Maize legume relay intercrops in Malawi - Microbial Ecology in Sustainable Agroecosystems - page 237

Phase 1: 1999-2001 Phase 2: 2003-2004

1) MZ+F 3) SS+F

1) MZ+F 2) SS+

1/2FF

5) MZ-F

2) MZ-F 4) SS-F

5) TV+F 7) PP+F

3) TV+
1/2FF

4) PP+
1/2FF

6) TV-F 8) PP-F

Figure 9.1 Project designs 1 (phase 1) and 2 (phase 2). Phase 1 treatments: (1) maize control plus
full fertilizer (MZ+F); (2) maize control without fertilizer (MZ-F); (3) maize intercropped with S.
sesban plus full fertilizer (SS+F); (4) maize intercropped with S. sesban without fertilizer (SS-F); (5)
maize intercropped with T. vogelii plus full fertilizer (TV+F); (6) maize intercropped with T. vogelii
without fertilizer (TV-F); (7) maize intercropped with pigeon pea plus full fertilizer (PP+F); and (8)
maize intercropped with pigeon pea without fertilizer (PP-F). Phase 2 treatments: (1) maize plus
full fertilizer (MZ+F); (2) maize intercropped with S. sesban plus the second fertilizer application
(SS+1/2F); (3) maize intercropped with T. vogelii plus the second fertilizer application (TV+1/2F); (4)
maize intercropped with pigeon pea plus the second fertilizer application (PP+1/2F); and (5) maize
receiving no fertilizer (MZ-F). When sampling for phase 2, we always sampled from the center of
the plot. In this way, we averaged the variation between previously fertilized and unfertilized plots.

During D1, one-half of each plot received an inorganic fertilizer treatment at a rate of
45 kg N ha -1 (see Figure 9.1 for cropping system abbreviations). As is typical for the region,
fertilizer was applied twice during the cropping cycle: one-half at maize planting (for-
mulated as 23N:21P:0K and 4S) and the other half as a side-dress of calcium ammonium
nitrate when maize was roughly 60 cm in height. For D2, the experimental design was
modified to determine if legumes could replace the first inorganic fertilizer application. At
the time, few farmers were able to afford the full recommended rate of fertilizer as used in
D1. Thus, in 2001 new treatments were superimposed over the same plots previously used;
each legume treatment remained in the same location, but the plots were no longer split
( Figure 9.1 ) . The three legume systems did not receive the first inorganic fertilizer appli-
cation, but each still received the second fertilizer application (22.5 kg N ha -1 ). The entire
maize control plot received the full rate of fertilizer (45 kg N ha -1 ) at the two standard appli-
cation timings. A fifth plot, MZ-F, was added at each farmer's field as the untreated control
with maize alone. Based on this design, there was no true control to determine the impact
of legumes versus fertilizer. Legume systems (SS+1/2F, TV+1/2F, and PP+1/2F) could only
be compared to one another and again unfertilized or fertilized maize controls. Data from
the 2001/2002 cropping cycle were not included in the analysis due to the recent change in
the experimental design.
To fit within smallholder farming systems and minimize competition with maize, the
perennial legume species were annually replanted. For both designs, nonwoody portions
of the legumes were cut and incorporated into the soil, typically in late September or early
October. Researchers applied fertilizer, provided much of the labor related to legumes
(sowing, transplanting, and incorporation), and harvested maize from subplots. While
this was done to ensure uniform management, it would have been preferable in retrospect
to have farmers directly involved in these activities (see further discussion). Farmers per-
formed land preparation, sowed maize, weeded, and harvested maize outside the experi-
mental subplots. D1 data were collected by researchers at the University of Malawi, Bunda