I used population dynamics models fitted to 16 years of annual census data from a grass-dominated herbaceous community to understand how co-occurring perennials may partition resources in time and thereby contribute to their coexistence. Temporal variation has almost always been incorporated in such models as stochastic variation in maximum population growth rate (R or its counterpart) or as residual variation. I took a different approach: incorporating temporal variation as variation in equilibrium population size (K or its counterpart). Allowing K to vary in time provided better fits to the data (lower AIC). And by doing so, I was able to estimate equilibrium population sizes in each time interval and compare patterns among species.

Results/Conclusions

There were congruences in temporal variation in equilibrium population size among some species and substantial differences among others. These commonalities and differences among species were not always evident in comparisons of observed population sizes. The differences among species included differences in the years in which each species experienced density-independent (geometric) growth. These results support the hypothesis that temporal niche separation is important in structuring this community, despite the species’ obvious synchronous responses to rainfall typical of arid grasslands.