Meta-analysis of maize yield response to woody and herbaceous legumes in sub-Saharan Africa

A number of studies have tested the effect of woody and herbaceous legumes on soil fertility and maize yields in sub-Saharan Africa. However, their effects on maize productivity are much debated because results have been variable. A meta-analysis was conducted with the aim of evaluating the evidence in support of yield benefits from woody and herbaceous green manure legumes. A total of 94 peer-reviewed publications from West, East and southern Africa qualified for inclusion in the analysis. Maize yield from herbaceous green manure legumes (54 publications), non-coppicing legumes (48 publications), coppicing woody legumes (10 publications), natural fallows (29 publications), and fully fertilized monoculture maize (52 publications) were compared. Mixed linear modelling using yield differences (D) and response ratios (RR) indicated that the response to legumes is positive. The mean yield increase (D) over unfertilized maize was highest (2.3 t ha(-1)) and least variable (CV=70%) in fully fertilized maize, while it was lowest (0.3 t ha(-1)) and most variable (CV=229%) in natural fallows. The increase in yield over unfertilized maize was 1.6 t ha(-1) with coppicing woody legumes, 1.3 t ha(-1) with non-coppicing woody legumes and 0.8 t ha(-1) with herbaceous green manure legumes. Doubling and tripling of yields relative to the control (RR > 2) was recorded in coppicing species (67% of the cases), non-coppicing legumes (45% of the cases), herbaceous green manure legumes (16% of the cases) and natural fallows (19% of the cases). However, doubling or tripling of yields occurred only in low and medium potential sites. Amending post-fallow plots with 50% of the recommended fertilizer dose further increased yields by over 25% indicating that legume rotations may play an important role in reducing fertilizer requirements. Except with the natural fallow, the 95% confidence intervals of D and RR were higher than 1 and 0, respectively indicating significant and positive response to treatments. Therefore, it is concluded that the global maize yield response to legumes is significantly positive and higher than unfertilized maize and natural vegetation fallows.