Document(s) 18 of 36

Abstract

The subhumid zone of West Africa has 180–270 growing d and receives 900–1 500 mm of rainfall. The zone is covered mainly with ferruginous tropical soils that have very low levels of N and P and a low cation exchange capacity. The common farming system is crop–livestock, with a predominance of cash crops wherever these have been introduced. The genus Stylosanthes, which originated mainly in South America, was tested in West Africa as early as the 1940s in an attempt to improve livestock nutrition and soil fertility. Some of the major attributes explaining the success of this genus are tolerance to the fungal disease anthracnose, adaptation to infertile soils, drought resistance, ability to fix N without special Rhizobium inoculum, and high seed yield.

Integration of Stylosanthes into the West African farming systems intensified with the opening of the International Livestock Centre for Africa's subhumid research site in Kaduna, Nigeria, in 1978, and since then the genus has been exploited to suit the domestic needs of various countries.

The management systems included sole crops of Stylosanthes ("fodder banks") to supplement natural range or improved grass–legume and legume–legume associations. Stylosanthes has also been successfully integrated in crop rotations as an intercrop and relay crop.

Tremendous variation occurs in the research–development efforts and diffusion of Stylosanthes in the West African region. For instance, in Nigeria, this legume is exploited by smallholders for agropastoral herds, small ruminants, and crop production. More recently, the concept of mixed cover crops involving Stylosanthes was introduced, and evidence suggests that such a management system leads to more profitable and sustainable crop–livestock production systems. In Cameroon, a lot of potential exists for the expansion of Stylosanthes-based enterprises into agropastoral herds, but the smallholders have not fully exploited such benefits. This is mainly because on-farm research and extension are in their infancy in this region. In Côte d'Ivoire, the use of Stylosanthes has been geared toward the dairy industry through the Eco-farms Project, a scheme jointly sponsored by the African Development Bank, Gesellschaft für Technische Zusammenarbeit (organization for technical cooperation), and the Ivorian government. The aim has been to enable smallholders to generate incomes equivalent to those of their counterparts in the cities. In Mali, Stylosanthes could improve cereal production and the performance of traction animals.

Some constraints to adoption, such as lack of labour and capital, plant disease, insufficient quantities of seed, and the expense of fencing, cut across all the four countries studied; other problems are specific to individual countries. The land-tenure constraint affecting agropastoralists, for instance, is more acute in Nigeria. In Côte d'Ivoire, pasture management is a serious problem in the Eco-farms where Stylosanthes is grown in association with aggressive grasses. Also, the scheme is too capital intensive. In Cameroon, the weak support for on-farm research and extension is unique, whereas in Mali, the successful cotton industry is providing capital for Stylosanthes adoption.

The potential of Stylosanthes for feed improvement, land reclamation, and control of noxious weeds should be further exploited by integrating the legume in crop rotations and promoting it to farmers with more profitable enterprises, such as those involving beef cattle, dairy cattle, traction animals, and cotton.

Other measures recommended for the long-term sustainability of Stylosanthes systems include the establishment of an adequate seed supply, identification of highly productive and disease-resistant varieties, use of live poles (for example, Ficus) for fencing, and the use of draft power. Government policies that promote loan schemes and protect land rights are also needed.

Résumé

La zone subhumide de l'Afrique occidentale se caractérise par une pluviométrie de 900 à 1 500 mm par année et par une période de croissance allant de 180 à 270 jours. Cette zone est principalement couverte de sols ferrugineux tropicaux qui ont un niveau très bas de N, de P et de capacité d'échange cationique. Le système d'exploitation agricole habituel est l'exploitation culture–élevage où prédominent les cultures commerciales là où celles-ci ont été introduites. Le genre Stylosanthes, qui provient surtout de l'Amérique du Sud, a été mis à l'essai en Afrique occidentale dans les années 1940, afin d'améliorer le fourrage ainsi que la fertilité des sols. Certaines des qualités essentielles attribuées au succès du genre Stylosanthes sont une tolérance à la maladie fongique, l'anthracnose, une adaptation aux sols infertiles, une résistance à la sécheresse, une capacité fixatrice de N sans inoculant rhizobien particulier, et un rendement grainier élevé.

L'intégration des Stylosanthes dans les systèmes d'exploitation agricole de l'Afrique occidentale s'est accrue depuis l'ouverture du site de recherche subhumide du Centre international pour l'élevage en Afrique à Kaduna, au Nigeria, en 1978. Depuis, le genre a aussi été exploité pour répondre aux besoins locaux de divers pays.

Les pratiques de gestion utilisées comprenaient les cultures de rangée de fourrage de Stylosanthes seules pour compléter la gamme naturelle ou les associations améliorées herbe–légumineuse et légumineuse–légumineuse. Les Stylosanthes ont aussi connu un succès dans leur intégration aux rotations de cultures comme culture intercalaire ou culture de relais.

Les efforts de recherche et de développement ainsi que les efforts de diffusion liés aux Stylosanthes en Afrique occidentale sont nombreux et variés. Par exemple, au Nigeria, cette légumineuse est cultivée par de petits exploitants pour la production de bovins, de petits ruminants et de cultures. Plus récemment, le concept de cultures de couverture mixtes utilisant les Stylosanthes a été introduit et semble démontrer qu'un tel système de gestion favorise une exploitation culture–élevage plus profitable et durable. Au Cameroun, il existe un grand potentiel pour l'expansion d'entreprises utilisant les Stylosanthes pour les troupeaux de pâturage, dont les petits exploitants n'ont pas encore complètement tiré parti. La raison principale est que la recherche en ferme et la vulgarisation en sont encore à leur début. En Côte d'Ivoire, l'utilisation des Stylosanthes est dirigée vers l'industrie laitière au moyen du Projet de fermes écologiques, une entreprise conjointement appuyée par le Groupe de la Banque africaine de développement, le Gesellschaft für Technische Zusammenarbeit ( agence allemande de coopération technique ) et le gouvernement ivoirien. Le but était de permettre aux petits exploitants de générer des revenus équivalant à ceux de leurs homologues des villes. Au Mali, les Stylosanthes offrent la possibilité d'améliorer le rendement des animaux de trait et celui de la production céréalière.

Les contraintes liées à l'utilisation de la légumineuse, telles que le manque de main-d'uvre et de capitaux, les maladies des plantes, le manque de semences et les coupures imposées aux quatre pays à l'étude, s'ajoutent aux problèmes spécifiques de chacun des pays. Par exemple, la contrainte de la tenure qui nuit aux agropasteurs fulanis est typique du Nigeria. En Côte d'Ivoire, l'exploitation des pâturages est un sérieux problème pour les fermes écologiques, où les Stylosanthes poussent aux côtés d'herbes envahissantes. De plus, la structure fonctionne à trop forte intensité de capitaux. Au Cameroun, il existe peu de soutien pour la recherche en ferme et pour la vulgarisation, alors qu'au Mali, le succès des nouvelles technologies dans l'industrie du coton nuit à la popularité des Stylosanthes.

Il est recommandé d'exploiter davantage les possibilités que représentent les Stylosanthes pour l'amélioration alimentaire, la récupération des terres et le désherbage, en intégrant la légumineuse dans les rotations des cultures et en visant les entreprises plus rentables telles que le buf, les bovins laitiers, les animaux de trait et le coton.

Pour favoriser la durabilité des systèmes utilisant les Stylosanthes, il est également recommandé de s'approvisionner suffisamment en semences, d'utiliser des variétés hautement productives et résistantes à la maladie et de recourir à des intrants à moindre coût, tels les engrais, pour le billonnage cloisonné, en plus d'utiliser une puissance de traction, des politiques gouvernementales favorables aux systèmes de prêts et axées sur la protection des droits terriens.

Introduction

Subhumid West Africa

The West African subhumid zone (SHZ) covers Burkina, Benin, Cameroon, Côte d'Ivoire, The Gambia, Ghana, Guinea, Liberia, Nigeria, Senegal, Sierra Leone, and Togo. The SHZ has a single growing season (180–270 d) and covers 45% of sub-Saharan Africa (SSA). The common farming system is crop–livestock In some areas, cash crops predominate. The SHZ offers the greatest potential in SSA for growing crops and producing livestock, both of which are currently in crisis.

A major constraint to meeting the food demands in the SHZ is that SSA is experiencing substantial land degradation, leading to decreasing total agricultural productivity (Lal 1989). Marginal lands are increasing, as a result of land pressure from rapid population growth. Traditional grazing lands are acquired for cultivation, and the long fallow periods (crucial for regenerating the soil's fertility) have become unfeasible (Ruthenberg 1980). Recent studies suggest that the population of SSA (0.5 billion in 1990) will reach 1.2 billion by 2025 (Winrock International 1992). The studies also predict a demographic shift, with urban dwellers increasing from 29% to 55% of the population, implying that the rural sector will have to produce more food to feed the urban population.

Some technical developments have improved the crop–livestock production systems in the region. An example is the dual use of Stylosanthes as feed for starving animals and as an amendment for poor soils. Over the years, West African countries have developed various scenarios for exploiting the potential of this genus, based on their own needs. For instance, in Nigeria and Cameroon, Stylosanthes is used mainly for agropastoral herds, small ruminants, and crop production. In Côte d'Ivoire, the legume is used to boost the dairy industry; in Mali, to improve soil fertility and the performance of traction animals.

This paper provides a synthesis of the ways Stylosanthes is used in four West African countries — Nigeria, Cameroon, Côte d'Ivoire, and Mali — and the constraints to adoption of the technology. We selected these countries because they are known to use Stylosanthes at all levels and stages, from research through to on-farm adoption. We discuss the critical points and future opportunities for the long-term sustainability of Stylosanthes-based systems in subhumid West Africa. To obtain information, we reviewed the existing literature, made field visits, and held detailed discussions with smallholder farmers, researchers, and extension workers. This information is probably applicable to other countries in subhumid West Africa.

Brief history of Stylosanthes

Stylosanthes, a genus of the subtribe Stylosanthinae, tribe Aeschynomenae, subfamily Papilioniodae, family Leguminosae, occurs naturally in the tropical, subtropical, and temperate regions of the Americas, tropical Africa, and southeast Asia (t'Mannetje 1984). The major centres of diversification are the southern neotropics, particularly Brazil; a secondary centre is in the Mexican–Caribbean basin (Stace and Cameroon 1984). About 45 species and subspecies belong to the genus; these are classified into two sections, Stylosanthes and Styposanthes. Stylosanthes is self-fertile and predominantly self-pollinating. The range of photoperiod response in the genus is wide: short day, long day, day neutral, and long–short day. Stylosanthes spp. differ from most tropical pasture legumes in other genera because of their nonclimbing growth habit. Growing points are often close to the ground, and this is advantageous under grazing. Another rare feature of Stylosanthes is its single seed in an indehiscent pod, which helps to regulate germination and improve seed survival (Gardener 1975). Stylosanthes is the genus that has received most attention in the search for tropical pasture legumes, and this has resulted in the release of a wide range of commercial cultivars, as summarized in Table 1.

Table 1. Released Stylosanthes cultivars.

Country and species	Common name	Cultivars	Year of release	Country of origin
Australia
S. guianensis var. guianensis	Common stylo	Schofield	—	Brazil
		Cook	1971	Colombia
		Endeavour	1971	Guatemala
		Graham	1979	Bolivia
S. guianensis var. intermedia	Fine-stem stylo	Oxley	1965	Argentina
S. hamata (2n = 40)	Caribbean stylo	Verano	1973	Venezuela
		Amiga	1991	—
S. humilis	Townsville stylo	Common type	—	Australia
		Gordon	1968	Australia
		Lawson	1968	Australia
		Paterson	1969	Australia
S. scabra	Shrubby stylo	Seca	1976	Brazil
		Fitzroy	1979	Brazil
Brazil
S. guianensis	Alfalfa de Nordeste	IRI 1022	1966	Brazil
S. guianensis var. pauciflora	Tardio stylo	Bandeirante	1983	Brazil
S. macrocephala		Pioneiro	1983	Brazil
China
S. guianensis		Pia Hua Dou = CIAT 184 = Pulcallpa	1987	Colombia
Colombia
S. capita		Capica	1982	Brazil
Peru
S. guianensis		Pulcallpa = CIAT 184	1985	Colombia
Thailand
S. humilis	Khon Kaen stylo	Khon Kaen	1984	Venezuela
Source: Peters (1992).

Overall, the genus is adapted to the tropics and subtropics. The natural habitats of Stylosanthes are usually areas of low soil fertility, especially where the soil has a low P content and an acidic nature, although forms adapted to alkaline soils are common in the Caribbean, Central America, and Mexico.

Stylosanthes has been shown to perform well under both drought and waterlogged conditions (Edye and Grof 1984). In contrast to most other tropical pasture species, Stylosanthes usually exhibits a high N content, combined with a very low P content, and the P content decreases as the plants age, especially under water stress. Although the amount of P is inadequate for the nutrition of grazing animals, other minerals seem to be available in sufficient amounts. In addition to improving natural rangeland and animal performance, Stylosanthes spp. have shown particular promise for inclusion in ley systems and as a cover crop in plantation agriculture (McCowan et al. 1986; Tarawali 1991).

The response to inoculation with Rhizobium varies largely among and within species; the N-fixation efficiency presumably depends on the environmental conditions of the collection sites. There is some evidence that tetraploid and allotetraploid plants of Stylosanthes tend to be of the promiscuously nodulating type, whereas diploid species collected from alkaline soils are more specific. In glasshouse studies, Stylosanthes has shown positive reactions to inoculation with Rhizobium (Saif 1987). However, infection with native strains of Rhizobium is likely to occur under most field conditions; therefore, inoculation is usually unnecessary (Howeler et al. 1987).

The most damaging disease, and thus one of the major constraints to propagation of the Stylosanthes, is the fungal disease anthracnose, caused by Colletotrichum gloeosporioides. An extensive pathogenetic specialization and variation for virulence can be found among strains of C. gloeosporioides. Stylosanthes shows some field resistance, although this varies widely between accessions and agroecosystems.

Major attributes of Stylosanthes hamata

The genus Stylosanthes has provided ample germplasm for a wide variety of agro-ecological situations in the tropics. Stylosanthes hamata cv. Verano was found to be particularly adaptable in agropastoral farming systems in the SHZ of West Africa. This could be attributed to the following characteristics (de Leeuw and Mohamed-Saleem 1994):

• Rapid germination of seeds (50–80% within 2 d);
• Requirement for high temperatures (>50°C) to break dormancy (which means that out-of-season rainfall does not cause a problem);
• Rapid root growth, leading to deep penetration and high soil-water extraction at an early age;
• Fast aboveground growth rates during periods of high soil-water content and high temperatures (>25°C);
• Facultative-perennial nature (some plants survive into the next growing season, further assuring sustained seed production in most growing seasons);
• Species indeterminacy (nonselective defoliations during flowering and seed setting have no serious effects on subsequent seed yield);
• Efficient seed-dispersal mechanisms (herbivores ingest seeds, which are then spread by feces and transported by ants and termites);
• High anthracnose tolerance; and
• Low relative palatability (compared with grass) early in the growing season, but high levels in the late rainy and early dry seasons.
  
  

The subhumid zone of Nigeria

Climate and soils

The studies were mainly conducted in the SHZ of central Nigeria (Figure 1), which has an average annual rainfall of 1 200 mm (more than 95% of this falls between April and October) and a growing period of 180–270 d. It has a long, 6-month dry season (October–April). The soils are essentially ferruginous, with low C and N contents, poor drainage, and a low cation exchange capacity (CEC).

The herbaceous cover of the SHZ of Nigeria consists mainly of annual grasses (Andropogon, Hyparrhenia, Pennisetum, Loudetia, etc.), with a low percentage of native legumes (Alysicarpus, Tephrosia, etc.) and trees such as Daniellia oliveri and Isoberlinia doka.

Figure 1. The subhumid zone of Nigeria, showing research–extension sites.

Socioeconomic conditions and cultural features

In the Nigerian SHZ, Stylosanthes interventions target three major categories of livestock and crop farmers (Waters-Bayer and Taylor-Powell, 1986):

• Pastoralists — Full-time livestock keepers, ranging from those with no consistent association with a particular area (nomads) to those based at one site (pure pastoralists);
• Agropastoralists — Livestock keepers who practice some cropping but as an enterprise subsidiary to animal husbandry:

  - Transhumant agropastoralists — those who grow crops at one site but seasonally move all or some of their cattle to other grazing areas;

  - Sedentary agropastoralists — those who keep cattle year-round close to the site of their cropping activities; and

• Crop farmers — Mostly indigenes who keep some livestock, mostly small ruminants, but as an enterprise subsidiary to cropping.

The agropastoralists are Fulani who no longer consider it necessary to move their small herds. These Fulani have settled close to farming communities, which provide markets for their meat, milk, and manure. In addition, they value the presence of public services, such as schools and dispensaries.

The settled Fulani live year-round at one site but shift every few years to another a few kilometres away, in contrast to the transhumant Fulani, who come into central Nigeria from the north each dry season. The influx of transhumant herds creates competition for grazing resources. The homesteads of settled Fulani are generally on marginal lands bordering hamlet areas and on fields that farmers have left fallow for several years. The Fulani own no land in central Nigeria and have no certificates of occupancy.

Most of the crop farmers in central Nigeria are from the Kaje, Kamantan, Ikulu, Aten, and Hausa ethnic groups. Their main crops are sorghum, millet, maize, cocoyam, and yam, and the animals they raise include cattle, sheep, goats, pigs, and poultry (Ingawa 1986).

Average household size is nine persons, who contribute virtually all the labour. Peak labour demands occur in May to August (cultivation) and in November (harvest). Seasonal labour shows some age- and gender-related differences.

Farming systems and Stylosanthes management

Adopting a farming-systems research approach and promoting the use of Stylosanthes in the region, the International Livestock Centre for Africa (ILCA, now the International Livestock Research Institute [ILRI]) identified animal diseases, poor nutrition, and difficulties arising from land-tenure systems as the main constraints affecting the livestock industry in the SHZ (ILCA 1979). Serious cattle diseases, such as rinderpest and contagious bovine pleuropneumonia, can be fairly well controlled with available vaccines and techniques. The land-tenure issues tend to be site specific and highly political; the best approach is therefore thought to be to accept this limitation and work with it. ILCA considered malnutrition, especially in the dry season, the area in which some improvement could be made.

Productivity-monitoring of traditionally managed herds in the SHZ revealed that poor nutrition leads to low milk offtake per lactating cow (700 mL d^-1), long calving intervals (2 years), drastic weight losses (15–20%) in the dry season, high calf mortality (30%), and low fecundity (50%), a result of nutritional anestrus (Mani et al. 1988; Rege, von Kaufman, and Mani 1993; Rege, von Kaufman, Mwenya et al. 1993). Obviously, the natural vegetation cannot adequately support the existing cattle population. The accepted minimum level of 7.5% crude protein (CP) in the ruminant diet (Crowder and Chheda 1982) is attained only from June to September, and the digestibility of the natural forage is also low.

Farmers in Nigeria try to overcome this feed constraint by providing the cattle with crop residues after harvest (mostly of maize, sorghum, millet, and rice), at the beginning of the dry season. In the late dry season, the animals browse and graze forage resources on fadama (lowland areas where residual moisture permits plant growth throughout the dry season). Agropastoralists also supplement their animals' diets with agroindustrial by-products and local salt lick (kanwa). Although the agropastoralists take advantage of this wide variety of feed resources, these measures are still inadequate, as the productivity of their cattle remains low.

Recently, it was found that the crop farmers' traditional practice of tethering goats in the wet season to prevent them from damaging crops creates feed stress. This feed stress leads to undernutrition and weight losses in breeding females, with consequent low reproductive performance (ILCA 1991), which poses a new problem for crop farmers who are landowners but have no interest in livestock other than small ruminants.

Crop yields per unit of land and per unit of labour are low mainly because fertilizers are not readily available to smallholders. For instance, average grain yields in farmers' fields were 1 800 kg ha^-1 for maize, 1 420 kg ha^-1 for sorghum, and 700 kg ha^-1 for millet (Powell 1984).

The poor nature of savanna soils contributes to the poor quality and low productivity of the herbage and crops in the SHZ. Any attempt to promote livestock production in the SHZ should, therefore, consider a program for maintaining soil fertility, as well as improving the nutritional value of the pasture. Herbaceous legumes offer an attractive option in this context, as they can provide both fodder for livestock and N to the soil.

Extension agents recommended the use of agroindustrial by-products — such as cottonseed cake, groundnut meal, urea, and molasses — to improve the productivity of lactating and pregnant cows. However, supplies of these products are not readily available, and prices are escalating. Similarly, a recommendation to use chemical fertilizers to boost crop production is problematic, as supplies of these chemicals are irregular. The fertilizers are too expensive for the small-scale farmer, and they also create environmental concerns.

In view of these ecological and financial constraints, ILCA considered a sustainable enterprise, such as planted forage legumes ("fodder banks"), as a more appropriate long-term option for improving cattle nutrition and soil fertility. This is because leguminous plants, such as Stylosanthes, can maintain a CP content of more than 8% in the dry season and the associated Rhizobium can fix N. It is against this background that low-input techniques have been developed for establishing Stylosanthes on natural range (Otsyina et al. 1987) and in cropped areas (Mohamed-Saleem 1985) to improve forage quality and soil fertility.

Stylosanthes on natural range

Sowing forage legumes can improve the nutritive value of the natural vegetation in rangelands. The low-input guidelines developed by ILCA (Otsyina et al. 1987) for the establishment, management, and use of these pastures are as follows:

Establishment

• Select an area close to the homestead (often 4 ha is adequate for an average-sized herd [40–50 animals per household]);
• Fence the area, using either metal or live poles, to prevent communal or untimely grazing;
• When the rains commence in April or May, prepare land for planting by confining animals as long as necessary or by using animal traction;
• Sow scarified seeds (10–12 kg ha^-1) after mixing with Single Superphosphate^TM (SSP) at a rate of 150 kg ha^-1;
  
  

Management

• Control grasses by early-season grazing, slash any shrubs, and destroy termite mounds;
• Leave forage to bulk up;
• Construct fire breaks at the beginning of the dry season; and
• Control dry-season grazing to ensure sufficient seed drop and adequate stubble for Stylosanthes regeneration in subsequent seasons.

ILCA's recommendation aims at feeding 15–20 lactating and pregnant cows for about 2–3 h d^-1, but herd owners tend to prefer strategic "survival feeding" for the whole herd.

Stylosanthes fallows in cropped areas

The undersowing technique is considered the most feasible method for introducing Stylosanthes into crop mixtures in the year before a piece of land is left to lie fallow. The understorey of stylo increases the nutritional value and quantity of the succeeding crop residue and improves soil fertility faster than a natural fallow. Undersowing exploits the land preparation primarily done for the crop and does not affect the cultural practices of the agropastoralists and farmers because it is based on intercropping, which is already the most common practice. When undersowing sorghum with Stylosanthes, a farmer must sow the legume 3–6 weeks after planting sorghum to avoid competition between the two crops (Mohamed-Saleem 1985).

Stylosanthes in legume–legume mixtures

Earlier evaluation work had identified several accessions of legumes that grew well in the West African SHZ, including Aeschynomene histrix, Centrosema brasilianum, Centrosema pascuorum, Centrosema pubescens, Chamaecrista rotundifolia, Stylosanthes guianensis, and S. hamata (Peters et al. 1994a, b). Tests made on collections of these species to identify material best adapted to the agroecological zone showed that none of these species is ideal. Centrosema pascuorum establishes well but soon disappears from a pasture. Centrosema brasilianum and C. pubescens establish slowly but stay green in the dry season. Aeschynomene histrix, S. guianensis, and S. hamata establish well and can persist for several seasons but do not stay green throughout the dry season. Individually, none of the legumes is ideal, but in the right combination, they might provide sustainable year-round grazing.

Therefore, our research efforts focused on developing legume–legume mixtures to replace the more commonly used grass–legume mixtures. A large-scale grazing trial was established to evaluate selected legume mixtures (such as C. pascuorum + S. guianensis + C. pubescens and C. pascuorum + S. guianensis + Centrosema macrocarpum) as supplementary pastures for young heifers (Tarawali et al. 1996).

Stylosanthes capitata, identified in similar environments in South America as a highly productive, drought- and disease-tolerant species adapted to soils with low fertility (Thomas et al. 1987), was evaluated for its potential in subhumid West Africa. This legume failed to nodulate in preliminary trials. However, later observations on abandoned plots in the same area revealed that several years after introduction, S. capitata nodulated and produced higher yields than S. hamata, the most widely used forage legume in subhumid Nigeria. An on-farm trial was therefore initiated to study the effects of S. capitata and S. hamata in various mixtures on forage and subsequent crop yields.

Animal evaluation

An on-farm study was conducted for 10 years (1977–87) at four locations in the SHZ of Nigeria (Mani et al. 1988). Fifty-eight herds of Bunaji cattle were involved, each with about 40–50 animals. In each herd, the animals were divided into two groups: those allowed to graze on Stylosanthes pastures for 2–3 h d^-1 during the dry season (October–March), in addition to grazing on natural pasture; and those grazing strictly on natural pasture all year round. Calves were weighed every 2 weeks until they were weaned, and adult animals were weighed periodically. All births, deaths, and milk yields were recorded for statistical analysis. Cattle with access to forage legumes in the dry season produced more milk, lost less weight, and had shorter calving intervals and a better rate of calf survival (Table 2).

Table 2. Effect of dry-season dam supplementation on the productivity of Bunaji cattle.

Variable	Grazing only	Fodder bank a	Improvement (%)	Significance (P)
Cow survival (%)	92.2	96.0	4.7	NS
Calving (%)	53.8	58.1	8.0	NS
Calf survival (%)	71.8	86.3	20.2	0.05
Calf weight at 1 year (kg)	98.1	103.4	6.6	0.05
Total milk yield (kg)	300.2	312.5	4.1	NS
Productivity index	51.5	69.1	34.2	NS
Source: Mani et al. (1988). Note: NS, not significant. a Planted forage legumes.

A similar study, involving West African Dwarf goats in 32 flocks owned by 45 smallholder farmers, was carried out for 30 months (Tarawali and Ikwuegbu 1993). At the beginning of the wet season, the goats were allowed to graze freely on one of two main treatments: natural vegetation and Stylosanthes pasture (miniature fodder bank). Animal performance was measured in terms of birth weights, deaths, stillbirths, abortions, etc. Weights were recorded fortnightly; the kids were weighed within 24 h of birth and weaned at about 5 months. A comparison of the wet-season liveweight (LW) changes of nonpregnant adults showed that those grazing on the Stylosanthes pastures had reduced (P < 0.05) weight losses (Figure 2). The kids' survival rate was also improved (P < 0.05) by legume supplementation (Ikwuegbu and Ofodile 1992).

Figure 2. Liveweight changes of West African Dwarf goats grazing on Stylosanthes and natural pasture. Source: Tarawali and Ikwuegbu (1993).

In experiments with legume–legume mixtures (Tarawali et al. 1996), the differences between heifers grazing on the legume mixtures and those grazing on unimproved pasture were dramatic. For instance, in the 1994/95 dry season, the former gained an average 140 g d^-1, whereas the latter lost an average 58 g d^-1 (Figure 3). The trial, which ran for two dry seasons (1994/95 and 1995/96) and one wet season (1995), compared the performance of heifers (three per treatment) grazing on one of two legume mixtures or on native range. The trial was run for two dry seasons (1994/95 and 1995/96) and one wet season (1995). Animals were weighed fortnightly and received routine veterinary care. The animals on the native pasture recovered on their own at the beginning of the wet season (this compensatory growth was due to the rapid improvement in the quantity and quality of pasture vegetation at the onset of the rains). However, this group was not as productive as those on improved pastures.

Figure 3. Effect of supplemental mixed-legume pastures on liveweight gains of Bunaji heifers. Source: Tarawali et al. (1996). Note: Mix 1 = Centrosema pascuorum + Stylosanthes guianensis +Centrosema pubescens; mix 2 = Centrosema pascuorum + Stylosanthes guianensis +Centrosema macrocarpum).

Agronomic evaluation

Pasture productivity

Yields from natural pasture and Stylosanthes-based fodder banks were compared at the end of the growing season, before grazing (Tarawali and Mohamed-Saleem 1994). Table 3 shows yields of about 4.3–7.9 t DM ha^-1 for fodder banks containing 52–68% Stylosanthes. These data were generated from researcher-managed on-farm trials. The yields were measured from randomly placed 1 m x 1 m quadrats, from which the herbage was cut and separated into legume, grasses, and forbs. The material was later dried at 60C and then weighed to determine productivity. The information presented in Table 3 is just for Stylosanthes pastures, but the corresponding yield for natural pasture (control) generated in a similar way was 2.4–2.9 t DM ha^-1. The productivity for farmer-managed pastures with 30–60% Stylosanthes was about 4–5 t DM ha^-1.

Table 3. Average DM yield and proportion of Stylosanthes in selected fodder banks in International Livestock Centre for Africa's study areas.

Location	DM yield (t ha-1)	Stylosanthes component (%)
Abet	4.28	58
Ganawuri	7.90	60
Kachia	7.11	68
Kontagora	6.12	52
Kurmin Biri	6.10	60
Avg.	6.30	60
Source: Tarawali and Mohamed-Saleem (1994). Note: Avg., average; DM, dry matter.

Table 4. Effect of different proportions of Stylosanthes capitata and Stylosanthes hamata on forage yields (total of 3 years), soil properties, and subsequent maize (fertilized with only P and K) production.

S. capitata– S. hamata ratio	Forage yield, 1990–92 (kg ha-1)	Maize yield, 1993 (kg ha-1)	Total soil N (g kg-1)	Organic C (g kg-1)
100 : 0	10 209	303 ± 70	0.15 ± 0.04	8.57 ± 1.2
75 : 25	14 822	587 ± 172	0.19 ± 0.01	9.85 ± 1.7
50 : 50	12 925	492 ± 114	0.22 ± 0.01	11.56 ± 0.5
25 : 75	11 959	509 ± 141	0.23 ± 0.18	10.48 ± 2.5
0 : 100	10 495	460 ± 190	0.20 ± 0.04	9.53 ± 2.2
Source: Tarawali and Peters (1997).

Results of an S. hamata–S. capitata compatibility trial (Tarawali and Peters 1997) showed that total forage-DM yields in the mixtures were higher than in the sole stands of either species (Table 4), with S. capitata increasing its contribution over time. This trial was initiated in 1990; S. capitata and S. hamata were sown at a seed rate of 10 kg ha^-1 each in 2 m x 3 m plots. Four to six weeks after planting, the plants were thinned to 100 seedlings m^-2 in various proportions (see Table 4). The five treatments were arranged in a randomized complete-block design, with four replicates. Plots were kept weed free and were fertilized with SSP at 150 kg ha^-1 at planting and at 100 kg ha^-1 in subsequent years (1991 and 1992). Forage parameters were studied for the first 3 years; in 1993, plots were cropped with maize to obtain information on the effect of various proportions of the two legumes on cereal production. The maize did not receive any N fertilizer, but basal dressings of P and K were each applied at 60 kg ha^-1 as SSP and muriate of potash, respectively. Before maize planting, soil samples were taken for determination of total N and organic C.

Total soil-N concentrations were higher in plots following 3 years of S. capitata–S. hamata mixtures than in plots that had had sole stands of S. capitata (see Table 4). Organic C concentrations were higher in plots after the mixtures than in plots after either sole S. capitata or sole S. hamata. Maize yields were correspondingly higher following the mixtures than after either of the sole stands. The higher soil-N and organic C contents and maize yields in the plots following the mixtures suggest the complementarity of S. capitata and S. hamata. Thus, although not recommended for sole-legume pastures, S. capitata could be used in mixtures with other complementary legume species, such as S. hamata, C. brasilianum, and C. pubescens.

Crop production

In another trial (Tarawali 1991), maize yields at three levels of applied N (0, 60, or 120 kg ha^-1) were greater on plots that had had a leguminous cover crop than on plots that had been natural pasture. Without fertilizer-N additions, the average grain yields were 1 700 kg ha^-1 in the leguminous area and 800 kg ha^-1 for the natural pasture (Figure 4). The trial, which was initiated in 1986, was conducted at four locations in central Nigeria to evaluate the fertilizer response of maize grown after at least 4 years of uncropped natural fallow or 3 years of Stylosanthes pasture (fodder bank). The experiment was a split-plot design, with the Stylosanthes and non-Stylosanthes areas as the main plots and the levels of N as subplots. Basal dressings of P and K were each applied at 60 kg ha^-1 in the form of SSP and muriate of potash, respectively. At the end of the growing season, in October, the crops were harvested, dried, and weighed to determine grain yield.

Figure 4. Effects of fertilizer-N on grain yield of maize inside and outside Stylosanthes plots. Source: Tarawali (1991).

In the first year of cropping, maize grown on the natural pasture needed 45 kg N ha^-1 to produce a yield equivalent to that of unfertilized maize grown on a good Stylosanthes pasture. In the second year, the yields were much lower, but the proportional increase in yield attributable to forage legumes was similar to that in the first year. This suggests that the legumes still had some positive residual effect, but this was insufficient for the optimum growth of maize.

In the case of acha (Digitaria exilis) grown after Stylosanthes, no response was shown to the supplemental addition of fertilizer-N. A multilocation trial (Tarawali and Pamo 1992) compared the performance of acha grown in two main plots, previously under Stylosanthes or natural fallow, at various levels of N (0, 40, 80, and 120 kg ha^-1). P and K were each applied to all the plots at 60 kg ha^-1. The experiment was a split-plot design, and each treatment was replicated four times. At the end of the trial, grain yield was determined. The highest acha-grain yield (560 kg ha^-1) was obtained on the Stylosanthes pasture, with no N application (Figure 5). The highest yield on natural fallow required 40 kg N ha^-1. This trial shows that maximum acha yield can be obtained if the grain is planted after a Stylosanthes pasture and receives no N fertilizer.

Figure 5. Response of acha inside and outside Stylosanthes plots to fertilizer-N. Source: Tarawali and Pamo (1992).

The positive impact of Stylosanthes on crop production has also been demonstrated for other crops, such as millet, sorghum, and soybean (Tarawali and Mohamed-Saleem 1995).

The higher crop yields on the Stylosanthes pastures were due to the legume's improvement of the soil's physical and chemical properties. For instance, a series of measurements (Tarawali and Ikwuegbu 1993) showed that Stylosanthes decreased the soil's bulk density and increased its porosity (capacity to retain moisture), its CEC, and its organic C and N contents (Table 5). No standard errors are presented for these data because the analysis was conducted for a limited number of fodder banks.

Table 5. Soil chemical and physical properties under Stylosanthes and natural fallow.

Property	Stylosanthes (3 years)	Natural fallow (>4 years)
N content (g kg-1)	1.14	0.87
CEC (cmol kg-1)	3.24	2.22
Organic C (g kg-1)	4.31	2.70
Bulk density (g cm-3)	1.51	1.66
Total porosity (%)	43.1	37.4
Macroporosity (%)	42.1	36.4
Microorganisms (n x 107 g-1)	34	12
Source: Tarawali and Ikwuegbu (1993). Note: CEC, cation exchange capacity.

Economic evaluation

The economic benefit of fodder banks was assessed by von Kaufmann and Mohamed-Saleem (1989), who compared the cost of producing a unit of CP from fodder banks (1.96 NGN kg^-1 CP) with the market price of a given unit of cottonseed cake (2.27 NGN kg^-1 CP) (in 1998, 75.2 Nigerian naira [NGN] = 1 United States dollar [USD]). It was shown that CP produced from fodder banks was cheaper than that from purchased cottonseed cake, an alternative form of dry-season feed supplementation (Table 6). Given the current high rate of inflation, coupled with the scarcity and high cost of cottonseed, the agropastoralists who established their fodder banks a few years ago should now be making some gain from their investments. Using a model to appraise the economic returns of fodder banks over 10 years, von Kaufmann and Mohamed-Saleem confirmed that fodder banks could be attractive investments (Table 7). For instance, their evaluation, which included capital and recurrent costs (such as those for fencing, seed, fertilizer, labour) of a 4-ha fodder bank and benefits (such as those from animal products and crop yield), showed an internal rate of return varying from 22.5 to 36.3% for fodder-bank-supplemented herds.

Table 6. Costs of obtaining crude protein from a 4-ha fodder bank and from cottonseed cake, subhumid Nigeria, 1989.

	Quantity
Fodder bank (4 ha)
DM produced (kg)	16 000
DM available (kg)	8 000
CP content (kg) a	720
Capital cost (NGN)	5 944
Recurrent cost (NGN kg-1 CP)	1.96
Cottonseed cake
CP	720
Required DM at 30% CP (kg)	2 400
Capital cost (NGN)	0
Recurrent cost b	2.27
Source: von Kaufmann and Mohamed-Saleem (1989). Note: CP, crude protein; DM, dry matter; NGN, Nigerian naira (in 1989, 7.3 NGN = 1 United States dollar [USD]; in 1998, 75.2 NGN = 1 USD). a Assumes 9% crude protein content available in dry matter. b Calculated as 680 NGN t-1 of cottonseed cake, at 30% CP.

Table 7. Economic returns on fodder banks in subhumid Nigeria over 10 years, 1989.

	Net present value a (NGN)	Internal rate of return (%)	10th-year herd value		10th-year incremental net revenue (NGN)
			Without fodder bank (NGN)	With fodder bank (NGN)
IHP	1 414	22.5	49 907	90 833	4 950
IHP + reduced forced sales	7 538	34.1	49 907	90 833	7 138
IHP + increased crop yields	9 395	36.3	49 907	90 833	8 544
Source: von Kaufmann and Mohamed-Saleem (1989). Note: IHP, improved herd productivity; NGN, Nigerian naira (in 1989, 7.3 NGN = 1 United States dollar [USD]; in 1998, 75.2 NGN = 1 USD). a Calculated at 20% discount rate.

Adoption of Stylosanthes and farmers' perceptions

Agropastoralists in the Nigerian SHZ originally exploited Stylosanthes for cattle, but crop farmers later took advantage of the soil-improving properties of this cover crop for small ruminants and crop production (miniature fodder banks). Formal and informal surveys were conducted by a multidisciplinary team of scientists and extension workers to record farmers' reactions to the innovations and the associated benefits. Both farmers and agropastoralists acknowledged the beneficial effects of the fodder-bank intervention in terms of improved agricultural productivity (and hence increased income) as well as increased environmental protection. This was reflected in an increase in the number of fodder banks from 2 in 1980 to about 620 in 1991 (Figure 6). The documented adoption trend (Ajileye et al. 1994) was influenced mainly by the ILCA–ILRI projects and the Nigerian Livestock Department's promotion of Stylosanthes among agropastoralists and farmers, though there was evidence of farmer-to-farmer dissemination. The data came from records and were verified on field visits by extension staff. According to more recent reports, the technology has continued to expand in the farming systems of Nigeria and other West African countries (de Leeuw et al. 1994; ICTA 1995).

Figure 6. Cumulative number of Stylosanthes pastures in Nigeria from 1981 to 1991. Source: Adapted from Ajileye et al. (1994).

Document(s) 18 of 36