Seed dispersal syndromes are simple and widely applied functional classes that may allow generalizations about plant
species' seed dispersal. Seed dispersal syndromes beg the question; how are species with the same seed dispersal syndrome similar, and how are they different from species with different seed dispersal syndromes (other than with regard to fruit morphology)? In other words, what is the significance of seed dispersal syndromes for plant ecology? This issue was considered for woody species in the vine-forests of south-east Queensland at both regional and local scales. Regional scale analyses assessed association of seed dispersal syndromes with habitat, seed size, life form and regional abundance within habitats. Analyses at the local scale focused on the importance of seed dispersal for recruitment.
Regional scale analyses were based on floristic data collated for 250 rainforest stands spread across the
south-east Queensland bioregion. These ranged in structure from semi-evergreen vine thicket to complex notophyll vine forest. Floristic classification of these 250 sites identified six broad floristic groups, incorporating 15 floristic subgroups. The primary axis of floristic ordination was strongly correlated with rainfall and the secondary axis was most strongly correlated with annual mean temperature. Species distributions across the six broad floristic groups were translated into seven species habitat classes, including a 'wide' habitat class, for species prevalent in most broad floristic groups.
Seed dispersal syndromes were differentiated for 566 species based on fruit morphology. Three primary syndromes were: vertebrate dispersal, characterised by animal enticing diaspores incorporating resources
for vertebrate consumers other than those within the seed (69% of species); wind dispersal, characterised by winged or plumed diaspores (14% of species); and, unassisted dispersal, characterised by unadorned diaspores lacking features promoting dispersal by animals or wind (13% of species). The frequency of vertebrate dispersal among species in sites averaged over 80% in the wetter broad floristic groups and declined to an average of 71% in the drier inland vine-forests. Lower vertebrate dispersal in drier forests was matched by increased frequency of species with unassisted dispersal. Average seed size in sites and the frequency of larger life forms were also positively correlated with rainfall.
Complex associations between dispersal syndrome, life form, seed size and habitat were apparent across
species. Widespread species were over-represented among vertebrate-dispersed species, tended to have small to intermediate sized seeds, and were underrepresented among shrubs and small trees. The vertebrate dispersal syndrome was overrepresented among species of wetter habitat groups. Shrubs and small trees were overrepresented among vertebrate-dispersed species and strongly under-represented among wind-dispersed species. Large seeds were over-represented among vertebrate-dispersed species, larger life forms and wetter habitat groups. Overall, patterns in the frequency of character classes across species paralleled patterns in their frequency across sites in floristic groups.
The regional abundance of species was measured as their incidence within each of the six broad floristic groups of sites. This
controlled some of the effects of habitat specialisation on regional abundance and allowed patterns to be compared across different forest types. Species regional abundance was much more variable within taxonomic groups than any of the other species variables. This variability suggests that any morphologically based characters, closely related to taxonomy, will have limited predictive value with respect to regional abundance. Vertebrate-dispersed species had greater regional abundance than unassisted species, but not wind-dispersed species. Species with characters associated with a habitat class in cross species analyses were generally also more frequently recorded in the corresponding broad floristic group of sites. The seed size of species with the greatest incidence declined from wet to dry broad floristic groups. Dispersal syndromes had a significant effect on incidence in wetter forests, but no effect in drier forests.
The importance of seed dispersal to recruitment at local scales was considered by examining the distribution of trees and shrubs more than 50cm tall in several one-hectare plots. Spatial patterns were qualitatively similar in four plots in structurally different forests. Most species were strongly clumped as saplings and tended toward random spatial distribution in larger size classes. Small saplings tended to be spatially associated with clumps of larger saplings. Sapling clumps were often randomly distributed with respect to larger conspecifics and several species in each plot had saplings repelled from large conspecifics. These patterns indicate that spatial heterogeneity in recruitment probability across the forests was temporally stable and poorly or negatively spatially-correlated with large conspecific plants, suggesting seed dispersal had
considerable importance for recruitment at local scales.
Overall these analyses suggest that dispersal syndromes are related to other species characters, species' habitat associations and regional abundance, and that seed dispersal is important for species' recruitment success at local scales. Although these associations imply considerable ecological significance for dispersal syndromes they were also both complex and weak. These patterns have considerable significance for the evolutionary ecology of seed dispersal.