African Crop Science Journal African Crop Science Society ISSN: 1021-9730 EISSN: 2072-6589 Vol. 7, Num. 4, 1999, pp. 355-363

African Crop Science Journal, Vol. 7. No. 4, pp. 355-363, 1999

INTERCROPPING PERENNIAL LEGUMES FOR GREEN MANURE ADDITIONS TO MAIZE IN SOUTHERN MALAWI

B.C.G. KAMANGA, G.Y. KANYAMA-PHIRI and S. MINAE¹
University of Malawi, Bunda College of Agriculture, P. O. Box 219, Lilongwe, Malawi
¹The First Financial Services, National Insurance Company Centre, P. O. Lilongwe, Malawi

Code Number: CS99026

ABSTRACT

A three-year study was conducted in Zomba, southern Malawi to assess agroforestry-based soil management technologies in smallhold farmers’ fields. Sesbania (Sesbania sesban), Tephrosia (Tephrosia vogelii) and pigeon peas (Cajanus cajan) provided green manure that was incorporated with maize stover. In the control plots, maize stover alone was applied. Forty-eight fields representing three landscape positions; dambo valleys (0 – 12% slope), dambo margins (0 – 12%) and steep slopes (> 12%) were selected. The agroforestry legumes were relay intercropped with maize at first weeding in every other furrow using seedlings for Sesbania and seeds for Tephrosia and pigeon peas. At the end of the growing season, legume leafy biomass and maize stover were incorporated into the soil. Half of each plot was supplemented with 48 kg N ha^-1 as Calcium Ammonium Nitrate. There were significant differences in maize yields (P = 0.018) from the organic inputs in all the landscape positions. Effects of mineral N additions also resulted in significant increases in maize grain yield (P= 0.001). Sesbania relay intercrops had the highest maize yields (2937 kg ha^-1) followed by Tephrosia (2592 kg ha^-1) and then pigeon peas (2122 kg ha^-1). Maize yields were highest in dambo margins (2912 kg ha^-1) followed by dambo valleys (2709 kg ha^-1) and steep slopes (1648 kg ha^-1). These results have shown that resource-poor farmers can obtain higher maize yields when organic inputs are combined with inorganic fertilisers, however, such innovations should target the lower landscape positions, as the upper slopes require different approaches to land resource improvement.

Key Words: Cajanus cajan, grain yield, landscape positions, nitrogen fertilizer, soil fertility replenishment, Sesbania sesban, Tephrosia vogelii

RÉSUMÉ

Une étude de trois ans a été conduite en Zomba au sud du Malawi pour évaluer les technologies de gestion de sols à base d’agroforesterie en champs de petits agriculteurs. Sesbania (Sesbania sesban), Tephrosia (Tephrosia vogelii) et le pois cajan (Cajanus cajan) ont fourni l’engrais vert qui a été incorporé avec les fanes du maïs. Les fanes du maïs étaient appliquées seules dans les parcelles témoins. Quarante huit champs représentant trois positions du paysage: les vallées dambo (0 – 12% de pente), les bords de dambo (0 – 12%) et les fortes pentes (> 12%) ont été sélectionnés. Les légumineuses d’agroforesterie étaient en relai intercalée avec le maïs au premier sarclage dans chaque autre sillon utilisant des plantules pour Sesbania et des semences pour Tephrosia et le pois cajan. A la fin de la saison agricole, la biomasse feuillue de légumineuses et de fanes du maïs a été incorporée au sol. La moitié de chaque parcelle était complémentée avec 48 kg N ha^-1 d’ammoniatre. Il y avait des differences significatives (P = 0.018) des rendements du maïs pour les intrants organiques dans toutes les positions du paysages. Des effets d’addition d’N mineral a abouti aussi à une augmentation significative (P= 0.001) du rendement en grains du maïs. Le Sesbania en relai intercalé a eu le rendement le plus élevé du maïs (2937 kg ha^-1) suivi par Tephrosia (2592 kg ha^-1) et du pois cajan (2122 kg ha^-1). Les rendements du maïs étaient plus élévés dans le bords du dambo (2912 kg ha^-1) suivi des vallées dambo (2709 kg ha^-1) et enfin de fortes pentes (1648 kg ha^-1). Ces résultats ont montré que les pauvres agriculteurs peuvent obtenir des rendements élevés du maïs quand les intrants organiques sont combinés avec les engrais inorganiques, cependant, telles innovations devraient viser les bas-fonds paysage, puisque les pentes supérieures nécessitent de recherche additionnelle ou autres approches pour l’amélioration de la resource-terre.

Mots Clés: Pois cajan, rendement en grains, positions de paysage, en grais d’azote, restauration de la fertilité, Sesbania sesban, Tephrosia vogelii

INTRODUCTION

Agriculture is the mainstay of Malawi’s economy, contributing 34% to the gross domestic product with smallholder agriculture contributing 25% between 1984 and 1994 (Malawi Government, 1995). It has also contributed 85% to the country’s export earnings over the same period (Msukwa, 1994). Agriculture is a principal occupation for more than 80% of Malawi’s population (Kachule, 1994; Malawi Government, 1995).

Maize is the major staple food in Malawi with little competition from secondary staples such as cassava, rice and sorghum. Despite the importance of maize in smallholder agricultural production, yields remain consistently low and food insecurity is a chronic problem. One of the contributing factors to low crop production is the increase in human population estimated at 3.5% per annum (Ng’ong’ola et al., 1992). This has led to reduced land holding sizes, deforestation and soil degradation. Decline in soil fertility ranks high among the factors limiting food production in Sub-Saharan Africa (Buresh et al., 1997), and there is a need to develop more nutrient efficient food production systems. Increased use of hybrid maize and mineral fertilisers are the most promising route for improving crop yields (Malawi Government, 1995; Snapp, 1995). Although fertilisers may produce greater maize yields (Smale, 1991), their use is not widespread because their prices are prohibitive to subsistence farmers.

Research on soil fertility has focussed upon traditional knowledge in using organic resources to improve crop yields (Blackie and Jones, 1993). In the past when land was plentiful, farmers maintained soil fertility through shifting cultivation. With increased unavailability of fallow lands, the system is no longer viable. Some farmers incorporate crop residues to improve soil fertility but these residues are often low in nutrients (MacColl, 1990). Fundikila is also common in Malawi, Tanzania and Zambia. In this system, dry grass is collected, buried and burnt in mounds where crops are subsequently grown. However, these practices do not contribute to long-term soil fertility improvements.

Onim et al. (1990) reported that the use of green manure increased maize yield. They found that 13603 kg dry matter ha^-1 of Sesbania contained 448 kg N, 31 kg P and 125 kg K. Addition of 4806 kg dry matter ha^-1 of pigeon pea residues provided 161 kg N, 4 kg P and 26 kg K while 7793 kg ha^-1 of maize stover added 120 kg N, 5 kg P and 7 kg K. Maize grain yield following these additions were largest with Sesbania (6667 kg ha^-1) followed by pigeon peas (6380 kg ha^-1) and maize stover (5156 kg ha^-1). Phombeya et al. (1989) reported that Tephrosia vogelii provided 2883 kg ha^-1 of maize yield when incorporated in the soil. It was also shown that the N content is substantial in Sesbania and pigeon peas such that when incorporated they can improve crop growth and yield (Giller and Wilson, 1991).

A study was performed in Songani, an area in Southern Malawi with a high percentage of resource poor farmers. Farmers predominantly practice continuous maize-based intercropping with little use of inorganic fertilisers. It has been shown that 1.0 ha is required for enough food production for a household of five people (World Bank, 1990) but average land holdings are often less than 1.0 ha making it difficult to produce sufficient food from the average farm. Farming practices that allow for increased use of organic inputs from nitrogen-fixing plants and tight recycling of crop residues is one means of improving food production under these impoverished smallholder conditions (Woomer et al., 1998). The objective of this study was to examine the feasibility of inter-planting nitrogen-fixing perennial legumes into maize fields as a source of periodic addition of green manures. Three candidate legumes were evaluated in farmers’ fields occupying different landscape positions over two consecutive cropping seasons.

MATERIALS AND METHODS

The study was conducted in Songani watershed, Malosa Extension Planning Area in Machinga Agricultural Development Division. The site is located at 15º 18.5' S and 35º 23.5' E and is 785 metres above sea level. Songani is relatively level but 10% of the farmers’ fields are on slopes >12%.

Mean annual rainfall averages 1150 mm. Soils are classified as Ultisols and Alfisols with a pH in calcium chloride ranging from 5.3 to 6.5. Nitrate-N content in the soil ranges from 0.09 to 0.14%.

Transect sampling was used in characterising the cropping systems. Forty-eight farmers were selected at random along six transects spaced at 0.6 km. Each transect runs in an east/west direction originating from the Zomba-Lilongwe road at the lowest elevation and terminating in the highest area of the catchment on the Zomba Mountain forest reserve boundary.

The baseline survey. A survey involving 101 farmers in eleven villages was conducted prior to implementation of the trials in the first year using semi-structured questionnaires. The baseline survey was performed to collect socioe-conomic characteristics about households and farming systems. Data from the baseline survey was analysed using the Statistical Package for Social Scientists. Descriptive analysis and cross tabulation were performed.

Field experimentation. Field experimentation was researcher-designed and farmer-managed. Three landscape positions were identified to determine the effect of slope on soil fertility and its management. These positions included the dambo valleys (0 - 12 % slope), the dambo margins (0 - 12 % slope) and the steep slopes (> 12 %). The experiments were arranged as a 3 x 4 factorial in a split plot design with 16 replicates. Slope positions were the main plots and legume inter-planting the subplot. Four 15 m x 15 m maize plots were demarcated on each of the 48 participating farmers’ fields. At the onset of rains, maize (Zea mays L.) hybrid variety MH18 was planted in the four plots at a population of 37,000 plants ha^-1. The plots were relay intercropped with Sesbania sesban (L.) Merr, Tephrosia vogelii and Cajanus cajan (L.) Mills at the first maize weeding while the fourth plot remained as a maize monocrop control.

The maize/Sesbania and maize/Tephrosia plots were new to farmers while the maize/pigeon peas and maize monocrop represented existing systems. Bare-rooted seedlings of Sesbania were transplanted at 75 cm x 180 cm providing 7,400 plants ha^-1. Tephrosia and pigeon peas were directly sown at 3 seeds per hole in ridge shoulders at 24, 691 plants ha^-1. These were later thinned to one per hole to provide the same plant density as Sesbania.

In the second year, the procedure was repeated in all plots. Before the onset of rains, each plot was split with one half recieving 48 kg N ha^-1 Calcium Ammonium Nitrate as top dressing. The splitting of the plots changed the design to 3 x 4 x 2 factorial in a split-split plot design. Researchers applied fertilisers in the plots as many farmers were inexperienced in its uniform application. The same planting and management patterns were repeated in the third year.

Data collection. Maize was harvested at physiological maturity from sample areas of 5 m x 5 m, cobs collected, shelled and weighed to determine total biomass and grain yield. Grain yields were adjusted by assuming 12.5% moisture content. The legumes remained in the field until shortly before the subsequent rains, then felled, fresh dead leaves and fine branches (< 0.05 cm) stripped and incorporated into the soil by hand ridging. Sub-samples of plant residues (500 g) were collected from maize at harvest and for the legumes at incorporation for nitrogen determination (Onim et al., 1990). The samples were oven-dried to a constant weight at 60 º C for 48 hours and a sub-sample retained for subsequent chemical analysis. The dry weight of the sample was used to adjust the fresh weights per net plot to dry weights per plot. Woody biomass from Sesbania was also collected and weighed from sample areas of 5 m x 5 m. The woody stems of Tephrosia and pigeon peas were also removed from the field to facilitate field operations.

Analytical procedure. Data were compiled onto a computer spreadsheet with columns representing experimental variables or measurements and rows consisting of individual cases (farms). The spread sheet was inspected and then imported into MSTATC computer statistical software.

Analyses of Variance were performed for a split plot in the first season and a split-split plot for the second with all interactions included into the model statement.

RESULTS AND DISCUSSION

Farm survey. Land holding sizes were small in the area as reflected by the average farm size (0.56 ha) as well as the proportion of farmers with land < 1 ha. Sixty-eight percent of farmers reported land holdings of < 1 ha and 25% cultivated between 1 and 2 ha. Concurring with these results, the Malawi Government (1991) reported that 79% of farmers in Machinga Agricultural Development Division, southern Malawi cultivate less than < 1 ha and only 6% had more than 2 ha.

Table 1 shows the dominant cropping systems in Songani, Zomba. Out of 163 fields surveyed, 57% had mixed intercrops of maize with pigeon peas, sorghum, cassava and other crops. Strip intercropping was observed in 32% of the fields where cassava formed the main strips with maize and other crops planted in between. Eight percent of the fields were planted to one crop while 3% of the cropping systems in the fields could not be defined.

TABLE 1. Cropping systems in Songani

According to Shaxson and Tauer (1992), 95% of the total area cultivated in Zomba was intercropped to various degrees, with 84% of the land planted to maize. These results support the findings of this study where 90% of the farmers in the study area practice maize-based intercropping. The large proportion of intercropping suggests a scarcity of land, forcing farmers to make optional use of limited field areas. In so doing, more nutrients are being mined from soils, identifying the need for farm interventions that contribute to nutrient inputs and recycling, as does interplanting with perennial legumes.

Table 2 presents information on household availability of food throughout the year. The results show that well over 80% of the households had no food by December, seven months after the most recent harvest. Alarmingly, 2% of the sample households exhaust food supplies by April, the month that maize is normally harvested. This is ominous considering that families generally have very limited or no income to purchase food. The months of critical food insecurity in the study area coincide with the peak period of high labour demands. Selling one’s labour is an important strategy for survival in the face of food shortage but this may in turn result in domestic labour shortages on one’s own farm. Furthermore, less than 1% of the farmers reported having no food problems throughout the year. In a similar study in Zomba, 80% of the farmers were reported to have depleted their food stocks by December (Malawi Government, 1991; Barbier, 1991).

Table 2. Maize provision ability of farm families in Songani

Biomass yields, nitrogen content and nitrogen equivalents. The slope positions and legume interplanting significantly (P<0.001) influenced biomass production. Yield of Sesbania (2217 kg ha^-1) was significantly higher than that of Tephrosia (850 kg ha^-1) and pigeon peas (579 kg ha^-1) (Table 3). Stover yield (1546 kg ha^-1) was highest (P = 0.094) in the maize/Sesbania interplanting (1714 kg ha^-1) and lowest in the maize/pigeon pea interplanting (1370 kg ha^-1). Biomass production on different landscape positions was highest in the dambo valleys followed by the margins and least in the steep slopes. High biomass obtained in the lower slopes may be attributed to residual moisture. In the steep slopes, production of Sesbania was poor. Tephrosia and pigeon peas produced low biomass probably due to poor growth which could be attributed to their susceptibility to waterlogging conditions (Nene et al., 1990) which occurred shortly after planting.

Table 3. Plant biomass, nitrogen content and partial economic analysis for the different production systems (n = 45)

^a Means followed by different superscript letters in a column are significantly different at P < 0.05. Column three was obtained by multiplying columns 1 and 2. Column four was obtained by using the percentage N (27%) in the CAN fertiliser. Column 5 is the value of the fertiliser equivalents while column 6 is the cost of producing the fertiliser

The nitrogen contained in both maize stover and legumes were combined to calculate total nitrogen inputs. Analysis for nitrogen contents showed that Sesbania biomass contained 3.0% N, Tephrosia and pigeon peas 3.1% and maize stover had 0.8%. These findings on N contents are in general agreement with Giller and Wilson (1991) and Gitteridge (1991) who reported N contents of 3.6% and 3.5%, respectively. Small difference may be attributed to the fertility of the soils where they were planted and other environmental conditions.

In comparison of N contributions, Sesbania provided 66.5 kg N ha^-1 and maize stover 15.4 kg N ha^-1. The maize/Tephrosia and maize/pigeon pea inter-planting contributed 38.7 and 27.3 kg N ha^-1, respectively. The maize monocrop provided 12.3 kg N ha^-1. Crop nitrogen use efficiency from legume biomass is reported to be approximately half as effective as inputs from mineral nitrogen fertiliser (Giller and Wilson, 1991). The factors regulating nutrient supply from an organic input are decomposition and mineralisation and their synchronisation with crop nutrient demands (Myers et al., 1994). Sesbania decomposes rapidly and as such provides more nutrients to the soil. Based on the effectiveness of the organic N, it follows that the nitrogen values reported in Table 3 might not all be available to the crop at the same time as the inorganic N. The advantage, however, is that the organic N accumulates slowly and it is less subject to loss with more biomass incorporation.

Comparing the nitrogen in plant biomass to commercial N fertiliser such as CAN (27% N), the biomass incorporated into the soil added the equivalent of 303.3, 143.6, 101.1 and 45.4 kg ha^-1 of CAN from the Sesbania, Tephrosia, pigeon pea and maize treatments, respectively. At the current US$14 per 50 kg bag, this is a substantial saving for the financially constrained smallholder. In a similar a study, Onim et al. (1990) measured nitrogen yields of 488 and 161 kg ha^-1 from 13.6 and 4.8 tonnes of Sesbania and pigeon pea biomass, respectively.

The first year production cost of establishing the Sesbania inter-planting was US$30 ha^-1 offering farmers’ savings of US$9 in terms of N purchases alone. This does not take into account the potential regrowth in subsequent seasons or the value of fuelwood. In this experiment, legume inter-plants were re-established during each cropping season, harvested late in the following dry season and woody biomass removed for use as fuel. Other options exist including regrowth of the perennial legumes, resulting in reducing the costs of establishing the legumes. This management option deserves attention, as it may prove more economical in the long-term.

Some difficulties exist in accounting for plant nutrients in the treatments. Root biomass of maize and the legumes was not measured. Presumably, legume root biomass contained considerable N inputs through symbiotic N fixation (Giller and Wilson, 1991). Many of the lower maize leaves were shed during grain ripening, resulting in unaccounted nutrient recycling as surface mulch. The legume inter-plants, however, remained green throughout the dry season and their biomass was monitored with reasonable accuracy.

Maize response to organic inputs and N fertiliser. The main effects on maize grain were all significant with fertiliser N levels being strongest (P< 0.001) followed by residues (P = 0.001) and slope positions (P = 0.038). The interactions were significant except for landscape positions x N addition (P = 0.018) in the second season, suggesting that mineral nitrogen inputs are more strategic on maize yields in dambo margins followed by dambo valleys and steep slopes. Table 4 shows maize yields at different seasons, landscape positions and N treatments. Dambo margins yielded 2912 kg ha^-1 followed by dambo valleys 2709 kg ha^-1 while the steep slopes gave 1648 kg ha^-1 when supplemented with 48 kg N ha^-1. In the first year of the study, maize yields were low. In the second year, there was a depression in maize yields despite the incorporation of legume residues. In the third year, yields were higher than the second year in all landscape positions (Table 4). The depression in yield might have resulted from the immobilisation of nitrogen in the soil. However, marked improvements were observed in the third year of the study. When biomass was supplemented with 48 kg N ha^-1 as CAN, a strong yield response of maize to nitrogen addition was noted.

Table 4. Maize grain yield at different landscape positions and nitrogen treatments

^aMeans followed by different superscript letters in a row or column are significantly different (P =0.0175) for 1997/98 and (P=0.002) for 1998/99 growing seasons

Table 5 shows the domestic impacts of maize yields resulting from the legume inter-planting. The Sesbania inter-planting with 48 kg N ha^-1 of mineral fertiliser provided sufficient food for the average family of 5.8 people from 1 ha and also resulted in a surplus of 20% that the household could market. Tephrosia also showed similar benefits of producing sufficient food with 6% surplus. Pigeon pea and maize inter-planting each produced 87% and 83% of the household maize requirements, respectively. However, the same inter-plantings without inorganic fertilisers provided insufficient food for the household.

Forty three percent of the total food requirement would be met from Sesbania with 40%, 25% and 24% from Tephrosia, pigeon peas and maize inter-planting, respectively. This implies that yield benefits obtained from the legume systems with mineral N fertilisers are sufficient to overcome food insecurity of many households, particularly within smaller land holdings.

Table 5. Maize yields and food security from the legume based systems for 1998

¹Maize production was based on the mean landholding of 0.56 ha

Fuelwood production is an additional benefit from legume inter-planting. Fuelwood consu-mption by households is presented in Table 6. The survey revealed that women required 4 hours to collect 40 kg of fuelwood that lasted about eight days. The mean distance covered was 6 km from homestead to the protected forest where only deadfall may be removed. A family of 5.8 people required about 1920 kg of fuelwood per year. These figures are similar to those reported by Jane et al. (1984) that per capita fuelwood consumption varies from 0.5 to 1.2 m^-3 yr^-1. The wood density of Sesbania is 432 kg m^-3. From Sesbania’s wood production (1777 kg ha^-1), 4.4 m³ were obtained (Table 7). Fuelwood consumption also depends on the availability of alternative sources of energy, as wood burning is not preferred in some households. However, when one considers the fraction of Sesbania fuelwood production to total fuelwood required, 0.92 of the domestic fuelwood needs were met from an hectare during the second season and 1.01 in the third season. These findings are important because greater domestic fuelwood sufficiency within smallholdings would impact positively on both labour allocation and neighbouring woodland ecosystems.

TABLE 6. Fuelwood consumption and production in Songani

¹ stands for standard deviation of the mean
²T = Tambala (100 T = 1 Kwacha, and 1 US$ = 42 Kwacha)

TABLE 7. Fuelwood production from sesbania inteplanting in maize in Songani

CONCLUSION

Potential exists for installing legumes into maize-based farming systems as sources of organic inputs but this intervention alone is insufficient to obtain domestic food security. The benefits are not realised in the first year but start in the second year. This is particularly true in dambos. Farmers who cultivate fields in these niches have high potential of improving domestic food security by using the technologies examined in this study, particularly mineral N fertilisation from the legume inter-plants. Sesbania with an additional benefit of fuelwood production sufficient to meet the needs of most households was the most promising. Farmers with limited resources to purchase fertilisers are able to improve maize production through the strategic use of inter-planted legumes.

ACKNOWLEDGMENT

The authors thank Paul Woomer for providing useful comments during the preparation of this paper. The Rockefeller Foundation under Forum for Agricultural Resource Husbandry is gratefully acknowledged for the financial assistance.

REFERENCES

Blackie, M.J and Jones, R.B, 1993. Agronomy and increased maize productivity in eastern and southern Africa. Biological Agriculture and Horticulture 9:144 -160.

Barbier, E. 1991. The Role of Smallholder Production Prices in Land Degradation. The case of Malawi. London Environmental Economics Centre Discussion Paper, pp. 91-105. London.

Buresh, R.J., Sanchenz, P.A. and Calhoon, F. 1997. Replenishing Soil Fertility in Africa. SSSA Special Publication Number 51, Soil Science Society of America, Madison, Winscosin. 251 pp.

Giller, K.E. and Wilson, K.J. 1991. Nitrogen Fixation in Tropical Cropping Systems. CAB International. Wallingford, U.K. 313 pp.

Gitteridge, R.C. 1991. Evaluation of the leaf of a range of tree legumes as a source of nitrogen for crop growth. Experimental Agriculture 28:195-202.

Jane, M., Vietmeyer, E.N.D and Ruskin, F.R. 1983. Firewood Crop; Shrubs and Tree Species for Energy Production. National Academy Press, Washington D.C. 2:6-21.

Kachule, R.N. 1994 . Economic Analysis of Bean Production in Northern Malawi: Implications for Agricultural Policy. M.Sc. Thesis, Bunda College, Lilongwe, Malawi, pp. 1-10.

MacColl, D. 1990. The effects of lime, minimum tillage and return of stover to the soil on yields of maize over seven years. Bunda College Journal of Agricultural Research 2:85 - 95.

Malawi Government, 1991. Household Food Security in Malawi. 1991 FSNM report 1. Ministry of Agriculture. Lilongwe, Malawi, pp. 7-19.

Malawi Government, 1995. Economic Report. 1995. Ministry of Economic Planning and Development. Zomba, Malawi, pp. 20 -34

Msukwa, L.A.H. 1994. Food Policy and Production towards Increased Household Food Security. Centre for Social Research. University of Malawi, Zomba, pp. 1-6.

Myers, R.J.K., Palm, C.A., Cuevas, E., Gunatilleke, I.U.N. and Brossard, M. 1994. The synchronization of nutrient mineralization and plant nutrient demand. In: The Biological Management of Tropical Soil fertility. Woomer, P.L. and Swift, M.J. (Eds.), pp. 81-116. John Wiley and Sons, Chichester, UK.

Ng’ong’ola, D.H., Mtimuni, B.H., Kanyama-Phiri, G.Y., Nothale, D.W. and Wiyo, K.A. 1992. Socio-economic Forecasts for Land Uses in Malawi. Bunda College, Lilongwe, Malawi, pp. 1- 15.

Nene, Y.L., Hall, S.D. and Sheila, V.K. 1990. The Pigeonpeas. International Crop Research Institute for the Semi-Arid Tropics. CAB International, Cambridge Press, U.K. 490 pp.

Onim, J. F.M., Mathuwa, M., Otieno, M. and Fitzhugh, H.A. 1990. Soil fertility changes and responses of maize and beans to green manures of leucaena, sesbania and pigeon peas. Agroforestry Systems 12:197-215.

Phombeya, H.S.K., Saka, A.R., Bunderson, W.T., Itimu, O.A. and Mtupanyama, Y. 1989. Comparison of Leaf Manure rates from Different Tree Species on Maize Yields. Agroforestry Commodity Team, Ministry of Agriculture, Chitedze Agricultural Research Station. Working paper. Lilongwe, Malawi, pp.15-17.

Shaxson, L. and Tauer, L.W. 1992. Intercropping and diversity. An economic analysis of cropping patterns on smallholder farms in Malawi. Experimental Agriculture 28:211- 228.

Smale, M. 1991. Chimanga Chamakono, Hybrids and Composites: An Analysis of Farmers’ Adoption of Maize Technology in Malawi. CYMMYT, Lilongwe, pp. 17 - 23.

Snapp, S.S. 1995. Improving fertilizer efficiency with small additions of high quality organic inputs. In: Report on the First Meeting of the Network for Maize-based Farming Systems in Selected Countries of Southern Africa. Waddington, S.R. (Ed.), pp. 60-65. The Rockefeller Foundation/CIMMYT, Lilongwe, Malawi.

Woomer, P.L., Bekunda, M.A., Karanja, N.K., Moorehouse, T. and Okalebo, J.R. 1998. Agricultural resource management by smallhold farmers in East Africa. Nature and Resources 34:22-33.

World Bank, 1990. Malawi Food Security Report. Agriculture Operations Divisions, Southern Africa Development, African Region. Washington D.C. U.S.A. 98 pp.

©1999, African Crop Science Society