Geographical gradients in seed mass in relation to climate

Aim To determine whether latitudinal and longitudinal gradients in seed mass are related to variation in climatic features including temperature, solar radiation and rainfall. Location Australia. Methods Seed mass was estimated from over 1600 provenances covering the latitudinal and longitudinal extents of 34 perennial Glycine taxa in Australia. Climatic data were obtained from ANUCLIM 5.1 for collection locations based on long-term meteorological records across Australia. These climatic data were subject to principal components analysis to extract three components as climatic indices. Generalized linear models were used in three separate sets of analyses to evaluate whether seed mass-latitude and seed mass-longitude relationships persisted after taking climatic variation into account. First, relationships were examined across species in analyses that did not explicitly consider phylogenetic relationships. Secondly, phylogenetic regressions were performed to examine patterns of correlated evolutionary change throughout the Glycine phylogeny. Within-species analysis was also performed to examine consistency across different taxonomic levels. Results Geographical variation in seed mass among species was related primarily to temperature and solar radiation, while rainfall was much less influential upon seed mass. Partialing out the influence of temperature and solar radiation in models resulted in the disappearance of significant seed mass-latitude and seed mass-longitude relationships. Patterns within species were generally consistent with patterns among species. However, in several species, factors additional to these climatic variables may contribute to the origin and maintenance of geographical gradients in seed mass, as significant seed mass-latitude and seed mass-longitude relationships remained after controlling for the influence of climatic variables. Main conclusions Our empirical results support the hypotheses that (1) seed mass is larger at low latitudes and in the interior of the Australian continent due to increased metabolic costs at high temperatures, and that (2) higher levels of solar radiation result in an increase in the availability of photosynthate, which in turn leads to an increase in biomass for the production of large seeds. In effect, our findings show that greater energy is available precisely where needed, that is, where high temperatures require large seed mass on the basis of metabolic requirements.