A spatially nested experimental design was used to survey reptiles in a sheep- and cattle-grazing landscape in southern New South Wales, Australia. Sixteen landscape units were chosen on the basis of their aspect, topography and amount of tree cover. Across these landscape units, 288 pitfall traps were established at 144 plots arranged in 48 sites. Each plot was searched for reptiles and pitfall-trapped in two separate survey periods. The four-fingered skink (Carlia tetradactyla) was the most commonly encountered reptile species, and detailed statistical habitat models were constructed for this species. The four-fingered skink responded to both landscape scale and microhabitat variables. The species was more likely to be detected in landscape units characterised by a high amount of tree cover. Adults were most likely to inhabit microhabitats dominated by box- or gum-type eucalypts, with a moderate amount of canopy cover and with a large number of spiders. Juveniles were most likely to inhabit microhabitats with a moderate amount of weed invasion where shrubs were present. Finally, the four-fingered skink was more likely to be found at sites with at least two other co-occurring small reptile species. Body condition and tail loss of adult skinks were not related to the habitat or microhabitat where a given individual was found. The results are discussed in relation to current herpetological research, and some long-term conservation implications are highlighted.