Science Collected Works - Theses
Permanent URI for this collection
Search Results
1 - 1 of 1
-
ItemSeed predation by insects in sclerophyllus vegetation at Wilson's Promontory, Victoria, with particular reference to ants(University of Melbourne, 1985)Seed-eating insects, mammals and birds can destroy large proportions of seed crops, both before and after dispersal, but their impact on seedling recruitment is poorly understood. This thesis examines seed predation by insects in sclerophyllous vegetation at Wilson's Promontory in southeastern Australia, a region in which insects appear to be by far the most important seed predators, and where studies of seed predation have been few. Pre-dispersal seed predation was investigated in four species of Leptospermum, three species of Eucalyptus and four species of Casuarina at many sites over several seasons. The major insect predators were the larvae of anobiid beetles (Dryophilodes spp.), chalcidoid wasps (mostly species of Megastigmus, Eurytoma and Bootanelleus) and unidentified cosmopterigid moths. In most cases plants suffered heavy seed losses, with the exclusion of insects from developing fruit resulting in increases from two- to five-fold in the number of viable seeds per fruit. In all cases insects caused far more damage than was obviously attributable to then, showing that estimates of predation intensity based only on observations of fruits are highly conservative. Seed-eating ants are by far the most important post-dispersal seed predators at Wilson's Promontory, and they were studied in detail. Ant communities from many sites were investigated to elucidate factors determining the composition and abundance of seed-eating species. All seed-eating ants were unspecialized omnivores, and their distributions were therefore controlled by factors influencing ants in general, rather than by seed availability in particular. The most abundant seed-eating ants were species of Rnytidoponera, Pheidole and Chelaner (kiliani group) in dry heaths and woodlands, and species of Prolasius and Chelaner (sculpturatus and leae groups) in cooler and wetter forests. Their overall abundance and therefore rates of seed removal were highest in woodlands, where total ant abundance was greatest. An artificial baiting technique was developed which provided reasonably accurate estimates of the rates of seed removal by ants. The proportions of seeds removed from baits ranged from 0 to 100% over ten days, and varied according to species, site and season. Studies of the spatial patterns of seed removal found that it was extremely patchy at many levels of scale. The fate of any given seed was determined by a hierarchy of factors operating at increasingly finer levels of scale. On a broad scale, overall rates of seed removal differed between sites because of differences in the composition and abundance of seed-eating ants, which were controlled by properties of the habitat. Seed removal was also patchy on a much smaller scale, so that even when overall rates of seed removal were very high, there were places where the probability of seed removal was consistently low. This small-scale patchiness was related to the foraging behaviour of ants. An experiment was conducted to test the effect of seed-eating ants on seedling densities. The elimination of ants by insecticidal treatment increased by 15-fold the seedling densities of Eucalyptus baxteri, whose seeds are removed by ants at particularly high rates. Seeds of Casuarina pusilla are removed less rapidly, and increases in seedling densities following ant elimination were less marked for this species. Fire plays an integral role in the reproductive biology of many Australian plants in sclerophyllous vegetation, and the effect of fire on seed removal by ants was investigated. A prescribed fire at a woodland site increased foraging populations of broadly-adapted ants, including seed-eating Rhytidoponera tasmaniensis. However the mass release of seeds from woody fruits resulted in predator satiation, with rates of removal from baits dropping by half immediately after fire, which might be an important factor contributing to successful recruitment. In addition to initial predator satiation, fire had a marked longer term effect on rates of seed removal, which might have important implications for recruitment at burnt sites. Foraging populations of R. tasmaniensis continued to rise for some weeks after fire, and, once seedfall returned to low levels, rates of seed removal rose rapidly. Weekly removal reached 100% seven weeks after fire, which was far higher than had ever previously been recorded. Results from a nearby heath site showed that unusually high rates of removal following fire persisted for at least two years. The significance of seed losses is a contentious issue, since they do not necessarily affect seedling recruitment. The likely, impact of seed predation on recruitment in the absence of fire was investigated for Eucalyptus baxteri, Casuarina pusilla, Leptospertmum juniperinum and L. myrsinoides at Tidal Overlook. Seed budgets quantifying seed production, seed losses, seedfall, soil seed stores, field germination, seedling densities and seedling survival were constructed for each species. It was estimated that only 3%, 13%, 4% and 2% respectively of baxteri, C. pusilla, L. juniperinum and myrsinoides seeds avoid predation, and are therefore available for recruitment. Although recruitment in sclerophyllous vegetation is generally believed to be extremely rare in the absence of fire, the results suggest that seedling establishment, although uncommon, regularly occurs in unburnt vegetation at Tidal Overlook. It is concluded that seed predation is likely to be an important factor limiting seedling recruitment in the absence of fire.