From 47 S2 lines which had been extracted from a random mated population of sorghum, eight lines for a glasshouse experiment and four lines for a field experiment were divergently selected for variation in osmotic adjustment, and were grouped into two, High and Low osmotic adjustment (OA). Both the glasshouse and field experiments examined whether osmotic adjustment modified the plants' response to soil water deficit and also whether grain sink demand for assimilates, varied by removal of 50% spikelets, affected osmotic adjustment. In each experiment, there were well-watered control and water stress treatments. In both experiments, the dawn osmotic potential in the High OA group was always lower than in the Low OA group under water limiting conditions, and the difference was significant after anthesis. The difference in osmotic potential was about 0.1 MPa in the field and up to 0.25 MPa in the glasshouse. In the glasshouse experiment, removal of 50% spikelets at anthesis significantly decreased osmotic potential during grain filling, suggesting that osmotic adjustment is influenced by the availability of assimilates in the leaves. Under well-watered conditions, the two groups behaved very similarly in terms of maximum leaf area, green leaf area retention during grain filling, total dry matter production, grain yield and grain number in both experiments. Under water-limiting conditions, the High OA group produced larger maximum leaf area and had better leaf retention during grain filling. Despite similar water use, total dry matter was also significantly higher in the High OA group though the difference was small. Grain number was also greater in this group in both experiments, whereas grain yield was significantly higher in the High OA group in the field, but not in the glasshouse where severe water stress developed more rapidly. It is concluded that the adverse effect of water stress can be reduced by adopting sorghum genotypes with high osmotic adjustment. However, selection for high osmotic adjustment needs to ensure that osmotic adjustment is not solely due to small head size.

Two sets containing large numbers (23 and 47 entries) of sorghum genotypes were grown in the glasshouse to examine the effect of osmotic adjustment on water extraction, dry matter growth and grain yield. Water stress was developed in two periods, one before and one after anthesis. The results were similar in the two experiments despite a large difference in the genetic background of the plant material. Since osmotic potential did not differ significantly among genotypes before water stress was induced, osmotic potential obtained under stress was used directly to indicate the genotype's ability to adjust osmotically. Osmotic adjustment was positively associated with green leaf area retention during grain filling and to root length density at 70 cm depth. Genotypes with high osmotic adjustment used more water during the second drying period. As a result, total dry matter was well related to osmotic adjustment during grain filling, but grain yield was negatively associated with osmotic adjustment in one experiment and not significantly related in the other. When comparison was made for lines which had similar leaf water potential during early stages of growth but which differed in osmotic adjustment during grain filling, there was still a positive effect of osmotic adjustment on total dry matter. This suggests that the positive effect was not caused by large plants extracting more water during early stages of growth, but was due to the difference in line's ability to extract water during grain filling.

Development of genotypic variation in osmotic adjustment was examined in two glasshouse experiments using two sets of sorghum material. In the first experiment, 47 S2 lines extracted from a randomly mated population were used, whereas in the other, inbred parents and their 15 hybrids were compared. In both experiments, water deficit was induced in two periods, one before anthesis and the other after anthesis for most genotypes. In both experiments osmotic potential at the beginning of the first drying period was similar among genotypes and therefore osmotic potential obtained under water deficit was used for the comparison of osmotic adjustment among genotypes. In the first drying period of both experiments, when stress was milder, about 40% of the variation in osmotic adjustment was accounted for by difference in leaf water potential. When the effect of water potential was removed by covariance analysis, there was significant genotypic variation in osmotic adjustment in the second experiment, but not in the first experiment. On the other hand, in the second drying period, when stress was more severe, the effect of leaf water potential on osmotic adjustment was small. There was significant genotypic variation in osmotic adjustment in both experiments after the water potential effect was removed by covariance analysis. Osmotic adjustment in the second drying period was also negatively correlated with grain sink/source ratio (number of grains/leaf area) in the first set of materials. The comparison of osmotic adjustment among hybrids and their parents showed that, in this particular set of genotypes, the female parents were more important than the male in determining osmotic adjustment of the hybrids. The genotypic variation was associated with performance under water deficit in the field. It is concluded that there is considerable genotypic variation in osmotic adjustment in the genetic material examined. Osmotic adjustment is, however, correlated with water potential and grain sink/source balance, and hence the selection for osmotic adjustment needs to ensure that high value is not due simply to low water potential or small head size.

No abstract