T. grandiflorum is self-incompatible in these populations, and relies entirely on insect pollination, primarily by bumblebees, for its reproductive success. We found that the degree of pollen limitation varied across populations, and was negatively correlated with population density (Knight 2003b). Although both herbivory and pollen limitation had strong effects on individual fitness components, only herbivory affected the population dynamics of T. grandiflorum, because it affected vital rates with high sensitivity values (Knight 2004). This study was the first to provide empirical evidence that elasticities of demographic transitions shift with an ecological factor.

We combined our demographic data with Mark Vellend's (U. British Columbia) data on deer-facilitated dispersal of T. grandiflorum, and we have found that deer herbivory at low frequencies can facilitate Trillium population expansion and explain post-glaciation movement of Trillium (Vellend et al. 2006).

In T. grandiflorum, we found that deer preferentially feed on individual T. grandiflorum that flower earlier in the growing season. With Bob Holt and Mike Barfield (U. Florida), we combined demographic matrix models with selection analyses to determine whether evolution of flowering time could occur fast enough to prevent extinction. We found that large T. grandiflorum populations might be able to rescue themselves from deer-driven extinction through the evolution of flowering time. This is because flowering time has a large affect on fitness of plants by reducing the incidence of herbivory, and because these fitness components contribute to vital rates that the population dynamics are sensitive to (Knight et al., in prep).