Pyrenophora semeniperda can infect nondormant Bromus tectorum seeds under optimal germination conditions, but most escape mortality. This reduces pathogen fitness relative to infection of dormant seeds, which are almost always killed. However, field experiments showed that a large proportion of seeds killed following inoculum addition were not accounted for as dormant seeds, but instead were likely nondormant seeds that would have germinated without inoculum addition. We hypothesized that widely fluctuating water availability to seeds would favor pathogenesis by delaying germination and allowing disease progression at water potentials below those that permit radicle emergence. To test this, nondormant host seeds were inoculated, imbibed for 8 or 24 h, subjected to controlled dehydration for 1 e21 d, rehydrated, and scored for mortality. With dehydration at -4 MPa, mortality increased with dehydration duration after short or long imbibition. At -40MPa, mortality increased with dehydration duration only after long imbibition. At -150MPa, there was no effect of dehydration duration on generally low mortality. These results illustrate that fluctuating moisture can cause high nondormant seed mortality, explaining how this pathogen kills nondormant seeds.