hide
Free keywords:
rapid surveys; amphibians; DNA barcoding; identification
Abstract:
We provide an estimate of identification reliability of Malagasy frog species based on different methods. According to our estimate, for 168 out of 358 species, a reliable identification based on morphology alone is not possible for reasonably trained researchers. By also considering colouration in life, this number went down to 116 species. Of 252 species for which calls are known, a reliable identification based exclusively on bioacoustics is not possible for 59 species. DNA barcoding performs distinctly better; problems with molecular identification are only known for 61 out of 347 species for which genetic data are available.
In a second approach we also present preliminary data on a comparative study of performance of various inventory techniques applied to three frog communities along eastern rainforest streams. At these streams tadpole collection and their subsequent identification via DNA barcoding allowed for an average detection success of 45% of all species per site, while standardized call surveys detected 28% and visual encounter surveys 29% of the species. However, these results varied widely among rough ecological guilds of frogs, with forest frogs that breed independently from open water, obviously, being undetectable in the tadpole surveys, arboreal frogs being poorly detectable in visual encounter surveys, and stream edge frogs being very poorley detectable in bioacoustic surveys. We suggest that a combination of methods is necessary to obtain a maximum of positively and reliably identified species records in a limited amount of time, and we emphasize the extreme importance of increasing data verifiability by listing voucher specimens, and as much as possible, including DNA barcoding, call recording, and photographs in life. For a public and easy access to such supplementary data to any amphibian survey in Madagascar, creation of a joint website is recommended.
