Three grass/legume pastures were evaluated under Northern Rhodesian conditions.

A plot trial measured the herbage yields of Stylosanthes gracilis, velvet beans (Stizolobium deeringianum) and giant Rhodes grass (Chloris gayana) grown as pure stands, and in a mixed association with Rhodes grass. The herbage was harvested for 3 years, and the plots were then planted with a maize crop to test any residual effects.

After the establishment year, the D.M. yields, and especially the C.P. yields of the legume and grass/legume treatments, were markedly superior to grass alone. The increased yield of the grass/legume mixtures was due to the net gain contributed by legume herbage. There was no evidence of an underground transference of N from the legume to the associated grass.

Although there was no measurable treatment effect on soil N and C, the maize following the legume and grass/legume treatments had a higher leaf N level, and approximately double the yield, compared with the maize following the grass-alone treatments.

There were no significant differences in cattle weight gains when Rhodes grass alone, Rhodes grass/Stylosanthes, Rhodes grass/velvet beans, and Rhodes grass/pigeon peas (Cajanus cajan) were utilized as hay and foggage in the dry season.

Digestibility trials showed mature Rhodes grass hay to be a submaintenance feed, deficient in dietary protein. The addition of Stylosanthes raised the herbage digestible C.P. fourfold. The Rhodes grass/Stylosanthes hay was a maintenance feed.

The N yield of the grass/legume mixtures, although low by temperate standards, was 2½ times the N yield of grass alone. The potential value of a tropical grass/legume pasture will depend on the efficiency of the legume to fix N.