PHYSIOLOGICAL-CHARACTERISTICS OF HIGH-YIELDING RICE INFERRED FROM CROSS-LOCATION EXPERIMENTS

Understanding the processes controlling yield in well-managed, high-in put crops may offer ways of increasing currently recognized yield pote ntials. In this study, 13 crops of Koshihikari rice were managed with different nitrogen strategies to achieve maximum yield in three enviro nments, Kyoto (6 t ha(-1)) and Ina (10 t ha(-1)) in Japan, and Yanco ( 13 t ha(-1)) in Australia. A common set of data on radiation intercept ion, growth, yield, yield components, N uptake and non-structural carb ohydrates was collected for each crop. There was a similar efficiency of dry matter production per unit of incident global radiation at the three sites, 0.78 g MJ(-1). At Yanco, incident and intercepted radiati on were high but were offset by a low conversion of intercepted radiat ion to dry matter (radiation-conversion efficiency, RCE) of 1.0 g MJ(- 1) compared to RCE = 1.4 at Kyoto and Ina. The RCE at Yanco was partic ularly low during the grain filling stage, apparently due to poor root function associated with low soil aeration and to low tiller survival and low minimum temperatures. If these constraints could be overcome and the RCE increased to the values in Kyoto and Ina, the estimated yi eld potential at Yanco would be about 18 t ha(-1). Given the relativel y high values of RCE at Kyoto and Ina, the most promising way to incre ase yield potential would be to increase the efficiency of converting growth to grain yield. At all three locations the number of grains was less than the number of spikelets, The quantity of assimilate availab le for grain production was estimated from the amount of non-structura l carbohydrate at heading plus the dry matter growth after heading. Th e potential sink for assimilate was estimated from the number of flore ts and the weight of a filled kernel. By comparing these amounts it ap peared that some of the crops at Ina were source limited and some were sink limited, while at Kyoto crops were neither clearly source Limite d nor sink limited. This paradoxical result is explained by a lack of assimilates soon after heading, leading to poor grain set, followed by a surplus of assimilates during late grain filling. On the basis of o ther evidence, there appears to be genetic variability for the ability to set grains in conditions of poor assimilate supply, suggesting tha t in the low-radiation environments in which much rice is grown, there may be scope for increasing yield by increasing the number of fertile florets.