OPTIMUM PLANT-POPULATION DENSITIES FOR LUPIN (LUPINUS-ANGUSTIFOLIUS L) IN THE WHEAT-BELT OF WESTERN-AUSTRALIA

Lupin (Lupinus angustifolius L.) seed yields at various plant population densities were studied in 33 separate experiments throughout the wheatbelt of Western Australia between 1987 and 1990. The experiments were designed to test the hypotheses that optimum plant population densities for lupins vary between environments and between cultivars. Another objective was the development of a framework for sowing rate recommendations from a large data set derived from sowing rate experiments. Two types of equation were fitted to each data set by nonlinear regression: one described an asymptotic response, the other a response where yield reached a maximum but declined at higher densities. The second type of equation was used to describe a data set if the residual mean square was significantly lower than for the asymptotic equation. In all, 122 individual responses were fitted, of these only 13 were not adequately described by the asymptotic model. Optimum density was chosen according to an economic criterion (when marginal revenue from an increase in plant population density equalled marginal cost). This was equivalent to choosing the point where the slope of the response curve was 0.004 t.m(2)/ha.plant (equivalent to 0.4 g/plant). Optimum density ranged from 14 to 138 plants/m(2) and was linearly related to yield potential, which we defined as either the asymptotic yield value, or the maximum yield for responses that did not approach an asymptote. Yield potential ranged from 0.13 to 4.1 t/ha. The relationship between optimum density and yield potential was the same for cvv. Danja, Gungurru, and Yorrel, and for a reduced branching breeding line (75A/329). It was also the same on soils classified as good or poor for lupins. We suggest that the relationship between optimum density and yield potential will be useful in determining target plant densities for lupins under a wide range of conditions in Western Australia, and that the techniques should prove useful in producing recommendations from density experiments in other agricultural regions.