Inferences about Population Movements of Red-winged Blackbirds from Morphological Data

Abstract

For species of birds that exhibit regular clines of geographic character variation in the breeding season, repeated morphological comparisons even among small samples of individuals can be used to infer the overall spatial organization of the wintering population in relation to that of the breeding population. Here we demonstrate how this approach can be applied by presenting an analysis of measurements of specimens of 667 red-winged blackbirds (Agelaius phoeniceus). In this species breeding adults in the eastern and central United States have a gradual but regular cline of increasing size northward and westward from southeastern Florida. This general pattern is maintained year after year even though dispersal between breeding seasons (either between hatching and breeding or between different breeding seasons) must increase variation in morphology within localities. Such effects of dispersal tend to mask the effects of factors that maintain clines. These factors are, presumably: (1) site fidelity of birds to their places of origin; (2) natural selection for optimal phenotypes, and (3) environmentally induced local morphological variation during development. Red-winged blackbirds that breed in the northern and N-central states form massive winter roosts at more southern localities. But geographic trends of larger size farther N can be detected in winter as well as in the breeding season. Within single southern localities, comparisons between wintering and breeding populations can indicate whether wintering birds are residents or migrants, and differences in winter between the sex-age classes can indicate differences in their geographic origin. In our winter samples from Tallahassee, Florida, most adult males were local permanent residents, but most adult wintering females were birds with wings longer than 97 mm, so they were probably from breeding populations N of Florida. Repeated seasonal comparisons of external measurements and of skeletons between the wintering and breeding populations at single localities in the southeastern states show that the seasonal differences are consistently greater in females than in males. We interpret this finding to mean that females generally migrate farther than males. In addition, at five widely separated localities, older wintering females tend to be larger than hatching-year females. This result is not just an effect of growth, because it is evident in measurements of bones that are fully ossified in the younger age class. We interpret this as evidence that older females migrate farther than younger females. This is a tentative conclusion because our sample sizes at single localities are small, but our repeated comparisons by age-class, sex and locality all show the same relationship. Most of the results are presented in box-and-whisker plots. We argue that for displaying patterns of geographical variation these plots are more appropriate than the more commonly presented Dice-Leraas diagrams or than sets of statistical tests. They are examples of exploratory data analysis, an important new development in statistics that looks for patterns in data and is cautious about the application of statistical tests in cases in which observations are not randomly drawn from a population.