Variable grain legume yields, responses to phosphorus and rotational effects on maize across soil fertility gradients on African smallholder farms

Abstract

Promiscuous soyabean varieties have potential to contribute significantly to income generation, food security and soil N budgets on smallholder farms. One of the major factors limiting this potential is farmers’ preference to allocate nutrient resources to food security cereal crops on the most fertile fields, leaving grain legumes to grow on residual fertility on infertile fields. Two experiments were conducted to: (i) compare the current farmer practice with targeting manure and single super phosphate (SSP) to soyabean in a three-year rotation cycle on two fields with different soil fertility: an infertile sandy soil and a more fertile clay soil; and (ii) assess the effects of variability of soil fertility within and across farms on productivity of soyabean and groundnut. In the first experiment, soyabean (<0.2 t ha⁻¹) and maize yields (<0.7 t ha⁻¹) without fertilizer were poor on a degraded sandy soil. Both crops responded poorly to SSP due to deficiency of other nutrients. Manure application significantly increased soyabean and maize yields, led to yield stabilization over three seasons and also significantly increased the proportion of N₂ fixed by soyabean (measured using ¹⁵N natural abundance) from 60% to 83%. On the sandy soil, P was used more efficiently and gross margins were greater when SSP and manure were applied to maize in a maize–soyabean rotation. Soyabean and maize yields without fertilizer inputs were larger on clay soil with moderate fertility (0.4–0.7 t ha⁻¹ and 2.0–2.3 t ha⁻¹ respectively) and were significantly increased by application of SSP and manure. Within rotations, P recovery was higher when manure and SSP were applied to maize (43 and 25%) than when applied to soyabean (20 and 19%). However, application of manure to soyabean on the clay was more profitable than application to maize for individual crops and within rotations. In the second experiment, soyabean and groundnut yields were largest (∼1 and ∼0.8 t ha⁻¹ respectively) on plots closest to homesteads on wealthy farms, which were more fertile due to good past management. Yields were poor (< 0.5 t ha⁻¹) on other fields which previously had received little nutrient inputs. Soyabean and groundnut yields correlated well with available P (R ² = 0.5–0.7) and soil organic C (SOC) contents (R ² = 0.4–0.6). For smallholder farmers to maximise benefits from legume production they need to focus attention on the more fertile plots, although production should be optimized in relation to maize. Targeting nutrients to maize as currently practiced by farmers was more efficient and economic under poor soil fertility conditions, whilst potential exists to increase income by targeting manure to soyabean on the more fertile soils.