I studied the conditions controlling the timing of breeding migration of the Japanese mountain stream frog Rana sakuraii, an explosive breeder, over 9 years (1992, 1993 and 1999-2005). I analysed two factors: the daily cumulative temperature (DCT) during hibernation and the triggering temperature on the day of migration onset. Frogs hibernated in shallow running water in streams in December. A total of 53,155 breeding migrating adults were captured by traps. Every year, breeding migration was induced by a rise in daily maximum water temperature to about 5 degrees C. However, its date was limited to 1 February at the earliest, and the onset needed a DCT (from 20 January, using 3 degrees C as the threshold for daily effective temperature) of at least about 15 degrees C. Earlier (e.g. in mid- or late January), even if the maximum temperature rose to 5-8 degrees C, migration did not begin. Moreover, even in early February, if the maximum temperature rose to 5 degrees C, if it had been too cold in January and the DCT was low, migration would not begin until mid- or late February. Thus, the earliest date of readiness for migration varied from 1 February to mid-February, depending on the winter DCT. Logistic regression analyses showed that both factors, the DCT and the daily temperature, were significant. I propose that in temperate-zone amphibian explosive breeders, both the passing of an essential number of days and an essential DCT during hibernation are prerequisites for the onset of breeding migration before the daily temperature rises to the threshold.

Studies on amphibian autumn migrations to hibernation sites are few, and those investigating the conditions controlling the timing of migrations are even fewer. I investigated an explosive-breeding frog, Rana sakuraii, in Japanese mountain headwaters in 6 different years, focusing on the start and end of autumn migrations. Autumn migrations began in late October. Initially, frogs moved into the streams from terrestrial sites around the stream sources, and thereafter moved downstream entirely underwater. Start days were always the first rainy days after the minimum air temperature had dropped to about 6°C or lower (minor start days) or about 4°C or lower (major start days) regardless of the air temperature on the start days. In early or mid-December, in-stream migratory movements ended when the maximum water temperature dropped below the 5°C threshold (the same threshold temperature initiating the winter breeding migration). I hypothesize that, in many temperate-zone amphibians, the timing of autumn migration to hibernation sites is basically dependent on three common conditions: (1) the experience of a prerequisite minimum air temperature (PMAT) before the start day; (2) the decrease in day length to a threshold value; (3) the first rainy (or high humidity) day after (1). Further, I hypothesize that the threshold ambient temperature inducing the end of the autumn migration is almost the same as that inducing emergence from hibernation, that is, the start of the winter breeding migration for explosive breeders, and that this is the daily maximum, and not the daily minimum.