Myrmecochory in Australia's seasonal tropics: Effects of disturbance on distance dispersal

Successful ecosystem restoration requires the re-establishment of fundamental ecological processes, many of which involve plant-animal interactions. Myrmecochory (seed dispersal by ants) is a particularly important plant-animal mutualism in Australia, but little is known about its response to either disturbance or restoration following disturbance. Here we investigate the effects of disturbance on seed dispersal by ants, and the extent to which the ant-seed relationship has re-established at sites undergoing rehabilitation, at Ranger uranium mine in the seasonal tropics of Australia's Northern Territory. We focused on the composition of seed-dispersing ant assemblages, rates of seed removal by ants, and the dispersal curves generated by ants, as determined by observations of removal from seed depots. Ten sites were studied, comprising four 'natural' (undisturbed) sites representing a range of savanna habitats occurring in the region, four disturbed sites representing a range of habitat disturbance but with intact soil, and two waste rock sites subject to preliminary revegetation trials. A total of 22 ant species from 10 genera were observed during 154 observations of seed removal, most commonly Rhytidoponera aurata (53 records), Monomorium (rothsteini gp) sp. 1 (14), Iridomyrmex sanguineus (13), Iridomyrmex pallidus (12) and Pheidole sp. 3 (10). Removal rates (over 3 h) averaged 29% across all sites and time periods, varying markedly both between and within sites. However, mean rates of removal were similar between natural, disturbed and waste rock sites (29%, 28% and 31%, respectively). A high incidence (62% of all depots) of 'aril robbing' by ants (primarily Monomorium spp.) eating arils in situ, without removal, was observed. Dispersal distances varied markedly between ant species, with Iridomyrmex sanguineus having both the highest mean (7.25 m) and maximum (13.08 m) dispersal distances. Species of Pheidole typically dispersed seeds less than 0.5 m, and Meranoplus, Monomorium and Tetramorium spp. only ever moved seeds a few centimetres, usually dropping and abandoning them before reaching the nests. The dispersal curves characteristic of each site varied markedly due to the different composition of seed-dispersing ants. The mean dispersal distance at disturbed sites (3.91 m) was significantly higher than at natural sites (2.19 m), and the curves were strongly skewed in the former, but relatively uniform in the latter. The implications of these differences for recovery following disturbance are unclear. At rehabilitated waste rock sites, all observed removals involved distances less than 0.5 m, with a mean of 17 cm. This lack of effective ant-seed relationships might represent a barrier to further vegetation development at rehabilitated sites.