FACTORS INFLUENCING THE OCCURRENCE OF MAMMALS IN RETAINED LINEAR STRIPS (WILDLIFE CORRIDORS) AND CONTIGUOUS STANDS OF MONTANE ASH FOREST INTHE CENTRAL HIGHLANDS OF VICTORIA, SOUTHEASTERN AUSTRALIA

Hairtubing was used to census mammals at 70 sites in 53 year old mount ain ash (Eucalyptus regnans) and alpine ash (Eucalyptus delegatensis) forest in the Central Highlands of Victoria, southeastern Australia. A total of 40 sites was located in linear strips or wildlife corridors that were retained in timber production areas. The remaining 30 sites were in stands where the adjacent forest had not been logged. Thirteen species of mammals were detected in the hairtubing surveys, but only three were detected frequently enough to allow detailed analysis. The detectability of these animals and other less commonly recorded specie s varied according to the placement and the type of hairtubes that wer e used. Although data on feral predators were very limited, no apparen t effect of the proximity of hairtubing plots to roads was detected. D etailed statistical analyses of the data were conducted at two levels: the site level, and for individual hairtubing plots. Logit regression models were used to explore the relationships between: (1) the probab ility of detecting brown antechinus (Antechinus stuartii), bush rat (R attus fuscipes) and swamp wallaby (Wallabia bicolor) and the measured dimensions (e.g. width and length), environmental attributes and veget ation structure of retained linear strips; (2) the probability of occu rrence of Antechinus stuartii, Wallabia bicolor and R. fuscipes in hai rtubing plots within areas of contiguous forest and retained linear st rips and the characteristics of the vegetation structure and plant spe cies composition in a 10 m area surrounding such plots. For the first group of models, the variation in topography within retained linear st rips significantly influenced the proportion of detections of Antechin us stuartii and R. fuscipes, although there was a different response f or each species. Other significant explanatory attributes in these mod els included the number of roads and tracks that disrupted vegetation cover as well as structural and floristic parameters such as the domin ant tree species on the site and the basal area of Acacia spp. Analyse s of the combined data for the two types of sites indicated that Walla bia bicolor was more likely to be detected in plots within E. regnans forest and where there was a limited cover of ferns on the forest floo r. The type of study site (retained strip vs. contiguous forest) was a significant variable in the models developed for Antechinus stuartii and R. fuscipes, reflecting a significant influence of the configurati on of sites on the probability of detection of these animals. Separate models for the two types of sites were developed both for Antechinus stuartii and R. fuscipes. The dominant species of tree was a significa nt factor in several of these models, and there were significantly few er records of Antechinus stuartii from the ends of strips of E. delega tensis forest. Most of the other explanatory variables were measures o f the vegetation cover and plant species composition which are likely to reflect the importance of the suitability of foraging substrates fo r Antechinus stuartii and R. fuscipes. For these species, a different set of variables was found in each of the various models that were dev eloped. These findings imply a degree of partitioning of the use of th e forest environment by Antechinus stuartii and R. fuscipes. They also suggest that patterns of microhabitat use by some species of mammals may change when areas of montane ash forest are confined to a linear c onfiguration (i.e. a retained strip).