Many roads to success: different combinations of life-history traits provide accurate germination timing in seasonally dry environments

Germination timing is determined by several plant life-history traits. Seed dormancy regulates the time and place of early plant development and spreads recruitment risks over time. Dispersal phenology and syndrome can influence germination timing and buffer spatial heterogeneity. The ecological requirements for germination (the germination niche) can also influence when and where germination takes place. To date, the relative importance of each of these four traits to ensure the phenological adaptation of individual species in diverse communities remains unexplored. Here, we investigated the functional interactions among them and their relevance in heterogenous, seasonally dry environments. We collected seed dispersal phenology and syndrome for 82 species of the Brazilian savanna (cerrado) and evaluated the dormancy and germination behavior of the seeds of every taxon. Based on these data, we developed two new ecological indexes to estimate the likelihood of a non-dormant seed to germinate upon dispersal ( increment G) and the overall variability of germination through time (sigma T). We then evaluated the influence of each trait on germination timing within a phylogenetically controlled framework. Our results show that even though germination is concentrated at the beginning of the rainy season, seed dispersal takes place year-round. Non-dormant seeds released during the dry season were characterized by high increment G values that delayed their germination until the onset of the favorable season. Simultaneously, seed dormancy and spatial dispersal (i.e., the two risk-reduction mechanisms) were negatively correlated as dormancy and high sigma T values were only prevalent in seeds with reduced spatial dispersal ability. We conclude that the timing of seed germination is ultimately the net outcome of adaptive interactions among life-history traits, which can result in multiple functionally equivalent phenotypes. It is possible that this might contribute to community diversity by providing opportunities for the coexistence of different species.