Moult nestedness and its imperfections: insights to unravel the nature of passerine wing‐feather moult rules

Passerine moult shows two well‐known although intriguing characteristics: 1) a varying subset of wing feathers are retained during certain moult episodes, and 2) the identity of the replaced feathers is not random. However, the rules that underlie these moult features are largely unknown. Non‐randomness could arise under time and energy constraints if feathers differed in their importance, such that they could be ordered following a hypothetical moult rank. This rule predicts a nested relationship among wing‐feather moult phenotypes, where phenotypes of smaller extent would consist of subsets of the wing feathers replaced in phenotypes of larger extent. Here, we analysed a dataset of 3909 preformative and prealternate moult phenotypes from 434 Holarctic and Neotropical passerine species. We found that passerine wing‐feather moult was nested (nodf = 84.31). However, systematic imperfections in the nestedness pattern led us to test whether the existence of different feather moult ranks may increase nestedness in homogeneous subsets of moult phenotypes. We tested this hypothesis between moult episodes and found that nestedness increased for the preformative moult (nodf = 90.99) and decreased for the prealternate moult (nodf = 67.28). We further analysed the adaptive meaning of different moult ranks associated with feather mass, feather exposure and feather contribution to flight. Overall, these three feather traits explained 53.74% of the variation in the priority of wing‐feather replacement, with exposure having a stronger explanatory power than the other two traits. Our findings provide a first model to interpret variation of wing‐feather moult phenotypes in passerines, and outlines the relevance of season for passerine moult, suggesting a dynamic link between moult expression and plumage requirements.