FATE OF SEEDS ADAPTED FOR DISPERSAL BY ANTS IN AUSTRALIAN SCLEROPHYLL VEGETATION

Influences of ant behavior on short- and long-term fate of elaiosome-bearing seeds were investigated in fire-prone sclerophyll vegetation near Sydney, New South Wales. The short-term fate of seeds was assessed by measuring dispersal distances, seed predation within ant nests, and seed burial. Most seeds were taken directly to ant nests. The mean dispersal distance was 1.1 m (range 0.06-11 m). Seed predation by three ant species was measured by feeding known quantities of seeds and inedible seed mimics, made from plastic beads, into nests. Nests were excavated 2, 7, and 30 d later to recover seeds and beads. The recovery rate of beads estimated the disappearance rate of seeds from nests due to factors other than predation. Pheidole sp. 1 appeared to eat most seeds soon after collection but also stored some seeds in small caches within the nest. Both Rhytidoponera "metallica" and Aphaenogaster longiceps removed most seeds from the nest intact (the elaiosome having been consumed) and discarded them on the surface, but left some within nests at depths < 12 cm. Seeds discarded on the soil surface may undergo secondary dispersal by other ants. A transition matrix analysis was used to assess the importance of this secondary dispersal route and to estimate the probabilities that seeds initially taken by a particular ant species would be either eaten or finally buried within a nest. The analysis indicated firstly that a significant proportion of seeds initially removed by A. longiceps and R. "metallica" will eventually be eaten by Pheidole spp. Secondly, some seeds will be discarded in the nests of species other than the original seed remover. The analysis also indicated that seeds initially taken by A. longiceps and R. "metallica" would be more likely to survive than those taken by Pheidole sp. 1. As many Australian ant-dispersed seeds are stimulated to germinate by fire, long-term seed fate was investigated by monitoring seedling emergence after fire from ant nests that had been fed a known quantity of seeds. Overall emergence was almost-equal-to 1% of seeds fed into nests. The greatest emergence was from nests of A. longiceps. This result is qualitatively consistent with those of the short-term experiments. However, the low overall emergence rate meant the power of this experiment to detect statistical differences between ant species was limited. The spatial pattern of emerging seedlings differed between nests of the different ant species. Seedlings from the nests of Pheidole sp. 1 were considerably more clumped than those from nests of the other three species. The seed fate experiments collectively indicated that seed-removing ants have different effects on seed fate, with A. longiceps and R. "metallica" being more beneficial than Pheidole.