al., 1994), the USA (Board et al., 1980), and Korea, Bangladesh, India, Nepal, and other countries (Kaneda Accurate prediction of spikelet sterility in rice (Oryza sativa L. ) and Beachell, 1974). Improved cold tolerance has been is a prerequisite for accurately predicting grain yield in cool climates, since severe yield losses frequently occur when spikelet sterility is a key goal of rice breeding programs, and management induced by cool temperatures during reproductive growth. Both cool practices such as changing the planting schedule, water air temperature (T a ) and cool water temperature (T w ) are detrimental management, and fertilization have been widely introfactors that can cause spikelet sterility, but a large discrepancy between duced to farmers in an effort to prevent yield losses T a and T w is often observed in paddy fields. The depth of water can (reviewed by Wada, 1992). However, there have been also affect spikelet sterility. We proposed a model that accounted for few attempts to evaluate the impact of these improvethe effects of T a , T w , and water depth on spikelet sterility, and was ments on yield and its stability (Yajima, 1994). For that based on panicle temperature (T p ), then tested the model using 23 purpose, a process-based growth model would be a powindependent sets of field data from northern Japan. We also quantified erful tool that could quantitatively and separately anathe role of daily amplitude (the difference between maximum and lyze the causal effects of weather and management facminimum temperatures) and differences in plant sensitivity to temperature in determining spikelet sterility. A cool-irrigation experiment tors, and that could also be used to predict spikelet revealed that spikelet sterility depended more strongly on T p than on sterility and yield losses as a result of cold weather.T w or T a . We also developed six models using "cooling degree-day" Spikelet sterility of rice grown under flooding is deterconcept. The model based on T p had higher accuracy than models mined by the temperatures of both the irrigation water based solely on T w or T a . In addition, average temperature was a (T w ) and the air (T a ), but the plant organ that is most better predictor than minimum temperature. Accounting for the difsensitive to low temperature and that determines spikeference in temperature sensitivity also improved the model's accuracy.let sterility is the panicle itself (Sakai, 1949; Satake et A model that considers these factors would thus improve prediction al., 1988). After panicles have differentiated at the base accuracy for spikelet sterility due to cool weather.
