Evolutionary comparative methods taking into account the potential effects of relatedness reveal that, among 202 species of animal-dispersed tropical woody angiosperms from Peru, large seeds may be an adaptation to: (1) dispersal by mammals rather than by birds and (2) greater plant height. Using the most powerful techniques currently available, appropriate comparisons were made between related species or species groups at all taxonomic levels, thus allowing use of all species in the data set. Even so, the results show no evidence of adaptation (i.e., a functional relationship) of seed mass to successional syndrome or to growth form (trees or vines). In previous cross-species analyses, seed mass has been linked to both of these ecological variables. The lack of relationship between seed mass and habitat use or growth form may be due to targeting the wrong trait to predict these two variables, or because seed mass is subject to exaptation, i.e., a particular seed mass has arisen for other reasons, but allows invasion into a habitat and currently serves to maintain that species in the habitat. Differentiation between these two explanations for the lack of correlation between seed mass and successional syndrome or growth form would entail species-by-species examination of the function of seed mass under natural conditions. The effect of relatedness on the ecological correlates of seed mass and plant height is determined for each of three taxonomic schemes. It is concluded that the three schemes do not differ greatly in the results obtained.
Keywords: Comparative ecology; Seed size; Seedling establishment conditions; Taxonomic relatedness; Tropical trees.