The Western Australian (WA) wheatbelt is ideal for studies of habitat fragmentation. Habitat alteration of the wheatbelt, involving the clearance of 140 000 km2 (Saunders, 1989) has had a massive and often unpredictable impact on endemic fauna and flora and habitat fragmentation has been implicated as the key mechanism responsible for the loss of many of Australia’s plant and animal species (Hobbs and Hopkins, 1990).
To investigate the effects of habitat fragmentation on amphibian populations of the WA wheatbelt, I sampled small fragments (~ 1km2) and non-fragments (~ 10 km2). All fragments were at least 5 km from other fragments and beyond the dispersal capcity of most amphibians. The primary aims of this study were to compare the abundance and morphometrics of amphibian populations in fragments and non-fragments.
Four species of frog were captured. These were, in order of decreasing abundance, Pseudophryne occidentalis, Neobatrachus kunapalari, Heleioporus albopunctatus and Neobatrachus pelabatoides. Statistical tests of abundance between non-fragments and fragments did not produce a significant result for any of the captured species. However, there were 303 specimens of P. occidentalis captured in non-fragments, compared to 66 captured in fragments and thus although not statistically different, there was a distinct effect of fragmentation on the abundance of this species. H. albopunctatus was only present in non-fragmented habitats, whilst N. kunapalari and N. pelabatoides were present at equally low densities across all sites. Interestingly, inclusion of a multivariate test using data for all species (such as H. albopunctatus which did not occur in fragments) produced a significant result. But because of the low densities of other species and the inclusion of H. albopunctatus, this result must be treated with caution.
Morphometric analysis showed that weight, head width and left tibia length did not vary significantly between fragments and non-fragments. Snout-vent length, however, was very close to significance with habitat type. In addition, bimodal distributions of snout-vent length were detected in non-fragments, suggesting the existence of two size classes.
Examinations of fluctuating asymmetry (FA) showed that there was no significant effect between habitat types. However, FA was apparent between sites within fragmented habitats. FA was manifested in populations with low densities and is likely to have been swamped in larger populations, such as those in fragments. Accordingly, FA was not detected in one of the three fragments which was characterised by relatively high abundances of P. occidentalis.
Finally, environmental influences including site geomorphology, vegetation composition and structure, disturbance of sites and climate, are considered to have imparted some influence on capture rates and morphometrics.
As a result of this research, I recommend more extensive study into the effects of habitat fragmentation on wheatbelt frog populations. These studies should include investigations of the contributions of climate, site geomorphology and vegetation to any observed effects of habitat fragmentation.