Four trials with spring wheat over two seasons examined the effect of seed numbers plant populations, and ear population at harvest on wheat yield. In four unreplicated blocks on a Marton yellow-grey earth in 1979-80, yield increased linearly with ear population until an optimum population (produced from 360 seeds per m2) was reached at around 680 ears per m 2 . A population greater than this led to a 7.5% significant yield reduction through a decrease in the number of grains per ear and in grain weight. Three replicated trials on a Kairanga sandy loam soil in 1978-79 and 1979-80 showed that ear populations of 700-800 per m 2 at harvest were supported with no detrimental effect on total yield, although grains per ear and grain weight tended to decrease with increasing ear populations. Conventional drilling produced great variation in plant numbers per unit area, which affected recorded yields. The significance of this is discussed and methods for achieving evently spaced optimum populations are considered.