Physiological Constraints and Their Genetic Manipulation

Abstract

Crop yield is the end product of a great many different processes operating within the plant for all or some part of the growing season. Its expression is controlled by a wide range of complex polygenic systems and is highly sensitive to environmental variation. Moreover, much of the variation in yield estimates from trials of cultivars conducted over different sites and seasons can often be attributed to genotype x environment interactions (Brennan and Byth 1979). Clearly, such interactions will greatly reduce the precision with which a breeder can identify high-yielding genotypes, especially in early segregating generations. Visual selection (McKenzie and Lambart 1961) and selection based on single-plant yields (Knott 1972) in early segregating generations of crosses in self-pollinating species have usually proved ineffective. Approaches to improving early generation selection for yield have centred on planting designs which reduce micro-environmental variation in the breeding block (Gardner 1961; Fasoulas 1973) or the use of alternative selection criteria (Bhatt 1988). The results of selection experiments involving special planting designs have been inconsistent, although some improvements have been proposed to improve their efficiency (Bos 1983). Alternative criteria of selection for yield must be simple to measure, significantly correlated with a plot yield and highly heritable. Most characters fulfilling these conditions would be morphological in type, but in certain cases are associated with physiological determinants of yield. The most widely used alternative criteria have been primary morphological components of seed yield such as the number of pods per plant, number of seeds per pod and weight per seed in the case of the oilseed brassicas. Although these characters would be appropriate in respect of the conditions listed above, they have often proved unreliable because of the existence of compensatory mechanisms resulting in reductions in one or more components in response to the selection for another specific component (Adams 1967). Various biometrical procedures have been proposed to overcome the problems associated with yield component compensation (Grafius and Thomas 1971), but their effectiveness in actual breeding programmes is yet to be properly evaluated. In certain instances, direct selection for one or other grain yield component has proved effective, as observed in the case of selection for seed weight in mustard (Gupta and Labana 1985).