AbstractGermination in 35 species from 15 legume genera of southeastern Australia was promoted by a heat treatment which broke the seed coatcaused dormancy. Once the critical temperature was reached, most seeds had their dormancy broken, independent of the duration of heating. Species fell into three classes according to whether their dormancy was broken by a temperature of 40, 60 or 80°C.Highest germination in all species was achieved by heating in the temperature range 80–100°C, although long durations (120 min) at 100°C caused seed death in several species. At 120°C, seeds of most species were killed at all but one minute's duration. A proportion of seeds from 7 species (Acacia myrtifolia, Pultenaea daphnoides, P. incurvata, P. linophylla, P. polifolia, Dillwynia floribundaandSphaerolobium vimineurn) was not killed at 120°C and had their dormancy broken. This proportion varied markedly and resultant germination levels were significantly less than those at 80 and 100°C, except inS. vimineum.Between‐site variations in the 4 species tested (A. myrtifolia, A. suaveolens, A. terminalisandA. ulicifolia) were small. These variations concerned: (i) the minimum temperature required to break seed dormancy in 2 species: 60°C in one population ofA. myrtifoliaandA. suaveolens, and 80°C in the other; and (ii) the intensity of the germination response.Duration of heating was less important than temperature as a determinant of germination. Ordination techniques revealed that results from one duration across temperatures were comparable with data from multiple durations. This has significant applications in studying rare species, where seed may be in short supply.Predicted germination levels after a moderate intensity fire should far exceed those after a low intensity fire. Little germination was predicted for many species after a low intensity fire and for one species,A. elongata, no germination was predicted. The potential role of indicator species in relation to the maintenance of species in a community i