Evolutionary transitions in seed dormancy in the Legume family: Consequences for global scale plant distributions

Background/Question/Methods

Convergence in form and function among lineages in similar environments is expected where traits are adaptive in those environments. Such adaptations can explain global-scale patterns in trait distributions. While much effort has been devoted to understanding large-scale patterns of leaf and phenological traits, less attention has recently focused on seed traits, particularly seed dormancy. We developed a simple plant model examining the probability of seed dormancy in perennials with decreasing length of the favorable period. The model assumes relationships between seed size, seed number and predation risk; seed predation risk and seedling mortality depend on the length of the favorable period. We explored the verisimilitude of the model with a large dataset of seed and climate data. Dormancy class and seed size were accessed for > 500 taxa in the Legume family along with climate factors, including growing season length, across species distributions. We tested for evidence of evolutionary transitions within the Legume family from non-dormancy to dormancy associated with climatic transitions from short to long favorable-season length.

Results/Conclusions

The model predicts very generally that the probability of seed dormancy should increase with the length of the unfavorable season and decrease with increasing seed size. Therefore, small dormant seeds should be selected for in climates with greater seasonality in precipitation and temperature, and large non-dormant seeds should be selected for in climates with low seasonality. Analyses of the dataset largely confirm these predicitions. Within the legumes, seed dormancy is significantly associated with seasonal environments and non-dormant seeds are significantly larger than dormant ones. Moreover, different dormancy classes seem to be associated with different climates, with physical dormancy more prevalent in arid environments that have pronounced variations in temperature and physiological dormancy more frequent in temperate, seasonally wet environments. These results implicate seasonality and environmental components as important drivers in the evolution of seed dormancy with opposing selection forces for large seeds vs seed dormancy. Seed evolution appears to respond to environmental factors in a nuanced manner, resulting in very specific matches between seed attributes and climate. This highlights the influence of seeds on plant ranges and has potentially important consequences for understanding plant distribution at a global scale.