Serial germination tests were carried out on dormant seeds of six rice varieties (four varieties of Oryza sativa L. and two varieties of O. glaberrima Steud.) stored at several different constant temperatures within the range 27° C to 57° C. Probit analyses of the results were carried out to determme the mean dormancy period for each variety at each temperature. Regression lines fitted to these data showed that there is a direct negative relationship between storage temperature and log mean dormancy period over the range 27° C to 47° C, thus confirming a previous result obtained on a single variety. At 7° C there were indications of a slight departure from this relationship in that the mean dormancy periods at this temperature were slightly longer than would have been predicted by extrapolation of the regressions calculated from the results obtained at lower temperatures. In all cases where the results were unambiguous (i.e. in all the sativa varieties and one of the glaberrima varieties) a constant Q10 of 3.13 was shown for the rate of loss of dormancy over the range of storage temperatures from 27° C to 47° C. In the remaining glaberrima variety, where the results were less reliable, a Q10 of 2.54 was found. Germination tests on all varieties were carried out at 32° C, but in the case of one sativa variety germination tests for all storage treatments were also duplicated at 27° C. This investigation showed that, in contrast to the effect of storage temperature, the higher temperature during the germination test consistently resulted in a lower percentage germination. In addition the results demonstrated that there is no interaction between storage temperature and germination temperature: consequently the storage-temperature coefficient has the same value irrespective of germination temperature. Some theoretical implications of the results are discussed.