Context . In many parts of the world, artificial refuges (ARs) are increasingly used to survey different kinds of herpetofauna. Despite gaining popularity, the merit of using ARs compared with standardised active searches remains poorly known, as does their application in regions that support species exhibiting varied life-form strategies.
Aims . We examined the effectiveness of using two survey methods, active searches and ARs (corrugated steel, roof tiles and three different timber refuges), to detect herpetofauna in a range of eucalypt-woodland communities in south-eastern Australia.
Methods . We collected count data over a 12-year period from three independent long-term biodiversity monitoring programs in southern New South Wales. We used generalised linear models to compare detection probabilities among survey methods (active searches versus ARs) and among ARs, for each study area.
Key results . In all study areas, active searches detected the highest mean species richness per site. However, both methods provided complimentary species, thereby maximising species inventory at a regional scale. Species more likely to be detected in active searches included diurnally active, terrestrial and arboreal heliotherms, whereas species detected more frequently using ARs included nocturnal thigmotherms.
Conclusions . A combination of active searches and AR types is required to provide regional-scale representative reptile assemblages, although more than five consecutive surveys may be needed before species accumulation curves reach plateaux. In future studies, we recommend using stacks of corrugated steel to detect heliothermic Scincidae and arboreal Gekkonidae, roof tiles to detect thigmothermic Pygopodidae and railway sleepers to detect cryptozoic Elapidae and tunnel-dwelling Gekkonidae.
Implications . Using a combination of ARs and active searches will increase the chance of detecting both common and cryptic species and deploying corrugated steel provides a cost-effective method for surveying reptiles in long-term studies.