Field Crops Research 115 (2010) 149–157 Contents lists available at ScienceDirect Field Crops Research journal homepage: www.elsevier.com/locate/fcr Indigenous legume fallows (indifallows) as an alternative soil fertility resource in smallholder maize cropping systems H. Nezomba a, T.P. Tauro a,b, F. Mtambanengwe a, P. Mapfumo a,c,* a Department of Soil Science and Agricultural Engineering, University of Zimbabwe, P.O. Box MP167, Mount Pleasant, Harare, Zimbabwe Chemistry and Soil Research Institute, Department of Research and Specialist Services (DR&SS), P.O. Box CY 550 Causeway, Harare, Zimbabwe c Soil Fertility Consortium for Southern Africa (SOFECSA), CIMMYT Southern Africa, P.O. Box MP 163, Mount Pleasant, Harare, Zimbabwe b A R T I C L E I N F O A B S T R A C T Article history: Received 8 May 2009 Received in revised form 12 October 2009 Accepted 13 October 2009 Widening the range of organic nutrient resources, especially N sources, is a major challenge for improving crop productivity of smallholder farms in southern Africa. A study was conducted over three seasons to evaluate different species of indigenous legumes for their biomass productivity, N2-fixation and residual effects on subsequent maize crops on nutrient-depleted fields belonging to smallholder farmers under contrasting rainfall zones in Zimbabwe. Under high rainfall (>800 mm yr1), 1-year indigenous legume fallows (indifallows), comprising mostly species of the genera Crotalaria, Indigofera and Tephrosia, yielded 8.6 t ha1 of biomass within 6 months, out-performing sunnhemp (Crotalaria juncea L.) green manure and grass (natural) fallows by 41% and 74%, respectively. A similar trend was observed under medium (650–750 mm yr1) rainfall in Chinyika, where the indifallow attained a biomass yield of 6.6 t ha1 compared with 2.2 t ha1 for natural fallows. Cumulatively, over two growing seasons, the indifallow treatment under high rainfall at Domboshawa produced biomass as high as 28 t ha1 compared with 7 t ha1 under natural fallow. The mean total N2 fixed under indifallows ranged from 125 kg ha1 under soils exhibiting severe nutrient depletion in Chikwaka, to 205 kg ha1 at Domboshawa. Indifallow biomass accumulated up to 210 kg N ha1, eleven-fold higher than the N contained in corresponding natural fallow biomass at time of incorporation. Application of P to indifallows significantly increased both biomass productivity and N2-fixation, translating into positive yield responses by subsequent maize. Differences in maize biomass productivity between indifallow and natural fallow treatments were already apparent at 2 weeks after maize emergence, with the former yielding significantly (P < 0.05) more maize biomass than the latter. The first maize crop following termination of 1-year indifallows yielded grain averaging 2.3 t ha1, significantly out-yielding 1-year natural fallows by >1 t ha1. In the second season, maize yields were consistently better under indifallows compared with natural fallows in terms of both grain and total biomass. The first maize crop following 2-year indifallows yielded 3 t ha1 of grain, significantly higher than the second maize crop after 1-year indifallows and natural fallows. The study demonstrated that indigenous legumes can generate N-rich biomass in sufficient quantities to make a significant influence on maize productivity for more than a single season. Maize yield gains under indifallow systems on low fertility sandy soils exceeded the yields attained with either mineral fertilizer alone or traditional green manure crop of sunnhemp. ß 2009 Elsevier B.V. All rights reserved. Keywords: Indigenous legumes N2-fixation Nutrient-depleted fields Maize productivity Smallholder farms 1. Introduction Escalating costs of mineral fertilizers and dwindling traditional soil fertility amendments calls for an increase in diversity of nutrient sources that can sustain productivity of smallholder farms in sub-Saharan Africa (SSA). Long-duration natural grass fallows, * Corresponding author at: Soil Fertility Consortium for Southern Africa (SOFECSA), CIMMYT Southern Africa, P.O. Box MP 163, Mount Pleasant, Harare, Zimbabwe. Tel.: +263 4 301807/11 803972; fax: +263 4 307304. E-mail address: p.mapfumo@cgiar.org (P. Mapfumo). 0378-4290/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.fcr.2009.10.015 traditionally used for fertility regeneration of unproductive fields (Nye and Greenland, 1960; Kwesiga and Coe, 1994), are no longer feasible due to land pressure resulting from increased population and competing land-use demands. Despite widespread promotion of legume-based technologies such as agroforestry (Jama et al., 1998; Mafongoya et al., 2006), adoption by resource-poor farmers has often been low primarily due to heavy demand on labour among other factors (Kiptot et al., 2007). The effectiveness of soil fertility ameliorants traditionally used by smallholder farmers, such as cattle manure and crop residues, has also been compromised by low application rates and poor quality of the organic resources (Giller et al., 1997; Vanlauwe and 150 H. Nezomba et al. / Field Crops Research 115 (2010) 149–157 Giller, 2006). For example, it has been estimated that less than 50% of smallholder farmers in Zimbabwe own cattle, and on average households have less than five head (Mapfumo and Giller, 2001; Nzuma, 2004). Probert et al. (1995) estimated that manure production rates in smallholder farming systems average 1 t per livestock unit per year, indicating that the usually recommended manure application rates of more than 10 t ha1 (Mugwira and Murwira, 1997) are invariably difficult to meet with small head sizes. Surveys carried out in smallholder farming areas of Zimbabwe showed that up to 40% of the households fallow part of their fields in any one season partly due to shortage of nutrient inputs (Chuma et al., 2000; Zingore et al., 2005). In some instances, because of continuous cultivation with little external nutrient inputs, fields that are persistently unproductive are increasingly being abandoned to natural fallowing (Mugwira and Murwira, 1997; Mtambanengwe and Mapfumo, 2005). Apart from mineral fertilizers, crop sequences and intercrops involving N2-fixing legumes have been shown to provide N to cereals through mineralization of the legume biomass (Rao et al., 2002; Makumba et al., 2005). In a groundnut–maize rotation, retained groundnut stover supplied N equivalent to 60–90 kg N ha1 of mineral N fertilizer (McDonagh et al., 1993). However, such data is usually derived from on-station studies where soils often have substantial amounts of residual soil phosphorus (P). When transferred to on-farm conditions, some of the ‘best bet’ legumes have failed to establish due to inherently low soil fertility. Hikwa et al. (1998) recorded biomass yields < 0.3 t ha1 from sunnhemp on P-deficient sandy soils in northern Zimbabwe. Chikowo et al. (2004) showed that Sesbania sesban and Acacia angustisimma failed to establish on sandy soils with less that 10% clay. Considering the predominance of such soils in smallholder farming areas of SSA (Buresh et al., 1997), an investment in P fertilization would be a prerequisite for successful establishment of traditionally cropped legumes. On the contrary, non-cultivated indigenous legumes have been shown to exhibit effective growth and root nodulation on nutrient-depleted soils (<10% clay, <5 ppm P, <0.4% organic carbon) (Mapfumo et al., 2005). Exploratory work showed the existence of over 36 indigenous legume species across different agro-ecological regions in Zimbabwe, dominated by species of the genera Crotalaria, Indigofera, Rothia and Tephrosia (Mapfumo et al., 2005). The work demonstrated the capacity of indigenous legumes to nodulate and grow relatively well on poor fertility soils when they are deliberately established in mixtures, suggesting their potential to restore soil productivity on nutrient-depleted fields. The main objective of this study was to evaluate the potential of indigenous legume fallows (indifallows) to enhance maize productivity on nutrient-depleted soils under smallholder farmer management conditions. The specific objectives were to (a) quantify biomass productivity of 1- and 2-year indifallows established on poor fertility soils; (b) quantify N2-fixation of indigenous legumes under 1-year indifallows; and (c) determine the effects of the 1- and 2-year indifallows on productivity of subsequent maize. sites are granite-derived sand to loamy sands that can be generally classified as Lixisols (World Reference Base, 1998). Nitrogen, P and sulphur (S) are the major limiting soil nutrients for crop production (Grant, 1981; Mapfumo and Giller, 2001). The farming system in all the study areas is predominantly maize-based, with only small patches of legume production. On-farm experiments were established on farmers’ least productive fields, which were identified using a criteria developed by Mtambanengwe and Mapfumo (2005), and farmer participatory research approaches that included transect walks and focus group discussions. According to Mtambanengwe and Mapfumo (2005), least productive fields are those which: (i) Give poor crop yields year after year regardless of management effort. (ii) Exhibit low response to external nutrient inputs. (iii) Often experience poor seed emergence due to surface crusting and have light coloured soils with poor moisture retention capacity. (iv) Generally experience crop failure if external nutrient inputs are low. In each study area, a field site was selected prior to the onset of the 2005–2006 rainfall season (late October 2005). The fields had been abandoned by farmers due to low crop productivity. At Domboshawa, the experiment was set up on land that had been previously under continuous cultivation of unfertilized maize. Soil physical and chemical properties prior to establishment of indifallows are shown in Table 1. 2.2. Experimental treatments—establishment of indifallows Field sites were ox-ploughed soon after the first effective rains in October 2005. The following treatments were effected on plots measuring 4.5 m  8 m: (i) indifallow; (ii) indifallow + P; (iii) natural fallow; (iv) natural fallow + P; (v) sunnhemp mixed fallow (simulating sunnhemp green manure with no weeding); (vi) continuous fertilized maize and (vii) continuous unfertilized maize. Phosphorus, in form of single super phosphate (SSP) (8% P and 12% S), was broadcast at 26 kg P ha1 in the +P treatments and incorporated before planting. The P application rate was based on recommended rates for legumes on sandy soils when grown in rotation with maize in Zimbabwe (Mtambanengwe et al., 2008). The continuous fertilized maize treatment received a basal fertilizer of PKS (0N:32 (P2O5):16 (K2O):5S) applied at 26 kg P ha1, 1, giving a potassium application rate of 24.5 kg K ha1. Nitrogen was applied at 90 kg N ha1 in form of ammonium nitrate (34.5% Table 1 Physical and chemical characteristics of soils at sites on which indigenous legumes were seeded. Soil parameter 2. Materials and methods 2.1. Study sites The study was conducted between 2005 and 2007 rainy seasons at Domboshawa Training Centre (Domboshawa) (178350 S; 318140 E) and Chikwaka (178440 S; 318290 E) and Chinyika (188130 S; 328220 E) smallholder farming areas. Domboshawa and Chikwaka are about 30 and 80 km northeast of Harare, respectively, while Chinyika is 250 km east of Harare. Both Domboshawa and Chikwaka are in agro-ecological region (NR) II which receives over 800 mm of rainfall annually between November and March. Chinyika is in NR III, receiving a mean annual rainfall of 650–750 mm. The soils at all Clay (%) Sand (%) Organic C (%) Total N (%) Available P (ppm)a pH (0.01 M CaCl2) Mineral N (mg kg1)b Ca (cmol(c) kg1) Mg (cmol(c) kg1) K (cmol(c) kg 1) a Site Domboshawa Chikwaka Chinyika 22 (0.5) 72 (0.6) 0.7 (0.08) 0.06 (0.003) 15 (0.3) 4.8 (0.15) 35 (1.2) 0.8 (0.06) 0.4 (0.01) 0.1 (0.005) 8 (0.1) 79 (1) 0.4 (0.06) 0.04 (0.006) 4 (1) 4.5 (0.11) 22 (1) 0.4 (0.01) 0.5 (0.01) 0.2 (0.001) 10 (0.6) 86 (1) 0.6 (0.04) 0.06 (0.002) 9 (0.2) 4.5 (0.12) 28 (1.3) 1.0 (0.02) 0.8 (0.04) 0.3 (0.004) Olsen P. Mineralizable N after 2 weeks of anaerobic incubation. Figures in parentheses denote standard error. b 151 H. Nezomba et al. / Field Crops Research 115 (2010) 149–157 Table 2 Field emergence (%) of indigenous legume species on sandy soils under different rainfall regions in Zimbabwe. Indigenous legume species Makoholi (450–650 mm yr1) Chinyika (650–750 mm yr1) Domboshawa (>750 mm yr1) Chikwaka (>750 mm yr1) Eriosema ellipticum Crotalaria ochroleuca Crotalaria pallida Crotalaria cylindrostachys Indigofera arrecta Indigofera astragalina Tephrosia radicans 40 3 4 4 1 2 2 45 (7.0) 15 (3.0) nd 3 (0.1) 8 (1.8) 5 (1.1) 5 (1.3) 18 13 9 8 10 0 2 22 8 10 7 10 3 3 (6.3) (0.4) (1.5) (0.4) (0.4) (0.8) (0.5) (4.0) (3.3) (2.6) (1.4) (1.8) (1.1) (4.0) (3.0) (4.1) (0.9) (1.4) (0.8) (1.5) Adapted from Tauro et al. (2009). nd = Not determined. Figures in parentheses denote standard errors. N). The experimental design was a randomized complete block design (RCBD), with 3 replicates per treatment. Seeds of the indigenous legumes: Crotalaria laburnifolia (L.), Crotalaria ochroleuca G. Don, Crotalaria cylindrostachys Welw. ex Baker, Crotalaria pallida (L.), Crotalaria glauca Willd., Indigofera arrecta Hochst. ex A. Rich., Indigofera astragalina DC., Neonotonia wightii (Wight & Arn.) J.A. Lackey, Eriosema ellipticum Welw. ex Baker, Tephrosia radicans Welw. ex Baker, Tephrosia purpurea Pers., Tephrosia longipes Meisn., Chamaecrista mimosoides Greene, and Macrotyloma daltonii (Webb) Verdc. were broadcast in mixtures at 120 seeds m2 species1, soon after land preparation. Collection of seeds had been done by farmers following training based on participatory methods developed by Mapfumo et al. (2005). The high seeding rates were used because past germination tests showed that without scarification, most of the species rarely exceeded 20% germination (Venge, 2003). In a complementary study, most of the species showed less than 15% emergence when established on a ploughed field (Table 2). None of the indigenous legume species used was known to be a noxious weed in the study areas. Natural fallows were established by leaving the ploughed land to naturally growing vegetation (mostly grasses). Sunnhemp, a commonly recommended green manure legume, was broadcast at a seed rate of 20 kg ha1 on separate plots as a second control. Rainfall received across sites during the experimental period is shown in Fig. 1. 2.3. Determination of biomass productivity and N2-fixation Fallow biomass productivity was quantified using random grid sampling (Mapfumo et al., 2005), at 3 and 6 months after seeding during February and May 2006 respectively. All standing legume and non-legume plants present within 3 randomly located 0.25 m2 quadrats were cut at soil level in each plot for biomass determination. At Domboshawa, biomass productivity under both indifallow and natural fallow treatments was also quantified in sub-plots measuring 4 m  4.5 m at 15 and 18 months after establishment (February and May, 2007). The harvested biomass for each component species was separately oven-dried to constant weight at 60 8C to determine dry matter. The biomass samples were analysed for total C and N using methods described by Anderson and Ingram (1993). Biological N2-fixation (BNF) by the different legume species was estimated at 3 and 6 months after seeding, using the N-difference method. Non-leguminous plants from adjacent natural fallow plots were used as reference plants. Since different legume species exhibited different growth patterns, different plants of corresponding growth habits were selected from preliminary studies and used as references (Giller, 2001). C. pallida and C. ochroleuca were compared with Bidens pilosa L. and Leucas martinicensis (Jacq.) R.Br. while C. cylindrostachys, C. juncea, Eriosema ellepticum, I. arrecta, I. astragalina, T. radicans and C. mimosoides were compared with Ageratum conyzoides L. and Nicandra physalodes (L.) Gaertn. Both legume and reference plant biomass samples were ovendried to constant weight at 60 8C and ground in a Wiley Mill to pass Fig. 1. Rainfall received at different study sites during evaluation of indigenous legume fallows for soil fertility management in Zimbabwe. through a 1 mm sieve followed by total N analyses using the modified micro-Kjeldahl procedure (Anderson and Ingram, 1993). The relative contribution of N2-fixation to total legume N uptake (Eq. (1)) and the absolute amount of N fixed per hectare were calculated (Eq. (2)). N from N2 -fixation ð%Þ ¼ total legume N uptake-total reference plant N uptake  100 (1) 152 H. Nezomba et al. / Field Crops Research 115 (2010) 149–157 total legume N uptake 3. Results 3.1. Biomass productivity of indifallows amount N fixed ðkg ha1 Þ ¼ legume N uptake ha1  reference plant N uptake ha1 (2) The estimated net N2 fixed in indifallow and sunnhemp systems were quantified using weighted averages of the total N2 fixed by each legume species, based on the relative contribution of the legume species to the overall system biomass. 2.4. Determining fallow treatment effects on succeeding maize Standing and fallen biomass in the fallows in Chikwaka and Chinyika were incorporated to a depth of 0.15 m using an oxdrawn plough, 2 weeks prior to planting of a maize test crop. At Domboshawa, the fallow biomass was incorporated using hand hoes, because only half of each fallow plot (4 m  4.5 m) was put under maize while the remaining half was left to develop into a 2year fallow. A maize cultivar SC 513 was planted with the first effective (in November) rains at a spacing of 0.9 m interrow and 0.3 m within rows. Two seeds were planted per station and subsequently thinned to one at 2 weeks after emergence (WAE). Each of the indifallow and natural fallow treatments received 45 kg N ha1 in form of ammonium nitrate (34.5% N), representing 50% of the N applied to continuous fertilized maize. The mineral N fertilizer was broadcast in three splits: 30% at 2 WAE, 40% at 6 WAE and the remaining 30% at 9 WAE (Mapfumo, 2000). The experimental plots were kept weed-free through hand hoeing and the maize crop received a single dose of Dipterex (1% Thiodan) at 6 WAE to control maize stalk borer (Busseola fusca Fuller). At thinning (2 WAE), 10 maize plants were randomly sampled and air-dried under shade, followed by oven-drying at 60 8C to a constant mass for biomass determination. At physiological maturity, maize ears (grain + cob) and stover were harvested from net plots of 2.7 m  7.4 m at Chikwaka and Chinyika and 2.7 m  3.4 m at Domboshawa. Grain yield was determined at a moisture content of 12.5% from net plots. The total dry matter yield was estimated after oven-drying of whole plant biomass to constant weight at 60 8C. 2.5. Determining maize yields after 2-year fallows During the 2007/2008 season, maize was planted to all plots, constituting a first test crop for the 2-year fallow treatments, and second maize test crop for the 1-year fallow treatments. Before planting of the second maize crop under 1-year fallow treatment plots, maize stover was removed as per the common smallholder farmers’ practice in Zimbabwe. Maize stover is a major source of feed for livestock during the dry season. Fertilizer application rates and general agronomic practices were similar to those in the first year. Maize biomass yield was quantified at physiological maturity in net plots of 2.7 m  3.4 m as described for the first year crop. 2.6. Data analyses Statistical analysis was done using GENSTAT for Windows, Discovery Edition 2 (2005). Analysis of variance (ANOVA) was used to determine treatment differences with respect to fallow biomass productivity, N2-fixation and maize biomass accumulation. Mean separations were done using least significant difference (LSD) at P < 0.05. Indifallows yielded significantly more shoot biomass than natural fallows across the three sites (Fig. 2). The highest biomass was recorded at Domboshawa where the indifallow produced 11.4 t ha1 of total biomass within 6 months, significantly (P < 0.05) out-yielding sunnhemp and natural fallows by 32% and 84%, respectively. Contrary to most of the indigenous legume species, sunnhemp reached its peak biomass after 3 months of sowing. At Chikwaka where the soil had low inherent fertility, biomass productivity was lower than at Domboshawa, with yields after 6 months of 5.8 and 2.2 t ha1 for indifallow and natural fallow, respectively. Under medium rainfall in Chinyika, indifallow yielded 6.6 t ha1 of total biomass over a period of 6 months; about three times the biomass accumulated in a corresponding natural fallow. On average, indigenous legumes contributed 67% and 83% to the total biomass generated under indifallows at 3 and 6 months, respectively. While sunnhemp produced biomass comparable to that of indigenous legumes at Domboshawa and Chinyika, the crop established poorly on nutrient-depleted soils at Chikwaka and accounted only for <10% of the total biomass. The effect of SSP on indifallow biomass productivity was significantly evident on soils of low background fertility. At Domboshawa where soils had relatively high clay (>20%) and available P (15 mg kg1) contents, addition of SSP marginally increased total biomass productivity of the indifallow. The biomass increased by only 0.9 t ha1 within 6 months following P application. On more depleted soils at Chikwaka and Chinyika, SSP treated indifallows accrued additional 1.7 and 1.5 t ha1, respectively, over corresponding treatments that received no P (Fig. 2). The effect of P on natural fallow biomass productivity was also evident. For example, natural fallow with SSP yielded 31% more biomass than the corresponding treatment without P over the same period of 6 months at Chinyika (Fig. 2). Cumulatively, over two seasons, indifallow produced significantly more biomass (28 t ha1) than natural fallow (7 t ha1) (Fig. 3). Biomass produced in 1-year indifallow was significantly higher than the 2-year cumulative biomass yield of natural fallow. There were no significant differences in cumulative biomass productivity between SSP treatments and corresponding non-P treatments for both indifallows and natural fallows. While legumes constituted most of the biomass in 1-year indifallows, the 2-year indifallows were dominated by non-leguminous biomass which accounted for >50% of the total fallow biomass (Fig. 3). However, the relative contribution of indigenous legumes was cumulatively over 50% in 2-year indifallows. 3.2. N2-fixation and amounts of biomass N at incorporation Total N2 fixed under indifallows ranged from 125 kg N ha1 at Chikwaka to 205 kg N ha1 at Domboshawa within 6 months (Fig. 4). With the exception of Chinyika, indigenous legumes in the indifallow systems fixed significantly (P < 0.05) more N2 than amounts fixed by sunnhemp in the sunnhemp fallows. Under high rainfall conditions at Domboshawa and Chikwaka, indifallows fixed an average of 164 kg N ha1 compared with 78 kg N ha1 in sunnhemp fallows over the same period. Overall, P addition increased N2 fixation. At Domboshawa, indigenous legumes fixed 9% more N2 with SSP than the corresponding treatment without P. A similar trend was observed at Chikwaka and Chinyika where addition of SSP to indifallows increased N2-fixation by 10 kg ha1 (Fig. 4). Consequently, the indifallows generated significantly higher biomass N than corresponding natural fallows at time of incorporation (Table 3). At Domboshawa, 1-year indifallow H. Nezomba et al. / Field Crops Research 115 (2010) 149–157 153 Fig. 2. Biomass productivity of indigenous legume fallows (indifallows) and natural (grass) fallows as determined at 3 and 6 months after establishment under different soils and rainfall environments in Zimbabwe. SED—Standard error of the difference of means. (198 kg N ha1) and 1-year indifallow + SSP (210 kg N ha1) treatments generated about 11 times more biomass N than 1-year natural fallow (18 kg N ha1). Similar results were obtained at Chikwaka and Chinyika where 1-year indifallows yielded 125 and 162 kg N ha1, respectively, compared with <9 kg N ha1 from 1year natural fallow treatments. The 2-year natural fallow at Domboshawa had 72% more biomass N than 1-year natural fallows. Indifallow and natural fallow treatments with added P had marginally higher biomass N than corresponding treatments without P. The indifallow biomass had C:N ratio ranging from 22 to 38 compared with >40 for natural fallow treatments (Table 3). 3.3. Maize productivity after 1- and 2-year indifallows Treatment effects became apparent as early as 2 WAE (Table 4). At Chikwaka, maize dry matter yield was highest (32 kg ha1) in 1year indifallow + SSP, which significantly out-yielded 1-year natural fallow, unfertilized maize and 1-year natural fallow + SSP treatments by 52%, 60% and 68%, respectively. Under medium 154 H. Nezomba et al. / Field Crops Research 115 (2010) 149–157 Fig. 4. Estimated total N2 fixed by indigenous legumes and sunnhemp under different smallholder farm environments in Zimbabwe. Vertical bars represent standard error of mean (SEM). Fig. 3. Cumulative biomass productivity and relative legume contributions under indigenous legume fallows (indifallows) and natural (grass) fallows 6 months (1 year) and 20 months (2 years) after establishment at Domboshawa in Zimbabwe. SED—Standard error of the difference of means. rainfall in Chinyika, the maize biomass ranged from 14 kg ha1 under 1-year natural fallow to 46 kg ha1 under 1-year indifallow + SSP. Continuous unfertilized maize yielded significantly lower than all the other treatments. At maturity, maize grain yields were significantly higher after indifallows compared with natural fallow treatments (Fig. 5). At Domboshawa, 1-year indifallow resulted in maize grain yield of 2.4 t ha1, about double the grain yield obtained after 1-year natural fallows (Fig. 5a). Overall, natural fallow treatments and continuous maize treatments gave no more than 2 t ha1 of maize grain. There were no significant differences between 1-year sunnhemp fallow and 1-year indifallow treatments and total biomass yields were consistent with grain yields. The highest maize total dry matter of 4.5 t ha1 was recorded after 1-year indifallow while the lowest biomass of 1.7 t ha1 was attained in unfertilized maize plots (Fig. 5a). Under farmers’ fields in Chikwaka, maize grain yields after indifallows averaged 2 t ha1 compared with <1 t ha1 for natural fallow, fertilized and unfertilized maize treatments (Fig. 5b). The indifallow treatments yielded two to five times the total maize biomass of natural fallow, continuous maize and sunnhemp fallow treatments (Fig. 5b). The dominance of indifallows over natural fallows in terms of maize productivity persisted into the second year of maize cropping (Fig. 6). Maize grown after 2-year indifallows yielded 3.2 t ha1 of grain with application of P, and a slightly lower yield of 2.7 t ha1 without P application. Under natural fallow treatments, maize yielded average grain and total biomass of 0.8 (Fig. 6). The second maize crop after 1-year indifallows yielded significantly higher than 1-year natural fallows although the yields were significantly lower than those attained by the first test maize crop following the 2-year indifallows. The maize grain yields averaged 1.9 t ha1 under 1-year indifallows compared with 0.8 t ha1 under corresponding natural fallow treatments (Fig. 6). 4. Discussion 4.1. Biomass productivity and N2-fixation under indifallows The presence of N2 fixing indigenous legumes accounted for the high biomass productivity of indifallows relative to natural fallows. Enhancing legume populations through deliberate seeding led to good stand establishment, and consequently high biomass yields of up to 10 t ha1 without mineral P fertilizer. However, on soils highly depleted of P due to continuous cropping, addition of mineral P significantly increased legume biomass yields and subsequently overall productivity of indifallows. Several studies have also shown the importance of P for biomass productivity of legumes in both cropping and natural ecosystems (Ndufa et al., Table 3 Estimated biomass N in indigenous legume fallows (indifallows) and natural (grass) fallows at time of field incorporation under smallholder farming conditions in Zimbabwe. Treatment 1-year 1-year 2-year 2-year 1-year 2-year 1-year 1-year indifallow indifallow + SSP indifallow indifallow + SSP natural fallow natural fallow natural fallow + SSP sunnhemp fallow Domboshawa Chikwaka Chinyika Biomass N (kg ha1) C:N Biomass N (kg ha1) C:N Biomass N (kg ha1) C:N 198 210 85 97 18 31 28 165 28 25 33 27 41 42 42 27 125 128 – – 8 – 10 8 38 33 – – 44 – 48 35 162 170 – – 9 – 17 112 22 31 – – 44 – 43 25 (8) (13) (0.7) (0.4) (0.3) (0.4) (0.5) (12) Figures in parentheses denote standard error. (3) (2) (0.1) (0.5) (0.6) (9) (5) (0.3) (0.4) (4) 155 H. Nezomba et al. / Field Crops Research 115 (2010) 149–157 Table 4 Maize biomass productivity at 2 WAE after 1-year indigenous legume fallows (indifallows) and natural (grass) fallows under smallholder farming conditions in the 2006/2007 season, in Zimbabwe. Treatment 1-year indifallow 1-year indifallow + SSP 1-year natural fallow 1-year natural fallow + SSP 1-year sunnhemp fallow Fertilized maize Unfertilized maize SED Chikwaka (kg ha1) Chinyika (kg ha1) Domboshawa (kg ha1) a 28 32a 21bc 19b 30a 21bc 20b a 39 46a 14b 33c 41a 22b 20b 54a 56a 41a 37a 48a 40a 33b 4 6 8 Means in the same column followed by the same superscript letters (a–c) are not significantly different. SED—Standard error of the difference of means. 1999; Vanlauwe et al., 2001; Rao et al., 2002). The dominance of non-legume species in terms of biomass productivity in 2-year indifallows suggests a response to improved N from mineralization of biomass submitted by annual legumes during the first year of fallowing. Most of the legume biomass accumulated in indifallows was from annual legume species such as C. cylindrostachys, C. pallida, C. ochroleuca and C. mimosoides, which produced their highest biomass in the first year of fallowing (Tauro et al., 2009). The other possibility is the failure of legume species to re-establish under already well established non-legume stands. Where soils were particularly low in fertility such as the continuously cropped fields in Chikwaka, indigenous legumes performed significantly better than sunnhemp. Such low sunnhemp biomass yields have also been reported in earlier studies on similar degraded soils (Hikwa et al., 1998). Given the high biomass productivity of indifallows on soils of inherent poor physical and chemical fertility, there is scope for widespread promotion of indifallows in resourceconstrained smallholder farming systems where inorganic fertilizers are expensive. Apart from the tolerance to soils of poor fertility, indifallows successfully established without weeding, providing an opportunity for saving labour. Indigenous legumes were able to generate relatively high biomass under the poor soils, and this resulted in substantial amounts of N being fixed. Up to 120 kg N ha1 were fixed on very poor soils in Chikwaka where sunnhemp green manure plots fixed a mere 5 kg N ha1. This amount of 120 kg N ha1 was notably higher than figures reported for grain legumes during previous studies on similar soils (Mapfumo et al., 1999; Chikowo et al., 2004). Although the N-difference method has its known limitations in accurately determining the amounts of N fixed (Witty and Giller, 1991), it has been known to provide credible estimates for practical purposes (McDonagh et al., 1993; Phoomthaisong et al., 2003). The BNF estimates reported in this study were solely based on shoot biomass, and this approach has been found to underestimate N2-fixation by 15–30% when compared with measurements that take into account the fixed N in stubble and roots (Dubach and Russelle, 1994; Høgh-Jensen and Kristensen, 1995). Fig. 5. Grain and total maize yield (t ha1) after indigenous legume fallows (indifallows) natural (grass) fallows and comparison treatments under contrasting soils at (a) Domboshawa and (b) Chikwaka in the 2006–2007 cropping season. Error bars represent LSD (P < 0.05) for comparison of treatment means. 156 H. Nezomba et al. / Field Crops Research 115 (2010) 149–157 Fig. 6. Maize yields after 1- and 2-year indigenous legume fallows (indifallows), natural (grass) fallows and comparison treatments at Domboshawa in the 2007– 2008 cropping season in Zimbabwe. Error bars represent LSD (P < 0.05) for comparison of treatment means. The N inputs at fallow termination were directly influenced by the quantity of above-ground biomass generated during the fallowing period. Although most of the biomass at fallow termination consisted of surface accumulated senesced litter, the predominantly N2-fixing indigenous legumes generated biomass amounts that translated into high N inputs at incorporation. In 1-year indifallows, significant amounts of N were also contributed by biennial legume species that included I. arrecta, E. ellipticum, T. radicans and N. wightii, which were still in vegetative phase at fallow termination. Similarly, re-sprouting perennial legume species such as C. laburnifolia increased biomass N input in 2-year indifallows. Contrary to indifallows, most of the natural fallow biomass at incorporation was composed of senesced grass litter with low N content. At most, natural fallows had accumulated 3 t ha1 of shoot biomass at the time of incorporation, confirming findings by Mapfumo et al. (2005) on fields abandoned by farmers due to poor soil fertility. 4.2. Effect of indifallows on succeeding maize crops The high maize biomass productivity after indifallows, as early as 2 WAE, suggests high soil N availability under this treatment relative to natural fallows. Mtambanengwe et al. (2006) reported a positive correlation between maize biomass productivity at 2 WAE and soil mineral N availability following incorporation of high and low quality organic nutrient sources on coarse sandy soils. Indigenous legumes constituting the indifallow systems had different growth habits and lifespan, and therefore submitted biomass at different times. For instance, legume species such as C. cylindrostachys, C. pisicarpa and C. glauca which had vegetative periods as short as 3 months could already have started decomposing during the fallow period. This provided a ready source of N during the early growing period of maize. McDonagh et al. (1993) reported that 50% of N in groundnut stover was released within 3 weeks of being incorporated into the soil. The relatively low maize biomass in natural fallows at 2 WAE was attributed to early season N immobilization by the low quality biomass generated in these fallows (C:N > 40). Across sites, indifallows produced significantly higher maize grain yields than natural fallows. Several other studies have also shown an increase in maize yields following incorporation of green manure legume biomass (Jeranyama et al., 2000), tree legume litter (Kwesiga and Coe, 1994) and grain legume residues (Kasasa et al., 1999). Split application of mineral N fertilizer could have helped to offset temporal N immobilization, ensuring that the maize crop was not critically starved of N. Application of mineral fertilizer in splits has been shown to improve N uptake by maize, especially in sandy textured soils where added N is prone to leaching (Piha, 1993). Consistently poor maize biomass productivity under natural fallows suggested that N was limiting. Promsakha na Sakonnakhon et al. (2005) showed that grass weeds with a C:N of 37–40 resulted in temporary microbial immobilization of applied fertilizer-N. Related laboratory studies on indifallows demonstrated occurrence of such immobilization (Nezomba et al., 2008). Although not addressed in this study, one other possibility is that other nutrients besides N were limiting. Balanced application of micro- and macronutrients is important in maximizing crop yields (Stoskopf, 1981). Despite being dominated by non-legume species in the second year of fallowing, 2-year indifallows produced higher maize yields than 1-year indifallows. This suggests that the non-legume species that emerged in the second year of indifallows could have effectively captured N released from decomposing biomass generated in the first year of fallowing and subsequently released it to the maize crop. Using 15N-labelled groundnut stover, Promsakha na Sakonnakhon et al. (2006) reported that broad leaf weeds that emerged after stover incorporation managed to capture and recycle 46–68 kg N ha1 of the groundnut stover N and subsequently increased maize yields compared with treatments where the weeds were removed. The biomass of re-sprouting perennial legume species such as C. laburnifolia, that were still at vegetative phase at termination of the 2-year indifallows, is also likely to have increased soil N availability upon mineralization leading, to improved maize productivity. 5. Conclusions The high biomass productivity of N2-fixing indigenous legumes (>2 t ha1) indicates that they can potentially be used in shortterm fallows to resuscitate productivity of nutrient-depleted sandy soils. Such soils are increasingly being left to the less productive natural fallows in smallholder farming systems of SSA. Application of P significantly increased biomass productivity and amounts of N fixed by the legumes. Biomass generated under 1- and 2-year indifallows can supply sufficient N to significantly increase biomass productivity and grain yield of subsequent maize crops, eliminating immobilization problems otherwise associated with the use of natural fallows when fertilizer N inputs are limiting. Natural fallows require significant mineral N fertilizer addition to offset immobilization. The capacity of indifallows to provide N early in the growing season provided the much needed crop vigour H. Nezomba et al. / Field Crops Research 115 (2010) 149–157 and subsequently high yields for maize in circumstances when N is a commonly known limiting factor under smallholder farming. Acknowledgements This work was funded by the Rockefeller Foundation under the Indigenous Legumes for Soil Fertility Project (Grant 2004 FS 105) granted to University of Zimbabwe. Additional support from SOFECSA under Grant 2005 FS 75 to CIMMYT by the Rockefeller Foundation is greatly appreciated. References Anderson, J.M., Ingram, J.S.I., 1993. Tropical Soil Biology and Fertility: A Handbook of Methods, second ed. CAB International, Wallingford, UK. Buresh, R.J., Smithson, P.C., Hellums, D.T., 1997. Building soil phosphorus in Africa. In: Buresh, R.J., Sanchez, P.A., Calhoun, F. (Eds.), Replenishing Soil Fertility in Africa. SSSA Special Publication 51, SSSA, Madison, WI, USA, pp. 111–141. Chikowo, R., Mapfumo, P., Nyamugafata, P., Giller, K.E., 2004. Maize productivity and mineral N dynamics following different soil fertility management practices on a depleted sandy soil in Zimbabwe. Agric. Ecosyst. Environ. 102, 119–131. Chuma, E., Mombeshora, B.G., Murwira, H.K., Chikuvire, J., 2000. The dynamics of soil fertility management in communal areas of Zimbabwe. In: Hilhorst, T., Muchena, F.M. (Eds.), Nutrients on the Move-Soil fertility Dynamics in African Farming Systems. International Institute for Environment and Development, London, pp. 45–64. Dubach, M., Russelle, M., 1994. Forage legume roots and nodules and their role in nitrogen transfer. Agron. J. 86, 259–266. GENSTAT, 2005. GENSTAT Edition 2 Reference Manual. Laws Agricultural Trust (Rothamsted Experimental Station). Clarendon Press, Oxford Science Publications, UK. Giller, K.E., 2001. Nitrogen Fixation in Tropical Cropping Systems. CABI Publishing, Wallingford, UK, p. 423. Giller, K.E., Cadisch, G., Ehaliotis, C., Adams, E., Sakala, W.D., Mafongoya, P.L., 1997. Building soil nitrogen capital in Africa. In: Buresh, R.J., Sanchez, P.A., Calhoun, F. (Eds.), Replenishing Soil Fertility in Africa. SSSA Special Publication 51, SSSA, Madison, WI, USA, pp. 151–192. Grant, P.M., 1981. The fertility of sandy soil in peasant agriculture. Zim. Agric. J. 78, 169–175. Hikwa, D., Murata, M., Tagwira, F., Chiduza, C., Murwira, H., Muza, L., Waddington, S.R., 1998. Performance of green manure legumes on exhausted soils in Northern Zimbabwe: a soil fertility network trial. In: Waddington, S.R., Murwira, H.K., Kumwenda, J.D.T., Hikwa, D., Tagwira, F. (Eds.), Soil Fertility Research for Maize-Based Farming Systems in Malawi and Zimbabwe. SoilFertNet/CIMMYT, Harare, Zimbabwe, pp. 81–84. Høgh-Jensen, H., Kristensen, E.S., 1995. Estimation of BNF in clover-grass system by 15 N dilution method and the total-N difference method. Biol. Agric. Hortic. 11, 203–219. Jama, B., Buresh, R.J., Place, F.M., 1998. Sesbania trees on phosphorus-deficient sites: maize yield and financial benefits. Agron. J. 90, 717–726. Jeranyama, P., Hesterman, O.B., Waddington, S.R., 2000. Relay intercropping of sunnhemp and cowpea into a smallholder maize system in Zimbabwe. Agron. J. 92, 239–244. Kasasa, P., Mpepereki, S., Musiyiwa, K., Makonese, F., Giller, K.E., 1999. Residual nitrogen benefits of promiscuous soybeans to maize under field conditions. Afric. Crop Sci. J. 9, 375–382. Kiptot, E., Hebinck, P., Franzel, S., Richards, P., 2007. Adopters, testers or pseudoadopters? Dynamics of the use of improved tree fallows by farmers in Western Kenya. Agric. Syst. 94, 509–519. Kwesiga, F., Coe, R., 1994. The effect of short rotation Sesbania sesban planted fallows on maize yields. For. Ecol. Manage. 64, 199–208. Mafongoya, P.L., Bationo, A., Kihara, J., Waswa, B.S., 2006. Appropriate technologies to replenish soil fertility in southern Africa. Nutr. Cycl. Agroecosyst. 76, 137– 151. Makumba, W., Janssen, B., Oenema, O., Festus, K.A., 2005. Influence of time of application on the performance of Griliricidia prunings as a source of N for maize. Exp. Agric. 42, 51–63. Mapfumo, P., Giller, K.E., 2001. Soil Fertility Management Strategies and Practices by Smallholder Farmers in Semi-Arid Areas of Zimbabwe. International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Bulawayo, Zimbabwe, p. 60. 157 Mapfumo, P., 2000. Potential contribution of legumes to soil fertility management in smallholder farming systems of Zimbabwe: the case of pigeonpea (Cajanus cajan [L] Millsp.). D. Phil. thesis. University of Zimbabwe, Harare, 198 pp. Mapfumo, P., Giller, K.E., Mpepereki, S., Mafongoya, P.L., 1999. Dinitrogen fixation by pigeonpea of different maturity types on granitic sandy soils in Zimbabwe. Symbiosis 27, 305–318. Mapfumo, P., Mtambanengwe, F., Giller, K.E., Mpepereki, S., 2005. Tapping indigenous herbaceous legumes for soil fertility management by resource poor farmers in Zimbabwe. Agric. Ecosyst. Environ. 109, 221–233. McDonagh, J.F., Toomsan, B., Limpinuntana, V., Giller, K.E., 1993. Estimates of the residual nitrogen benefit of groundnut to maize in Northeast Thailand. Plant Soil 154, 267–277. Mtambanengwe, F., Mapfumo, P., 2005. Organic matter management as an underlying cause for soil fertility gradients on smallholder farms in Zimbabwe. Nutr. Cycl. Agroecosyst. 73, 227–243. Mtambanengwe, F., Mapfumo, P., Vanlauwe, B., 2006. Comparative short-term effects of different quality organic resources on maize productivity under two different environments in Zimbabwe. Nutr. Cycl. Agroecosyst. 76, 271–284. Mtambanengwe, F., Nezomba, H., Nyamangara, J., Tauro, T.P., Nyagumbo, I., Mukungurutse, C., Murwira, M., Dhliwayo, D.K.C., Chikowo, R., Mapfumo, P., 2008. Promoting efficient mineral and organic fertilizer resource use through integrated soil fertility management (ISFM) in Zimbabwe. A SOFECSA-Zimbabwe Final Report submitted to the Soil Fertility Consortium for Southern Africa (SOFECSA). CIMMYT-Southern Harare, Zimbabwe, 31 pp. Mugwira, L.M., Murwira, H.K., 1997. Use of cattle manure to improve soil fertility in Zimbabwe: past and current research and future research needs. Network Working Paper No. 2. Soil Fertility Network for Maize-Based Cropping Systems in Zimbabwe and Malawi. CIMMYT, Harare, Zimbabwe, 33 pp. Ndufa, J.K., Shepherd, K.D., Buresh, R.J., Jama, B., 1999. Nutrient uptake and growth of young trees in P-deficient soil: tree species and phosphorus effects. For. Ecol. Manage. 122, 231–241. Nezomba, H., Tauro, T.P., Mtambanengwe, F., Mapfumo, P., 2008. Nitrogen fixation and biomass productivity of indigenous legumes for fertility restoration of abandoned soils in smallholder farming systems. S. Afr. J. Plant Soil 25, 161–171. Nye, P.H., Greenland, D.J., 1960. The soils under shifting cultivation. In: Technical Communication No. 51, Commonwealth Agricultural Bureau, Harpenden, UK. Nzuma, J.K., 2004. Manure management options for increasing crop production in the smallholder sectors of Zimbabwe. D. Phil thesis. University of Zimbabwe, Harare, Zimbabwe, 173 pp. Phoomthaisong, J., Toomsan, B., Limpinuntana, V., Cadisch, G., Patanothai, A., 2003. Attributes affecting residual benefits of N2-fixing mungbean and groundnut cultivars. Biol. Fertil. Soils 39, 16–24. Piha, M.I., 1993. Optimizing fertilizer use and practical rainfall capture in a semiarid environment with variable rainfall. Exp. Agric. 29, 405–415. Probert, M.E., Okalebo, J.R., Jones, R.K., 1995. The use of manure on smallholders’ farm in semi-arid eastern Kenya. Exp. Agric. 31, 371–381. Promsakha na Sakonnakhon, S., Cadisch, G., Toomsan, B., Vityakon, P., Limpinintana, V., Jogloy, S., Patanothai, A., 2006. Weeds-friend or foe? The role of weed composition on stover nutrient recycling efficiency. Field Crops Res. 97, 238–247. Promsakha na Sakonnakhon, S., Toomsan, B., Cadisch, G., Baggs, E.M., Vityakon, P., Limpinintana, V., Jogloy, S., Patanothai, A., 2005. Dry season groundnut stover management practices determine nitrogen cycling efficiency and subsequent maize yields. Plant Soil 272, 183–199. Rao, M.R., Mathuva, M.N., Gacheru, E., Radersma, S., Smithson, P., Jama, B., 2002. Duration of Sesbania fallow effect for nitrogen requirement of maize in planted fallow-maize rotation in Western Kenya. Exp. Agric. 38, 223–235. Stoskopf, N.C., 1981. Understanding Crop Production. Reston, VA, USA, p. 433. Tauro, T.P., Nezomba, H., Mtambanengwe, F., Mapfumo, P., 2009. Germination, field establishment patterns and nitrogen fixation of indigenous legumes on nutrient-depleted soils. Symbiosis 48, 92–101. Vanlauwe, B., Giller, K.E., 2006. Popular myths around soil fertility management in sub-Saharan Africa. Agric. Ecosyst. Environ. 116, 34–46. Vanlauwe, B., Aihou, K., Hounguandan, P., Diels, J., Sanginga, N., Merckx, R., 2001. Nitrogen management in ‘‘adequate’’ input maize-based agriculture in the derived savanna benchmark zone of Benin Republic. Plant Soil 228, 61–71. Venge, R., 2003. Population dynamics of selected non-cultivated indigenous legume species in Zimbabwe. BSc dissertation. University of Zimbabwe, Harare, 30 pp. Witty, J.F., Giller, K.E., 1991. Evaluation of errors in the measurement of biological nitrogen fixation using 15N fertilizer. In: Stable Isotopes in Plant Nutrition, Soil Fertility and Environmental Studies, FAO/IAEA, Vienna, Austria, pp. 59–72. World Reference Base for Soils, FAO/ISRIC/ISSS, 1998. World Soil Resources Report No. 84. Food and Agriculture Organization, Rome. Zingore, S., Manyame, C., Nyamugafata, P., Giller, K.E., 2005. Long-term changes in organic matter of woodland soils cleared for arable cropping in Zimbabwe. Eur. J. Soil Sci. 57, 727–736.