A methodological workflow for quantitative colouration and colour pattern comparison reveals taxonomic and habitat-level differences in the polymorphic fire salamander (Salamandra salamandra)

Abstract

The study of colour, including physical properties and patterns, is an essential step in comprehensively understanding the role, evolution and diversification of this trait involved in functions like survival, performance, reproductive success and communication. While quantitative studies on colour have primarily focused on colour attributes, such as hue or brightness, the quantitative assessment of how colours are distributed across an organism's body (i.e. pattern) has received less attention despite its high functionality irrespective and together with colour. This gap is particularly noticeable in amphibians, especially in highly polymorphic, continuously variable species. In this study, we use the urodele Salamandra salamandra as a study model to apply an analytical approach for extracting and quantifying colour proportions and patterns variation using the recently developed R packages patternize and recolorize to precisely describe and compare colour distribution at both taxonomic and habitat levels. We first assessed the potential application of this workflow in taxonomic studies by examining its accuracy and sensitivity in differentiating morphs, using S. s. bernardezi and S. s. gallaica subspecies. Subsequently, we applied the same analytical approach to compare colour and colour pattern shifts among population of S. s. bernardezi within an urban–forest system. Our results demonstrated that this methodological workflow effectively detects differences in colour patterns at the taxonomic and habitat levels, confirming its utility in different fields, from systematics to eco-evolutionary studies. Furthermore, our results revealed an increased prevalence of melanistic colouration and a less homogeneous colour pattern within urban settings, contributing to the limited number of studies about amphibian colouration in anthropic habitats. We outline and discuss alternative hypotheses to explain these patterns, although further investigation is necessary to address several new questions that arise from our results.