Colonization, tolerance, competition and seed-size variation within functional groups David A. Coomes and Peter J. Grubb Department of Plant Sciences, University of Cambridge, Downing Street, Cambridge, CB2 3EA, UK Ecologists interested in seed size have generally con- trasted functional groups of plants but, recently, some have focussed on explaining the large range of seed size found within a functional group. A potentially important theoretical advance was the idea that seed number –seedling survival tradeoffs could explain the coexistence of species, in particular colonization – competition tradeoffs where smaller-seeded species are superior colonizers and larger-seeded species are superior competitors. However, recent models have placed limits on the potential of this approach, chiefly by showing that the asymmetry of competition must be strong. Also, although there is evidence from studies within functional groups that seed size does trade off against number of seeds and dispersal of those seeds, and that seed size is correlated with competitive ability among seedlings and tolerance of hazards during establishment, the available evidence suggests that SNSS tradeoffs do not make possible long-term coexistence without other forms of niche differentiation. Ever since E.J. Salisbury published his seminal work of 1942 [1], the significance of seed size has fascinated ecologists [2–4]. Most ecologists have contrasted func- tional groups, such as shade-tolerators and light- demanders or bird- and mammal-dispersed species, but some ecologists have begun to focus on explaining the range of seed size found within functional groups, such as particular groups of annuals, or of rainforest trees [5–13]. The range of mean seed size among species within a functional group is commonly much greater than the difference between the means of contrasted groups. For example, in tropical lowland rainforests, the mean seed dry mass of shade-tolerant tree species is 10–100 times greater than that of light-demanding tree species [14], whilst the smallest and largest mean values within both these groups differ by 10 5 –10 6 . At the most basic level, seed mass is considered to be pivotal to the fitness of a plant because of its influence on two key components of life history [4]. First, the seed mass of a plant is negatively correlated with the number of seeds that it can produce. Consequently, smaller-seeded species are considered to be superior colonizers. Second, seed mass is positively associated with seedling survival, because larger seeds generally develop into larger seedlings, which are potentially better able to withstand either lack of resources (light or nutrients) or the various hazards that face them (dry spells, partial damage, etc). Therefore, smaller- and larger-seeded species differ in their life- history strategies as a result of a seed number versus seedling survival tradeoff. A simple mathematical model that incorporates the seed number–seedling survival SNSS tradeoffs predicts that, under the forces of natural selection, a single seed size should evolve in a given habitat. However, when appropriate biological details are included in mathemat- ical models, they start to reproduce something of the patterns observed in real communities. Specifically, the performance of a seedling is not simply correlated with seed size, but varies in response to environmental hetero- geneity, the effects of established vegetation, and compe- tition with surrounding seedlings. Also, the ability of a species to reach suitable regeneration sites does not depend only on how many seeds it produces, but also on how well the seeds are dispersed. We start by considering these issues in turn, seeing how their inclusion influences the predictions of models, and illustrating our major points with recently published examples. Then we consider for a few contrasted functional groups, whether the available evidence supports the idea that SNSS tradeoffs occur, and whether they maintain coexistence of species. Seedling responses to environmental heterogeneity The strongest case for an advantage of greater seedling size is the tolerance of hazards faced by young seedlings, reviewed most recently by Leishman et al. [4]. Large seeds usually produce large seedlings, which produce deeper roots and are more likely to survive dry spells; they are also likely to be better anchored in the face of frost-heaving or disturbance by animals. They have a better chance of pushing down through or up through a layer of litter and of surviving partial destruction by herbivores or falling plant parts. The greater tolerance of larger-seeded species would not be enough to promote coexistence in homogeneous environments (Box 1). However, in spatially and tem- porally varying environments, mixtures of species can persist provided that each species is fittest under some circumstances, and each species can disperse propagules to sites where conditions are optimal (Box 1). Corresponding author: David A. Coomes (david.coomes@plantsci.cam.ac.uk). Review TRENDS in Ecology and Evolution Vol.18 No.6 June 2003 283 http://tree.trends.com 0169-5347/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0169-5347(03)00072-7