The lateral stripe — a reliable way for software assisted individual identification for Hyla arborea

From Firenze University Press Journal: Acta Herpetologica

Simeon Lukanov, Institute of Biodiversity and Ecosystem Research, Bulgarian Academy of Sciences

Andrey Kolev, Faculty of Biology, Sofia University “St. Kliment Ohridski”

Angel Dyugmedzhiev, Institute of Biodiversity and Ecosystem Research, Bulgarian Academy of Sciences

Miroslav Slavchev, Institute of Biodiversity and Ecosystem Research, Bulgarian Academy of Sciences

For several decades, individual identification of animals has been a crucial aspect of conservation biology and behavioural ecology. Studies focused on life-history traits and long-term monitoring require to repeatedly and reliably identify captured individuals. Traditional mark-recapture methods for amphibians and reptiles involve the application of physical markers with varying degrees of invasiveness — from paint marking to toe clipping (review in Ferner, 2007). For amphibians in particular, popular markings include toe clipping and subcutaneous injections/pit tags, with toe clipping being the most widespread for many decades because of its affordability. However, toe clipping does inflict temporary (i.e., newts regenerate toes) or permanent physical harm to the ani-mal, and it is well-established that it could cause dimin-ished survival and altered behaviour in some smaller species (e.g., Clarke, 1972; Guimaraesh et al., 2014). Alternatives to toe clipping have been implemented since the turn of the century, incl. visible implant elastomers (VIE) (Pittman et al., 2008; Antwis et al., 2014) and vis-ible implant tags (Pittman et al., 2008). VIE provide reli-able and rapid identification, but are expensive so their usefulness is often limited by their cost (Le Chevalier et al., 2017). During the past decade, there has been a grow-ing number of studies investigating the potential of using the natural colouration pattern of various amphibian species in photographic identification methods (PIM), as the rapid advance in technology allows for inexpensive, reliable and non-invasive identification (Kenyon et al., 2009; Bendik et al., 2013; Elgue et al., 2014; Schoen et al., 2015; Sannolo et al., 2016; Kim et al., 2017; Renet et al., 2019; Lukanov, 2022). The relatively small species of the Hylidae family are susceptible to toe clipping, as it has a negative effect on their survival and capture probability, especially when multiple toes are removed (Waddle et al., 2008) — so dorsal, ventral or leg patterns have been used in PIM studies on some hylid species (Kim et al., 2017). While manual visual identification is very useful for smaller samples, it becomes unreliable and time-consuming for larger catalogues, and in recent years, there have been developed some specialized software packages aimed at image pattern recognition (Speed et al., 2007; Bolder at al., 2012; Crall et al., 2013; Matthé, 2018). As the usefulness of software-assisted image recognition for Hyla sp. has not been investigated, the present study aims to test whether photographs of the lateral stripe in the Common tree frog Hyla arborea (L., 1758) would allow for a reliable and time-efficient individual identification. The Common tree frog is distributed from the Southern Balkans to North-Western Europe. In Bulgaria, it is present in the Struma river basin, whereas the Eastern tree frog Hyla orientalis (Bedriaga, 1890) occurs across the rest of the country (Dufresnes et al., 2015). The current study is part of an ongoing project on tree frog morphology and distribution in Bulgaria, which aims to compare populations across the country.

DOI: https://doi.org/10.36253/a_h-15519

Read Full Text: https://oaj.fupress.net/index.php/ah/article/view/15519