Maize legume relay intercrops in Malawi - Microbial Ecology in Sustainable Agroecosystems - page 253

dried soil, making them appropriate for sampling large numbers of fields in distant sites.
It would be interesting to see if this suite of indicators effectively captures management
impacts on soil quality across a range of locations and if the permanganate measure proves
to be a good indicator of soil biological changes.
It is interesting to note that there is close correspondence between the systems recom-
mended and what farmers actually adopted during the course of the study. Essentially all
farmers planted some pigeon pea with their maize, about 20% planted Tephrosia, and only
6%, all wealthier farmers, planted Sesbania . Overall, our results concur with others (Snapp
et al., 2010; Kamanga, Waddington, et al., 2010; Kanyama-Phiri et al., 1998), which shows
that legumes could play an important role in increasing sustainability of maize production
systems in Malawi, but that the benefits of using legumes were greatly enhanced by their
use in combination with moderate amounts of inorganic fertilizer ( Table 9.1 , Figure 9.2 ) .
The higher productivity under D1 as compared to D2 suggests that farmers should ideally
use as much fertilizer as they can afford up to the recommended amount and, if possible,
apply it in two applications as done in D1. However, it is unclear at this point when to
recommend application of more limited amounts of fertilizer or indeed at what stage it is
best to incorporate the legume biomass. Our understanding of the dynamics of N release
from different age legume residues is currently limited, as is our understanding of effects
of amounts and timing of fertilizer application on growth and nitrogen fixation by differ-
ent legume species.
Legume productivity was less than optimal in these on-farm trials given the low bio-
mass obtained ( Table 9.2 ) relative to other values in the literature, suggesting there is con-
siderable potential to further improve system productivity. There is a real opportunity
for researchers or extension agents to work with farmers in an iterative and adaptive way
to optimize legume management for different landscape positions and resource levels
(Shennan, 2008). Research is needed on basic agronomic questions (optimizing seeding
rates, planting arrangements, fertilizer amounts and timing, timing of biomass incorpora-
tion, etc.), assessment of rhizobium levels and BNF, and the use of improved legume and
maize cultivars as suggested by Snapp, Blackie, et al. (2003). Work is also needed to better
understand legume and fertilizer effects on labile organic matter pools, microbial activity,
and patterns of mineralization/immobilization from different age residues to optimize
legume management.
Focus groups held at the end of the project showed that farmers had ideas for improv-
ing legume management and suggestions to make the system work better in terms of tim-
ing of labor demands. Farmers frequently suggested earlier incorporation of the legumes
so that the biomass was further along in decomposition and would not interfere with land
preparation (creating ridges) for the upcoming cropping season. Sirrine, Shennan, and
Sirrine (2010) noted that this would require farmers to forgo a small second harvest of dry
pigeon pea but would also incorporate residue earlier before the leaves begin to senesce,
perhaps minimizing N loss from decomposing leaves. Farmers also suggested that less
woody material be incorporated because they felt it decomposed too slowly, which was
primarily an issue for S. sesban, which has smaller leaves and stems that were more chal-
lenging to thresh. As discussed, changing the timing and composition of residue incor-
poration will impact N release dynamics and perhaps increase vulnerability to leaching
losses. Reducing total biomass inputs by not incorporating woody material may also have
an impact on long-term SOM changes. These trade-offs would need to be evaluated along
with the economics of the different incorporation strategies.
In many on-farm projects, farmer involvement is limited, as in this one, due to the
trade-off between including farmers in all facets of cropping system management and