Three experiments were conducted using combinations of the following legumes: velvetbean (Mucuna aterrima), crotalaria (Crotalaria striata), zornia (Zornia latifolia), jackbean (Canavalia ensiformis), Calopogonium mucunoides, pigeon pea (Cajanus cajan), and tropical kudzu (Pueraria phaseoloides). At the site 95% of the annual rainfall, about 1500mm, falls between October and April. The soils are Red Yellow or dark Red Latosols of the suborder Ustox, which have excellent physical conditions but are very acid and highly deficient in phosphorus. The site was cleared of Savanna vegetation and cropped several years prior to the experiments.
The green manure crops were grown and incorporated into the soil. Fallow plots were maintained to follow N mineralization and subsequent N accumulation in the soil. Maize was planted in other plots to measure N uptake and yield response. Fallow plots were sampled periodically and inorganic N was measured.
Measurements were taken of the N content of total above-ground matter grown on cropped plots, and grain yields. The relationship between inorganic N accumulation in fallow soil, and grain yield and N content of total above-ground matter was determined.
In experiment 1, maize grain yields up to 6,800 kg/ha were achieved using the legume green manure mucuna as the only N source. Using an adjacent fallow plot to measure N mineralization from legume green manures, the net increase in inorganic N resulting from mineralization varied from 60 to 165 kg N/ha. Approximately 40% to 50% of the total mineralized N was taken up by the maize crop.
In experiment 2, zornia, mucuna, Canavalia ensiformis and crotalaria were compared. Crotalaria produced much more total above-ground dry matter but contained only slightly more total N than did either mucuna or canavalia. Yield of the first maize crop grown during the dry season following zornia (4,700 kg/ha) reflected the lower amount of N fixed, while maize following crotalaria yielded slightly less (5,800 kg/ha) than maize following either mucuna or canavalia (6,100 and 6,300 kg/ha, respectively).
The third experiment included several green manures in addition to those used in the first two. The same procedure was used. Yield and N uptake related to N supplied by the green manure, and maize yields of up to 7,000 kg/ha were achieved. Crotalaria contained the most N but yield and N uptake did not reflect the additional N. The residual effect of these plant materials was rather small as shown by either yield increase or N uptake.
Other experiments show that not more than 60% of the inorganic N in the rooting zone could be absorbed by the crop.
These results show that when conditions are favorable for the growth of maize, there will be a rapid mineralization of fixed N and large responses to legume green manure N resulting in yields obtained of up to 7 T/ha. Yields as high as those obtained with 200 kg/ha of fertilizer N were found, an indication that green manure can be as effective as fertilizer N sources for maize.
Fifty-two varieties of legumes were also screened for drought tolerance at the CPAC site. Legumes were planted at the end of the rainy season; 37 of the 52 varieties survived a five-month dry season and began growing when the rains started again (Burle et al., 1988). With the best-adapted legumes, regrowth over a six-week period beginning at the start of the rainy season reached 5000 kg/ha, and total N in the aboveground dry matter exceeded 100 kg/ha. Yields of maize grain following incorporation of regrowth of legume green manures surviving the dry season reached 7.4 t/ha without added fertilizer N (Burle et al. 1989). Five legumes were tested for drought tolerance in an acid savanna to determine their survival capability in a dry season with no irrigation. Cajanus cajan, Canavalia brasiliensis, Mucuna aterrima, Tephrosia candida and Stylosanthes guianensis all survived the first year. All except Stylosanthes guianensis survived the second year. Above-ground dry matter ranged from 0.6 and 4.3 T ha-1. The legumes that produced greatest above-ground dry matter also extracted water from deeper in the soil profile. The effect of this system on maize yields was also tested and reached a maximum of 7.4 T ha-1 after legume incorporation.
To maximize N fixation, legumes must be adapted to prevailing climatic conditions, and suitable soil conditions favoring dry matter accumulation are required. Also, mineralization of fixed N should occur at a rate commensurate with the growth of the succeeding crop and in quantities that are equivalent to crop requirements at the time of release. According to these experiments, the rate of release of inorganic N from the rapidly oxidizable fraction of the plant material varied little among the various legume green manures tested.
Box 7113, Raleigh, NC 27695-7113 USA.