Maize was grown from three batches of seed, with mean iooo-grain weights of 225, 432 and 649 g. Initial plant size was larger when grown from large seed, but development rates were similar for all three sizes and relative growth rates were similar during the early stages. Relative differences in plant size became smaller as the crop matured. Crop growth rates during the linear phase of dry matter production were the same, and there were no significant yield differences. Comparisons of leaf and spikelet initiation, and individual leaf size are also reported.Several sizes of maize seed are available in Kenya for planting, largely due to the advent of mechanical planters which require graded seed. We could find no published work on the effect of seed size on subsequent yield of maize in Kenya, despite the obvious importance of this relationship for the small farmer. Indeed, even on a world scale this aspect of maize agronomy has been neglected compared with the smaller grain cereals. Kiesselbach (1937), Bruce Hunter and Kannenburg (1972), Sharma (1974) and Hicks etal. (1976) found little correlation between maize yield and seed size, but Rammana (1967) obtained significant increases in yield from larger seeds.Other work (Hammes, 1969;Gubbels, 1974) showed that seed size does affect seedling vigour, and earlier studies by Kidd and West (1919) led to the suggestion that 'recent work is tending to establish that ordinary variations observed in the size and vigour of seedlings are reflected in the size of the final yield independently of the environmental conditions obtained throughout the main period of the plant's life'. Furthermore Cooper and Law (1977) have shown that variations in seedling vigour in the Highlands of Kenya, induced by soil temperature differences, were closely correlated with final grain yield according to the equation: Y= IO-I W 5 + 11-9 (r = o-94) where Y = grain yield (g/plant), and W 5 = plant dry weight at 5 weeks (g/plant).It therefore seemed pertinent to assess what effect seed size had on seedling vigour, and to establish whether differences in vigour related to seed size were also associated with yield differences in the Highlands of Kenya.
MATERIALS AND METHODSMaize seed (Kenya hybrid 613c) was sieve-graded into three sizes: large round (iooo-grain weight 649 g), medium flat (432 g) and small (225 g), designated L, M and S respectively. The small seed represents cleanings, usually rejected by the f Present address: Department of Physiology and Environmental Studies, School of Agriculture, Sutton Bonington, Loughborough LE12 5RD.