in the sample area. For the whole sample area, 54 species including the most preferred tree species for timber, poles, wooden items and medicinal materials, were recorded as large and medium trees. Based on the total number of individuals recorded in the sample area (1097 large trees and 995 medium trees), tree species occurring 10 or fewer times were defmed as 'rare species', those occurring 10 - 40 times as 'average species' and 41 times or more (nearly twice the average number of trees) as 'common species'. Tables 3.24a and 3.24b and Figure 3.14 show the occurrence categories of the tree species recorded in all the study sites. For timber tree species, five species were classified as rare. Tree species with average occurrence were four. Timber tree species of common occurrence were Strombosia scheffleri, Carapa grandiflora and Symphonia globulifera. Of the 16 medicinal tree species, rare tree species in the large tree category were six (37.5%) whereas rare ones recorded as medium trees were nine. Five medicinal tree species had 'average occurrence' . ... .~ 10 .,c, 8 '" 6 ..0 4 ""'...0 ., El =' Z 2 0 Average Rare Abundance [] Timber • Medicinal. large trees Common classes • Poles • Medicinal, medium trees Figure 3.14. Occurrence categories of preferred tree species in the sample area Stellenbosch University 79 http://scholar.sun.ac.za Table 3.24a. List of preferred timber and pole species and their occurrence Tree species Timber Carapa grandiflora Dasy/epsis racemosa Entandrophragma exce/sum Hagenia abyssinica Newtonia buchananii Podocarpus /atifolius Parinari exce/sa Podocarpus fa/catus Po/yscias ju/ya Strombosia scheffleri Symphonia g/obulifera Syzygium parvifolium categories in the study area Occurrence value Occurrence category 126 4 18 1 12 2 35 3 8 217 47 21 Common Rare Average Rare Average Rare Average Rare Rare Common Common Average 1 82 47 2 76 17 5 13 3 41 Rare Common Common Rare Common Average Rare Average Rare Common Common Average Poles Alchornea hirtella Carapa grandiflora Casearia runssorica Cassipourea gummiflua Cassipourea ruwenzoriensis Macaranga neomildbraediana Olea hochstetteri Parinari excelsa Podocarpus latifolius Strombosia scheffleri Symphonia globulifera Syzygium parvifolium Table 3.24b. List of preferred 99 39 medicinal species and their occurrence Tree species Medicinal sl!ecies, lar~e tree cate~orl: Bridelia bride/iifolia Canthium oligocarpum Carapa grandiflora Entandrophragma exce/sum Maesa lanceo/ata Neoboutonia macroca/yx Newtonia buchananii Olea hochstetteri Oxyanthus troupinii Parinari exce/sa Pauridiantha paucinervis Rapanea melanoph/oes Symphonia globulifera Syzygium parvifolium Tabernamontana johnstonii Zanthoxyjum gil/etii Medicinal sl!ecies, medium tree cate~on: Bridelia brideliifolia Canthium oligocarpum Carapa grandiflora Entandrophragma exce/sum Maesa /anceolata Neoboutonia macroca/yx Newtonia buchananii Olea hochstetteri Oxyanthus troupinii Parinari exce/sa Pauridiantha paucinervis Rapanea me/anoph/oes Symphonia g/obulifera Syzygium parvifolium Tabernamontana johns toni i Zanthoxylum usambarense Zanthoxy/um gil/etii categories in the study area. Occurrence value Occurrence category 1 4 126 18 21 12 12 1 20 35 13 9 47 21 19 2 Rare Rare Common Average Average Average Average Rare Average Average Average Rare Common Average Average Rare 1 21 82 5 Rare Average Common Rare Rare Rare Average Rare Common Average Average Rare Common Average Rare Rare Rare I 1 32 5 79 13 II 3 99 39 5 3 3 Stellenbosch University 80 http://scholar.sun.ac.za Preferred tree species for building poles were identified among the species in the medium tree category and consisted of trees of2.5 - 9.9 cm OBH. Their distribution into occurrence categories revealed that Alchornea hirtella, Podocarpus latifolius and Olea hochstetteri were rare. Syzygium parvifolium, Macaranga neomildbraediana and Parinari excelsa had average occurrence while at least five of the tree species preferred for building poles occurred frequently in the sample area and were defined as 'common species'. They included Carapa grandiflora, Casearia runssorica, Symphonia globulifera and Cassipourea ruwenzoriensis. 3.2.5 Abundance of the most preferred tree species in the sample area The stem densities of the preferred tree species were calculated relative to the stem densities of all species per hectare. The stem densities ranged from 0.05 to 7.59 stems per hectare for timber tree species and 1.35 to 5.77 stems per hectare for poles. For medicinal trees, the stem densities ranged from 0.59 to 3.52 and 0.72 to 4.05 stems per hectare for the preferred tree species recorded as large and medium trees, respectively. Tables 3.25a to 3.25c show the relative densities of the preferred tree species by tree categories. Table 3.25a. Relative densities of the preferred timber tree species. Tree species Strombosia schejJleri Carapa grandiflora Symphonia globulifera Parinari excelsa Newtonia buchananii Syzygium parvifolium Entandrophragma excelsum Polyscias fulva Dasylepsis racemosa Podocarpus falcatus Podocarpus latifolius Hagenia abyssinica Mean stem density Standard error Confidence level (95%) Stem density Relative stem per ha density 86.8 19.77 11.48 50.4 4.28 18.8 14 3.19 8.4 1.91 8.4 1.91 7.2 1.64 3.2 0.73 0.36 1.6 0.27 1.2 0.8 0.18 0.09 0.4 3.82 1.71 3.77 81 http://scholar.sun.ac.za Stellenbosch University Table 3.25b. Relative densities of the preferred pole tree species. Tree species Symphonia globulifera Carapa grandiflora Cassipourea ruwenzoriensis Casearia runssorica Strombosia scheffleri Syzygium parvifolium Macaranga neomildbraediana Parinari excelsa Olea hochstetteri Podocarpus latifolius Alchornea hirtellaf glabrata Cassipourea gummiflua Mean stem density Standard error Confidence level (95%) Stem density Relative stem per ha density 140.03 9.95 115.98 8.24 107.50 7.64 66.48 4.72 57.99 4.12 55.16 3.92 24.05 1.71 18.39 1.31 7.07 0.50 4.24 0.30 1.41 0.10 2.83 0.20 3.56 1.00 2.21 Table 3.25c. Relative densities of the preferred medicinal tree species. Medium trees Stem density/ha Symphonia globulifera 140.03 Carapa grandiflora 115.98 Oxyanthus troupinii 111.74 Syzygium parvifolium 55.16 Newtonia buchananii 45.26 Canthium oligocarpum 29.70 Parinari excelsa 18.39 Pauridiantha paucinervis 15.56 Entandrophragma excelsum 7.07 Olea hochstetteri 7.07 Tabernamontana johnstonii 7.07 Rapanea melanophloeios 4.24 Zanthoxyllum gilletii 4.24 Zanthoxyllum usambarense 4.24 Bridelia brideliifolia 1.41 Maesa lanceolata 1.41 Neoboutonia macrocalyx 1.41 Mean stem density Standard error Confidence level (95%) Tree species Relative stem density 9.95 8.24 7.94 3.92 3.22 2.11 1.31 1.11 0.50 0.50 0.50 0.30 0.30 0.30 0.10 0.10 0.10 2.38 0.78 1.66 Large trees Stem densit}:/ha 18.80 50.40 8.00 8.40 8.40 1.60 14.00 5.20 7.20 0.40 7.60 3.60 0.80 Relative stem density 4.28 11.49 1.82 1.91 1.91 0.36 3.19 1.19 1.64 0.09 1.73 0.82 0.18 0.40 8.40 4.80 0.09 1.91 1.09 2.11 0.69 1.46 Using a density of equal or greater than 1.0 stems per hectare (at the lower limit of the 95% confidence interval) as a cut off limit for harvestable timber tree species Stellenbosch University 82 http://scholar.sun.ac.za following Davis and Johnson (1987), there were seven tree species which could be considered abundant enough for utilisation. Of these species, commonly available tree species were Strombosia scheffleri and Carapa grandiflora. Among medicinal tree species, there were 11 species that were considered abundant enough to allow collection of medicinal materials from large trees. Carapa grandiflora, Symphonia globulifera and Parinari excelsa were the most abundant medicinal tree species in the large tree category. Within the medium tree category, only eight medicinal tree species were abundant enough to be listed as harvestable species. Carapa grandiflora and Symphonia globulifera were predominant in the Forest and were classified among the abundant medicinal tree species. Finally, seven tree species of the medium tree category were abundant for potential supply of poles. The five most abundant were, in decreasing order of their abundance, Symphonia globulifera, Carapa grandiflora, Cassipourea ruwenzoriensis, Casearia runssorica and Strombosia scheffleri. 3.2.6 Dominance of the most preferred tree species All abundant preferred timber species had a relative dominance greater or equal to 0.01 and contributed more to the forest standing biomass. Of the most abundant and preferred pole species, the only tree species which showed a relative dominance of less than 0.01 was Macaranga neomildbraediana. However, many of the medicinal tree species represented by large trees showed a low contribution to the total biomass of the Forest. Tree species with sufficient contribution owing to their relatively higher densities of their stems included Carapa grandiflora, Syzygium parvifolium, Parinari excelsa, Symphonia globulifera and Newtonia buchananii. The medicinal tree species of the medium tree category contributed sufficiently to the standing biomass, with the exception of Canthium oligocarpum. The relative dominance of the species in the large tree category is presented in Appendix 10. 83 http://scholar.sun.ac.za Stellenbosch University 3.2.7 Size-class distribution and regeneration of most abundant preferred tree species. The distribution of stem densities into different diameter size classes differed among the preferred tree species (Figure 3.15). Of the 11 species, only Carapa grandiflora and Symphonia globulifera approximated the 'inverse J' curve that is characteristic of a mature and stable system or a population that is self-perpetuating (Bazzaz, 1998; Luken, 1990). regeneration. This suggests that there were shade-tolerant species with abundant For some species including Cassipourea ruwenzoriensis, Parinari excelsa, Casearia runssorica, Oxyanthus troupinii and Syzygium parvifolium, the infrequency of seedling establishment caused a reduction in the smaller size class, followed directly with a greater number of saplings or individuals of 2.5 - 9.9 cm DBH. Similarly, Newtonia buchananii and Pauridiantha paucinervis showed no overall pattern, with intermediate classes that contain no individuals at all. This suggests that the existing level of saplings were insufficient to re-stock these classes. The shape of the curve of Entandrophragma excelsum was also irregular, with the lack of regeneration among the smaller size class and the presence of saplings and few large trees in other diameter size classes. Generally, E. excelsum was found at lower density compared to other species. It can be seen from the curves of E. excelsum and Parinari excelsa that their size-class distributions exhibit relatively more large trees at the largest size class (i.e. z50 cm DBH) than those found at the intermediate sizes. This could indicate temporarily variable recruitment in the stages oftheir life history. The observed pattern in the size-class distribution of Oxyanthus troupinii represents a population that has very few large trees and relatively abundant seedlings and saplings. This pattern illustrates a population of a long-lived tree species that experienced a period of very poor recruitment or past intense harvesting of large trees. Figure 3.16 shows that the regeneration of the species is very variable across the study sites. The regeneration appears to be least at Nshili and Mudasomwa for some preferred tree species. Only Symphonia globulifera produced seedlings and saplings in all five study sites while seedlings of Newtonia buchananii were only recorded in the part of the Forest next to Kagano Commune. Stellenbosch University 84 http://scholar.sun.ac.za Carapa 1000 800 600 400 200 oE q il,i 1 I J, o ABC DBH Cassipourea oE 500 .~ 400 c 300 ~ 200 ~ I O~ C. runssorica grandiflora I T D E size classes, T T F G T T T D E F G size classes, cm Newtonia ABC E F size classes, ,.I,L C o DBH J:...__"r'.L .... ,'~J:......_T'~.L"'_r, ..J:..._, E F size classes, cm ABC D troupinii 500 :g 300 ~ 200 E 100 ~ (/) 0 I I D E F size classes, cm ABC DBH Pauridiantha .I. I F size classes, cm I G paucinervis , 1. , .1 .. G A B DBH Symphonia scheffleri G 400 cm E bosia excelsum j G excelsa B Strom I , .I. G Oxyanthus D Parinari F cm D DBH T DBH A 40 buchananii • E DB H size classes, f : l-+l----r, ...1--., ... oE q , .I Entandrophragma .I , .I DBH T ABC cm oE ABC 240 :g 180 ~ 120 E 60 s (/) 0 r u w e n z o r ie n s is i, , 300 C D I E F size classes, cm G g lo b u life r a 1000 , .- .- , ~ , .L , .L E F G size classes, cm , ABC D DBH llHI,I" T D T T E F G size classes, cm ABC DBH T Sy zy g iu m p a rv ifo liu m T A B DBH T I eDE size classes, A = <2.5 em; B = 2.5-9.9cm; C F = 40.0 - 49.9cm; G = ;;::50cm = T T F G cm 10.0-19.9cm; 0 = 20.0 - 29.9cm; E = 30.0-39.9cm; Figure 3.15. Size-class distribution of the most abundant and dominant trees preferred for timber, poles and medicines. Stellenbosch University 85 http://scholar.sun.ac.za Carapa :!? -:c ~ Casearia grandiflora ;J< ;: 30 c " 25 "0':> 20 " 15 <!: 10 ::" 5 1ii 0; 0 40 30 g- 20 :> <!:: .e" '" ~ runssorica 10 0 A B D C c.:: E A B Sites Cassipourea :!? ruwenzoriensis Newtonia 30 25 20 ec u :> :!? ~ g- g- IS u 10 .e 5 Ii 0 ~ <!:: u .~ r-l.-, A B Oxyanthus u :> 0- u <!:: u .~ 1ii "ii et: D C ~ et: E D E D E 30 20 IS 10 5 0 A B Parinari troupinii ;t. ~ c s 0- u <!:: u .~ ee l, B ~ 30 25 20 15 10 5 0 E D C excelsa A B C Sites Sites Pauridiantha C Sites 30 25 20 15 10 5 0 A Strombosia paucinervis ;t. ;t. ~ ~ e 30 25 sg. 20 u 15 <!: 10 u .e 5 1ii "ii 0 et: scheffleri JOL 25 20 u 15 <!:: u 10 .<: 5 u :> 0- ~ o et: A B A E D C Symphonia B C D E Sites Syzygium globulifera ;t. ;t. 30 25 ~ t: s 20 u 15 <!:: 10 .e" 5 1ii ~ 0 parvifolium 30 -: 25 " ~ gt: 20 IS 10 "<: 5 '" 0 0; et: 0- <!:: A B D C A E B = Nshili; B C D E Sites Sites = Musebeya; II ... Sites A E buchananii Sites ~ c D -:o 25 c <!:: ;t. C Sites C = Kagano; D = Mudasomwa; E = Busozo Figure 3.16. Relative abundance of seedlings and saplings of the abundant and dominant tree species across the five sites. Stellenbosch University 86 http://scholar.sun.ac.za 3.2.8 Quality of harvestable tree species for timber and poles Figures 3.17 and 3.18 show the quality of harvestable trees for timber and poles. Figure 3.17 shows that the proportion of good quality stems suitable for timber was higher for Symphonia globulifera, Entandrophragma excelsum, Parinari excelsa, and Strombosia scheffleri. Boles of these tree species were cylindrical in form, from the ground to the first branch, and therefore were good enough for use as sawn timber. Carapa grandiflora and Newtonia buchananii had irregular taper and protuberances and could only be used with a considerable proportion of waste. Syzygium parvifolium, for which many stems ranged from fair to acceptable, would be better used for a product where stem straightness is not a major requirement. preferred tree species for building poles, Symphonia globulifera quality stems, followed by Cassipourea ruwenzoriensis Of the most produced many (Figure 3.18). However, less than 20% of the individuals of Strombosia scheffleri were of good quality stems. 87 Stellenbosch University http://scholar.sun.ac.za 100% <Il 80% Il) ~ ....... 0 c: 0 Ot 0 0- 60% 40% e 0- 20% 0% Species Figure 3.17. Proportion of trees with stem suitable for timber. 100% <Il Il) 80% ~ ....... 0 60% s:: 0 Ot 0 00 0- ... 40% 20% 0% Species • Good 0 Fair • Acceptable Figure 3.18. Proportion of trees with stem suitable for building poles. 88 http://scholar.sun.ac.za Stellenbosch University 4 DISCUSSION In this chapter, both the results obtained from the household survey and from the forest resource assessment are discussed. The participatory assessment was carried out as a preliminary study aimed at collecting information on resource preferences and uses and the perceptions of local people regarding the appropriate institutions for the management of utilisation in Nyungwe Forest. Participatory techniques are commonly used to strengthen any management decisions. For example, Scott (1992) used PRA techniques to collect information on which to base recommendations multiple use activities at Bwindi Impenetrable National Park in Uganda. for In the resource assessment, the areas of the Forest next to the Communes surveyed were assessed for their potential to provide the preferred tree products to the communities. Cunningham (1996) carried out a similar forest survey prior to recommending use of resources at Mgahinga Gorilla National Park in Uganda. Preferred tree products from Nyungwe Forest are discussed in relation to their availability, abundance, size-class distribution and regeneration. The discussion focuses on the tree species most preferred for timber, poles and medicinal materials as identified during the forest assessment. Other uses of the Forest are also discussed, including the production of wooden goods, fuelwood, charcoal, beekeeping, fodder, honey, bean stakes and land. A focus on these products is justified by their economic potential and the role they play in the livelihoods of forest-dependent communities (Neumann and Hirsch, 2000; Prasaad, 1999). Further, having discussed the resources available in Nyungwe Forest, the long-term potential of the Forest to provide these resources is presented together with recommendations on the management of the Forest for utilisation. 4.1 4.1.1 RESOURCE PREFERENCES, USES AND AVAILABILITY IN NYUNGWE FOREST Wood products 4.1.1.1 Timber The survey established that 29 tree species were preferred for use as timber, of which the most sought after were Podocarpus latifolius, P. falcatus and Entandrophragma 89 http://scholar.sun.ac.za Stellenbosch University excelsum. These timber trees were found in low densities in the Forest, suggesting that they were targeted for pitsawying. selective and unmanaged, Logging of valuable trees has been very and it has resulted in some tree species becoming endangered (Gapusi and Mugunga, 1997; Habiyambere, 1999; Sorg, 1978). As an alternative to the loss of access to indigenous timber, the respondents were found to manage fast growing exotic timber species to cope with the short-term demand for furniture and construction materials or else to generate income needed by the households. The lack of indigenous timber tree species on the respondents' farms may reflect past social forestry programmes that emphasised the planting of exotic species, without a focus on the indigenous tree species preferred by the local people. Illegal logging for indigenous timber by neighbouring communities has been reported to be a major threat to the conservation of the Forest (MINITERRE, 2000). Since the local communities favour certain species for timber, harvesting is likely to be selective. Selective logging affects mature forests and species composition, as a result of canopy gap formation (Howard, 1991). distributions In addition, logging results in patchy of both the species and size-class categories harvested (Cunningham, of trees and shrubs 1996). Timber species in the Forest showed this pattern, with a limited abundance of marketable sizes of timber tree species. These species were found at low frequencies, with fewer harvestable sizes and an increasing number of young trees. The small number of large sized hardwoods was probably a result of overexploitation. Sustainable harvesting of timber is based on the methods of regeneration, silvicultural systems and yield regulation (Geldenhuys, 1998; ITTO, 1990). The ecological management of natural forest emphasises the extraction of resources that show a low impact of removal. Selective logging is thought to be a viable option if it is practised outside the conservation area of a forest (Struhsaker, 1997). Therefore, a management system could be developed for sustainable logging of timber over 10% of the total area of Nyungwe Forest as was established by a management developed in 1988 (Fimbel and Kristensen, effectively implemented. 1994; MINAGRI, plan 1984) but not Stellenbosch University 90 http://scholar.sun.ac.za 4.1.1.2 Building poles Nyungwe Forest is surrounded by human settlements. collected building poles from the Forest. In the past, local people had In the study area, managed woodlots are source of building materials. This is in common with other parts of the country where communities have managed trees for various uses, including small to medium sized poles which can be sold (den Biggelaar, 1996). The loss of access to poles and other wood products from the Forest is compensated by tree planting on farms (Olson et al., 1995). Ample quantities of poles appeared to be cultivated, but indigenous poles from Nyungwe Forest were apparently more preferred for their straightness and durability. From field observation, many houses were found to be built from cultivated Eucalyptus poles but in the south-eastern zone bamboo from the Forest was also used, with the use of the species decreasing with distance away from the edge of the Forest. The common use of cultivated poles in the study area was a result of the development of community nurseries and the supply of seedlings to local people by past donor-funded forest projects in order to reduce dependence on the Forest (MINAGRI, 1990). There are no available data on the consumption of indigenous poles. Due to the high population density (350 people/km') and the number of households around the Forest, the demand for poles is expected to be very high, leading to increased illegal logging for poles. Programmes of tree planting on farms with suitable indigenous species could partly alleviate illegal cutting of poles from the Forest in the long-term. 4.1.1.3 Fuelwood and charcoal In common with other rural areas in Rwanda, fuelwood provides the major source of household energy (MINECOFIN, 1998). Due to its high conservation status, fuelwood is the only commodity that a few people access from the buffer zone of Nyungwe Forest. Similar to poles, there have been attempts by foresters to provide fuelwood to adjacent communities to the Forest by encouraging tree planting on farms (Olson et al., 1995). The production of fuelwood for household consumption was the second priority next to pole production in the survey area. 91 http://scholar.sun.ac.za Stellenbosch University Charcoal production was considered as an income generating activity in the study area around the Forest. woodlots. The respondents produced charcoal from their own Eucalyptus Despite the availability of trees in farmers' fields, charcoal making has been reported as a threat to the conservation of the Forest (MINITERRE, 2000). This implies that there is a demand for charcoal from indigenous trees. Other reasons for exploitation of wood for charcoal from the Forest may be related to the shortage of land, which prevents some households from establishing woodlots. charcoal in cities also provides an economic The demand for incentive that encourages some individuals to encroach on the Forest. The annual increase in the rate of fuelwood consumption in Rwanda is estimated at 2%, translating to an annual increase of three million m3 per year, exceeding the limits of sustainable exploitation (Habiyambere, 1999). With an estimated forest cover of 7% of the country's area and the increasing demand for fuelwood and charcoal, the survival of the remnant indigenous forests is uncertain (Mitchell, 1997). It will likely depend on the development of suitable alternatives such as the use of biogas, the use of efficient cooking stoves and an active programme of community forestry in Rwanda (GEC, 2000; Percival and Dixon, 1995) 4.1.1.4 Wooden items Households adjacent to Nyungwe Forest rely on locally produced materials. The local producers are highly selective of species since certain attributes are required of the wood for specific purposes. The most prominent characteristics are flexibility, durability, strength and resistance to splitting. The selective harvesting of these species can result in genetic erosion of the species as vigorous and quality individuals are removed from the population (Cunningham, 1996). The current rarity of these species could be partly explained by the effect of illegal harvesting. Bamboo is used by local people in the south-eastern zone of Nyungwe Forest where it is predominant. Considerable quantities of materials made of bamboo are sold both locally and outside the area, or used for building purposes. The major incentive for the unlawful cutting of bamboo from the Forest is the economic advantages for individuals to benefit from the sale of goods. The use of bamboo for commercial and Stellenbosch University 92 http://scholar.sun.ac.za subsistence needs might result in overexploitation if not regulated or cultivated on farms for alternative sources of basketry and granary materials (Butynski, 1984). Bamboo can easily grow in people's gardens as shown by Cunningham (1996) as part of the management of Bwindi Impenetrable National Park in Uganda. Other tree species including Markhamia lutea and Polyscias fulva could receive attention in the development of agroforestry in Rwanda since they are indigenous species traditionally planted or conserved on farms (Niang and Styger, 1990) and valued for the making of various wooden items. 4.1.1.5 Bean stakes Beans, both bush and climbing varieties, are among the staple foods produced in the zone surrounding the Forest. In the survey area, climbing beans were preferred for their high yield and resistance to heavy rains during the growing seasons. Bean stakes are essential for climbing bean production, and huge quantities of bean stakes are required. In the past, stakes were a common product obtained from Nyungwe Forest by neighbouring farmers. The most favoured tree species was Alchornea hirtella, which is presumably the reason for its low frequencies in the study area. Many farmers in the area use branches of Pinus patula, Eucalyptus spp. and Grevillea robusta from their own farms or from the buffer zone for Nyungwe Forest as bean stakes. Far from the edge of the Forest, the shortage of adequate staking materials is reflected by the use of Pennisetum purpureum and stalks of sorghum and maize. As the population depends on agriculture for its subsistence, the demand for bean stakes is expected to increase. The demand cannot be met from the supplies available within the agricultural systems. It is recommended that in addition to assessing ways for the community to use bean stakes from Nyungwe Forest, development programmes should encourage the local people to cultivate their own supplies of bean stakes by establishing fast growing shrubs including Mimosa scabrella and Chamaecytisus palmensis identified by the Agroforestry Research in Rwanda to be suited on local soils and climate. Other species with high coppicing abilities could be introduced for the production of large quantities of stakes. For instance, communities surrounding Bwindi Forest in Uganda use bamboo for stakes (Wild and Mutebi, 1996). Stellenbosch University 93 http://scholar.sun.ac.za 4.1.2 4.1.2.1 Non-wood products Land Nyungwe Forest forms a physical barrier between two agro-ecological regions, with different farming systems found on either side (Renner, 1991). The sample area falls within the two agro-ecological zones, and so the farming conditions of the different groups of respondents vary. These differences were reflected in farm sizes, crop types and soil fertility. The decreasing availability of new farmland means the population around the Forest is unable to support itself from agriculture. Problems of soil infertility result from acidic soils. These soils are severely degraded, which leads to low crop production and increased poverty (Olson, 1994). The communities adjacent to Nyungwe Forest lack access to production resources including livestock for manure production, fertilisers, and other means to manage soil fertility. This result supports the finding by Blaikie and Brookfield (1987) that access to production resources is felt differently between households within the same region depending on their ability to afford inputs. In the past, adjacent communities used land outside and inside Nyungwe Forest in an extensive manner (Olson et al., 1995) until they were relocated without compensation onto land outside the Forest (Renner, 1992). With the current process of land acquisition through inheritance, the farm size is unable to meet the subsistence needs of the local population. There is also a lack of alternatives to agriculture, including diversified sources of income. opportunities Newbury (1992) attributed the lack of off-farm in the southwest of Rwanda to the past government regimes that concentrated government funds on the development of their home region". Local perception of Nyungwe Forest as a potential land resource could therefore be attributed to the historical background, including the influx of farmers looking for 2 Newbury (1992) argued that the southwest of Rwanda had fewer opportunities than other areas as the former President Habya/imana Juvenal and the people he selected for important government posts concentratedfunds in the North of the country where they originated. Stellenbosch University 94 http://scholar.sun.ac.za fertile lands to the eastern edge of the Forest (Olson, 1990; Cambrezy, 1984). The preference for farmland from within the Forest stems from the past practice of clearing the Forest for food production (Barabwiliza, 1980). Overuse of forest products would result in loss of biodiversity in the long-term, but clearance for agriculture is the greatest threat (MINAGRI, 2000; Mitchell, 1997). This pressure is driven by poverty and a growing population in adjacent areas. Subsistence agriculture was the major livelihood strategy amongst the survey population. Revenues from crops are relatively low, due to low prices given to agricultural produce (Ben Chabanne and Cyiza, 1992; World Bank, 1993). Income accrued by people around Nyungwe Forest can not meet their economic and financial requirements, resulting in an increased dependence on natural resources. Unless adequate strategies are adopted, increased land for agriculture and livestock farming will continue to be a major concern for local communities. 4.1.2.2 Gold mining Local villagers have high expectations of earning cash income (Barakabuye, 2001) and therefore illegal mining is considered to be a major law enforcement and conservation problem (MINITERRE, 2000). Mining in defined zones was listed by MINAGRI (1993) with the priority activities that were considered for authorisation to organised groups. Recognising access of local groups to mineral reserves can help in improving incomes to local households. Habimana (1982) found that 28% and 10% of the sample population in Mudasomwa and Kivu Communes next to the Forest obtained cash income from gold mining, respectively. Although the sample population was too small to make inference about the large adjacent population as a whole, it is assumed that access to minerals in the Forest could be an alternative source of income to the neighbouring communities. Due to the negative impact of mining on the conservation of biodiversity, it is necessary to define and demarcate mining areas and to monitor activities by miners' groups. 95 http://scholar.sun.ac.za Stellenbosch University 4.1.2.3 Wildlife and ecotourism In the past, neighbouring communities extracted bushmeat from Nyungwe Forest through hunting and the use of traps (Storz, 1983). The survey established that wildlife is not managed for utilisation as part of the livelihoods of the people adjacent to the Forest. Hunting for household consumption is regarded as a threat to the conservation of wildlife, including rare and protected animals such as chimpanzees and buffaloes (MINITERRE, 2000). The value of the resource in money terms cannot be estimated without an in-depth survey, as off-take from the Forest is illegal and bushmeat trade said to be non-existent. In Equatorial Africa, hunting and bushmeat trade occupy a central place in the economy, and the volumes traded are substantial. For instance, the annual bushmeat trade in Gabon is valued at US $ 22 million on informal markets and at US $ 3 million on formal markets (Bailon 1996, quoted by Klein and van der Wall, 1998). In the Republic of Congo, 26 000 animals are said to be sold in Pointe Noire monthly, implying that half a million animals are killed for this market each year (Wilson and Wilson, quoted in Colchester, 1994). In Cameroon and in the Democratic Republic of Congo, the bushmeat consumption is estimated at 21 and 41 kg per person per year (Wilkie and Carpenter, 1999). Wildlife can be managed as a source of subsistence or income. For example, the annual income from hunting and trapping in villages around the Korup National Park in Cameroon amounted to approximately US $ 1050 per hunter and accounted for 50% of the total village income (Infield, 1988). In Nyungwe Forest, the non- consumptive use of wildlife through ecotourism is favoured by the government managing institution, ORTPN (Barakabuye, 2001). The communities around Nyungwe Forest do not enjoy the benefits of ecotourism, apart from the employment of a few local guides, because all the economic benefits accrue to Government. In the year 2000, Nyungwe Forest registered 777 visitors from different foreign countries, from which US $ 6 722 were obtained by ORTPN (Barakabuye, 2001). This revenue is low compared with the revenues obtained before 1994. In the early 1990s, ecotourism in the Forest raised US $ 15 000 per year, but this amount was only enough to pay the staff and the upkeep of the Forest (Offutt, 1992). Since all of these 96 http://scholar.sun.ac.za Stellenbosch University revenues go to government funds, they have little impact on communities around the Forest. When carefully planned and managed, an ecotourism development in a tropical forest can provide a sustainable return, much of which can remain in the local economy (Horwich, 1988). In many countries, attempts have been made to involve local people in the management of forest reserves and to create opportunities for local communities to benefit from the forests. This has been the case for Budongo Forest in Uganda (Langoya and Long, 1998), Serengeti National Park in Tanzania (Emerton and Mfunda, 1999), woodlands in Zimbabwe (Hasler, 1999) and wildlife management in South Asia (Kothari et al., 2000). The potential for community involvement in wildlife management in Nyungwe Forest exists and it is justified for the following reasons: • Local people around the Forest have knowledge of the resources, including rare and endangered wildlife; • They are aware of the value of wildlife as bushmeat and as an ecotourism attraction; • They have knowledge of the reproductive cycles and the habitats of wildlife. Nevertheless, with the present condition of poverty and absence of alternative livelihoods strategies, the sustainability of the wildlife resource is threatened. It is essential for the Government to consider approaches to natural resource management taken by many other governments which offer an immediate value to themselves and benefits to local communities. 4.1.2.4 Medicinal plants Medicinal plants are an important element of traditional medicine on which 80% of the world population currently depends (WHO et al., 1993). Of the 250 000 higher plants worldwide, one fifth are used as medicinal plants (WWF, 1994). The use of medicinal plants is one of the most important functions of non-timber forest products (Fedlmeir, 1998). Due to its illegal status, medicinal plants used by the local communities are supposed to be taken from farms. 97 http://scholar.sun.ac.za Stellenbosch University Although traditional healers do exist in the region, it appears that most people in the region treat themselves or seek advice from friends, in particular from elder persons. No data are availed on the number of traditional healers in Rwanda. With an estimate of more than 500 medical practitioners who were registered to treat livestock in 1992 (Mbarubukeye, 1992), there is no doubt that traditional healers could number more than medical doctors. Medical clinics are found in the Communes next to the Forest. However, they are difficult for some people to reach, lack adequate physicians, and the drugs prescribed are often too expensive. This explains the dependence on locally collected medicinal plants for the treatment of various illnesses including headaches, diarrhoea, and dental cavities. The leaf material is the major component of medicinal plants commonly collected on farms. Due to the low diversity of the species available on farms, additional material could be collected from the Forest. The majority of medicinal trees found in Nyungwe forest were preferred for their bark material. The debarking of trees for the gathering of medicinal material can result in death of the tree resource or its susceptibility to fungal attack and other diseases. In Madagascar, for instance, the collection of the bark of Prunus africana has resulted in its death (Kollert, 1998). Cunningham and Mbekum (1993) reported that stripping the bark off two opposing quarters of the trunk of Prunus africana in Cameroon enabled the tree to resist total injury. Thus the use of improved extraction methods of the bark of medicinal trees is a means of achieving sustainable utilisation of medicinal trees. Some tree species are unable to re-grow the bark removed and thus more vulnerable than those that can. Other factors that affect the sustainable harvesting of medicinal plants include the amount harvested, the level of commercial trade and the part used (Cunningham, 1996). Since medicinal trees in the Nyungwe Forest could be exploited largely by debarking, these trees are at high risk of death if the collection methods are not improved. Medicinal plants have been shown to be economically important for sustaining local livelihoods. At present, in Nyungwe Forest, illegally collected medicinal plants are used for subsistence needs instead of commercial use. In some countries, however, medicinal plants are produced in large quantities, sold at local and international markets and contribute to the national economy. In Nepal, for instance, official Stellenbosch University 98 http://scholar.sun.ac.za records indicated an average of 50 tons per year of aromatic and medicinal plants commercially exported from the country (Aumeeruddy, 1998). In Ethiopia, seven medicinal plants are sold by vendors on local markets (Deffar, 1998). In South Africa in 1993, the annual trade of 20 000 tonnes of medicinal plants worth US $ 60 million were annually traded at the national level (Cunningham, higher today. 1993), and this figure is In the year 2000, the economic value of medicinal plants in South Africa approximated a value of US $ 41 million (approx. R 270 million) per annum (Mander, 1998, quoted by Lawes et al., 2000). The survey of medicinal plant use showed that local communities have considerable knowledge about many aspects of medicinal plants, including their abundance, uses and methods of harvesting. There is no doubt that medicinal plants cannot be managed effectively without the co-operation of the local communities, who are in the best position to protect and manage them. Although PCFN is seeking ways of managing the local collection of medicinal plants, agreements are needed between traditional healers' associations and the Nyungwe Forest authorities to confer certain rights and responsibilities. The agreements would define which medicinal plants should be used, the amounts to be collected, and by whom. This could result in an increased interest among local collectors in the protection and conservation of the resources for their continuing benefits. 4.1.2.5 Fodder plants and pastures The forest legislation in Rwanda states that livestock must not be grazed in state owned forests (JORR, 1989). An exception has been made for the Acacia melanoxylon and Pinus patuIa plantations in the buffer zone for Nyungwe Forest, where cattle, goat and sheep were grazed on the herb layer (Barbier, 1992). In the survey, evidence was found that livestock is still being grazed in the buffer zone with the permission of the agricultural extension agents in the area. Available fodder grass and shrub species mentioned by the respondents are similar to those reported by Gasana (1988), the most common belonging to the families of Poaceae, Asteraceae, Cyperaceae and Tiliaceae. 99 http://scholar.sun.ac.za Stellenbosch University Livestock keepers lack diversified means of coping with shortages of fodder during the dry season, the most common solution being free range grazing. Fodder is not stored, and currently there is no market available for fodder. The study did not focus on fodder productivity, abundance and composition in Nyungwe Forest. Field observation and household interviews revealed that, apart from patches of fodder along roads and in some gaps inside the Forest, N yungwe Forest was poor in fodder quality and quantity. This was so because the herb layer is densely occupied by ferns and lianas, most of which are not eaten by livestock. The insufficient grazing resources outside the Forest could explain the perception of the respondents of the Forest as a source of fodder and pasture. reported cases of uncontrolled Gapusi (1999) and excessive grazing in the Forest. Livestock population around N yungwe Forest is on the increase owmg to the demand for manure (from cattle) and income (from goats and sheep). There is a need for a welldefined national policy on grazing to sustain fodder supply by non-forest sources and by instituting change in livestock composition and grazing practices. 4.1.2.6 Beekeeping Beekeeping has a low impact on the forest, as has been the case at Bwindi Impenetrable and Mgahinga Gorilla National Parks of Uganda (Wild and Mutebi, 1996). Beekeeping in Nyungwe Forest was controlled by forest projects, and activities of beekeeper's groups were guided by contracts and regulations (Barbier, 1992). The only concern was possible fires that could be set by the honey collectors. Due to low production, honey is used for subsistence needs of the household in the study area, with only small quantities sold on the local market. In some countries, large quantities of honey are produced by a considerable number of beekeepers. In Malawi, some 8 000 beekeepers produce 1 000 tonnes and 150 tonnes of honey and beeswax per annum respectively (Chanyenga, 1999). The average export of honey and beeswax in Ethiopia is estimated at 3.05 tonnes and 270.34 tonnes per year, respectively (Deffar, 1998). Stellenbosch University 100 http://scholar.sun.ac.za The low productivity of honey in Rwanda could be due to the use of traditional production technologies. Production can be increased if improved methods, including the use of modem beehives, are used. Honey production also depends on the availability of certain plant species that are associated with bees. Known bee forage species listed by the respondents were less diverse compared to those found in the Forest and reported by Habiyambere (1999). Forest disturbances could be one of the reasons for the loss of tree species favoured by bees. Increasing bee forage species would certainly boost honey production. between beekeeper's Past experiences based on agreement associations and forest projects could guide decisions over beekeeping in the Forest for the benefit of the local communities. 4.1.2.7 Wildfoods Although an assessment of the availability of wild foods has not been conducted in the Forest, the households adjacent to Nyungwe Forest mentioned a number of seasonal wild foods that supplement the diet or are used as a source of income. The ranking of these non-wood products in relation to the other resources showed that these products were not of great importance to the population. Practical examples demonstrate how fruit and mushroom harvesting could be integrated into local forest management plans and can constitute sources of income for the population. harvesting of wild foods appears to be sustainable or to have low impact. The This is because: • Harvesting of wild vegetables, mushrooms and fruit is a seasonal activity, and practised by the poorest people. For wild vegetables, edible leaves of Amaranthus spp. and Solanum nigrum are collected during the rainy season; • The use of the fruit of Myrianthus holstii encourages the conservation of the female tree (Cunningham, 1996) and the population around Nyungwe Forest has planted this species on their farms; • Farmers in the area produce fruit and vegetables in their gardens (for example: Amaranthus spp., cabbages, carrots, onions), and thus do not rely entirely on the foods from the Forest. Stellenbosch University 101 http://scholar.sun.ac.za As the gathering of wild foods has relatively little impact on the forest dynamics, this is one category of resources that the management of Nyungwe Forest could allow local people to access, as a way of their gaining benefit from the conservation of the Forest. Mutual trust between parties would be essential here, however, to ensure that other, restricted, resources are not harvested at the same time. 4.2 CHARACTERISTICS OF THE PREFERRED TREE SPECIES Due to time and financial constraints, the study concentrated on the assessment of the standing stock of the most preferred tree species for timber, poles, and medicinal plants. These species were assessed for their density, size-class distribution and regeneration using plots in areas of the Forest adjacent to local people. were available for the population biology of the preferred As no data tree species, their distribution and responses to exploitation, the maximum sustainable yield could not be determined (see Lund, 1998; Ruiz Pérez and Arnold, 1996). The information provided by this study can only be used to understand the importance of Nyungwe Forest in the livelihoods of local communities as well as to justify the management policies of the Forest. To be able to determine sustainable levels of resource use requires information about the density, the distribution of the resources within the forest, the population structure, the productivity of these resources and the ecological impact of different harvesting levels (Cunningham, 1999). In this study, plots were localised, and therefore particular tree species that are restricted to certain habitats could have been missed. However, as the plots were located in areas of the Forest near the local communities, the study provides baseline data on the availability, abundance, size-class distribution and regeneration of the tree species of relevance to local people. This study can also be used to inform managers about the impact of illegal use of the species and as an aid to developing potential management strategies for future use. The impact of illegal logging in Nyungwe Forest could be partly reflected in differences in densities of seedlings and medium trees between the study sites. These differences could also be attributed to environmental conditions and types of disturbances including fire and grazing. The periphery of the Forest has been altered; 102 http://scholar.sun.ac.za Stellenbosch University this has been substantiated by the chi-square analysis of the densities of trees of different sizes among and within the study sites. The presence of many small individuals of the preferred tree species at many sites suggests that the periphery of Nyungwe Forest is in the process of secondary succession. species including neomildbraediana Polyscias fulva, Neoboutonia It is likely that some macrocalyx and Macaranga which are characteristic of secondary forests (ORINFOR, 1978) will become more common where human use is heavy. The total stem density of all species and the density of the trees of the preferred species determined the relative abundance of each species within each category of product. The density of preferred trees can only convey an impression of available resources because the majority of the species were preferred for various products. The densities provided could be used as conservative criteria to set harvesting quotas which do not rely on an estimate of maximum sustained yield. A resource plant may be abundant but for some reason not sufficiently appropriate for use for different products. For instance, preferred tree species for timber and poles do exist in the Forest but the densities of quality stems were sometimes in low proportion relative to the stem densities of some species classified as potentially utilisable for timber and poles. Other characteristics including size-class distribution and regeneration should be examined in order to determine sustainable utilisation of the species. Size-class distributions of trees may be used to predict the age structure and state of a forest (Lorimer, 1980), to determine the regeneration status of dominant tree species (Geldenhuys and Muray, 1992), and as an indicator of natural or human disturbance history (Lorimer, 1980). Size-class distributions of trees prior to harvesting provide important baseline information for future monitoring work (Cunningham, 1999). A comparison of the size-class distribution prior to and after harvesting indicates changes in the rate of seedling and sapling establishment in the harvested population (ITFC-EMP, 1999). The analysis of size-class distributions of the preferred tree species indicated that many populations examined were unstable, with variable transitions from one size class to the next. between successive A stable population is characterised size classes (Taylor and Walker, by low ratios of change 1984), corresponding to Stellenbosch University 103 http://scholar.sun.ac.za marginally convex curves without clear inflections (Knowles and Grant, 1983). Usually size-class information is analysed in terms of a negative exponential distribution (e.g. Midgley et al., 1990). With the exception of Carapa grandiflora and Symphonia globulifera, most of the species did not conform to this ideal. Since they showed a negative exponential distribution, the populations of Carapa grandiflora and Symphonia globulifera had more juvenile than adult trees, indicating they were more stable in densities and were self-replacing (Geldenhuys, 1992; Hall and Bawa, 1993). The depicted size classes for the remaining species represented unstable populations. Some species exhibited either the loss of reproductives or had temporarily variable recruitment. Peters (1994) attributed the resulting size-class distribution in undisturbed stands to sporadic or irregular seedling establishment where the level of regeneration maintains the population with the 'infrequency' of occurrence causing 'peaks' and 'valleys' in the size class distribution. The populations which exhibit this pattern are commonly species that depend on canopy gaps for regeneration or whose regeneration has been interrupted by seed mortality, physical damage to seedlings or lack of pollinators or dispersal agents (Armesto and Fuentes, 1988; van Wyk et al., 1996). Entandrophragma excelsum has shown exception in that it lacks seedlings. Midgley et al. (1995) expressed concern at the apparent absence of regeneration of some trees in Southern African forests. They speculated that there might be some recent changes in disturbance regimes in these areas. In retrospect, the lack of regeneration of E. excelsum could be due to the lack of major disturbances like those caused by elephant and buffalo when their herds were present in Nyungwe Forest. The low density of large trees of E. excelsum could also be due to intensive pitsawying of the species for valuable timber in the past (Habiyambere, 1999). Being a primary species, the population of this species could permanently disappear from the Forest if the current conditions do not change. Size-class distributions of the preferred tree species indicated predominantly that a few number of species were in natural or unharvested condition. Many species exhibited a size-class distribution frequently encountered among light demanding, 104 http://scholar.sun.ac.za Stellenbosch University early pioneer species, which require large canopy gaps for regeneration. absence of such a disturbance, these species whose regeneration Parinari excelsa, Pauridiantha paucinervis, In the is limited (e.g. Casearia runssorica and Cassipourea ruwenzoriensis) may temporarily disappear from the Forest, their populations being represented only by seeds lying dormant in the soils. The results of the study concur with the findings of Zimmermann et al. (1994) that regeneration in tropical wet forests is an important life history attribute of the species. Significant differences in the abundance of seedlings of different species across sites indicated poor regeneration and recruitment of some species in some sites. This may be caused, in addition to differences in environmental conditions (e.g. soils, rainfall) and the ability of the species to reseed, by differences in canopy cover or light gaps in the Forest (Armesto and Fuentes, 1988). Fire is reported to be the most common disturbance in Nyungwe Forest (Gapusi, 1999) and could have affected differently the part ofthe Forest being surveyed. 4.3 POTENTIAL SUPPLY AND VULNERABILITY OF THE PREFERRED TREE SPECIES TO EXPLOIT ATION The discussion in section 4.2. has attempted to show that Nyungwe Forest has different species with a variety of ecological characteristics that make sustainable harvesting a very difficult objective. The major problems identified by the study are the high diversity and low population density, the unstable size-class distribution and the low density of regeneration of many species. Ignoring these characteristics could cause an unsustainable exploitation of timber, poles and medicinal materials. According to Hall and Bawa (1993), extraction is considered sustainable if the impact of harvesting has no significant effect on the reproduction and regeneration of the population being harvested. A combination of criteria including density, size-class distribution and regeneration could convey an impression about the vulnerability of tree species to exploitation (Tables 4.1a and 4.1b). 105 http://scholar.sun.ac.za Stellenbosch University Table 4.1a. Classification of the preferred timber and pole species into 'vulnerable' and 'non-vulnerable' categories. Vulnerable Tree resources Timber Carapa grandiflora Entandrophragma excelsum Newtonia buchananii Parinari excelsa Polyscias fulva Strombosia scheffleri Symphonia globulifera Syzygium parvifolium Non-vulnerable x x x x x x x x Poles Carapa grandiflora Casearia runssorica Cassipourea ruwenzoriensis Macaranga neomildbraediana Strombosia scheffleri Symphonia globulifera Syzygium parvifolium x x x x x x x Table 4.1b. Classification of the preferred medicinal tree species into 'vulnerable' and 'non-vulnerable' categories Tree resources Medicinal tree s[!ecies2 laq:;e tree catelj:;ory Carapa grandiflora Entandrophragma excelsum Maesa lanceolata Newtonia buchananii Oxyanthus troupinii Parinari excelsa Pauridiantha paucinervis Symphonia globulifera Syzygium parvifolium Tabernamontana johnstonii Medicinal s[!ecies2 medium tree catelj:;o[Y Canthium oligocarpum Carapa grandiflora Newtonia buchananii Oxyanthus troupinii Parinari excelsa Pauridiantha paucinervis Symphonia globulifera Syzygium parvifolium Vulnerable Non-vulnerable x x x x x x x x x x x x x x x x x x As the information provided is a baseline data, the grouping of these species into vulnerable and non-vulnerable categories could serve as a benchmark for describing the initial ecological conditions that have to be monitored in the process of resource exploitation. Obviously, abundant species with many seedlings (e.g. Carapa grandiflora, Symphonia globulifera and Strombosia scheffleri) are less vulnerable to 106 http://scholar.sun.ac.za Stellenbosch University exploitation. More important than the overall abundance of the preferred tree species is the size-class distribution of individuals of each species. Stem densities of some species including Cassipourea ruwenzoriensis and Newtonia buchananii were dominated by the preponderance of small trees, large trees being very few. The exploitation of such species could not be maintained on a sustained yield basis and therefore species which approximate an inverse J-shaped curve are recommended for exploitation. Entandrophragma Felling of large trees of species including excelsum, Newtonia buchananii and Syzygium parvifolium for timber could eliminate these species from the Forest in a much shorter period of time because of low densities of their stems and instability of their populations. For many species including Carapa grandiflora, Symphonia globulifera, Strombosia scheffleri and Cassipourea ruwenzoriensis, poles were more abundant compared to the densities of harvest able individuals for timber. Another criteria, part of plant used, could be added to the three criteria in order to determine the vulnerability of harvesting medicinal materials. As no data was available on the impact of harvesting medicinal materials from the species, the information about vulnerability of medicinal species to exploitation is incomplete. In addition, it is clear that many of the tree species are favoured for more than one products, which might lead to overexploitation as observed by Fedlmeier (1998). Harvesting methods also have an impact on the sustainability of the resources. Geldenhuys (1998) proposed methods of timber harvesting based on single tree selection in the management of natural forest in South Africa. The methods include harvesting trees selected from a stand on the basis of mortality rates and harvesting of dead and dying trees of commercially valuable species. The harvesting of over- mature and dying trees in Nyungwe Forest can be applied with the aim of maintaining a stable population of the preferred tree species. As a way of achieving this, a silvicultural system that could be employed would exploit products from overstocked size-classes as it has been applied in the South African forests (Geldenhyus, 1992). However, as for all natural forests, the system should aim to increase growing space for faster growth of useful stems and to establish the regeneration of the desirable species to replace the stem removed in logging (Baur, 1964; Roach, 1974). If the Stellenbosch University 107 http://scholar.sun.ac.za removal of trees across the different size classes is proportional to the relative densities of the stems in each size class, then there will be no change in the size-class distribution of the species at a time when absolute densities are declining. Information on the life history of the species can guide decision over exploitability of species, as the removal of some trees might not enhance the regeneration of shadetolerant species. Other factors such as growth rate and coppice response to exploitation of products need to be considered in order to determine species that are either vulnerable or not It can be hypothesised that vulnerable species are those vulnerable to exploitation. highly preferred for timber and poles, with slow growth rate and weak coppicing ability. Species that are likely to be less vulnerable would be those with relatively fast growth rate. No data are available on growth rate and coppicing ability of the species studied, and this need to be investigated. Monitoring operations and harvesting adjustments enable to determine the sustainability of harvesting levels (Peters, 1994). The most important characteristic that must be monitored is the density of regeneration. densities, Peters (1994) recommends remedial In the case of low seedling treatments including enrichment planting, selective weeding, cutting and removal of woody vines from the crown of adult trees. These treatments could result in reducing competition and increasing of light levels into the understorey. The periphery ofNyungwe Forest surveyed requires such treatments in order to improve the availability and recruitment of seedlings of the majority of the species including for instance Casearia runssorica, Cassipourea ruwenzoriensis, parvifolium. Parinari excelsa, Pauridiantha paucinervis and Syzygium Species such as Entandrophragma excelsum require enrichment planting or the re-introduction of locally extinct herbivores (e.g. elephant, buffalo) that will induce disturbances needed for the regeneration of the species. 4.4 POTENTIAL FOR JOINT MANAGEMENT OF NVUNGWE FOREST After recounting the needs, problems and preferences of local communities as well as the potential supply of preferred tree species from Nyungwe Forest, the potential for community participation in the protection and management of the Forest is assessed. Stellenbosch University 108 http://scholar.sun.ac.za The discussion is guided by the experiences of joint forest management (JFM) or collaborative management in countries where these techniques have been applied. In Rwanda, JFM or collaborative management of protected forests has not previously existed. In the case of Nyungwe Forest, however, there have been programmes of social forestry that dealt with development activities in the rural areas adjacent to the Forest (Barbier, 1992; Labrousse, 1992). These were provided as incentives for conservation and to reduce dependence of local communities to the forest resources. Deforestation has been reported to continue at an alarming rate (e.g. Gapusi, 1999; MINITERRE, 2000) to provide in the daily food needs of the adjacent household to the Forest. Cutting of trees for various uses is an important cause of deforestation, arising from poverty and insufficient land to establish sources of tree products. This degradation causes a decline of species suitable for wooden goods, furniture, and basketry materials. The consultancy work by Gapusi (1999) indicated a number of wrong perceptions which relate to the meaning of sustainability and the dependence of adjacent households to the Forest. This accentuates the need for implementation of collaborative resource management strategies in the adjacent areas to the Forest. The effective and meaningful involvement of local communities in natural resource management has been attempted under JFM in India by linking socio-economic incentives and forest management (Badola, 1999; Tewari, 1992). Many countries in the world have undertaken a series of initiatives aimed at devolving management, control and ownership of natural forest and woodland resources to local communities. In India, resource degradation has led the government to enter into partnership agreements with communities whereby the responsibility for forest management is placed with communities. In return, these communities are allowed to harvest certain species in certain areas and to receive a fixed percentage of revenues from harvesting by the state forest department (Mukerji, 1992). Other countries community-based in Africa have changed their national forest management. forest policies In Ghana, for instance, toward a collaborative management scheme has been adopted in order for the community to benefit from the forest resources (Prah, 1997). In Zimbabwe, the CAMPFIRE programme insures that 109 http://scholar.sun.ac.za Stellenbosch University revenues derived from wildlife reach councils and communities (Scoones and Matose, 1993). Elsewhere in Africa, community based approaches to forest management have been adopted: in Tanzania in the Duru-Haitemba and Mgori forests (Sjoholm and Wily, 1995; Wily and Haule, 1995), in the Bwindi and Mount Elgon forests of Uganda (Wily, 1993), the Kilum and Ijim montane forests of Cameroon (Nurse et al., 1994), the forests of the Senegal Valley (Diouf, 1994), the Toumousseni forest in Burkina Faso (Hagberg, 1992), the Gshaka Gumti Park in Nigeria (Dunn, 1994), woodlands near Mopti in Niger (Kerkhot, 1990), the El Ain woodlands of Sudan (El Din and Shanks, 1992) and the woodlands in Mali (Skinner, 1988). JFM is also emerging as a strategy for conservation and sustainable use of forests as in Sri Lanka (Carter et al., 1994), Thailand and the Philippines (Poffenberger and McGean, 1998). Joint forest management approaches that work elsewhere may not be applicable to Rwanda in the context of the current legal and policy provisions particularly for protected areas, including Nyungwe Forest. of forests, Currently, the major focus is on the conservation of biological diversity and integrity. The presence of a unique wealth of biodiversity makes Nyungwe Forest an extremely sensitive area. The high conservation status, however, does not imply the exclusion of local communities to access certain resources and to participate in the management of the Forest. The collaborative management at Bwindi Impenetrable and Mgahinga Gorilla National Parks in Uganda is an example of community use of wood and non-wood products (Wild and Mutebi, 1996; Cunningham, 1996). The conservation and management of Nyungwe Forest could follow the examples of the Uganda National Parks and allow access to some resources, including those identified by this study. Many factors favour the development of JFM in Nyungwe Forest. Firstly, communities adjacent to Nyungwe Forest have used products from the Forest in the past. Secondly, the Action Plan for the conservation and management of natural forests in Rwanda adopted in 1984, conceived a three-zone system as follows (Fimbel and Kristensen, 1994): • Forest fringe zone where some timber harvesting was permitted (10% of all the forest area); 110 http://scholar.sun.ac.za Stellenbosch University • Natural reserve zone where minimal use was allowed (40% of all the forest area); • High protection zone, where resource use was restricted (50% of all the forest area). The subdivision of Nyungwe Forest into different zones approximated the system of buffer zones adopted in Uganda (see Wild and Mutebi, 1996). The restriction of resource access to the buffer zone can imply the loss of access to some resources, for example bamboo, which is found in the south-eastern part of Nyungwe Forest. However, since bamboo can be grown in gardens, it is possible to produce this resource outside the Forest. Thirdly, the decentralisation of power and authority by the central government aims to involve local people and their leadership in the management of all affairs related to socio-economic development in their regions. The role of community leadership is one of the important factors affecting the success of collaborative management (Wild and Mutebi, 1996). The survey of the respondents' perceptions of potential institutions for the management of utilisation found different organisations that could operate at local level for the management including administrative of the Forest. These organisations, District, Sectors and Cells, work with or have a direct relationship with the government structures and interact already with the forest authorities and conservation services. Due to its power, authority, organisational structure and mandate, a District could be one of the actors involved in management and conservation of the Forest. respondents The various categories of institutions listed by the as being more appropriate show the importance of having vanous stakeholders including both users and non-users in forest management. Fourthly, the conservation services do not exercise adequate harvesting of resources, which leads to overexploitation control over the and loss of biodiversity. Since 1994, there has not been effective management of the Forest. The conservation project staff is too small, without resources and therefore ineffective (WCS and ORTPN, 1995). There is patrolling of the Forest, but still there is intense illegal activities going on. Without effective management, the Forest will continue to be exploited, albeit illegally, by adjacent communities. The lack of an approach toward Stellenbosch University III http://scholar.sun.ac.za developing partnership between the conservation services and local organisations could exacerbate current conflicts over resource use and access. The local use of resources from the Forest can be based on the subdivision of the Forest into different management zones and on the vulnerability of some resources to exploitation (Peters, 1996). Because collaborative forest management or JFM involves agreement between resources users and the owner of the forest (in this case the government), further clarification of who can use the resources and who will be responsible for the management of the Forest may be required. Obviously, as has been the case in other countries, the foremost users are those people who live next to the Forest, but they need to be organised into user groups. People living in the environs of the Forest already work in associations or groups in various sectors such as crop production, livestock farming and beekeeping, among others. These associations provide opportunities for the formation of users groups for collaborative management. However, because the users are many and the Forest is large, the use of the resources could be demarcated following administrative boundaries, instead of natural boundaries as has been the case in Tukucha and Banskharka panchayats of Nepal (Hobley, 1987). This could enable the involvement of local administration in decision-making regarding utilisation and management of the part of the Forest falling within its boundaries. In order to make use of the Forest sustainable, it is desirable to link the forest produce with local enterprises, for both wood and non-wood products (Encarnacion, 1999). The development of alternatives to resource use could be an important complementary strategy for improving the livelihoods of the adjacent communities. Enabling access to some resources in the Forest alone cannot solve the livelihood problems of the population. Substitution activities can provide for some of the preferred and requested resources as a remedial strategy to low resource availability. Alternative activities could include local cultivation of indigenous tree and shrub species for the production of medicinal plants, basketry materials, poles, and wooden and woven products, following the example of the Guira Banso project in Ghana (Prah, 1997). 112 http://scholar.sun.ac.za Stellenbosch University Incentives for community involvement in forest protection and management can also come in the form of income, goods, and power to control the forest by local communities. Ecotourism in Nyungwe Forest could potentially provide sustainable financial activity, thus encouraging local people in the sustainable management of forest resources as has been the case for Budongo Forest Reserve in Uganda (Langoya and Long, 1998). In the long-term, the communities that live in the area surrounding the Forest can benefit directly from wildlife. This may be either through protein offtake or from financial benefit from revenue taken from tourists visiting to view wildlife. Such strategies are akin to the CAMPFIRE programme set up in Zimbabwe or the community conservation programmes successfully running in Kenya. In return for the benefit that local communities gain from wildlife, they can act to conserve the wildlife. A revenue sharing scheme has to be installed as a one of the means for resolving conflicts between forest authorities and adjacent communities. Money that is accrued from the Forest, once it has moved towards self-sustainability, can be used to answer to the immediate development needs of the local communities. Until the conservation service of the Forest has sufficient income from ecotourism revenues, the Rwanda Government in partnership with the donor community can have to work closely with the local human populations. In response to wood requirements, local communities could be encouraged, trained and provided with appropriate materials to establish their own sources of wood products. Forest projects in the region would, for example, aim at reducing the fuelwood and pole requirements. An alternative to fuelwood can be kerosene. The provision of efficient kerosene cookers and a campaign to familiarise people with its use would reduce the fuelwood requirements. If charcoal is used, then people could be trained how to produce efficient cookers. The most difficult, but most important aspect for the adjacent communities can be for the projects to induce a change from agriculture and livestock based economy to a monetary economy. The reasons for this change are as follows: (a) many aspects of households will increasingly need to acquire things that require money (e.g. clothing, school fee, house construction, medical fee); (b) the long-term impacts of the present 113 http://scholar.sun.ac.za Stellenbosch University economic system will prove to lead to reductions in farm sizes and heavy use of pastures, resulting in the households to become increasingly poor. Micro-economic projects that facilitate this move could be encouraged and subsidised. Such projects can also provide incentives for protection and conservation of the Forest through the development of local enterprises such as the production of bricks and roofing tiles, and the production of efficient ceramic cookers that could reduce the amount of fuel and building wood that is required by the households adjacent to the Forest. Incentives can be in the form of credits provided by projects to support the management of the Forest, as was done by the SAFIRE (Southern Alliances for Indigenous Resources) project in Zimbabwe (Grundy and Le Breton, 1998). The management of Nyungwe Forest should seek to diversify livelihood opportunities as much as possible among the communities living adjacent to the Forest. In the process of collaborative management, communities could be employed in the rehabilitation and enrichment planting of the degraded areas and in other silvicultural treatments inside the Forest. This tactic has shown success in some projects in India where local people were used as a source of labour (Anderson et al., 1998; Poffenberger et al., 1990). In return, communities could benefit from employment and rights to harvest certain abundant species sustainably. In order to avoid overexploitation, agreements between communities and forest managers are needed and could establish resource users, species and quantities to be harvested by each user and monitoring activities for managing resource use. It is possible to develop a joint venture between some resource user groups and the forest management authority to harvest some resources on sustainable basis, integrated with other forest management activities. The Forest authority can control the harvesting of resources through a permit system, but ensuring that permit rates do not exceed what the people can afford to pay, which could reduce illegal activities. Stellenbosch University 114 http://scholar.sun.ac.za 5 CONCLUSION AND RECOMMENDATIONS This research has shown that Nyungwe Forest is important to the livelihoods of many people living on its periphery, through the provision of wood and non-wood forest products. The preference and demand for forest products result from declining levels of soil fertility and the lack of livelihood alternatives in the study area. As a result of the demand for more livelihood alternatives, the managers of Nyungwe Forest are faced with a difficult situation arising from the effect of population pressure and rural poverty (Mitchell, 1997). This has led to illegal cutting of trees for various uses, and poaching of animals for food, gathering of wild vegetables and fruit, the harvesting of medicinal plants and mining in the valleys of the Forest. Although communities adjacent to the Forest have farmed trees, the products are less diverse and not enough to cater for their needs, implying increased reliance on the Forest for their livelihoods. However, not all resources preferred by adjacent communities are found in the Forest, and many available resources are present in insufficient abundance to be recommended for sustainable utilisation. As more than 50% of the preferred tree species for various uses were vulnerable to exploitation, the periphery of Nyungwe Forest is now considered to be an area of low resource potential for timber, pole and medicinal materials. Only a few tree species can be considered utilisable due to the fact that populations of many species were unstable. The uneven distribution of the range of size classes found for many species could be a cause for concern implications. and has management The removal of harvesting pressure for poles and timber may result in more stable populations of the preferred tree species (Shackleton, 1993). The effect of this is the development and the establishment of new seedlings at regular intervals leading to a standard size-class distribution of the species. The management of the preferred species for the sustainable supply of timber and poles can be achieved through appropriate silvicultural systems that maintain the characteristics of a stable population. As the density of harvestable trees for timber was very low, a silvicultural system based on single tree selection could allow the production of timber from overstocked Stellenbosch University 115 http://scholar.sun.ac.za diameter size classes in a well-defined zone of the Forest. However, timber harvesting should be controlled in order to avoid overexploitation and adverse impact on the growth dynamics of the harvested population. As the densities of poles were more than those of timber trees, the area of the Forest surveyed had more potential for the supply of poles than timber. Many tree species are favoured by local people for multiple uses. It is important to determine the use of each tree species if harvesting is to be sustainable. medicinal tree species have had their bark harvested for medicinal Many materials. Debarking of trees for the supply of medicinal material and using the same species for timber harvesting can adversely affect the regeneration and the genetic composition of the future population. Therefore tree species indicated for medicinal purposes should not be used for timber exploitation except if medicinal material is gathered only from the small trees or if the medicinal material is gathered before timber harvesting. Other resources of importance to the livelihoods of local communities gathered from Nyungwe Forest and are used for household consumption. are also Sustainable utilisation of these products by local communities can be one of the possible means of establishing positive linkages between local households and conservation of Nyungwe Forest. The gathering of these resources should also be organised, regulated and monitored. Substitutions have always management of forests. been considered an important component of the Some of the preferred tree species can be propagated on farms (example Entandrophragma excelsum for timber, Arundinaria alpina for basketry material and building purposes, Alchornea hirtella for bean stakes, Casearia runssorica for poles), which would reduce the pressure on these resources inside the Forest. Identification of alternatives to resources, which are less abundant and more vulnerable to exploitation, could be done in co-operation with the local communities. This study suggests a way of achieving a balance between community utilisation and the conservation ofNyungwe Forest. Collaborative management of the Forest can be a viable arrangement given well-structured local institutions in which local people have confidence, the government policy of decentralising power and authority to the Stellenbosch University 116 http://scholar.sun.ac.za lowest administrative level, and past experiences in the management of Nyungwe Forest. This management option could draw on approaches adopted in other countries such as India, where JFM emerged, and Uganda's Bwindi Impenetrable National Park where resource use by local people is permitted on specified zones of the Park. The use of resources in the Forest requires an enabling policy environment with a view to addressing the conflicts over resource use and access, silvicultural problems, productivity, and benefit sharing, through participation of local communities. Below is a summary of some suggestions regarding the use of Nyungwe Forest and areas for future research: • The management of Nyungwe Forest could permit the gathering of edible wild foods in areas outside the strict conservation zone of the Forest. Studies aimed at providing information on the availability, identification, abundance, seasonality, distribution and harvesting impact for these products have to be carried out in order to guide the frequency and quantities of removals. • Beekeeping can be a source of income to local communities. The current initiative of PCFN to stimulate the formation of beekeepers' associations which should use the resource should be encouraged. the responsibility In this way, beekeepers can take for preventing fires and avoiding the damaging use of the Forest. The traditional methods of honey production should be improved upon. • Bamboo is mostly used by people in the southwest ridge of the Forest, where it is dominant. A limited number of harvesters could be licensed to use the resource until sufficient supplies can be grown on farms. Research is needed to provide information on the effect of continued harvesting on the production of biomass in order to ensure long-term sustainability of bamboo extraction. • Mining can be carried out, but in accordance with the regulations of the Ministry of Commerce, Industry and Tourism and in areas defined by the conservation services. Local miners' groups can have access to the resource upon obtaining a mining title. Stellenbosch University 117 http://scholar.sun.ac.za • The illegal use of wood for the production of wooden items for household use has resulted in the scarcity of some tree species. This is a consequence of highly selective harvesting that removes the most fit individuals. Felling of these trees in the Forest should be stopped. Attention should be focussed on providing alternative sources of wood for wooden items, including, for instance, the planting of Polyscias fulva and Markhamia lutea outside the Forest or on farms. • Timber harvesting can be done for those tree species which have more individuals at the larger diameter size class than expected (e.g. Parinari excelsa, Strombosia scheffleri). The removals should be regulated and based on selective harvesting. • Indigenous pole species are preferred to exotic pole species. The indigenous poles are likely to be much sought after by the adjacent communities, overharvesting. leading to The management of Nyungwe Forest can determine harvesting cycles and the quantity of poles to be harvested for a particular species. The removal of poles as well as the harvesting of timber should ensure that the population of the species will perpetuate over time. • Traditional healers' groups can be permitted to collect medicinal plants. Due to the effect of debarking, the gathering of leaf materials should be encouraged along with the use of growth form categories other than trees (e.g. climbers, shrubs). Research is needed to evaluate the response of preferred tree species to disturbance, including harvesting of fruits and bark. • The management and conservation of Nyungwe Forest should promote the establishment of substitute resources, including the cultivation of preferred tree species, and assist the local communities to practice sustainable agriculture through the maintenance of productive cropping systems and the development of small, micro and medium enterprises based on cultivated resources. • The periphery of Nyungwe Forest is a secondary forest as a result of disturbances. Highly disturbed sites need to be identified and studied in terms of species Stellenbosch University 118 http://scholar.sun.ac.za composition and size classes. This can provide useful information for forest rehabilitation and restoration. • Studies on the ecology and the status of the preferred tree species should be carried out. This may involve the study of the dynamics, the constraints of life stages, productivity distributions. and growth rates, habitat requirements and spatial This can provide information on the management of the species for a sustainable supply of resources. The methods to be used should combine both plots and transects and should allow the evaluation of the spatial distribution of the preferred tree species along gradients and in different habitats. Stellenbosch University 119 http://scholar.sun.ac.za REFERENCES ADB (African Development Bank), 2001. African Development Report 2001. Fostering Good Governance in Africa. Oxford University Press, USA. AGRIFOR CONSULT, 1997. Réhabilitation du sous secteur forestier au Rwanda. Rapport de démarrage 27 Janvier - 30 Avril 1997. AGRIFOR CONSULT, Kigali. Anderson, J., Clément, 1., and Crowder, L.V., 1998. Accommodating conflict interests in forestry: concepts from pluralism. Unasylva 194: 1-56. ARDI, 1998. Stage de formation sur la culture des champignons. Agenee Rwandaise de Développement Intégré, Kigali. Armesto, J.J., Fuentes, E.R., 1988. Tree species regeneration in a mid-elevation, temperate rain forest in Isla de Chiloé, Chile. Vegetatio 74: 151-159. Kluger Academic Publishers, Dordrecht. Arnold, J.E.M., 1990. Social forestry and communal management in India, Social Forestry Network, Paper 11b Overseas Development Network, London. Aumeeruddy, Y.T., 1998. People and Plants applied ethnobotany project at Ayubi National Park, Pakistan. WWF, UNESCO and Royal Botanical Garden, Kew. Babbie, E. and Halley, F., 1998. Adventure in Social Research: Data analysis using SPSS for Windows 95. Pine Forge Press, California. Badola, R., 1999. People and protected areas in India. Unasylva 199: 12-14. Bahigiki, E. and Vedder, A., 1987. Etude socio-économique et proposition écologique sur la Forêt de Nyungwe (UGZ3). Min. Agric. Elev. et des Forêts, Direction des Forêts, Kigali. Barabwiliza, R. 1992. Problématique des réserves écologiques dans un pays ft économie essentiellement agricole - les forêts de montagnes et des volcans du Rwanda. Unpublished MSc. thesis. Fondation Universitaire Luxembourgeoise, Centre Interuniversitaire des Sciences de l'Environement, Luxembourg. Barabwiliza, R., 1980. Impact de l'homme sur les formations forestiêres des zones marginales au Rwanda. Bulletin Agricole du Rwanda 13: 14 -23. Barakabuye, N., 2001. Integration of local communities in the conservation of Nyungwe Natural Forest. ORTPN, Kigali. Barbier, C., 1992. Evaluation du Plan d'Action pour la conservation et l'aménagement des forêts naturelles de la Crête Zaïre-Nil. Rapport du consultant en sylviculture et aménagement. Min. Agric. Elev. et des Forêts, Direction des Forêts, Kigali. Baur, G.N., 1964. The ecological basis of rain forest management. Forestry Commission of New South Wales, Sydney. Bazzaz, F.A., 1998. Plant in changing environments: linking physiological, population and community ecology. Cambridge University Press, Cambridge. Ben Chaabane, H. and Cyiza, P., 1992. Méthodologie d'élaboration des données sur le revenu rural des communes Rwandaises. Direction de la Planification, Division des Stratégies de Développement Communal et Régional, Ministêre du Plan, Kigali. Bitariho, R., 1999. The abundance, distribution and use of montane bamboo in Bwindi Impenetrable and Mgahinga Gorilla National Parks, S.W. Uganda. Unpublished MSc. thesis. Makerere University, Institute of Environment and Natural Resources, Kampala. 120 http://scholar.sun.ac.za Stellenbosch University Blaikie , P.M. and Broofield, H.P., 1987. Land degradation and society, Methuen, New York. Bosshard, W. 1971. Rapport sur la situation forestiere du Rwanda et proposition pour la suite du projet. Birmensdorf. Bouxin, G., 1974. Etudes phytogéographique des plantes vasculaires du marais Kamiranzovu (Forêt de Rugege, Rwanda). Bulletin Jardin Botanique Belge 44: 141-159 Bryman, A., and Cramer, D., 1997. Quantitative Data Analysis with SPSS for Windows: a Guide for Social Scientists. Rowtledge, New York. Budowski, G., 1976. Proposition pour un programme de sauvegarde de la Forêt de Nyungwe. Ronéo, Min, Agric., Elevage, Kigali. Butynski, T.M., 1984. Ecological survey of the Impenetrable (Bwindi) Forest, Uganda, and recommendations for its conservation and management. Wildlife Conservation International, New York Zoological Society, USA. Cambrezy, L., 1984. Le surpeuplement en question: organisation spatiale et écologie des migrations au Rwanda. ORSTOM, Paris. Carter, J. Connelly, S. and wilson N., 1994. Participatory forestry in Sri Lanka: Why so limited. Change on the Horizon. Rural development Forestry Network, Document 17b, London. CGF, 1993. Rapport Annuel 1992. Département de Foresterie, ISAR-RUHANDE, Butare. Chanyenga, T.F., 1999. Non-wood forest products in Malawi. From: http://www.fao.org/forestry. Colchester, M., 1994. Slave and enclave: The political economy of Equatorial Africa. World Rainforest Movement, Penang. Cunningham, A.B., 1993. African medicinal plants: Setting priorities at the interface between conservation and primary healthcare. People and plant Initiative, Division of Ecological Sciences, Paris. Cunningham, A.B., 1996. People, park and plant use. Recommendation for multipleuse zones and development alternatives around Bwindi Impenetrable National Park, Uganda. People and plants Initiatives. Division of Ecological Sciences, UNESCO, Paris. Cunningham, A.B., 1999. Woodcarvings, drums and monitoring tree resources. A test-case of local-level monitoring in Mpanga Forest, Uganda. Field workshop, 8 - 12 March 1999. Unpublished report. WWFIUNESCO/Kew People and Plants Initiative. Paris. Cunningham, A.B., 2000. Methods for selection and monitoring of plants species in multiple-use zones. A training course at the Institute of Tropical Forest Conservation, Bwindi Impenetrable National Park, Uganda. WWFIUNESCO/Kew people and Plants Initiative. Unpublished Report. Cunningham, A.B., and Mbenkum, F.T., 1993. Sustainability of harvesting Prunus africana bark in Cameroon. People and Plants Working Paper No.2, UNESCO, Paris. Davis, L. S., Johnson, K. N., 1987. Forest management. Third Edition. McGrawHill, New York. Deffar, G., 1998. Non-wood forest products in Ethiopia. From: http://www.fao.org/forestry den Biggelaar, C., 1996. Farmer experimentation and innovation: a case study of knowledge generation processes in agroforestry systems in Rwanda. FAO, Rome. 121 http://scholar.sun.ac.za Stellenbosch University Dessouter, S., 1991. Pharmacopée humaine et véterinaire au Rwanda. Institut National de Recherche Scientifique, Butare. DGF, 1985. Projet de la loi forestiere. Direction Générale des Forêts, Kigali, Rwanda. Diouf, B., 1994. The responsibility of local people for the management of forest resources in North Senegal. Rural Development Forestry Network, Network Paper l8d, , ODI, London. Dubois, 0., 1998. Rights and wrongs of rights to land and forest resources in SubSaharan Africa: Bridging the gap between customary and formal rules. Forest Participation Series No.lO. Forestry and Land use Programme, lIED, London. Dunn, A., 1994. Gashala Gumti, Nigeria - from game reserve to national park., Rural Development Forestry Network, Network Paper 25d ODI, London. El Din, A. and Shanks, E., 1992. Natural forest management project, Kozdofan, Sudan. Mid-term review. ODI, London. Emerton, L., Mfunda, 1., 1999. Making wildlife economically viable for communities living around the Western Serengeti, Tanzania. lIED Biodiversity Group, London. Encarnacion, C.S., 1999. Community based enterprises and conservation: the Kalahan Forest Farms Development Projects. Unasylva 199: 15-17. FAO, 1989. Food and Agricultural Legislation 38: pp. 122 -125, FAO, Rome FAO, 1992. Produits forestiers non-ligneux : Quel avenir? FAO, Rome. FAO, 1996. Projet de Stratégie pour le développement agricole national - Horizon 2000 Rwanda. Suivi du Sommet Mondial de I' Alimentation. FAO, Kigali. FAO, 1999. Etudes des resources forestiêres naturelles et les plantations forestiêres du Rwanda. FAO, Rome. Fedlmeier, e., 1998. The development of secondary forest on abandoned grazing areas, using Costa Rica as an example. Plant Research and Development, Volume 47/48. Institute for Scientific Co-operation, Tubingen. Fimbel, R.A. and Kristensen, K.A., 1994. Gold mining activities within the UGZ4 management Zone, Nyungwe Forest Reserve, Rwanda. Report to WCS. Natural Resources Management Project, Kigali. Fischer, E., et Hinkel, H., 1992. La nature du Rwanda: Apercu sur la flore et la faune. Mayence. Fisher, R.J., 1999. Devolution and decentralisation of forest management in Asia and the Pacific. Unasylva 199: 3-5. Gapusi, R.J. and Mugunga, P.e., 1997. Espêces rares et en extinction dans la flore Rwandaise. Revue bibliographique. ISAR-RUHANDE, Butare. Gapusi, R.J., 1999. Identification et Analyse des Options pour la Conservation de la Biodiversité dans la Forêt Naturelle de Nyungwe, Direction de l'Environnement, Kigali. Gasana K.J., 1988. La foresterie sociale au projet Crête Zaire-Nil. CZN, Gikongoro GEC, 2000. Conservation programs. Green Environmental Conservation, Kigali. Geldenhuys, C.J. and Murray, B., 1992. Floristic and structural composition of Hanglip forest, nothern Transvaal. Report No. FOR-DEA 432, Division of Forest Science and Technology, CSIR, Pretoria. Geldenhuys, C.J., 1992. The use of diameter distributions in sustained-use management of forests: examples from South Africa. The Ecology and Management of Indigenous Forests in South Africa. Oxford University Press, Cape Town. Geldenhuys, C.J., 1998. Development of a forest management system for sustainable resource use: basic principles with an example from South African evergreen 122 http://scholar.sun.ac.za Stellenbosch University forest. Paper presented at International Workshop on 'Sustainable management of African Rainforests today', Libreville. Grison, F., et Braesco, A, 1984. Forêt de Nyungwe: Factibilité du Projet Forestier. Rapport Scientifique aux actions forestiêres, Caisse Centrale de Coopération Economique, CTFT. Grundy, I. and Le Breton, G., 1998. Le programme MIT! de la SAFlRE - une nouvelle approche en matiere de gestion des resources naturelles dans les zones communales au Zimbabwe. Rural Development Forestry Network 22e, ODI, Londres. Habimana, P., 1982. Etude géographique de l'impact de la Forêt de Nyungwe. Influence socio-économique sur les Districts de Mudasomwa et Nyamagabe. Mémoire présenté en vue de I' obtention du grade de licencié es sciences. Université Nationale du Rwanda, Fac. de Lettres, Butare. Habiyambere T., 1999. Etude pour l'élaboration d'un plan d'action stratégique pour la conservation et la gestion de la biodiversité des ecosystémes. Cas du Rwanda. Direction des Forêts, Kigali. Habiyambere, T., 1997. Mot d'ouverture il l'occasion du séminaire-atelier sur « La Nécessité d'une agriculture durable: Contribution il l'agroforesterie ». in: La Nécessité d'une agriculture Durable: Contribution il l'agroforesterie. Rapport du Séminaire- Atelier. ISAR-RUHANDE, Butare. Hagberg, S., 1992. Management of natural forests. The gazetted forest reserve of Toumousseni, Burkina Faso. Working Paper 215, Swedish University of Agricultural Sciences, Uppsala. Hall, P. and Bawa, K., 1993. Methods to assess the impact of extraction of nontimber tropical forest products on plant population. Economic Botany 47: 234247. New York Botanical Garden, New York. Hasler, R. 1999. Overview of the social, ecological and economic achievements and challenges of Zimbabwe's CAMPFIRE Programme. Evaluating Eden EE DP03, Biodiversity Group, lIED, London. Hayes, W.AF., 1991. La culture des champignons au Rwanda. TECHNO SERVE Inc., Kigali. Hecq, J. and Froment, D., 1961. Contribution il l'étude des problêmes de reboisement et de· la conservation des sols - Structures agraires traditionnelles, rationalisation de l'agriculture, conservation du sol et amélioration de la productivité - le Rwanda et le Burundi. Institut National pour l'Etude Agronomique du Congo. Brussels. Hobley, M., 1987. Involving the poor in forest management: Can it be done? The Nepal - Australia Project experience. ODI, Agricultural Education Unit, London. Horward, P.C., 1991. Nature conservation in Uganda's tropical forest reserves. Forest Conservation Programme, IUCN, Gland. Horwich, R.H., 1988. The Community Baboon Sanctuary. An approach to the conservation of private land in Belize. In: Bradwold and Greenberg, eds. Saving the tropical forest. Earthscan. Infield, M., 1988. Hunting, trapping and fishing in villages within and on the periphery of the Korup National Park. WWF, Yaoundé. Inglis, AS., 1999. Implication of devolutiion for participatory forestry in Scotland. Unasylva 199: 45-51. ITFC-EMP, 1999. The potential supply of weaving and medicinal plant resources in the proposed Remera multiple-ues zone of Bwindi Impenetrable National Park, Stellenbosch University 123 http://scholar.sun.ac.za S.W. Uganda. Institute of tropical Forest Conservation, Ecological Monitoring Programme, Kabale. ITFC- EMP, 2001. Protocol for monitoring harvesting impacts of key utilised plant species in Bwindi Impenetrable National Park, S.W. Uganda. Institute for Tropical Forest Conservation, Ecological Monitoring Programme, Kabale. ITTO, 1990. Guidelines for the sustainable management of natural forest. Technical series No.5, International Tropical Timber Organisation, Yokohama. IUCN, 1979. Action Strategy for Protected Areas the Afrotropical Realm. IUCN, Gland, Switzerland and Cambridge. IUCN, 1987. Action strategy for protected areas in the Afromontane Realm. IUCN, Gland, Switzeland and Cambridge. IUCN/WWF, 1985. Rapport d'une mission au Zaïre et Rwanda. IUCN, Gland and Cambridge. JORR (Journal Officiel de la République Rwandaise), 1989. Loi No.4711988 du 5 Décembre 1988 portant organisation du régime forestier. 23eme Année. Journal Officiel de la République Rwandaise 3; 192 - 220. Kamatenesi, M., 1997. The conservatiion and utilisation status of Nyakibazi (Rytigynia spp.) in the multiple-use areas of Bwindi Impenetrable National Park, Uganda. Unpublished MSc. thesis. Makerere University, Institute of Environment and Natural Resources, Kampala. Kayinamura T., 1986. La Médecine TraditionnelIe Rwandaise: Centre de Médecine TraditionnelIe de Bare. Centre de Médecine TraditionnelIe de Bare, Kibungo. Kent, M. and Coker, P., 1992. Vegetation Description and Analysis: A Practical Approach. CRC Press, Boca Raton, Florida. Kerkhof, P., 1990. Agroforestry in Africa: A Survey of Project Experiences. Panos, London. Klein, M. and Van der Wal, M., 1998. About tropical hardwood, hunters and gorillas: Conservation of forest fauna in South Cameroon. In: The Congo Bassin. IUCN, The Netherlands. Knowles, P., and Grant, M.C., 1983. Age and size class structure analyses of Engelmann spruce, Ponderosa pine, Lodgepole pine, and Limber pine in Corolado. Ecology 64: 1-9. Kollert, W., 1998. Sustainable resources management in Sabah, Malaysia: Licensing policies and economics. Plant Research and Development, Volume 47/48. Institute for Scientific Co-operation, Tiibingen. Kothari, A., Patha, K.., and Vania, F., 2000. Where communities care. Community based wildlife and ecosystem management in South Asia. Evaluating Eden EE03, Biodiversity Group, lIED, London. Kruger, L.M., Midgley JJ. and Cowling, R.M., 1997. Resprouters vs reseeders in South African forest trees: a model based on forest canopy height. Functional Ecology 11: 101 - 105. British Ecological Society, Great Britain. Labrousse, R. 1992. Evaluation conjointe du Plan d'Action pour la conservation et l'aménagement des forêts de la Crête Zaire-Nil. Section 2. Rapport sectoriel: Economie. Caisse Francaise de Development, Kigali. Langoya, C.D. and Long, C., 1998. Local communities and ecotourism development in Budongo Forest Reserve, Uganda. Rural Development Forestry Network, Network paper 222. ODI, Portland House, Stag Place, London. Lawes, M.l., Mander, M., and Cawe, S., 2000. The value and uses of natural forests. South African Forestry Handbook, Vol.2, pp. 613-624. South African Institute of Forestry, Pretoria. 124 http://scholar.sun.ac.za Stellenbosch University Lonely Planet, 2001. Rwanda Map. Available at: www.lonelyplanet.com. Lorimer, CJ., 1980. Age structure and disturbance history of a southern Appalachian virgin forest. Ecology 61: 1169 - 1184. Luken, J.O., 1990. Directing Ecological Succession. Chapman and Hall, Great Britain. Lund, H.G., 1998. IUFRO guidelines for designing multiple resources inventories. IUFRO World Series Vol. 8. IUFRO Secretariat, Vienna. MacKinnon, J. and MacKinnon, K. 1986. Review of the protected areas system in the Afrotropical Realm. IUCN, Gland, Switzerland and Cambridge, UKlUNEP, Nairobi. Mbarubukeye S., 1992. La médecine vétérinaire traditionnelle : Rapport du 1er atelier des guérisseurs traditionnels du bétail au Rwanda. LVNR, Kigali. Mbarubukeye, S. et Niang, A.I., 1996. Les usages particuliers des plantes ligneuses autochtones du Rwanda: Utilisation en médecine traditionnelle humaine et vétérinaire, en alimentation humaine, en apiculture et en artisannat. LVNR, Kigali. Mbonyintwari, A., 1989. L'évaluation écologique des terres dans le cadre de l'aménagement des forêts naturelles de la Crête Zaire-Nil du Rwanda. Mémoire présenté en vue de I'obtention du grade de maitre es sciences (MSc.), Ecole des Gradués, Université Laval Mead, R., Curnow, R.N. and Hasted, A.M., 1993. Statistical Methods in Agriculture and Experimental Biology. Second Edition. Chapman & Hall, Great Britain. MICOMART, 1998. Proposition d'un projet d'étude de réalisation d'une unité pilote de production d'huiles essentielles au Rwanda. Min. Commerce, Artisan et Industrie, Kigali. Midgley, J., Seydack, A., ReyneIl, D., and McKelly, D., 1990. Fine-grain pattern in Southern Cape plateaux forests. Journal of Vegetation Science 1: 539-546. Opulus Press, Uppsala. Midgley, J.J., Everard, D.A., and van Wyk, G., 1995. Relative lack of regeneration of shade-intolerant canopy species in some South African forests. South African Journal of Science 91: 7-8. Mihigo, A., 1999. Situation du secteur forestier et des statistiques forestiéres. In Workshop on data collection and analysis for sustainable forest managementLinking national and intrnational efforts EC - FAO partnership programme (1998-2000). Projet GCP/INT/679/EC, Lambarene, Gabon, 27 September - 1 October 1999. Ministêre de l"agriculture, de l'Elevage et des Forêts, Kigali. MINAGRI, 1984. Plan d'Action pour la conservation et l'aménagement des forêts naturelles de la Crête Congo-Nil. Direction Générale des Forêts, Kigali. MINAGRI, 1990. Rapport annuel 1990. Projet Crête Zaïre-Nil. Min. Agric. Elevage, Kigali. MINAGRI, 1991. Synthése des réalisations des UGZ. Min. Agric. Elevage, Kigali. MINAGRI, 1992. Evaluation du Plan d' Action pour la conservation et l'aménagement des forêts naturelles de la Crête Zaïre-Nil. Section 1. Ministêre de l'agriculture, de l'Elevage et des Forêts, Direction des Forêts, Kigali. MINAGRI, 1993. Orientation prioritaires ft prendre en compte pour la poursuite de la mise en oeuvre du Plan d' Action pour la conservation et l'aménagement des forêts naturelles de la Crête Zaïre Nil. Document de Travail. Direction des Forêts, Kigali. MINAGRI, 1997. Agenda 21 National. Min. Agric. Elevage, Environ. Dev. Rural, Kigali. Stellenbosch University 125 http://scholar.sun.ac.za MINAGRI, 1998. Strategy and action plan for food security. Ministry of Agriculture, Livestock, Environment and Rural Development, Kigali. MINAGRI, 2000. Food Security Survey: Phase I. Agricultural production and land use. FSRP and DSA, Kigali MINECOFIN, 1998. Socio-Demographic Survey of Rwanda 1998. Office National de la Population, Kigali. MINITERRE, 2000. Stratégies et Plan d' Actions pour la Conservation de la Biodiversité dans les Aires Protégées. MINITERRE, Direction de l'Environnement, Kigali. Mitchell, T., 1997. Rwanda and Conflict. ICE Case Studies, No. 23. Muderevu, F., 1974. Estimation de la régression de la forêt naturelle du Rwanda par photo-interprétation. Projet Pilote Forestier, Kibuye. Muhwezi, 0., 1997. The biology, use and impact of harvesting on the Afromontane liana (Loeseneriella apocynoides) in and around Bwindi Impenetrable National Park, Uganda. Unpublished MSc. thesis. Makerere University, Institute of Environment and Natural Resources, Kampala. Mukerji, I.F.S., 1992. Greening of Himalayas through people's participation in the State of Himachal Pradesh (Western Himalayas). Paper prepared for the conference on wise management of tropical forests. Oxford. Murekezi, A., 1999. Les données statistiques sur les produits forestiers non-ligneux au Rwanda. FAO, Kigali. Murererehe, S., 1999. Etudes sur les ressources naturelles et les plantations forestiêres du Rwanda, FAO, Kigali. Mvukiyumwami, J., 1987. Les páturages du Mutara et leurs problématiques. In: Bulletin Agricole du Rwanda 12: 214 - 232, Min. Agric. Elevage, Kigali. Mwima, P.M., 2000. Forest regeneration and ecological recovery within the Mbwa River Tract in Bwindi Impenetrable National Park, Uganda. Institute of Tropical Forest Conservation, Kabale. Ndayambaje, J.D., 1999. Identification et analyse des options pour la conservation de la biodiversité dans le Pare National de l' Akagera. Min. Agric. Elevage, Environ. Dev. Rural, Direction de l'Environnement, Kigali. Neuman, W.L., 2000. Social Research Methods: Qualitative and Quantitative Approaches. Allyn and Bacon, Massachusetts. Neumann, R.P., and Hirsch, E., 2000. Commercialisation of Non-Timber Forest Products: Review and Analysis of Research. Center for International Forestry Research, Bogor. Newafrica, 2000. Rwanda National Parks Maps. Available at: www.newafrica.com. Newbury, C., 1992. Rwanda: recent debate over governance and rural development. In: Hyden, G. and Bratton, M, eds. Governance and Politics in Africa. Nhira, c., and Matose, F., 1996. Joint forest management and resource sharing: Lesson from India for Zimbabwe. Forest Participation Series No.5, Forestry and Land use Programme, ODI, London. Niang, A.I. et Styger, E., 1990. Les systêmes d'utilisation des terres et leur potentiel agroforestier au Rwanda. ICRAF/ISAR, Butare. Nurse, M., McKay, c., Young, J., and Asanaga, C., 1994. Biodiversity conservation through community forestry in the montane forests of Cameroon. Rural Development Forestry Network, Network Paper 18d, ODI, London. Nyilimana, S., 1997. Politique de développement apicole au Rwanda. ARDI, Kigali. Offut, M., 1992. Tourism as a conservation strategy in Nyungwe Forest Reserve, Rwanda. IUCN/WCMC, Cambridge. 126 http://scholar.sun.ac.za Stellenbosch University Ogwal, S.F., 1998. A study of abundance, utilisation and management of smilax anceps in Bwindi Impenetrable National Park, S.W Uganda. Unpublished MSc. thesis. Makerere University, Institute of Environment and Natural Resources, Kampala. Olson, J.M., 1990. The impact of socio-economic factors on migration patterns in Rwanda. M.A. thesis. Department of Geography, Michigan State University. Olson, J.M., 1994. Farmer responses to land degradation in Gikongoro, Rwanda. Ph.D. dissertation, Department of Geography, Michigan State University. Olson, J.M., Manyara, G., Campbell, D.J., Lusch, D.P., and Hu, J., 1995. Rwanda Society Environment Project. Department of Geography, Michigan State University. ORINFOR, 1978. La Forêt de montagne du Rwanda. Office de l'lnformation, Kigali. ORTPN, 1989. Rapport Annue11989. Min. Agric. Elevage, Kigali. ORTPN, 1991. Plan de conservation de l'éléphant au Rwanda. Unpublished report. Office Rwandais du Tourisme et des Pares Nationaux, Kigali. Owiunji, I., 1997. The biodiversity survey of Bwindi Impenetrable National Park. Makerere University, Institute of Environment and Natural Resources, Kampala. Palm, C., Izac, A.M. and Vosti, S., 1993. Procedural Guidelines for Characterisation and Diagnosis. ICRAF, Nairobi. PCFN, 1989. Proceedings from the first international workshop for the conservation and management of afromontane forests, 19 - 23 June 1989 Cyangugu, Rwanda. Project de conservation de la Forêt de Nyungwe, Cyangugu. Percival, V. and Dixon, T.H., 1995. Environmental scarcity and violent conflict: The case of Rwanda. American Association for the Advancement of Science and the University of Toronto, Washington D.C. Perlez, J., 1988. Compromise keeps Rwanda's dwindling families of Gorillas alive. The Globe and Mail. P.A. 15. Peters, e.M., 1994. Sustainable harvest of non-timber plant resources in tropical moist forest: An ecological primer. USAID Biodiversity Support Programme, Washington D.C. Peters, C.M., 1996. Beyond nomenclature and use: A review of ecological methods for ethnobotanists. In: Alexiades M.N. eds. Selected guidelines for ethnobotanists research: Afield manual, pp. 241 - 276. Philip, M.S., 1983. Measuring trees and forests: A textbook written for students in Africa. University of Dar-es-Salaam, Dar-es-Salaam. Poffenberger, M. and McGean, B., 1998. Village voices, forest choices. Joint forest management in India. Oxford University Press, New Delhi. Poffenberger, M., Bhatia, K., and McGean, B., 1990. Forest management partnerships: regenerating India's Forests. Ford Foundation, New Delhi. Prah, E.A., 1997. Joint Forest Management - the Gwira- Banso experience. Commonwealth Forestry Review 76: 171 - 174. Prasaad, R., 1999. Joint forest management in India and the impact of state control over non-wood forest products. Unasylva 198: 58-62. Renner, D., 1991. Recherche Action: Initiation d'un processus pour la mise en oeuvre du programme de sensibilsation - vulgarisation concernant la conservation et l'utilisation durable de la Forêt Naturelle de Nyungwe. Premiere étape. Décembre 1989 - Décembre 1990. UGZl, Kibuye. Stellenbosch University 127 http://scholar.sun.ac.za Renner, D., 1992. Discussion sur la phase initiale 1991/92 du programme de sensibilisation-vulgarisation de l'UGZl concernant la conservation et l'utilisation durable du massifforestier de Nyungwe. AUGZ1, Kibuye. Roach, B.A., 1974. Selection cutting and group selection. Applied Forestry Research Institute Miscellaneous Report No.5, State University of New York, New York. Ruiz Pérez, M. and Arnold, J.E.M., 1996. Current issues in non-timber forest products research. Proceedings of workshop 'Research on NTFP'. Hot Springs, Zimbabwe 28 August - 2 September 1995. Center for Internatioanl Forestry Research, Bogor. Sarin, M., and SARTHl, 1996. The view from the ground: community perspectives on joint forest management in Gujarat, India. Forest Participation Series No.4, lIED, London. Scoones, 1. And Matose, F., 1993. Local woodland management: constraints and opportunities for sustainable resource use. In: Bradley, P.N. and McNamara, K., eds. Living with trees: policies for forestry management in Zimbabwe, pp. 157 - 198. World Bank Technical Paper 210, World Bank, Washington DC. Scott, P.J., 1992. Fringe benefits: Minor forest product collection within buffer zones as a potential tool for conflict resolution in Bwindi Impenetrable National Park. MSc. Thesis. Agricultural University of Norway. Shackleton, e.M., 1993. Demography and dynamics of the dominant woody species in a communal and protected area of the eastern Transvaal Lowveld. S. Afric. J Bot. 59: 569-574. Sjoholm, H. and Wily, L., 1995. Finding a way forward in natural forest management in Tanzania. The emergence of village forest reserves. IRDC Currents 10, International Rural Development Centre, Swedish University of Agricultural Sciences, Uppsala. Skinner, J., 1988. Toward better woodland management in the sahelian Mali. Pastoral Development Network, Network Paper 25d" ODI, London. Smith, LR., Wells, D.A., and Welsh, P., 1985. Botanical survey and monitoring methods for grasslands. Focus on Nature Conservation 10. Natural Conservation, Council. Snedecor, G.W., and Cochran, W.G., 1964. Statistical Methods. Sixth Edition. The Iowa State University Press, Iowa. Sood, M.P., 1994. New forest initiatives in Himachal Pradesh. Forest Participation Series No.3, lIED, England. Sorg, J.P., 1978. La Forêt de Nyungwe: Exploitation ou protection? In: Schweizerischen, zeitschrift fur forstwesen 129: 445-452. Spinage, C., 1972. The ecology and problems of the Volcano National Park, Rwanda. Biological Conservation 4: 194-204. Stanley, A.C., and de Olieviera Castro,G.M., 1959. Manual of Vegetation Analysis. Harper & Brothers, Publishers, New York. Stebler, J.D., Darricau, Y., Kassianoff, P., Combe, J., Sentenac, A. et D'Huart, J.P., 1984. Mission d'évaluation sur la conservation et l'aménagement des forêts naturelles de la Crête Zaire-Nil. Rapport de Synthêse, Direction Générale des Forêts, Kigali. Storz M., 1983. La Forêt de Nyungwe et sa Faune. Projet Pilote Forestier, Direction des Eaux et Forêts, Kigali. Struhsaker, T.T., 1997. Ecology of an African rain forest: logging in Kibale and the conflict between conservation and exploitation. Gainesville, University Press of Florida. Stellenbosch University 128 http://scholar.sun.ac.za Sun, C., Kaplin, B., Kristensen, K.A., Munyaligoga, V., Mvukiyumwami, J., Kajondo, E., and Moermond, T.e., 1996. Tree phenology of a tropical montane forest in Rwanda. Biotropica 28: 668-681. Taylor, S.E., and Walker, B.H., 1984. Autecology of an invading population of the cactus Cerus peruvianus (queen of the night) in the central Transvaal. South African Journal of Botany 52: 397-402. Tewari, D.N., 1992. Joint (participatory) forest management. The Indian Forester 119: 687-698. Troupin, G., 1978. Flore du Rwanda: Spermatophytes. Volume I. Institut National de Recherche Scientifique, Publication No. 18, Butare. Troupin, G., 1983. Flore du Rwanda: Spermatophytes. Volume II. Agenee de Cooperation Culturelle et Technique, Musée Royal de l"Afrique Centrale, Tervuren. Troupin, G., 1985. Flore du Rwanda: Spermatophytes. Volume III. Agenee de Cooperation Culturelle et Technique, Institut National de Recherche Scientifique, Butare. Troupin, G., and Bridson, D.M., 1982. Flore des plantes ligneuses du Rwanda. Institut National de Recherche Scientifique, Publication No. 21, Butare .. Troupin, G., and Donneaux, A., 1982. Plantes ligneuses du Rwanda: Lexique des Noms Kinyarwanda. Institut National de Recherche Scientifique, Butare. Twarabamenye, E., 1999. Identification et Analyse des Options pour la Conservation de la Biodiversité dans le Pare National des Volcans. Min. Agric. Elevage, Environ. Dev. Rural, Direction de l'Environnement, Kigali. UFRIC, 2000. Joint forest management. A community training manual for forest communities. Uganda Forestry Resources and Institution Centre, Makerere University, Kampala. UGZ1, 1987. Directives pour les interventions dans la zone tampon. AUGZ1, Kibuye. van Wyk, G.F., Everard, D.A., Midgley, J.J., and Gordon, l.G., 1996. Classification and dynamics of a southern African subtropical coastal lowland forest. S. Afric. J Bot.62: 133-142. Vedder, A. and Weber, B. 1984. Forest Conservation in Rwanda and Burundi. Swara 7: 32 - 35. Wambugu, E.W., 1999. Women's role in agroforestry: a case study of Embu District, Kenya. Unpublished MSc.thesis. University of Stellenbosch, South Africa. WCS and ORTPN, 1995. Nyungwe Forest Conservation Project. Progress Report April- September 1995. Office Rwandais du Tourisme et des Pares Nationaux, Kigali. Weber, A.W, 1992. Evaluation du Plan d' Action pour la Conservation et l' Aménagement des Forêts Naturelles de la Crête Zaïre-Nil. Rapport Sectoriel sur les Interactions Forêt - Population. Section 2. Ministêre de l' Agriculture, de l'Elevage et des Forêts, Direction des Forêts, Kigali. Weber, A.W., 1989. Conservation and development on the Zaire-Nil divide: An analysis of value conflicts and convergence in the management of afromontane forests in Rwanda. Ph.D. Thesis. University of Wisconsin, Madison. WHO, IUCN, and WWF, 1993. Principes directeurs pour la conservation des plantes médicinales. World Health Organisation, Gland. Wild, R.G., and Mutebi, 1., 1996. Conservation through community use of plant resources: Establishing collaborative management at Bwindi Impenetrable and Mgahinga Gorrilla National Parks, Uganda. People and plants Initiatives. Division of Ecological Sciences, UNESCO, Paris. Stellenbosch University 129 http://scholar.sun.ac.za Wilkie, D.S. and Carpenter, IF., 1999. Bushmeat hunting in the Congo Bassin: An assessment of impacts and options for mitigation. Biodiversity and Conservation. Williams, S.D., and Ntayombya, P., 1999. Akagera Biodiversity Project. Min. Agric. Elevage, Environ. Dev. Rural, Direction de l'Environnement, Kigali. Wilson, R., 1991. The conflict in Rwanda and its impart on the mountain gorilla. Oryx 25: 119 - 120. Wily, L. and Haule, H., 1995. Good news from Tanzania. Village forest reserves in the making: the story of Duru-Haitemba. Forest trees and People Programme, Newletter No. 29, Uppsala. Wily, L., 1993. Establishing the basis of joint forest management. Integrating information collection with participatory planning process in regard to Mount Elgon, Uganda. IUCN, Gland. Wollenberg, E., 1999. A conceptual framework and typology for explaining the outcome oflocal forest management. Journal of Forest Management 9:. 1-35. World Bank, 1993. Inventoring and monitoring biodiversity in the Nyungwe Forest. Mimeo. World Bank, Washington D.C. World Bank, 1998. World Development Indicators. World Bank, Washington, USA. WWF, 1994. La diversité végétale: une richesse vitale. WWF, Gland. Zar, J.H., 1999. Biostatistical Analysis. Fourth edition. Prentice - Hall, Inc. New Jersey. Zimmermann, J.K., Everham, E.M., Waide, R.B., Lodge, D.J., Taylor, C.M., and Brokaw, N.V.L., 1994. Responses of trees to hurricane winds in tropical wet forests in Puerto Rico: Implications for tropical life histories. Journal of Ecology 82: 911-922. Stellenbosch University 130 http://scholar.sun.ac.za APPENDICES Appendix 1. Questionnaire for household interviews I. Background information of the respondent 1. Respondent Ref. No. _ 3. Commune: Sector: 2. Date of interview: -----Cell: _ 4. Gender: --------5. Marital status: -----6. Age: years II. Socio-economics 7. What is the family size of your household? member(s) 8. What is your highest level of education? _ 9. Mention, in priority order, the enterprises undertaken by the household 1. 2. 3. 4. 5. 10. What is the most important source of your livelihood (survival strategy)? 11. Could you rank the sources of income for the household? 1. 2. 3. 4. 5. 12. How do you access to financial resources? 13. Do you belong to any group or association? If yes, which one? _ III. Land tenure and land use 14. What is the size of your farm? 15. What is the process through which land has been acquired? _ Stellenbosch University 131 http://scholar.sun.ac.za 16. How do you rate the fertility of your farm? 1. Very poor: _ 2. Poor: _ 3. Moderate: ----- 4. High: _ 5. Veryhigh: _ 6. Don't know: --- 17. What is the status of soil fertility in the region during the last 5 years? 1. Static 3. Decreased --2. Increased 4. Don't know --18. What are the causes of soil fertility decline, increase or stability? 1. Stability: a b. c. 2. Increase: a b. c. 3. Decrease: a b. c. ------4. Don't know: --- _ 19. How does the household solve land availability problems? IV. Crops 20. What are the types of crops grown by the household? 1. Cash crops: 2.Foodcrops: 3. Crops grown for food and cash income: _ _ 21. For the most important, whether food or cash crops, could you tell us the number of bags yielded at harvest? Most important Crops 1. ------------2. _ 3. _ 4. ------------5. _ Number of bags at harvest 22. Do you also farm fruit trees, medicinal plants and fodder plants? If yes, what are they? 1. Fruit trees: a. b. c.-------2. Medicinal plants: a. b c. _ 3. Fodder plants: a. b c.------4. Other horticultural materials: a. b. c.-----23. How do you compare revenues from crops with those from trees? Stellenbosch University 132 http://scholar.sun.ac.za V. Livestock farming 24. What are the types of livestock kept and how many are they? Livestock types Numbers 1. _ 2. _ 3. __ 25. What is the most important purpose oflivestock kept? Livestock types Purpose 1. _ 2. _ 3. __ 4. _ 5. _ 26. What is the average income from different animal products? Animal products Income (RWF) 1. _ 2. _ 3. ___ 27. How is grazing practised? (grazing practices) 28. What are the sources of fodder during critical periods? 29. List, in order of preference, the most three palatable tree and grass species that are found in Nyungwe Forest Tree species 1. 2. 3. Grass species 1. 2. 3. Stellenbosch University 133 http://scholar.sun.ac.za 30. What are the particular locations where these plants can be obtained in Nyungwe Forest? Why are they found there? Tree species 1. 2. 3. Grass species 1. 2. 3. _ _ _ _ _ _ Location 1. 2. 3. Location 1. 2. 3. Reason for location _ _ _ _ _ _ VI. Availability and uses of natural resources 31. Do you own any tree crop? If yes, what is its size and species planted? Area: speCIes: _ 32. What is the most important purpose of the planted trees? _ 33. What is the average income from tree products? Tree products 1. 2. 3. Average income (RWF) 34. What are the resources available in Nyungwe Forest? 1. 2. 3. 5. 6. 7. 4. -----8. _ 35. Before Nyungwe Forest becomes a forest reserve, mention in order of importance the resources to which local people have had access. 2. 5. 2. 6. 3. 7. 4. 8. ------ _ 36. To which resources do you have access to and use in Nyungwe Forest and in the Buffer zone? 1. Nyungwe Forest: a. b. c. _ 2. Buffer zone: a. b. c. ------37. List in order of importance for your livelihood the resources present in Nyungwe Forest and in the buffer zone. 1. 2. 3. -------- Stellenbosch University 134 http://scholar.sun.ac.za 38. For the above resources, could you tell how much is available using the scale of 1 to 5, where 1 means "very little and 5 "very abundant"? Resources Grade 1. _ 2. 3. 4. _ _ _ 39. What could be the different uses of these resources? Resources 1. _ 2. _ 3. 4. _ _ 40. For timber, poles, medicinal plants, fodder and fruits, could you list in order of importance the preferred species present in Nyungwe Forest? Timber: --------------------------------------------------Poles: ___ Medicinal plants: _ Fruits: ---------------------------------------------------Fodder: __ 41. How is each of the products above harvested? Timber: --------------------------------------------------Poles: ---------------------------------------------------Medicinal plants ___ Fruits: -------------------------------------------------Fodder: -------------------------------------------------42. If timber, poles, medicinal plants, fodder and wild fruits from Nyungwe Forest are used for commercial purposes, where do you normally sell them? Products 1. _ 2. _ 3.--------------4. _ 5. _ Market location 43. For marketed product, how much is sold and at what cost? Products 1. Quantity Cost (RWF) _ 2. -------3. _ 44. Where are these resources most abundant in Nyungwe Forest? Stellenbosch University 135 http://scholar.sun.ac.za Timber: Poles: ----------------------------------------------------Medicinal plants: Fruits: --------------------------------------------------Fodder: __ _ __ 45. Who do the collection of timber, poles, medicinal herbs and fruit from Nyungwe Forest? Timber: __ Poles: ----------------------------------------------------Medicinal herbs: _ Fruit: ----------------------------------------------------Fodder: __ 46. Using a scale of 1 to 5, where 1 means "very little" and 5 "very high" could you tell us how much each key resource is harvested at a time? Timber: Poles: Medicinal herbs: Fruit: Fodder: 47. For medicinal plants, which parts of plants are collected or harvested? Medicinal plants Parts of plant used 1. ------------2. _ 3. _ 4. _ 5. ------------6. _ 7. _ 8. _ 48. In which season, medicinal plants, fruit and fodder are most abundant? Medicinal plants: _ Fruits: ---------------------------------------Fodder: --------------------------------------- 49. Are there any changes that have occurred in Nyungwe Forest resources? If yes, why? Changes Resources concerned Reason for change 1. _ 2. _ 3. _ 4. _ 5. ------------6. _ Stellenbosch University 136 http://scholar.sun.ac.za VII. Wildlife 50. What wild animals occur in Nyungwe Forest and how abundant are they? Animal types 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Rare -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 Moderate -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 Abundant -3 -3 -3 -3 -3 -3 -3 -3 -3 -3 Very abundant Most abundant -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 Don't know -6 -6 -6 -6 -6 -6 -6 -6 -6 -6 51. What types and for what most important purposes are animal hunt? (Mention one for each) Game Most important purpose 1. ___ 2. ----------------------3. ___ 4. ___ 52. What kind of change has occurred in animal population size and why? Change Wildlife species Reason for change 1. ---2. _ 3. _ 4. _ 53. For the major games, say when hunting is done? Animal types Time for hunting 1. 2. 3. 4. 5. VIII. Institutions 54. In the past, where there any local institutions that were responsible management ofNyungwe Forest? If yes, which ones? 1. 2. 3. for the _ Stellenbosch University 137 http://scholar.sun.ac.za 55. What was the role of local leadership in the management and conservation of Nyungwe Forest resources during that time? (before Nyungwe became a forest reserve) 1. 2. 3. 56. Give examples of some conflicts that arise in Nyungwe resource access and use. 1. __ 2. _ 3. ___ 4. _ 5. __ 6. _ 57. Could you tell how each conflict is important using the scale of 1 to 4 where 1 means "not important" and 4 " most important"? Nature of conflict 1. _ 2. ------------------------------3. __ 4. _ 5. -------------------------------- Importance of conflict 58. For each conflict, what is the most appropriate strategy to avoid it? Conflict 1. 2. 3. 4. 5. Use fine -1 -1 -1 -1 - 1 of Consultation! meeting -2 -2 -2 -2 -2 59. How are the civil structures / institutions conservation of natural resources a. b. c. courts Customary approach -4 -4 -4 -4 -4 -3 -3 -3 -3 -3 involved In sustainable None -6 -6 -6 -6 -6 use and 60. According to your opinion, which institutional arrangement could be responsible for management of the Forest for the benefit of local people and why? Stellenbosch University 138 http://scholar.sun.ac.za 61. What might be the role of local institutions in the management ofNyungwe Forest resources? 1. 2. 3. 4. 5. 62. What would you do to improve and maintain availability of Nyungwe Forest resources in the future? 1. 2. 3. 4. 5. Stellenbosch University 139 http://scholar.sun.ac.za Appendix 2. List of on-farm medicinal plants mentioned by the respondents in the survey area. Scientific name Local name Parts used Asteracae spp. Idoma Leaves, whole plant Barleria grandicalyx Bugangabukari Leaves Cenecio cydoniifolius Irarire Leaves Clematis simensis Umunkamba Leaves Cynura scandens Ikizimyamuriro Leaves Dychoriste tricocalyx Akanyamapfundo Leaves Iboza riparia Umuravumba Leaves Lysimachia ruhmeriana Umuyobora Leaves Mitragyna rubrostipulosa Umuzibaziba Leaves, roots Momordica foetida Umwishya Leaves Oeimum swave Umwenya Leaves Pentas decora Isagara Leaves Physalis peruvian a Iperi Leaves Plecthranthus barbatus Igicunshu Leaves Scrabera alata Umubanga Leaves Solanum aculeastrum Umutobotobo Leaves Solanum nigrum Urusenda Leaves Strychnos usambarensis Umuhoko Leaves Tephrosia vogelii Umuruku Leaves, roots Vernonia amygdalina Umubirizi Leaves Stellenbosch University 140 http://scholar.sun.ac.za Appendix 3. Palatable tree and grass species mentioned by 5 or more respondents in the survey area, n = 200. Scientific name Local name Occurrence Panicum hochstetteri Triumfetta cordifolia Mariscus thomaiophyllus Digitaria hackelii Ipomea invo/ucrata Exotheca abyssinica Cynodon aethiopicus Vireetaria major Pennisetum purpureum Botriocline /ongipes Ferns Arundinaria a/pina Sericostachys scandens Eragrostis olivacea Cynodon n/emfuensis Syzygium parvifolium Rubus rigidus Asystacia gangetica Hyparrhenia div. spp. Ikirumbi Umunaba Igikerakezi Urwili Everywhere Bottomland Swamps Forest gaps Bush Forest gaps Everywhere Hilltops Disturbed area Everywhere Hilltops Hilltops Everywhere Forest gaps Everywhere Disturbed area Everywhere Everywhere Forest gaps s: Inyovu Umunigi Urukiryi ""_, Urubingo Igihehe Ib ish ihe Umugano Umukipfu Ishinge Umucaca Umugote Umukeri Ijojwe Umukenke %of res onses 73 25.5 17.5 16.5 14 13.5 11.5 11.0 10.5 8.5 6.5 5.0 4.5 4.0 4.0 3.5 3.0 3.0 3.0 Stellenbosch University 141 http://scholar.sun.ac.za Appendix 4. List of preferred timber tree species by the respondents in the study area, n = 200. Scientific names Vernacular names Afrocrania volkensii Umujegeshi Bersama abyssinica Umukaka Carapa grandiflora Umushwati Cassipourea ruwenzoriensis Ingongo Chrysophyllum gorungosanum Umutoyi Croton megalocarpus Umunege Dasylepsis racemosa Umugoma, imbayu Dichaetanthera corymbosa Umuhube Entandrophragma excelsum Umuyove Faurea saligna Umutiti Hagenia abyssinica Umugeti Macaranga neomildbraediana Umusekera Maesopsis eminii Umuhumuro Markhamia lutea Umusave May tenus acuminata Inembwe Newtonia buchananii Umukereko Ocotea michelsonii Umuganza Ocotea usambarensis Umutake Oxyanthus troupini Ingendajoro Parinari excelsa Umunazi Podocarpus falcatus Umufu Podocarpus latifolius Umuhulizi Polyscias fulva Umwungo Prunus africana Umwumba, rwamba Rhamnus prinoides Umusasa Strombosia schefJleri Umushyika Symphonia globulifera Umushishi Syzygium parvifolium Umugote Zanthoxyllum gil/etii Umuturirwa Stellenbosch University 142 http://scholar.sun.ac.za Appendix 5. List of preferred medicinal plants found in Nyungwe Forest, as mentioned by the respondents. Scientific name Vernacular name Asc/epiadacae div. spp. Indondori Begonia meyeri-johannis Irebe Bridelia brideliifolia Umugimbu Brillantaisia cicatricosa Icyunga Canthium glabriflorum Imvuvu Canthium oligocarpum Umushabarara Carapa grandiflora Umushwati Clerodendron div. spp. Umukuzanyana Crabbea velutina Sinkangwinguma Croton megalocarpus Umunege Cyathula polycephala Igifashi Dichaetanthera corymbosa Umuhube Dichrostachys cinerea Umunkamba Embelia schimperi Umukaragata Entada abyssinica Umusange Entandrophragma excelsum Umuyove Eucalyptus sp. Inturusi Fagaropsis angolensis Umugomera Fagodia obovata Umutanoga Faurea saligna Umutiti Iboza riparia Umuravumba Indigofera arrecta Umutarabana Lantana trifolia Umuhengeri Maesa lanceolata Umuhanga Mikaniopsis tedlei Umuhokoro Mitragyna rubrostipulosa Umuzibaziba Momordica foetida Umwishywa Myrica kandtiana Isubyo Neobutonia macrocalyx Umwanya Newtonia buchananii Umukereko Ocotea michelsonii Umuganza Ocotea usambarensis Umutake Olea hochstetteri Intobo Oxyanthus troupini Umutovu Parinari excelsa Umunazi Pauridiantha paucinervis Umusibya Pentas decora Isagara Prunus africana Umwumba Rapanea melanophloes Uruneke Sericostachys scandens Umukipfu Symphonia globulifera Umushishi Synadenum grantii Umukoni Syzygium parvifolium Umugote Tabernamontana johnstonii Umuronzi Vernonia amygdalina Umuravumba Zanthoxyllum usambarense Intareyirungu Zanthoxylum gilletii Umubavu, umuturirwa F- Forest; 0 = Outside Forest; ND = Not determined. Part used Occurrence Roots Whole plant Roots Roots Bark Bark Bark Leaves, roots Leaves Bark Leaves Bark Leaves Roots, wood Leaves, roots Bark Leaves Bark Leaves, roots Bark Leaves Roots Leaves Leaves, roots Bark Leaves, bark Leaves Wood Leaves Bark Bark Bark Bark Bark Bark Leaves Leaves, bark Bark, leaves Bark, leaves Leaves Bark, leaves Roots Bark, leaves Leaves, bark Leaves Bark Bark 0 NO F NO NO F F 0 0 NO NO NO 0 NO NO F 0 NO NO NO 0 NO NO F NO F,O 0 NO F F NO NO F F F F 0 NO F F F NO F F 0 F F Stellenbosch University 143 http://scholar.sun.ac.za Appendix 6. List of scientific and vernacular names of tree species recorded in all study sites. Scientific name Agauria salicifolia HOOK Alangium chinense REDHER Albizia gummifera C.A. SMITH Alchornea hirtella BENTH. Allophylus kiwuensis GILG Bersama abyssinicaFRESEN. Bridelia brideliifolia FEDDE Canthium oligocarpum HIERN Carapa grandiflora SPRAGUE Case aria runssorica GlLG Cassipourea gummiflua 1. LEWIS Cassipourea ruwenzoriensis ALSTON Chassalia subcordata ROBYNS Chrysophyllum gorungosanum ENGL Chrysophyllum rwandense TROUPIN Cleistanthus polystachyus HOOK Combretum orophilum LIBEN Cremaspora triflora SCHUMANN Dasylepis racemosa OLIVER Dichapetalum heudelotii BAILON Diospyros gabonensis GUERKE Ekebergia capensis SPARRMAN Entandrophragma excelsum SPRAGUE Galiniera coffeoides DELILE Grewia mildbraedii BURRET Hagenia abyssinica J.F. GMELIN Harungana montana SPIRLET Ikinyarubabi Ilex mitis RADLK. Lindackeria kivuensis BAMPS Macaranga neomildbraediana LEBRUN Maesa lanceolata FORSSKAL Magnistipula butayei BRENAN May tenus acuminata LOES Memecylon walikalense A. and R. FERNANDES Millettia psilopetala HARMS Myrianthus holstii ENGL. Neoboutonia macrocalyx PAX Newtonia buchananii GILBERT and BOUTIQUE Ochna afzelii R. BR. ex OLIV. Olea hochstetteri BAKER Oricia renieri G. GLIBERT Oxyanthus troupinii BRIDSON Parinari excelsa SABINE Pauridiantha paucinervis BREMEK. Pentadesma reyndersii SPIRLET Pleiocarpa pycnantha STAPF P odocarpus fa/catus PILGER Podocarpus latifolius RENDLE Polysciasfulva HARMS Rapanea melanophloeios MEZ Rhamnus prinoides L 'HERIT. Rinorea gracilipes ENGL. Ritchiea albersii GILG Rytigynia kiwuensis ROBYNS Strombosia scheffleri ENGL. Symphonia g/obu/ifera L. f. Syzygium parvifolium MILDBR. Tabernamontanajohnstonii PICHON Xymalos monospora BAILLON Zanthoxylum gilletii WATERMAN Zanthoxylum usambarense KOKWARO Local name Umukarakara Intogota, umurangara, umuvugangoma Umusebeya Bwizabwishyamba Imbayu Umukaka Umugimbu Umushabarara Umushwati Umuhanda Intiritiri Ingongo Ikibonobono Umutoyi Urushehe Umusamba Urubabara Kanyabwoya Ingomwa, urubumburi Umumenamabuye Umunyakayumbo Umufumba, umujuga Umuyove Umubonobono Urushingati Umugeti Umushayishayi Ikinyarubabi Umunywande Umunywamazi Umusekera Umuhanga Intambasha Inembwe, umunembwe Umusuri Umunaniranzovu Umwufe Umwanya Umukereko Isanzu, urushingo Intobo Umuzo Umutovu, umuyebe Inkungu, umunazi Umusibya Umwasa Ikinesha Umufu Umuhulizi Umwungo Uruneke Umusasa Umwicaranyenzi Umuhu Urusarabatemyi Umushyika Umushishi Umugote Umuronzi Umuhotora Isoyo, umuturirwa lntareyirungu Stellenbosch University 144 http://scholar.sun.ac.za Appendix 7. Relative frequencies of trees recorded within the large tree category. 5 Total 4 Tree species 1 2 3 Agauria salicifolia 0.00 0.00 0.00 1.28 5.60 0.00 Alangium chinense 0.47 0.85 0.46 0.80 0.00 0.00 Albizia gummifera 2.14 0.40 2.21 0.00 0.47 1.00 Alchornea hirtella 0.00 1.10 0.90 0.95 0.00 0.55 Allophylus kiwuensis 0.00 1.20 0.00 0.00 0.95 0.46 Bersama abyssinica 1.71 0.40 0.00 0.00 0.00 0.46 Bridelia brideliifolia 0.00 0.43 0.09 0.00 0.00 0.00 Canthium oligocarpum 0.90 0.95 0.00 0.36 0.00 0.00 Carapa grandiflora 12.71 13.12 14.69 8.12 11.49 9.60 Casearia runssorica 4.80 7.18 3.17 8.53 5.13 5.65 Cassipourea ruwenzoriensis 2.80 3.87 4.07 5.21 4.70 4.10 Chassalia subcordata 0.95 0.00 1.37 3.20 0.55 1.81 Chrysophyllum gorungosanum 2.84 2.99 2.40 4.42 0.00 2.46 Chrysophyllum rwandense 0.40 0.00 0.00 0.00 0.00 0.09 Cleistanthus polystachyus 0.00 3.31 24.43 0.00 3.42 6.20 Combretum orophilum 0.00 1.10 0.45 0.47 0.85 0.55 Cremaspora triflora 2.40 0.00 2.71 4.27 0.00 1.91 Dasylepis racemosa 0.40 1.10 0.00 0.47 0.00 0.36 Diospyros gabonensis 0.40 2.21 3.62 4.27 0.43 2.10 Ekebergia capensis 0.00 0.55 0.00 0.00 0.00 0.09 Entandrophragma excelsum 2.80 1.66 0.00 1.90 1.71 1.64 Galiniera coffeoides 0.00 0.55 0.00 0.00 0.43 0.18 Grewia mildbraedii 2.40 0.00 0.00 2.37 0.43 1.09 Hagenia abyssinica 0.40 0.00 0.00 0.00 0.00 0.09 Harungana montana 0.00 0.00 0.00 0.00 0.43 0.09 Lindackeria kivuensis 0.40 0.00 0.00 0.00 0.00 0.09 Macaranga neomildbraediana 9.60 1.66 0.45 0.00 1.28 2.83 Maesa lanceolata 7.20 0.00 0.00 0.00 1.28 1.91 Magnistipula butayei 0.80 0.00 0.00 0.47 0.00 0.27 Memecylon walikalense 1.20 0.55 0.00 1.42 0.43 0.73 Myrianthus holstii 6.80 4.97 2.26 8.06 8.55 6.20 Neoboutonia buchananii 0.00 0.55 4.07 0.00 0.85 1.09 Newtonia macrocalyx 0.00 0.00 0.00 4.27 1.28 1.09 non identified 0.40 0.00 0.00 0.47 0.00 0.18 Ochna afzelii 0.00 1.10 0.00 0.00 2.56 0.73 Olea hochstetteri 0.40 0.00 0.00 0.00 0.00 0.09 Oricia renieri 0.40 1.66 0.90 1.90 1.28 1.19 Oxyanthus troupiniii 4.00 0.00 4.27 0.45 0.00 1.82 Parinari excelsa 1.20 4.97 4.07 4.74 1.71 3.19 P auridiantha pauc inerv is 2.80 0.00 1.36 1.42 0.00 1.19 Pentadesma reyndersii 1.60 0.00 0.45 0.00 0.00 0.46 Pleiocarpa pycnantha 0.00 0.55 0.45 0.47 0.00 0.27 Podocarpus falcatus 0.00 1.66 0.00 0.00 0.00 0.27 Podocarpus latifolius 0.80 0.00 0.00 0.00 0.00 0.18 Polyscias fulva 0.40 0.55 0.00 0.47 2.14 0.73 Rapanea melanophloeios 3.60 0.00 0.00 0.00 0.00 0.82 Rhamnus prinoides 2.80 0.00 0.00 1.90 0.00 1.00 Rinorea gracilipes 0.00 6.08 0.90 0.00 1.71 1.55 Ritchiea albersii 0.00 0.00 0.45 0.00 0.00 0.09 Strombosia scheffleri 6.00 25.41 24.89 17.06 27.78 19.78 Symphonia globulifera 3.60 3.31 1.36 2.84 9.83 4.28 Syzygium parvifolium 5.60 2.21 0.90 0.47 0.00 1.91 Tabernamontana johnstonii 0.00 1.66 1.81 0.00 5.13 1.73 Zanthoxylum gilleiii 0.00 0.55 0.00 0.00 0.43 0.18 Total 100.00 100.00 100.00 100.00 100.00 100.00 .. 1 - Musebeya. 2 = Nshili; 3 = Kagano; 4 = Mudasomwa; 5 = Busozo Stellenbosch University 145 http://scholar.sun.ac.za Appendix 8. Relative frequencies of trees recorded within the medium tree category Tree species Alangium chinense Albizia gummifera Alchornea hirtella Allophylus kiwuensis Bersama abyssinica Bridelia brideliifolia Canthium oligocarpum Carapa grandiflora Casearia runssorica Cassipourea gummiflua Cassipourea ruwenzoriensis Chassalia subcordata Chrysophyllum gorungosanum Cleistanthus polystachyus Combretum orophilum Cremaspora triflora Dasylepsis racemosa Dichapetalum heudelotii Diospyros gabonensis Ekebergia capensis Entandrophragma excelsum Grewia mildbraedii Unidentified 'Ikinyarubabi' llex mitis Lindackeria kivuensis Macaranga neomildbraediana Maesa lanceolata Magnistipula butayei May tenus acuminata Memecylon walikalense Myrianthus holstii Neoboutonia macrocalyx Newtonia buchananii Ochna afzelii Olea hochstetteri Oricia renieri Oxyanthus troupiniii Parinari excelsa Pauridiantha paucinervis Pentadesma reyndersii Pleiocarpa pycnantha Podocarpus latifolius Polyscias fulva Rapanea melanophloeios Rinorea gracilipes Ritchiea albersii Rytigynia kiwuensis Strombosia scheffleri Symphonia globulifera Syzygium parvifolium Tabernamontana johnstonii Xymalos monospora Zanthoxylum gilletii Zanthoxylum usambarense Total I = Musebeya; 2 .. = Nshili; 3 = Kagano; 2 1 0.00 2.34 0.00 0.00 0.00 0.00 7.42 8.59 2.34 0.39 0.39 9.38 3.52 0.00 0.00 3.13 0.00 0.00 0.00 0.39 1.17 1.17 0.78 0.78 0.39 2.34 0.39 0.39 4.69 0.39 0.78 0.00 3.13 0.00 1.17 3.52 11.33 1.17 0.78 0.00 0.00 1.17 0.00 1.17 0.00 2.34 6.25 1.17 2.73 5.86 0.39 6.64 0.00 0.00 100 4 = Mudasomwa; 5 = Busozo Total 4 5 3 0.40 2.30 0.35 0.50 0.00 3.02 12.64 0.00 2.50 4.71 0.10 0.00 0.00 0.50 0.00 0.10 0.00 0.35 0.00 0.00 1.18 0.80 5.75 0.35 0.00 0.10 0.00 0.00 0.50 0.00 2.11 0.00 0.00 1.00 0.00 8.24 13.79 6.03 8.50 8.24 2.30 5.67 8.00 4.12 4.72 0.00 0.00 0.50 0.00 0.20 7.64 4.60 16.31 4.50 9.41 4.32 1.15 2.48 1.00 5.29 3.42 6.90 1.77 4.50 2.94 1.76 1.81 2.30 4.61 0.00 0.00 0.35 0.00 8.82 1.61 0.00 1.42 2.50 0.00 1.71 0.00 0.35 0.00 0.00 0.10 2.31 0.00 0.00 6.50 5.88 1.18 0.00 1.42 2.50 1.11 0.00 0.00 0.00 0.00 0.10 0.00 0.00 1.00 0.00 0.50 0.00 0.35 1.50 1.76 1.01 0.00 0.00 0.00 0.00 0.20 0.00 0.71 0.50 0.00 0.50 0.00 0.00 0.00 0.00 0.10 2.30 1.42 1.00 1.76 1.71 0.00 0.00 0.00 0.00 0.10 0.00 0.71 0.50 0.00 0.40 0.00 0.35 0.50 0.00 1.41 0.00 0.00 0.00 0.59 0.20 1.15 0.35 2.50 5.29 1.81 0.00 0.00 0.00 0.59 0.10 4.60 0.00 5.00 5.88 3.22 4.60 0.00 0.50 0.00 0.50 0.00 0.35 0.50 0.00 0.50 6.90 0.00 11.50 3.53 4.42 1.15 10.64 9.00 0.59 7.94 0.00 1.42 3.00 0.00 1.31 0.00 1.42 1.50 1.18 1.11 0.00 0.71 0.00 0.00 0.20 3.45 3.55 0.50 1.18 1.61 0.00 0.00 0.00 0.00 0.30 6.90 0.35 0.50 0.59 0.90 0.00 0.00 0.00 0.00 0.30 5.75 1.06 0.00 0.00 0.80 0.00 4.26 3.00 0.59 2.51 0.00 0.00 0.00 0.00 1.61 0.00 5.67 3.00 9.41 4.12 11.49 15.25 8.50 12.94 9.95 0.00 8.16 0.50 0.00 3.92 0.00 0.35 1.00 0.59 0.50 0.00 0.00 0.00 0.00 1.71 0.00 1.06 0.00 0.00 0.30 0.00 0.35 1.00 0.00 0.30 100 100 100 100 100 Stellenbosch University 146 http://scholar.sun.ac.za Appendix 9. Relative frequencies of small trees recorded within the small tree category. Tree species Albizia gummifera Alchornea hirtella Carapa grandiflora Casearia runssorica Cassipourea ruwenzoriensis Chassalia subcordata Chrysophyllum gorungosanum Cleistanthus polystachyus Combretum orophilum Dichapetalum heudelotii Diospyros gabonensis lkinyarubabi Macaranga neomildbraediana Magnistipula butayei Milletia psilopetala Myrianthus holstii Newtonia buchananii Ochna afzelii Olea hochstetteri Oricia renieri Oxyanthus troupiniii Parinari excelsa Pauridiantha paucinervis Pentadesma reyndersii Pleiocarpa pycnantha Polyscias fulva Rhamnus prinoides Rinorea graci/ipes Ritchiea albersii Strombosia scheffleri Symphonia globulifera Syzygium parvifolium Tabernamontana johnstonii Xymalos monospora Total I = Musebeya; 4 1 2 3 0.00 11.11 0.00 0.00 0.00 0.00 0.00 0.00 20.00 0.00 16.07 28.57 1.79 0.00 3.33 7.41 0.00 0.00 3.57 0.00 20.00 0.00 0.89 0.00 3.33 0.00 8.93 17.86 0.00 7.41 6.25 0.00 0.00 14.81 0.00 0.00 0.00 3.70 0.00 3.57 0.00 3.70 0.00 0.00 3.33 0.00 0.00 0.00 0.00 0.00 0.89 0.00 0.00 3.70 0.00 14.29 0.00 0.00 0.00 3.57 3.33 7.41 0.00 0.00 0.00 0.00 30.36 0.00 0.00 3.70 0.00 0.00 0.00 0.00 0.89 0.00 0.00 0.00 0.00 3.57 23.33 0.00 2.68 0.00 0.00 0.00 0.89 0.00 3.33 0.00 0.00 3.57 0.00 0.00 1.79 0.00 0.00 3.70 0.00 0.00 0.00 3.70 1.79 0.00 0.00 7.41 0.00 0.00 0.00 7.41 0.00 0.00 0.00 0.00 0.89 0.00 0.00 3.70 0.89 7.14 6.67 11.11 16.96 17.86 10.00 0.00 4.46 0.00 0.00 0.00 0.00 0.00 3.33 0.00 0.00 0.00 100 100 100 100 .. 2 = Nshili; 3 = Kagano; 4 = Mudasornwa; 5 = Busozo Total 5 0.00 1.21 0.40 2.00 12.00 15.38 2.00 2.43 4.00 2.43 4.00 3.64 2.00 6.88 0.00 3.64 18.00 5.26 4.00 1.62 2.00 0.81 0.00 0.40 2.00 0.81 0.00 2.02 0.00 0.40 10.00 3.24 0.00 13.77 0.00 0.40 0.00 0.40 6.00 1.62 2.00 4.45 0.00 0.40 2.00 1.21 0.00 0.81 0.00 0.40 0.00 1.21 2.00 1.21 0.00 0.81 0.00 0.40 8.00 3.24 16.00 14.98 0.00 3.24 2.00 0.40 0.00 0.40 100 100 Stellenbosch University 147 http://scholar.sun.ac.za Appendix 10. Basal areas of tree species within the large tree category in the sample area. Tree species Frequency basal area, m" Relative basal area mean basal area, nr' Agauria salicifolia Alangium chinense Albizia gummifera Alchornea hirtella Allophylus kiwuensis Bersama abyssinica Bridelia brideliifolia Canthium oligocarpum Carapa grandiflora Casearia runssorica Cassipourea ruwenzoriensis Chassalia subcordata Chrysophyllum gorungosanum Chrysophyllum rwandense Cleistanthus polystachyus Combretum orophilum Cremaspora triflora Dasylepis racemosa Diospyros gabonensis Ekebergia capensis Entandrophragma excelsum Galiniera coffeoides Grewia mildbraedii Hagenia abyssinica Harungana montana lkinyarubabi Lindackeria kivuensis Macaranga neomildbraediana Maesa lanceolata Magnistipula butayei Memecylon walikalense Myrianthus holstii Neoboutonia macrocalyx Newtonia buchananii Ochna afzelii Olea hochstetteri Oricia renieri Oxyanthus troupinii Parinari excelsa Pauridiantha paucinervis Pentadesma reyndersii Pleiocarpa pycnantha Podocarpus falcatus Podocarpus latifolius Polyscias fulva Rapanea melanophloeios Rhamnus prinoides Rinorea gracilipes Ritchiea albersii Strombosia scheffleri Symphonia globulifera Syzygium parvifolium Tabernamontana johnstonii Zanthoxylum gilletii Total 14 5 Il 6 5 5 1 4 126 62 45 15 27 I 68 6 21 4 23 I 18 2 12 1 I 2 1 31 21 3 8 68 3 21 8 1 13 20 35 13 5 3 3 2 8 9 Il 17 1 217 47 21 19 2 1097 1.9761 0.1251 0.4490 0.0797 0.0605 0.1075 0.0839 0.0468 9.2044 4.4855 2.0773 0.6195 3.9447 0.0094 7.9985 0.0837 0.3294 0.1261 1.7815 0.1988 3.9164 0.0247 1.9279 0.0314 0.0326 0.0198 0.0082 1.5163 0.7440 0.0888 0.2945 2.5679 0.0386 15.4484 2.2266 0.0452 0.1820 0.7252 12.3693 0.8350 0.3354 0.0656 0.2635 0.1433 0.9756 1.0812 2.7289 0.5200 0.0143 17.8591 6.8172 2.6544 0.9918 0.9033 112.1841 0.0176 0.0011 0.0040 0.0007 0.0005 0.0010 0.0007 0.0004 0.0820 0.0400 0.0185 0.0055 0.0352 0.0001 0.0713 0.0007 0.0029 0.0011 0.0159 0.0018 0.0349 0.0002 0.0172 0.0003 0.0003 0.0002 0.0001 0.0135 0.0066 0.0008 0.0026 0.0229 0.0003 0.1377 0.0198 0.0004 0.0016 0.0065 0.1103 0.0074 0.0030 0.0006 0.0023 0.0013 0.0087 0.0096 0.0243 0.0046 0.0001 0.1592 0.0608 0.0237 0.0088 0.0081 1.0000 0.1412 0.0250 0.0408 0.0133 0.0121 0.0215 0.0839 0.0117 0.0731 0.0723 0.0462 0.0413 0.1461 0.0094 0.1176 0.0140 0.0157 0.0315 0.0775 0.1988 0.2176 0.0123 0.1607 0.0314 0.0326 0.0099 0.0082 0.0489 0.0354 0.0296 0.0368 0.0378 0.0129 0.7356 0.2783 0.0452 0.0140 0.0363 0.3534 0.0642 0.0671 0.0219 0.0878 0.0716 0.1219 0.1201 0.2481 0.0306 0.0143 0.0823 0.1450 0.1264 0.0522 0.4517 0.1023

RELATED PAPERS

RELATED TOPICS

Log In

Potential for Joint Management and Multiple Use of Nyungwe Forest, Rwanda

Potential for Joint Management and Multiple Use of Nyungwe Forest, Rwanda

Related Papers

RELATED PAPERS

RELATED TOPICS