Recruitment of new individuals from seed is a critical component of plant community assembly and reassembly, especially in the context of ecosystem disturbance and recovery. While frameworks typically aim to predict how communities will be filtered on the basis of traits influencing established plant responses to the environment, assembly from seed is more complex: the responses of seeds (affected by dormancy and germination function) and establishing plants (affected by root and leaf function) can both influence outcomes within a single growing season. This creates a potential role for a more diverse set seed and seedling traits, and for environmental variability on shorter timescales (e.g. seasonal versus annual dynamics), than are typically considered. We followed thousands of individual seeds comprising eleven herbaceous grassland species through the first growing season, seeking to uncover critical environmental (precipitation amount and timing) and trait‐based filters on seedling emergence and survival in assembling communities. We saw the biggest recruitment limitation when seeds failed to emerge, driven independently by a dry spring and interspecific variation in seed mass (positive effect) and seed dormancy (negative effect). Seedling survival rates were higher than emergence, with weaker predictive roles for traits like seedling root mass allocation (positive effect) and seed mass (positive under spring drought), and lesser impacts of summer rainfall on soil moisture and survival. Interestingly, most trait relationships were not conditional on rainfall, suggesting water‐independent mechanisms of their respective advantages. Although recruitment is a complex process, our findings suggest that trait‐based assembly frameworks can be a useful way to anticipate outcomes, particularly if dynamic early‐stage conditions (e.g. spring rainfall) and attributes (e.g. seed dormancy) receive greater attention. Given the importance of recruitment for community turnover in the context of global change and land management efforts, this is an area ripe for continued expansion in trait‐based and applied ecology.

No abstract