Low temperature-induced sterility in rice: Evidence for the effects of temperature before panicle initiation

Spikelet sterility of rice (Oryza sativa L.) results from low temperatures during panicle development. However, this temperature alone cannot fully explain the fluctuations in sterility observed in the field, since the susceptibility of rice plants to low temperature often changes according to its physiological status during sensitive stages. In the present study, we examined whether temperatures before the panicle initiation stage (PI) influence the plant's susceptibility to sterility during panicle development. To test this, we conducted a 2-year pot study and an analysis of field data using a model of cooling degree-days (CDD). In the pot study, the air temperature (T_a) and water temperature (T_w) were controlled independently for 3 weeks during the vegetative growth stage before PI. After PI, the plants were submerged in a cool water bath at a critical temperature of 19.5 °C to induce floral sterility. We found that low T_w during vegetative growth significantly increased the sterility. Low T_a during vegetative growth also significantly increased the sterility, but this effect was diminished by warm T_w even at low T_a. There was a close and negative correlation between sterility and T_w during vegetative growth. In the analysis of field data, we introduced CDD using temperatures below a threshold level of 20 °C to represent the magnitude of the exposure to low temperature from PI to the heading stage. Data of T_a was used for this analysis because data of T_w was scare. The CDD model was applied to 77 independent data sets collected at nine Agricultural Research Centers during four typical cool summers (1980, 1988, 1993, and 2003) in northern Japan. Year-to-year variations in sterility at one site were roughly accounted for by the variations in CDD, but large deviations were observed among the years. The deviations were related to T_a averaged over the 30-day period before PI. For a similar level of CDD, the lower the T_a before PI, the greater the sterility. Similar deviations were observed in the between-site relationships between sterility and CDD, and these deviations were related with the T_a before PI. These results suggest that temperatures before PI, and especially T_w, change the susceptibility of a rice plant to low temperatures during panicle development.