Abstract
Sprouting dates of potted mulberry trees were examined by applying high temperature treatment at various times of the year. Multiple regression analysis was also carried out using mulberry sprouting dates collected over 21 areas of Japan as independent variable, while local temperature every ten days as explanation variable. High temperature treatment in November, December and January corresponded with a slight delay in sprouting, while treatment in February and March corresponded with slight to extreme promotion of sprouting. Additionally, the partial regression coefficient which was positive through December turned negative in January, and then continuously declined through March. So, thermosensitivity through January was thought to be different from that after February, and this might be due to an insufficient dormancy awakening. Next, time-dependent models with five and seven parameters were constructed based on the above information. To determine the best model and to select optimum values for parameters, each model was examined by calculating the sprouting dates over Japan under various temporal values for parameters. The best model obtained here consisted of three reactions : one for dormancy awakening and two for sprouting. Three reactions proceed every day and sprouting occurs when the product of the last reaction accumulates to a threshold point. Temperature dependency of these reactions was determined as being similar to Arrhenius'equation. Making allowance of ±3 days, 87 to 92% of estimated dates accorded with actual sprouting dates under the determined optimum parameters.
