Wheatear molt and assignment tests: ongoing lessons in using stable isotopes to infer origins

Abstract

In an earlier paper (Rocque et al. 2006), we attempted to assign two generations of feathers in three species of intercontinental migrants to their respective breeding and wintering areas using stable isotopes. We found that while two species of Pluvialis plovers that breed in Alaska and winter on different sides of the Pacific Ocean (South America versus Asia and the South Pacific) showed significant differences between summerand winter-grown feathers, we could not statistically separate feathers grown on the two continentally different wintering areas. Nor could we assign summer-grown feathers to within thousands of kilometers of their actual origin using the common approach of a North American isotope (dD) isocline map. In Alaska-breeding Northern Wheatears (Oenanthe oenanthe), which winter in Africa, we found no significant differences between summerand (presumed) wintergrown feathers. Larson and Hobson (2009) helpfully point out a molt-related problem with our study, specifically with respect to where our presumed winter-grown Northern Wheatear feathers were actually grown. We welcome the opportunity to draw attention to wheatear molt, in which we made an error: the body plumage color changes in O. oenanthe between autumn and spring must be one of the most dramatic to occur in highlatitude birds due to feather wear and not to molt (Fig. 1). Measurements of 10 back feathers from five males in fresh basic plumage showed that approximately one-third (average 36.5%) of the feather length is worn off to achieve this color change. Thus, most if not all of our presumed winter-grown back feathers from adult Northern Wheatears were grown in Alaska, not Africa. This is mea culpa for DAR and KW. We overestimated the extent of the prealternate molt and appreciate the input from other colleagues and now Larson and Hobson (2009) that provides us the opportunity to correct this error. This recognition helps us understand why this group of feathers did not differ significantly from secondaries known to have grown during the previous breeding season in Alaska on these same birds. We hope this prevents others from committing a similar error, for such mistakes only add to the pitfalls associated with using stable isotope analysis to infer origins and movement patterns of animals. However, the presumed winter-grown wheatear back feathers represented only one of six datasets in our study, and we consider that the other conclusions remain robust. In addressing our Pluvialis data, Larson and Hobson (2009) do not dispute the spatial location of feather growth (in the tropics on different sides of the Pacific Ocean in the respective species). Instead, they provide a series of reasons why winter-grown feathers might have high variance in dD (as we did also). They suggest that we should have excluded samples for which dC was less negative than -20% because these samples represent feeding in the Communicated by F. Bairlein.