Survey methods included opportunistic searching, transect searching, pitfall trapping and acoustic recording. Transect searches were repeated during the day and night to account for any diel differences in activity, taking place in the morning and evening. Searching took place approximately two metres either side of the transect and up to two metres in height, and for amphibians were directed towards vocalising males. Searches in Ankarafa occurred in both the dry and wet season (during the 2011 period) and followed the same routes where possible. Sites were sampled in a randomised order and all searches were conducted by the same two individuals to avoid systematic observer bias. Location was logged using a handheld GPS receiver (Garmin eTrex Vista HCx; Garmin International Inc., Olathe, USA). Representative individuals were photographed to document their coloration, using a digital camera; tissue samples were collected, as were call recordings of amphibians. An integrative taxonomic approach was taken to assess species identification of both amphibians and reptiles; utilising the keys provided by Glaw and Vences (2007, and subsequent publications), personal photographic and acoustic catalogues, the application of molecular taxonomic identification as well as the comparative material hosted in the herpetological collection of the Museo Regionale di Scienze Naturali, Torino, Italy.
Pitfall traps with drift fences were made by sinking plastic buckets (270 mm deep, 220-250 mm internal diameter) into the ground at 6 m intervals along a 30 m drift fence, 0.4 m high, and buried 50 mm deep. Plant detritus was placed in the bottom of each bucket to act as a refuge for animals and holes punched in the bottom to allow water to drain. The pitfalls were checked each morning and evening for captured animals, and non-target animals were released. An initial four pitfall lines constructed in Ankarafa Forest in October 2011 were checked for a period of 13-15 days; these proved to be ineffective and inefficient, so a large scale expansion of pitfall trapping was discounted. A further three pitfall lines were constructed in Ankarafa Forest along a ridge, a slope and a valley bottom, for two periods of 14-15 days in October/November 2011 and December/January 2011-2012, covering the dry and wet seasons.