Seed germination rates (GR =inverse of time to germination) are sensitive to genetic, environmental, and physiological factors. We have compared the GR of tomato (Lycopersicon esculentum Mill.) seeds of cultivar T5 to those of rapidly germinating L. esculentum genotypes PI 341988 and PI 120256 over a range of water potential (ψ). The influence of seed priming treatments and removal of the endosperm/testa cap enclosing the radicle tip on germination at reduced ψ were also assessed. Germination time-courses at different ψ's were analysed according to a model that identified a base, or minimum, ψ allowing germination of a specific percentage (g) of the seed population (ψb(g)), and a ‘hydrotime constant’ (θH) indicating the rate of progress toward germination per MPa.h. The distribution of ψb(g) determined by probit analysis was characterized by a mean base ψ (b) and the standard deviation in ψb among seeds (σψb). The three derived parameters, b, σψb) and θH, were sufficient to predict the time-courses of germination of intact seeds at any ψ. A normalized time-scale for comparing germination responses to reduced ψ is introduced. The time to germination at any ψ(tg(ψ)) can be normalized to be equivalent to that observed in water (tg(0)) according to the equation tg(0)=[l−(ψ/ψb(g))]tg(ψ). PI 341988 seeds were more tolerant of reduced ψ and had a more rapid GR than T5 seeds due to both a lower b and a smaller θH. The rapid germination of PI 120256, on the other hand, could be attributed entirely to a smaller θH. Seed priming (6 d in −1.2 MPa polyethylene glycol 8000 solution at 20 °C followed by drying) increased GR at all ψ>ψb(g), but did not lower the minimum ψ allowing germination; i.e. priming reduced θH without lowering ψb. Removing the endosperm/testa cap (cut seeds) markedly increased GR and lowered the mean ψ required to inhibit germination by 0.7 to 0.9 MPa. However, this resulted primarily from downward adjustment in b during the incubation of cut seeds at low ψ in the test solutions. The difference in b between intact and cut seeds incubated at high ψ was much less (0.l MPa), indicating that at the time of radicle protrusion, the endosperm had weakened to the point where it constituted only a small mechanical barrier. In the intact seed, endosperm weakening and the downward adjustment in embryo b ceased at ψ < −0.6 MPa, while the reduction in θH associated with priming proceeded down to at least −1.2 MPa. Based on these data and on the pressure required to push the embryos from the seeds at various times after imbibition, it appears that the primary effect of priming was to shorten the time required for final endosperm weakening to occur. However, as priming increased GR even in cut seeds, priming effects on the embryo may control the rate of endosperm weakening.