Evolution of recruitment features in plants: a comparative study of species in the Rosaceae

Theory predicts that life history traits that reduce the impact of environm ental variation show patterns of negative covariation (trade-offs), In plan ts, seed size, seed dormancy and dispersal (in space) interact to reduce ri sk in a temporally and spatially variable environment. Dispersal in time an d space permits escape from unfavourable conditions, whereas a large seed s ize may improve establishment under unfavourable conditions. However, large seeds may suffer a high rate of loss due to seed predators. The objective of this study was to examine relationships between seed size, seedling esta blishment and seed dormancy by combining data from field experiments with i nformation of phylogenetic relationships among I I species in the subfamily Rosoideae (Rosaceae). The predicted relationship between seed size and see d predation was also examined. Emergence from seed, survival and recruitmen t of seedlings, effects of seed predators (post-dispersal predation), and s eed dormancy were studied experimentally in the field, All species generate d seedlings in the experimental plots. Overall, the emergence of seedlings was promoted by disturbance. Several species possessed seeds that were able to germinate after being buried in the soil column for three years. Despit e a small data set (ten phylogenetically independent contrasts), the compar ative analysis confirmed an expected positive effect of seed size on establ ishment ability. The emergence of seedlings increased with seed size. The r esults did not demonstrate the hypothesised positive association between se ed size and predation risk, or negative association between seed size and s eed dormancy. However, the contrast analysis showed a significant negative relationship between seed dormancy and survivorship of seedlings, Le, betwe en quantitative data of dormancy and establishment capacity of the species obtained under natural conditions in the field. When not controlling for ph ylogeny (without contrasts), the only significant effect was a negative ass ociation between seed size and seed dormancy.