Aim To enhance our understanding of the evolutionary interactions between seed-dispersal syndromes, life-forms, seed size, and habitat characteristics by studying their association with the regional-scale distributions of subtropical rain-forest plants in the context of climatic gradients.
Location South-east Queensland, subtropical eastern Australia (152° E, 26° S).
Methods We classified 250 rain-forest sites into six floristic site-groups based on their woody plant composition. The resulting classification was strongly associated with variation in rainfall. The distribution of species across the floristic site-groups was used to assign 568 species to seven habitat classes (one class for ‘widespread’ species, with all other species classified according to the site-group within which they were most frequent). Species were also classified for three other categorical life-history factors: three dispersal syndromes based on diaspore morphology (fleshy, wind-assisted, and unadorned); four life-forms (trees, shrubs and small trees, tall climbers, and short and shrubby climbers); and four seed-diameter classes (< 3 mm, ≥ 3 and < 4.5 mm, ≥ 4.5 and < 7 mm, and ≥ 7 mm). We used a basic comparative approach augmented by simple phylogenetically constrained comparisons to assess association between dispersal syndrome, seed size, life-form, and habitat class.
Results Across the rain forests of south-east Queensland, the proportion of species with fleshy diaspores or of large stature increases with rainfall. High-rainfall sites also have larger average seed sizes, but the difference in average seed size between high- and low-rainfall sites is small compared with variation within sites. Among species, those with fleshy fruit tend to have larger seeds and to favour high-rainfall sites. Very few small trees produce diaspores adapted for wind-assisted dispersal. On average, species with unadorned diaspores have smaller seeds than those with fleshy diaspores. However, within sites, species with unadorned and fleshy diaspores have similar average seed sizes, and some species with unadorned diaspores from high-rainfall habitats have extremely large seeds.
Main conclusions Commonly observed associations between fleshy fruit, larger plants, larger seeds, and productive habitats are apparent within the rain-forest flora of south-east Queensland. However, these associations are generally weak and involve complex interactions. For example, the strong tendency for species with fleshy fruit to have larger seeds than those with unadorned diaspores concealed a significant group of species from wetter forests that produce extremely large seeds and unadorned diaspores. The most widespread species in this study tend to be large plants (particularly robust lianes) and to produce fleshy fruit, but they tend not to have relatively large seeds. The association between large seeds, large plants, fleshy fruit and productive habitats is discussed as part of an evolutionary strategy favouring fitness in populations close to carrying capacity. We review some problems with focusing on establishment chances per seed as the driver towards association between large seeds, large plants and productive rain-forest habitats (the difficult-establishment hypothesis). Instead we suggest that production of large, short-lived seeds by long-lived plants in temporally stable, closed habitats may reflect the limited evolutionary potential for strategies enhancing colonization (e.g. producing large numbers of dormant seeds), thus allowing the establishment benefits of large seeds greater selective influence (the slow-replacement hypothesis). The association of fleshy fruit with large seeds probably reflects the difficulty of dispersing large seeds by other means (the difficult-dispersal hypothesis).