Improvement of rice plant type concepts: systems research enables interaction of physiology and breeding

Abstract

Varietal selection is mostly based on convention and breeders' intuition, with little of the knowledge of quantitative plant-environment interactions. Yet, growth models can predict the performance of existing rices. Arecent application of models is the design of alternative plant types. Two studies are presented. One resulted in a new ideotype for direct seeded, flooded rice in favourable environments; the other addresses genotypic requirements for adverse thermal environments in the Sahel. Assimilate and nitrogen partitioning, gas exchange and yield of direct seeded and of transplanted IR64 rice were studied in the Philippines in 1986–1990. The growth model L3QT was deve1oped, simulating crop growth and production under nitrogen limitation. Partitioning patterns were then modified to simulate hypothetical genotypes. In IR64, leaf area index limited growth prior to panicle initiation, but during later stages, the leaf area appeared to be excessive, the leaf nitrogen content low and growth limiting, particularly for direct seeded rice where a dense population and the absence of a transplanting shock led to a large vegetative biomass. Simulated yields increased if initialleaf growth was rapid at the expense of tillering, and leaf area was kept low later on. This required enhanced leaf assimilate export and storage in stems. The lower leaf area corresponded with a higher leaf nitrogen concentration, and hence a higher assimilation, and delayed monocarpic senescence. Transfer of more nitrogen to upper leaf strata, and a short vegetative and long ripening phase also improved yield.

The model L3QT is being adapted for ideotype development for Sahelian, irrigated rice environments. Focus is on seasonal and diurnal thermal variation and their effect on crop duration and yield potential.