Mitochondrial DNA Variation and the Phylogeny of Zonotrichia

Abstract

-We used 19 restriction endonucleases to analyze patterns of cleavage site variation in the mitochondrial DNA (mtDNA) of the five recognized species of Zonotrichia. Each species possesses a unique mtDNA profile. Relative to most congeneric avian comparisons, these species are closely related, with an average percent nucleotide divergence of 4.1%. The Golden-crowned Sparrow (Z. atricapilla) and White-crowned Sparrow (Z. leucophrys) differ by a single restriction site (out of 122), which yields an estimate of 0.11% sequence divergence. These species are sister species in all analyses (Wagner and Dollo parsimony, bootstrapping, distance analyses), and White-throated Sparrow (Z. albicollis) is the sister to these, followed by Harris' Sparrow (Z. querula) and Rufous-collared Sparrow (Z. capensis). The phylogeny based on mtDNA restriction sites differs from that based on allozymes (Zink 1982) in that Harris' and White-throated sparrows are reversed in this sequence. In general, however, the allozyme and mtDNA phylogenies are highly concordant. Similarities in song between White-throated and Golden-crowned sparrows can be interpreted as ancestral retentions. The song of the White-crowned Sparrow is derived relative to other north temperate congeners, and song dialects appear to have evolved independently in White-crowned and Rufous-collared sparrows. Based on the mtDNA data, patterns of hybridization are inconsistent with relationships, and we suggest that hybridization should not be used in taxonomic decisions. Received 4 September 1990, accepted 26 December 1990. A PHYLOGENY not only provides a trace of a lineage's evolutionary diversification, but it also serves as a framework for the interpretation of the evolution of characters and suites of characters. Molecular methods have been widely heralded because they provide a set of characters that have a simple, known genetic basis, and are genetically independent. Furthermore, it is likely that molecular characters are selectively neutral (e.g. Barrowclough et al. 1985), and thus their evolution can be predicted by explicit evolutionary models (Nei 1987) and can serve as an approximate molecular clock. Although molecular characters have disadvantages for phylogeny inference (Hillis 1987), they are used with increasing frequency as a sort of null hypothesis, both for estimation of evolutionary history and interpretation of character evolution. Whether molecular analyses actually produce superior estimates of phylogeny remains to be documented, and tests will likely involve comparison of independently derived phylogenies (including other molecular estimates) of common sets of organisms. In avian systematics, protein electrophoresis has been used to produce a number of estimates. In many interspecific comparisons of avian protein evolution (e.g. Avise et al. 1980, Johnson and Zink 1983), there was little variation partitioned among species. This does not inspire confidence in any pattern (without independent corroboration from other data sets). Other studies of allozyme variation that used phylogenetic and phenetic methods often produced results that surprised many taxonomists, including the biochemical systematists themselves (e.g. Johnson et al. 1988, Dittmann et al. 1989). However, few protein-based estimates for birds were tested for robustness or confidence with data-resampling techniques such as bootstrapping (Felsenstein 1985b). Therefore, the confidence in these protein-based branching diagrams as phylogenies is unknown. One method of assessing confidence is to compare estimates derived from independent data sets. Zink and Avise (1990) compared allozyme and mitochondrial DNA (mtDNA) evolution in the genus Ammodramus and found a high degree of congruence. Zink and Dittmann (1991) found a generally high correspondence between allozymes and mtDNA in towhees (Pipilo). In these examples, the signal appeared in both of the genetically independent data sets. However, in Ammodramus and Pipilo, unlike many avian studies, there was considerable allozymic dif578 The Auk 108: 578-584. July 1991 This content downloaded from 157.55.39.208 on Fri, 14 Oct 2016 04:21:30 UTC All use subject to http://about.jstor.org/terms July 1991] mtDNA Evolution in Zonotrichia 579 ferentiation, which rather strongly supported a particular phylogeny. Congruence of mtDNA and allozyme results implies confidence in conclusions, whereas confidence in a single data set is difficult to assess (Felsenstein 1985b). The genus Zonotrichia contains five species, several of which have been studied extensively by behaviorists and ecologists. For example, vocal dialects in the White-crowned Sparrow (Z. leucophrys) have been studied thoroughly (Kroodsma et al. 1985). Vocalizations and dialects have also been studied in the Rufous-collared Sparrow (Z. capensis; Handford and Nottebohm 1976). Interpretation of the evolution of vocal as well as morphological and ecological conditions in Zonotrichia requires an independently derived phylogeny (Felsenstein 1985a, Mickevich and Weller 1990). Zink (1982) used allozymes and morphometrics to assess the systematic status of species in Zonotrichia, and he concluded White-crowned and Golden-crowned (Z. atricapilla) sparrows were sister taxa, followed by Harris' Sparrow (Z. querula), the White-throated Sparrow (Z. albicollis), and the Rufous-collared Sparrow. This arrangement conflicted with previous opinions (Paynter 1964, Short and Simon 1965). For example, Mayr and Short (1970) stated that Golden-crowned and White-throated sparrows were sister species, and together with White-crowned Sparrows formed a species group. By default, Harris' and Rufous-collared sparrows would be outliers to this group. Evidence cited for the close relationship between White-throated and Golden-crowned sparrows was an essentially allopatric distribution; these species also exhibit similarities in male primary song. Mayr and Short (1970) suggested that hybridization between the White-throated Sparrow and Darkeyed Junco (Uunco hyemalis) revealed a close relationship between these taxa. For birds, mtDNA has been shown to provide more characters for systematic analysis than allozymes (Avise and Zink 1988, Zink 1991). This corresponds with the relatively fast rate of mtDNA evolution observed in birds (Shields and Helm-Bychowski 1988) and other vertebrates (Avise 1986, Moritz et al. 1987). We used restriction endonuclease analysis of purified mtDNA to provide information for the inference of relationships among species of Zonotrichia. In so doing, we test the allozyme-based tree of Zink (1982) and the taxonomic conclusions of others, provide data on evolution of this organellar piece of DNA, and comment on the evolution of behavioral and distributional characteristics in the genus.