A long-standing goal in ecology is to understand how abiotic and biotic factors interact to influence individual performance. One well-known outcome of this effort is the wealth of studies demonstrating that plant-plant interactions change from competitive to facilitative along gradients of increasing physical stress, i.e. the pattern articulated by the stress gradient hypothesis. This research has revealed that a single individual can experience both facilitation and competition over its lifetime, and that much of the variation in this facilitation-competition balance comes from physical stress and the sizes or ages of plants during the interaction. Germination phenology determines both physical stress and size-structured interactions, but to my knowledge it has not been explicitly considered as a mediator of the facilitation-competition balance. I performed a field experiment over two years using a naturalized population of Arabidopsis thaliana(Brassicaceae) to test whether an individual's germination phenology alters the effects of the local plant community on its survival, fecundity, and total fitness. Specifically, I tested the hypothesis that germinating during periods of stress leads to facilitation from neighbors, whereas germinating under conditions that are favorable for growth and reproduction leads to competition from neighbors.
Overall, neighbors increased focal plant survival early in ontogeny, particularly in the most stressed cohort in the more stressful year. However, neighbors reduced the fecundity of adults, especially in the cohorts that otherwise would have been the most fecund. In each year, neighbors affected total fitness in the later-emerging cohort, but the direction of this effect depend on the year: facilitation was observed in the more stressful year while competition predominated in the less stressful year. These effects of neighbors acted to relax selection on germination time imposed by the physical environment, essentially expanding the temporal dimension of the germination niche. Notably, the net effect of neighbors on the total fitness of focal plants was neutral when phenological variation was not considered. Therefore, accounting for temporal stress gradients created by phenological, ontogenetic, and seasonal environmental variation reveals fitness and demographic effects of neighbors that may cancel each other. Further, these results suggest that variation in the magnitude and sign of neighbor effects observed in natural populations and communities may be explained, in part, by unmeasured temporal stress gradients.