Chaffinch Imitates Canary Song Elements and Aspects of Organization

commonly occur in winter in large flocks that exploit similar patchy resources, for example locally abundant Toyon berry (Heteromeles arbutifolia) crops. Unlike the above pairs, some other pairs of significantly covarying species show no obvious ecological relationships. Examples of such pairs are the Cedar Waxwing with Dark-eyed Junco or the Hermit Thrush with Purple Finch. In fact, close study of the correlation matrix shows that about half of the significant relationships are between ecologically rather dissimilar species pairs. The results of the randomization procedure for testing the "significance" of the various group structures are presented in Fig. 1. The most "realistic" ecological group structure, which incorporated information about diet and microhabitat association, was significant (P = 0.049). The simple foraging-site grouping and the rather arbitrary taxonomic structure did a poor job of combining concordant species. The simple migratory-habit group structure was highly significant (P = 0.015). This latter result probably reflects the overall "boreal/montane bird year" phenomenon. If these results verify the existence of a boreal/ montane bird invasion phenomenon, what does that pattern mean? One possible explanation for the pattern is that the fluctuations represent actual changes in boreal/montane bird populations. Perhaps years with excellent summer recruitment are followed by a widespread occurrence of these species in the lowlands of southern California during winter. This could explain why ecologically similar species tend to covary. Alternatively, co-occurrence patterns might resuit from similar responses to local food or weather conditions. If the distribution of wintering density varies across the range of these species, certain years may be characterized by high densities in the Los Angeles basin. Such years need not be years with high recruitment but could represent years of locally favorable food supplies or weather conditions (or unfavorable conditions elsewhere). Identifying the relative influence of each factor on local population densities would require a broader scale analysis. Such a study should consider recruitment success, migratory pathways, and winter population densities across the entire range of the species in question. We thank Carl Bock, Charles Collins, Glen Ford, Kimball Garrett, Terry Root, John Rotenberry, Cliff Crawford, Aida Parkinson, John Wiens, Bruce Woodward, Jared Verner, and an anonymous reviewer for reading an earlier draft of the manuscript and making many helpful suggestions. Special thanks are due Dennis Heinemann and Beatrice Van Horne