The Herpetological Journal

https://doi.org/10.33256/hj31.1.19

pp 1-9

Authors: Izabela M. Barata, Richard A. Griffiths, Deborah J. Fogell & Andrew S. Buxton

Abstract: Surveys of rare or cryptic species may miss individuals or populations that are actually present. Despite the increasing use of environmental DNA (eDNA) analysis to survey species in ponds, rivers, and lakes, very few studies have attempted to use eDNA for the detection of species using very small water bodies such as those accumulated within plants. Our aim was to investigate the feasibility of an eDNA sampling method for detecting Crossodactylodes itambe, an endemic bromeliad-dwelling frog from a remote location in Brazil. We collected water samples from 19 bromeliads for which we had observational data from direct visual surveys. We compared occupancy estimated from direct observations with the results from quantitative real-time PCR based eDNA assays. For observational surveys, we used a single season occupancy model. We applied a novel Bayesian occupancy model to estimate occupancy from eDNA samples, as well as false positives and false negatives at different stages of the workflow. eDNA from bromeliad tanks provided reliable estimates, with very low error levels and improved detection when compared to detectability from direct observation. Estimated occupancies using eDNA and visual survey methods were similar. The method is feasible for species restricted to small water bodies and exposed to direct UV radiation, and particularly useful to survey remote locations and confirm species presence. eDNA analysis provides a viable alternative to destructive sampling of bromeliads or direct observation methods that require logistically challenging repeated observations. Therefore, eDNA methods may be widely applicable to sampling programmes of other amphibians that live in plants.

Keywords: bromeliad, eDNA detection, false-positive, amphibian, occupancy, phytotelm