A common view of biologists for animal coloration is that this characteristic has various ecophysiological functions. Snakes exhibit a variety of colorations in their simple elongated body. Stimulated by their diversified color morphs, various types of studies attempting to elucidate functions of snake colorations have been conducted. Although a few exceptions exist, most of these studies classified color morphs into discrete categories. However, when we treat species having heterogenous color patterns in a single individual and species exhibiting continuous individual variations of coloration, it is often difficult to categorize samples into discrete morphs based on its coloration because we must ignore some characteristics and continuous variations of coloration. Principal components analysis (PCA) is an appropriate method to treat such species because each coloration can be represented by a numerical value as an independent point within a continuum of color variants. However, we cannot know, a priori, how many and what kinds of variables are at least necessary to adequately represent features of the whole sample. The aim of this study is, by adopting variable selection in PCA, to identify and provide a “least set of variables”, which is the least number of characteristics of color pattern necessary to measure when we evaluate and rank individual snakes based on their coloration. Using Elaphe quadrivirgata as a model subject, we found that 10 variables can represent features of the whole sample of the subject to the same degree as 53 original variables. Possible applications of this proposed method to behavioral studies are discussed.
How to translate text using browser tools
1 December 2007
Quantitative Evaluation of Individual Snake Coloration by Use of Principal Components Analysis with Variable Selection
Koji Tanaka,
Akira Mori
ACCESS THE FULL ARTICLE
color pattern
principal components analysis
Quantification of categorical data
Striped snakes
Variable selection