Systematics

No phenotypic variation described. See below for geographic differences in genetic introgression.

None described.

Although not sequenced in their study, Vermivora chrysoptera doubtless fits neatly within the narrow, monophyletic Parulidae, the American wood-warblers, identified by Lovette and Bermingham (2002). This supposition is borne out in subsequent studies (Klein et al. 2004, Lovette and Hochachka 2006). Vermivora sensu lato has long been treated as close to Parula, and the two genera are merged sometimes. Nonetheless, recent genetic studies (e.g., Lovette and Bermingham 2002, Klein et al. 2004, Lovette and Hochachka 2006) argue against recognition of a broad genus Vermivora, as part of that broad genus is more closely related to Parula and part is more distantly related to that genus. As a result, a Vermivora sensu stricto includes only three species (Chesser et al. 2010): V. chrysoptera (Linnaeus, 1766), V. cyanoptera (see Olson and Reveal for use of this name instead of V. pinus), and the likely extinct V. bachmanii (Audubon, 1834). Other species formerly placed in Vermivora, as well as two species formerly placed in Parula, are now placed in genus Oreothlypis (Chesser et al. 2010).

Within the restricted genus Vermivora, frequent and geographically extensive hybridization between V. chrysoptera and V. cyanoptera-see McCarthy 2006 for numerous references-argue for their status as sister species. Such crosses yield fertile offspring, and two principal hybrid phenotypes actually were described as species in their own right: Brewster's Warbler, V. leucobronchialis (Brewster, 1874), and Lawrence's Warbler, V. lawrencii (Herrick, 1874). Mitochondrical DNA (mtDNA) markers differ to an extent between the species and the geographic distribution of these markers implies variation in the extent of introgression among different zones of sympatry (Vallender et al. 2009). Neutral markers in the nuclear DNA differ little (Vallender et al. 2007), but there must be distinct differences somewhere in the nuclear genome given what is known of plumage inheritance (Parkes 1951). The black auricular and throat of V. chrysoptera are thought to be the homozygous recessive condition at a single locus, whereas the black transocular line and plain throat of V. cyanoptera are dominant traits. Under Parkes' (1951) Mendelian model of plumage inheritance, a Brewster's Warbler is the F1 offspring of pure parents that inherited the dominant transocular line and plain throat of V. cyanoptera but the dominate V. chrysoptera body color. The rarer Lawrence's Warbler exhibits recessive V. chrysoptera traits of the black auricular and throat and the recessive V. cyanoptera yellow body color. Some hybrids exhibit subtler forms of introgression, such as variability in body color, degree of separation of the wing patch into two bars, and varying degrees of yellow to white color of the wing. These traits span a complete range between the parental phenotypes and therefore may make hybrids difficult to distinguish from parental types, especially in the field. Such variation could result from expression of multiple genes, incomplete dominance, or modifier genes (Parkes 1951). Regardless, phenotypic introgression has been characterized fully by means of hybrid indices (Short 1969, Gill 1980).

As noted above, researchers can distinguish the species by means of mtDNA markers, which are particularly sensitive to the accumulation of small genetic differences among closely related populations. Interspecific comparisons revealed 3.0-3.2% divergence in the cytochrome-b gene (Gill 1997) and 4.3-4.9% divergence in more rapidly evolving ND2 gene (Shapiro et al. 2004; Dabrowski et al. 2005). This degree of differentiation equals or exceeds that between other sister taxa considered to be species (Avise and Zink 1988, Tegelström and Gelter 1990). Amplified fragment length polymorphisms (AFLPs) were the first nuclear markers found to differ significantly in frequency between the species: even though no species-specific markers were identified, frequency differences in 7 characters allowed Vallender et al. (2007) to assign individuals to the correct parental populations with >99% precision. Allozyme differences yielded a genetic distance (Nei's D) of 0.0001 (Gill 1987), meaning the species are less distinct than some congeneric species and even some avian subspecies (Barrowclough and Corbin 1978, Barrowclough 1980 and Christidis 1987). Similarly, recent examination of 13 microsatellite and 4 intron loci found no unique alleles and little to no interspecific differentiation between the species (Vallender et al. 2007), and searches for private alleles and assignment tests uncovered no combination of microsatellite or intron markers that could diagnosis the species.

Direction and degree of genetic introgression vary throughout the range, likely depend on time since secondary contact, and possibly depend on habitat segregation in some portions of the range. Gill (1997) first reported cryptic mtDNA introgression into pure V. chrysoptera phenotypes in ne. Pennsylvania. Introgression was asymmetrical, with little reciprocal introgression of V. chrysoptera mtDNA into V. cyanoptera phenotypes. Later studies reported bidirectional introgression of mtDNA (Dabrowski et al. 2005, Vallender et al. 2007, 2009). Introgression has been associated with an influx of hybrid and, later, V. cyanoptera phenotypes in Ontario (Vallender et al. 2007), West Virginia (Shapiro et al. 2004), and Kentucky (PJH). In s. upstate New York, where the two species have coexisted for over a century (Confer and Tupper 2000), phenotype-haplotype mismatches account for up to 33% of the survey population (Dabrowski et al. 2005, Vallender et al. 2009). Nest success and phenotypic purity of V. chrysoptera is significantly higher in swamp forests than in uplands (Confer et al. 2010). Fecundity of birds in swamp forest is much higher than birds in uplands, so swamp forests are a population source for V. chrysoptera that sustains coexistence in the uplands (JLC).

These genetic studies generally have revealed widespread introgression of V. cyanoptera haplotypes into V. chrysoptera phenotypes. By contrast, no population has exhibited a predominance of introgression of V. chrysoptera mtDNA into V. cyanoptera phenotypes. In essence, then, the V. chrysoptera is being replaced by V. cyanoptera. The only known phenotypically and genetically pure populations of V. chrysoptera breed in Manitoba (Vallender et al. 2009). Even phenotypically pure V. chrysoptera in Minnesota and Tennessee carry V. cyanoptera genes (Vallender et al. 2009).

Apart from extensive hybridization with V. cyanoptera, V. chrysoptera may have hybridized once with Mniotilta varia, the Black-and-white Warbler (McCarthy 2006:314).

• Golden-winged x Blue-winged Warbler (hybrid) Vermivora chrysoptera x cyanoptera
• Brewster's Warbler (hybrid) Vermivora chrysoptera x cyanoptera (F1 hybrid)
• Lawrence's Warbler (hybrid) Vermivora chrysoptera x cyanoptera (F2 backcross)

Not known.