Outside the envelope: rare events disrupt the relationshipbetween climate factors and species interactions.

The order in which species arrive during community assembly can be an important driver of community composition and function. However, the strength of these priority effects can be variable, in part because of strong site and year effects. To understand how priority effects vary in importance with abiotic conditions, we initiated identical community assembly experiments in which we varied the timing of arrival of native and exotic grass species in each of 4 yr across three grassland sites in northern California. This uniquely replicated experiment tested the power of priority to determine initial community structure in a restoration context across a natural range of conditions. There were large and significant differences in both total seeded cover and the strength of priority across sites and years of initiation, confirming the suspicion that most ecological experiments may lack spatial and temporal generality. On the other hand, much of the variation in strength of priority could be related to climate. Strikingly, however, the model fit across the three sites and the first 3 yr of the study (the first nine experiments) was radically altered when we included the fourth year, which was characterized by an unusual weather pattern with higher temporal variability in rainfall (a rainfall pattern predicted to increase with climate change). This year produced relatively low strength of priority, supporting the suggestion that highly variable climates may be associated with lower strength of priority effects. Experiments that examine community assembly over a range of naturally occurring abiotic conditions enhance our ability to predict when priority effects will be important, allowing us to explore shifting patterns of community assembly in the face of climate change and optimize restoration strategies based on environmental conditions.