Hybridization can have profound effects on biological diversity. However, predictable inheritance of plumage traits remains poorly understood, especially for rare hybrids. We reviewed the literature and compiled a comprehensive list of hybrids from the New World warbler family Parulidae, a diverse radiation of songbirds with divergent plumage traits. We used our compilation to analyze modes of inheritance in wing bar patterns and carotenoid coloration. Finally, we describe an unusual hybrid from the University of New Mexico in Albuquerque, New Mexico, southwestern USA. We identified evidence of hybridization in 44 of 47 (93%) North American parulid species, with the highest number of hybrids found in the genus Setophaga. Plumage patterns between hybrid offspring and parental forms in our 2 focal traits were predictable, supporting the identification of our hybrid as a Yellow × Black-throated Blue Warbler (S. petechia × S. caerulescens). We based our identification on the extent and pattern of white in the tail, a prominent white wing flag, and our ability to confidently rule out all other alternative parentals. Our results suggest that phenotypes of rare hybrid warblers likely have some degree of predictability.
Resultados predecibles de la hibridación de reinitas: síntesis y una pareja excepcional de Setophaga petechia × S. caerulescens
La hibridación puede tener efectos profundos en la diversidad biológica. Sin embargo, la herencia predecible de los rasgos del plumaje es poco conocida, especialmente en el caso de los híbridos raros. Revisamos la literatura y recopilamos una lista completa de reinitas híbridos de la familia Parulidae, una radiación diversa de aves cantoras con rasgos de plumaje divergentes. Utilizamos nuestra compilación para analizar los modos de herencia en los patrones de las barras alares y la coloración carotenoide. Por último, describimos un híbrido inusual de la Universidad de Nuevo México en Albuquerque, Nuevo México, al suroeste de los Estados Unidos. Identificamos evidencia de hibridación en 44 de las 47 (93%) especies de parúlidos norteamericanos, siendo el mayor número de híbridos encontrados en el género Setophaga. Encontramos una herencia predecible de los patrones de plumaje entre la descendencia híbrida y las formas parentales en nuestros dos rasgos focales, lo que apoya la identificación de nuestro híbrido como una reinita amarilla × reinita azul y negra (Setophaga petechia × S. caerulescens). Basamos nuestra identificación en la extensión y el patrón de blanco en la cola, una prominente bandera de ala blanca y nuestra capacidad para descartar con seguridad todos los demás orígenes parentales alternativos. Nuestros resultados sugieren que los fenotipos de los parúlidos híbridos raros probablemente tienen un cierto grado de predictibilidad.
Palabras clave: hibridación, identificación de híbridos, Parulidae, plumaje, reinita, Setophaga.
Hybridization is a phenomenon that has long captivated birders and evolutionary biologists alike (Mayr 1942, Mallet and Barton 1989, Mallet 2005, Toews et al. 2016, Baiz et al. 2020). Until recently, the role of hybridization in evolution was contentious (Dowling and Secor 1997); however, we now understand the consequences of hybridization on adaptive introgression and speciation, and how these forces can have profound effects on biological diversity (Seehausen 2004, Mallet 2005, Meier et al. 2017). Notably, hybrid zones often encompass recently diverged species pairs that inform our understanding of the drivers and patterns of speciation and diversification (i.e., Pearson 2000, Vallender et al. 2007, Irwin et al. 2009). Research on hybridization has also been extended to deepen our understanding of diverse topics ranging from migratory behavior (Toews et al. 2014, Lundberg et al. 2017) to the genomic basis for pigmentation (Toews et al. 2016, Brelsford et al. 2017).
Although hybridization and hybrid zones have been studied across the tree of life (McEntee et al. 2020), rare hybrids are difficult to study because they occur and/or are detected infrequently. Low abundance of rare hybrids can limit understanding of the dominance patterns of plumage and structural traits and it can prevent robust tests of genomic differentiation, as is possible in hybrid zones (e.g., Toews et al. 2016, Brelsford et al. 2017). Despite their overall scarcity, rare hybrids can be used to study evolutionary questions related to reproductive isolation and ecological divergence (Willis et al. 2014), and documenting rare hybrids is one method by which we can continue to expand our knowledge of evolutionary patterns.
One family with notably high instances of hybridization, and several well-studied hybrids, is the New World warbler family, Parulidae. The Parulidae radiation began in the late Miocene, ∼7 million years ago, and was quite rapid despite a relative lack of ecological differentiation typical of other rapid radiations (Lovette et al. 2010, Oliveros et al. 2019). Despite ecological niche overlap, many warbler species finely partition resources and are able to coexist in sympatry with congenerics (MacArthur 1957), and even closely related and ecologically similar species display exceptional variation in plumage color (Baiz et al. 2020). It has been estimated that hybridization occurs in ∼72% of parulid warbler species that breed in North America (Willis et al. 2014). Hybridization within Parulidae is diverse and includes many instances of intrageneric and intergeneric crosses, as well as one exceptional intergeneric three-way hybrid (Toews et al. 2018). The family boasts a number of rare hybrids (Willis et al. 2014) and thoroughly researched hybrid zones (Krosby and Rohwer 2010, Toews et al. 2016, Brelsford et al. 2017).
The rate of hybridization within Parulidae is notably higher than in most other avian families (Ottenburghs 2019). Although propensity for hybridization has not been formally quantified within Parulidae, it is thought that the overall high incidence of hybridization, combined with striking plumage divergence, have resulted in parulid hybrids being generally easier to identify than more cryptic hybrids (e.g., Cronemberger et al. 2020). These same characteristics make warbler hybrid zones some of the most productive for identifying genomic regions responsible for pigmentation differences between species (Funk and Taylor 2019). In particular, previous work has revealed that 2 key genes have a pronounced effect on pigmentation in hybridizing warblers: the agouti signaling protein (ASIP) is associated with black patterning, including face mask, throat patch, and streaking patterns (Toews et al. 2016, Wang et al. 2020); and beta-carotene oxygenase 2 (BCO2) is responsible for a variety of carotenoid-based phenotypes, including dramatic full-body yellow coloration (Toews et al. 2016, Baiz et al. 2020).
The high frequency and diversity of hybrid pairings, the long-term importance of gene flow, and a small number of known pigmentation genes of large effect make Parulidae an ideal group in which to study the predictability of hybrid phenotypes.
Here, we synthesize documented New World warbler hybrids from past to present to understand patterns of inheritance in variable traits, thereby building upon previously proposed trait-based approaches to hybrid identification (Graves 1990, Rohwer 1994). We focused our analysis on 2 key traits: carotenoid pigmentation and wing bar patterns. These traits were chosen due to their high variability within Parulidae and relevance to our focal individual. Specifically, we sought to understand patterns of inheritance in presence vs. absence and extent of wing bars, and in degree of pigmentation and extent of whole-body carotenoids, of hybrid offspring relative to parental forms. We then applied our results to present and identify a striking Yellow × Black-throated Blue Warbler (Setophaga petechia × S. caerulescens) hybrid from New Mexico, southwestern USA.
We compiled a list of parulid hybrids and their attributes from published literature, websites, books, guides, photos, and natural history observations (Table 1). In our search, we considered well-substantiated parulid hybrids whose parentage is/was largely undisputed, hybrids with disputed parentage, and several instances of partial uncertain parentage where the identity of at least one parent was known. We classified a hybrid as having undisputed parentage based on cumulative strength of evidence, including certainty of the original identification, certainty of later publications to elaborate on or correct the initial identification, and/or consensus among experts. We acknowledge that future genetic analysis may result in identification revisions for hybrids presently considered to have undisputed parentage. We included records from any time in the past to roughly February 2021 from a variety of sources, including written documentation and descriptions, photos, audio recordings, genetic evidence, and vouchered museum specimens (Table 1). Additionally, we limited our search to those hybrids whose parents spend at least part of the year in North America, including Canada, the United States, and Mexico. Although several hybrids in our compilation were identified outside of North America, they are included because the parentals occur in North America for at least a portion of their annual cycle. We focused on North American hybrids due to their extensive history of study and the higher likelihood that they will be photographed and posted to online sites such as eBird and Facebook Advanced Bird ID.
We reviewed all possible hybrids from published peer-reviewed sources, including Dunn and Garrett (1997), McCarthy (2006), and Willis et al. (2014), as well as web sources, including the Avian Hybrids website (https://avianhybrids.wordpress.com/parulidae/), eBird (https://ebird.org/home), the Facebook Advanced Bird ID group (https://www.facebook.com/groups/357272384368972), the Facebook Bird Hybrids of North America group (https://www.facebook.com/groups/2408162756138343), and others (Table 1). In eBird, we searched for all possible parulid hybrids by typing “warbler hybrid” (as well as comparable searches for species without “warbler” in the common name, e.g., “yellowthroat hybrid”) into the species search box and accessing all records for each hybrid combination. We additionally conducted a nonsystematic, ad hoc eBird search for obvious hybrids using the keyword “warbler sp.”; however, most birds with this designation are not hybrids, but rather species that observers were unable to identify for a variety of reasons (i.e., poor lighting, poor photos, observer inexperience) and this search yielded unusable and/or non-novel results only. Data and hybrid combinations suggested from Facebook bird groups allowed us to incorporate crowdsourced information from biologists and community scientists around the world; in most cases, crowdsourced suggestions were supported by published documentation and/or written descriptions, photos, and audio recordings available in eBird and/or on xeno-canto (https://www.xeno-canto.org/).
We coupled database and open platform search efforts with targeted searches in Google and Google Scholar for specific hybrid pairs of interest. For example, we searched “Connecticut Warbler and Mourning Warbler hybrid” and/or “Connecticut × Mourning Warbler” to attempt to retrieve more information about specific hybrids. Searches in Google Scholar focused our efforts on academic texts, whereas Google searches focused on bird observation websites and personal accounts (e.g., http://amazilia.net/images/Birds/NewWarblers/Hybrid_Warbler.htm). When possible, we obtained information about the presence of vouchered museum specimens and their identifiers by searching published sources and museum search engines (e.g., Arctos, www.arctosdb.org), or specific museum collections websites. We present our full compilation of parulid hybrids in Table 1, the evidence used to substantiate the hybrid identification, identifiers for vouchered museum specimens, and references. We acknowledge that other hybrid combinations likely exist that we failed to detect and that more parulid hybrid combinations await discovery.
We used hybrid pairings from Table 1 to summarize and analyze patterns of carotenoid and wing bar variation. Carotenoid coloration and wing bars are 2 variable traits relevant to our hybrid case study (Fig. 1a–d), and multiple studies have linked a causal gene (BCO2) to carotenoid pigmentation (Toews et al. 2016, Baiz et al. 2020).
For carotenoid analysis, we chose hybrids with notable body-wide or patch-specific carotenoid differences, as estimated by presence of yellow and orange colors, and we qualitatively catalogued and compared the extent of the carotenoid pigmentation among parentals and hybrid offspring (Table 2). EFG scored whether the extent of carotenoid distribution was more or less than “intermediate” in the hybrid compared to parental forms (Fig. 2a). The classification of “intermediate” was made with respect to both the spatial extent and saturation of yellow coloration, based on the assumption that “intermediate” coloration should have approximately midpoint color saturation and midpoint spatial extent of yellow, as described in Thompson et al. (2021). An individual was described as “more” or “less” than intermediate if either the saturation or the spatial extent was more or less than the midpoint between parents, respectively (Fig. 2a). Many cases were clear-cut, but in instances where scoring was unclear, we defaulted conservatively to the “intermediate” designation. We used the same evaluation criteria regardless of age and sex and attempted to infer what would constitute “intermediate” based only on hypothesized intermediate hybrid plumage relative to the 2 parentals. It is worth noting that, due to the potential importance of male pigmentation in speciation, bias may exist for divergent selection resulting in stronger mismatch in male-specific pigmentation mechanisms (Sibley 1957). This could result in hybrid males, as opposed to females and duller-plumaged immatures, with stronger parental bias. When available, we preferred photos for scoring, but well-described plumage traits were used when needed.
We chose wing bar pairings based on 3 criteria: (1) presence of wing bars on one parent and absence of wing bars on the other; (2) notable differences in wing bar extent between parental forms; and (3) notable differences in wing bar color between parental forms (e.g., yellow in one parent, white in the other). We defined wing bars as a line or lines of contrasting color (including white, yellow, or tan) across the middle of a bird's wings caused by markings on the coverts. Specifically, we considered wing bars to be any contrasting light edges of wing covert tips, notably differentiated from the color at the center of the coverts and edges along the lengthwise axis of the feather. EFG scored relevant pairings based on presence/absence of wing bars and whether wing bars had reduced width in hybrid offspring relative to the parent with more pronounced wing bars (Table 3; Fig. 2b). Width of wing bars was qualitatively defined based on the extent of pale tips in the relevant wing covert feathers, as demonstrated in Figure 2b. These evaluation criteria were chosen because in most cases it was difficult to score whether wing bars on hybrid offspring were “more than” or “less than” intermediate.
Hybrid field encounter
On 16 September 2020, MJA briefly observed an unusual warbler on the University of New Mexico (UNM) Main Campus at approximately 35.0831N, –106.6208W. Based on plumage patterns, particularly the presence of a pronounced collar and the bird's overall muted color, MJA tentatively identified the bird as an odd Cerulean Warbler (Setophaga cerulea), a species for which only 3 New Mexico records existed at the time. On 17 September 2020, MJA returned to the same location and found the bird foraging actively in a mixed flock of Wilson's (Cardellina pusilla) and Townsend's (Setophaga townsendi) warblers. The bird was observed and photographed for ∼1.5 h by all coauthors. The authors analyzed plumage features, coloration, and behavior extensively, debating many hybrid parent combinations, as well as the possibility that the bird was a hypoxanthic (color-depleted) Yellow Warbler or hypoxanthic member of the Black-throated Green (Setophaga virens) complex.
We noted a fairly large warbler with a thick and relatively long bill that was overall blue gray/green dorsally and white with darker streaking ventrally (Fig. 1a–d). We observed pronounced green/gray tones on the bird's head, nape, and back; pale and hardly noticeable white edging to the wing coverts; a bold white wing flag on each wing that extended down the primaries; a gray necklace of dense streaks separating the throat from the breast with a faint chestnut wash to parts of the base coloration of the necklace; gray flank streaks with a chestnut wash to the base coloration of the upper flanks; and faint yellow patches on an overall white breast, belly, and vent. The bird had white undertail coverts, and importantly, we noted that at least 50% of the 5 outermost rectrices were white (Fig. 1b–c). The bird seemed to respond to several species' calls and songs (including Black-throated Blue, Yellow, and Townsend's warblers) that we played back to it by approaching the sound source, but we felt like any conclusions drawn from these unsystematic playback attempts were equivocal. The original description of our encounter and photos taken by MJA can be found in the respective eBird checklists: https://ebird.org/checklist/S73710816 and https://ebird.org/checklist/S73717047. The full set of high-resolution images taken by MJA and KDO to document the hybrid is accessible on Dryad at https://doi.org/10.5061/dryad.4b8gthtbw.
The authors are affiliated with the Museum of Southwestern Biology (MSB) at the University of New Mexico and discussed the ethics of collecting the hybrid to obtain genetic samples. For permitting and logistical reasons, we did not collect the individual, nor was it possible to obtain feather samples for genetic analysis. We attempted to collect noninvasive fecal samples; however, we could not identify and distinguish fecal samples of our putative hybrid from other species in the flock. As such, we rely on detailed observations, photos, and analysis of plumage traits to substantiate our identification. A comparison of wing and tail colors and patterns between the hybrid and its putative parents is presented (Fig. 3) with comparable photographs of both the hybrid and specimens from the MSB (Table S1).
Hybridization in Parulidae
We documented 61 hybrid crosses among 44 North American parulid warbler species from the 1900s to present (Table 1). Of these, 56 hybrids had undisputed parentage, an additional 3 hybrids had disputed parentage or parentage that was called into question after the initial hybrid description (e.g., Parkes 1995), and 2 had uncertain parentage (Table 1). Consistent with intrageneric species richness within Parulidae, the genus Setophaga had the highest number of hybrid crosses (n = 40), followed by Geothlypis (n = 6), and then Vermivora (n = 5). We found 40 intrageneric hybrids and 16 intergeneric hybrids. Of hybrids with undisputed parentage, the species with the highest number of crosses was Yellow-rumped Warbler (S. coronata; n = 11), including 10 intrageneric crosses and one intergeneric cross. Magnolia Warbler (S. magnolia) had the second-highest number of hybrid crosses with n = 7 each. Most parulid species displayed higher rates of intrageneric than intergeneric crossing, but hybrid crosses involving the monotypic Black-and-white Warbler (Mniotilta varia) were necessarily intergeneric (all with Setophaga species) and 3 of 4 crosses involving Blue-winged Warbler (V. cyanoptera) were intergeneric (2 with Setophaga species and one with G. formosa; Table 1). One unusual hybrid was identified as a “triple hybrid” cross between a Brewster's Warbler (V. chrysoptera × V. cyanoptera) and a Chestnut-sided Warbler (S. pensylvanica), providing an example of both intrageneric and intergeneric hybridization (Table 1; Toews et al. 2018).
The most common method of diagnosing hybrids from past to present has been analysis and comparison of plumage types through field observations, photos, and handling of mist-netted birds; >96% (n = 59) of hybrid identifications have been made, at least in part, through plumage analysis (Table 1). More than 36% (n = 22) of hybrid crosses have been identified in part by morphological comparisons or formal analysis of morphology (i.e., principal component analysis and other statistical analyses), and ∼20% (n = 12) have used vocalizations to determine parental identity. Only ∼33% (n = 20) of the hybrids in our compilation have had identification confirmed by molecular analysis and/or observations of nesting behavior by known parents (Table 1). In a few instances, certain hybrids have been subjects of extensive and formal hybrid zone study, including crosses such as Mourning × MacGillivray's Warbler (Geothlypis philadelphia × G. tolmiei; Irwin et al. 2009), Hermit × Townsend's Warbler (Setophaga occidentalis × S. townsendi; Pearson 2000), and Golden-winged × Blue-winged Warbler (Vermivora chrysoptera × V. cyanoptera; e.g., Vallender et al. 2007; Table 1). A total of ∼36% (n = 22) of the hybrid crosses listed in Table 1 have representative specimens in museum collections; these vouchered specimens represent one method by which past hybrid descriptions might be extended in future research.
Carotenoid plumage pattern inheritance
Hybrid carotenoid inheritance was highly variable (Table 2). Of 23 pairings studied, 48% were scored as “less than intermediate” (however, the majority included examples where one parent lacked carotenoids), 39% were scored as “intermediate,” 9% were classified as “more than intermediate” but less than parentals, and one hybrid combination was variable (2 different individuals of Nashville × Tennessee Warbler [Leiothlypis ruficapilla × L. peregrina] had different levels of carotenoids present in different individuals; Parkes 1995). One prominent trend was that when one parent had carotenoid pigmentation and one parent lacked carotenoid pigmentation, hybrid offspring showed a significant reduction in carotenoids (all n = 6 pairings with “complete” magnitude of difference; Table 2). Notable examples of carotenoid reduction in hybrid offspring compared to parentals included genetically confirmed combinations such as Cerulean × Blue-winged Warbler (Toews et al. 2020) and Black-and-white × “Myrtle” Yellow-rumped Warbler (Vallender et al. 2009).
A unique pattern seen in hybrids whose parents had a strong difference in carotenoid pigmentation on the underparts was presence of asymmetrical and patchy pale-yellow coloration on a whitish background. Examples of this pattern include Black-throated Blue × Magnolia Warbler, Cerulean × Blue-winged Warbler (Tables 1–2), and our case study (Fig. 1a–d). Generally, offspring phenotypes were not predictable for pairings where one parent had far more carotenoid pigmentation than the other (i.e., n = 11 scored as “strong” magnitude of difference; Table 2). For these pairings, 45% were less than intermediate, 27% were intermediate, 18% were more than intermediate, and one hybrid combination was variable (2 individuals of Nashville × Tennessee Warbler, see above).
Wing bar inheritance
Based on 11 pairings examined, we found that hybrids strongly and consistently inherited wing bar patterns from parents (Table 3). Wing bars were present but reduced in hybrid offspring for all pairings where one parent had wing bars and the other did not (n = 9). The relative width of the wing bars on hybrids was always intermediate between the width of the 2 parents (as shown in Fig. 2b), but without specimens of hybrids and parents it was not possible to quantify the intensity of hybrid offspring relative to the parental with wing bars. In the 2 cases where both parents possessed wing bars differing only in extent, Blackburnian Warbler × Kirtland's Warbler (S. fusca × S. kirtlandii) and Blue-winged × Prairie Warbler (V. cyanoptera × S. discolor), hybrid wing bars were clearly more similar to those of one parent than the other (Table 3).
Identification of the UNM hybrid
Based on extensive analysis of size, shape, plumage features, color pattern, and general “gestalt,” we put forth the identification of Yellow Warbler × Black-throated Blue Warbler for the UNM hybrid individual (Fig. 1a–d). Our observation is the first photo-documented record of a hybrid between these 2 species, with one previous written description from Quebec, Canada (Ducharme and Lamontagne 1992).
Our synthesis revealed evidence of hybridization in >93% of New World warbler species (44 of 47), an increase from the previously estimated 72% (34 of 47 species; Willis et al. 2014). We found predictable inheritance of plumage patterns between hybrid offspring and parental forms in 2 focal traits, carotenoid coloration and wing bar patterns, and these results support our identification of an unusual Yellow × Black-throated Blue Warbler hybrid observed in Albuquerque, New Mexico, USA (Fig. 1a–d). Our analysis suggests that the phenotypes of rare hybrid warblers likely have some degree of predictability.
Inheritance of plumage characteristics
Warbler hybrids have long been central to studies of patterns and modes of inheritance (e.g., work with Vermivora species; Toews et al. 2016), and recent work has highlighted the influence of a few genes that have large effects on phenotype. For example, the high prevalence of BCO2 introgression (Baiz et al. 2020) and the apparent near dominance of “no-carotenoid” phenotypes (Table 2) suggests that whole-body carotenoid deposition may be governed by few nearly recessive genes. One warbler species with elevated levels of BCO2 evolution and gene tree discordance is the Yellow Warbler (Baiz et al. 2020), one of the parentals of our putative Yellow × Black-throated Blue Warbler hybrid, and these evolutionary patterns could partly explain plumage coloration patterns in this hybrid cross.
Although BCO2 may be an important gene underlying plumage patterns, its effects are by no means universal. In our analysis we found many cases where pairings with strong differences in parental carotenoids produced intermediate hybrids. This is exemplified by Black-throated Gray × Grace's Warbler (Setophaga nigrescens × S. graciae), where both parentals have similar BCO2 genes but major differences in levels of carotenoid pigmentation (Baiz et al. 2020). Similarly, despite American Redstart (Setophaga ruticilla) having “Yellow Warbler–like” BCO2, the hybrid with Magnolia Warbler shows very little yellow on the breast and belly (Brennan et al. 2021). Additionally, admixture mapping and inferred polygenic inheritance of phenotypes in “Audubon's” × “Myrtle” Yellow-rumped Warblers (Setophaga coronata complex) has revealed a dozen other genetic regions associated with throat coloration, including a candidate gene for carotenoid transport (SCARF2), but not BCO2 or ASIP (Brelsford et al. 2017). Polygenic modes of inheritance for carotenoid genes have also been found in the Northern Flicker (Colaptes auratus; Hudon et al. 2015, Aguillon et al. 2021).
Our carotenoid coloration analysis revealed that pairings where only one parent possessed carotenoids resulted in hybrids with substantially less yellow than the bright parent (Fig. 2; Table 2). This pattern is consistent with predictions of parental bias, where a hybrid's bivariate phenotype tends to resemble one parent about 50% more than the other, or mismatch due to different traits having dominance in conflicting directions (Thompson et al. 2021). Even in the most extreme examples (e.g., Blue-winged × Cerulean Warbler; Toews et al. 2020), we never observed complete dominance of non-carotenoid pigments. In pairings where both parents had carotenoid pigmentation, even when one parent had limited carotenoid pigmentation and the other had extensive yellow, coloration of hybrid offspring varied substantially (e.g., Magnolia × Palm Warbler and Magnolia × Chestnut-sided Warbler; Table 2). In such cases, it was therefore difficult to consistently predict carotenoid extent in hybrid phenotype. Hybrids did not show carotenoid pigmentation beyond what was apparent in parentals, consistent with the contradictory character approach (Rohwer 1994).
Wing bar inheritance followed a more consistent pattern; in most cases, the wing bars of hybrid offspring appeared close to strictly intermediate between parental forms (Fig. 2; Table 3). We found 2 apparent exceptions where wing bar width in hybrid offspring was closer to one parent than the other: one was Blue-winged × Prairie Warbler, which appeared more phenotypically similar to the Prairie Warbler parent (Tables 1 and 3); the other was Blackburnian × Kirtland's Warbler, which was much more similar to the Blackburnian Warbler parent (Table 3). Notably, we did not find any examples where 2 wing-barred parentals produced hybrid offspring without wing bars; even in cases where one parent had wing bars and the other lacked wing bars, offspring always possessed some degree of wing bars. This suggests that while wing bars may be reduced in hybrid offspring relative to parents, the presence or absence of wing bars in hybrids is governed by the presence or absence in one of the parental forms. The observed patterns of wing bar inheritance may be important for ornithologists and bird watchers to use when identifying hybrids in the field, especially in cases where the use of molecular markers is not possible.
In our analysis, we used a relatively coarse and subjective method to estimate degree of carotenoid saturation and wing bar extent. Many technical methods exist for precisely quantifying plumage coloration (e.g., Mason and Bowie 2020); however, we did not pursue such methods, as most of the hybrids we studied are not associated with vouchered museum specimens. We note that future studies could build upon our work by using computational methods to determine the relative extent of color and level of carotenoids in hybrid plumage, as well as modes of inheritance of both carotenoids and wing bars.
Identification of a striking Yellow × Black-throated Blue Warbler hybrid
The identification of the UNM hybrid was made by extensive study and consideration of many structural and plumage features, including overall size and shape of body and bill, coloration, and distinctive tail and wing patterns. Despite considerations of several structural and plumage features, 2 traits were critical in shaping our identification of this hybrid. The first trait was the extent of white in the tail. Our observed hybrid showed at least 50% white in the 5 outermost rectrices (Fig. 1b–c). Although no warbler species has extensive white on 5 rectrices, the tail pattern is identical to that of a Yellow Warbler lacking carotenoid pigmentation (Fig. 1b–c, e; Fig. 3a–b). The angle and shape of the coloration in the rectrices is also consistent with the angle and shape of rectrix coloration in the Yellow Warbler (Fig. 3a–b). Additionally, our bird's overall size, large bill, and “blank” face are consistent with the general appearance of Yellow Warbler. We therefore consider the combination of traits observed, particularly tail plumage patterns, to be diagnostic of a Yellow Warbler parent.
The second trait instrumental in cementing our identification of this hybrid was the white “wing flag” concentrated at the base of the primaries and faintly extending onto the secondaries, as seen on the folded wing (Fig. 1a, d). Among the Parulidae, this distinctive plumage feature is unique to Black-throated Blue Warbler (Fig. 1f). Although American Redstart has a similar wing flag colored with carotenoids that is evenly distributed across the bases of both the primaries and the secondaries, these patterns were not present in our hybrid. We note that the feather-specific pattern of the wing flag on our hybrid is not perfectly consistent with the wing flag pattern of Black-throated Blue Warbler, and we attribute this to mixed parentage with Yellow Warbler (Fig. 3c-e). This observed reduction of wing flag extent is additionally consistent with both documented examples of American Redstart × Magnolia Warbler (Brennan et al. 2021). The extent of smudgy yellow on our hybrid suggests one parent had extensive carotenoid pigmentation while the other had little to none, and the results of our analysis of wing bar inheritance strongly support that both parents of the UNM hybrid lacked wing bars (Tables 2–3). Additionally, the white instead of yellow coloration in the outer rectrices of the UNM hybrid relative to Yellow Warbler is consistent with the hybrid pairing of American Redstart × Northern Parula (Burleigh 1944), and the general pattern of carotenoid suppression in hybrid offspring when one parent lacks carotenoids. Taken together, several lines of evidence support Black-throated Blue Warbler as the second parent to our bird.
It is important to consider all possible parental combinations and alternatives. We initially debated whether the UNM hybrid was a hypoxanthic Yellow Warbler; however, a hypoxanthic individual would not display a pronounced white wing flag or the Black-throated Blue–like pattern of white on the underwing (Fig. 3f–h). A few other subtle plumage features appear to superficially complicate parental identification. First, the faint chestnut wash on the throat and flanks appears consistent with Bay-breasted Warbler (Setophaga castanea) or Chestnut-sided Warbler (Fig. 1). However, pigmentation that appears to be purely eumelanin-based (i.e., black feathers) can conceal pheomelanin (i.e., rufous). This pheomelanin-derived coloration may then be visible in hybrid offspring, as has been shown in titmouse hybrids (Paridae: Baeolophus; Curry and Patten 2014). The chestnut in the flanks on our hybrid may also be explained by the presence of pheomelanin chestnut streaks on the breast and flanks of the Yellow Warbler parent, the distribution of which may have been altered by unknown pigmentation mechanisms of the other parent. Additionally, lack of wing bars and presence of a wing flag, as well as gray to black coloration of the underpart streaking of the UNM hybrid, further rule out Bay-breasted Warbler as a parental candidate. A number of these same features also rule out Chestnut-sided Warbler. Finally, although the overall appearance and “necklace” of streaks on the throat of the UNM bird are strongly reminiscent of Cerulean Warbler, the position of throat streaking on the UNM bird is higher on the throat than that of Cerulean Warbler, and the lack of wing bars and presence of a wing flag rule out Cerulean Warbler as a possible parent. Thoughtful discussion about the UNM hybrid, including rationale for and against many parental candidates, can be accessed in our permanently archived Facebook Advanced Birding ID post (Table 1).
Likelihood of Yellow × Black-throated Blue Warbler hybridization
Hybridization of Yellow × Black-throated Blue Warbler has been reported once before from Havre-Saint-Pierre, Quebec, Canada, in 1992, where the 2 species' breeding ranges overlap (Ducharme and Lamontagne 1992). However, this pair seems especially unlikely in New Mexico, southwestern USA. Although the Yellow Warbler is a common migrant and regular riparian-breeding species in parts of New Mexico, the Black-throated Blue Warbler does not migrate through or breed in the state; it is considered casual in spring and very rare in fall, with most reports from the Rio Grande Valley eastward (Parmeter et al. 2002). It is therefore remarkable that a Yellow × Black-throated Blue Warbler hybrid was found so far outside of the core range of one of its putative parents, although this pattern is consistent with other rare hybrids, which often result from breeding events between one parent within its core range and one parent on the edge or outside of its breeding range (Short 1969). Accordingly, the only other report of Yellow × Black-throated Blue Warbler is from Havre-Saint-Pierre, Quebec, Canada, a locality northeast of the expected core range of the Black-throated Blue Warbler (Ducharme and Lamontagne 1992).
Based on known distributions, breeding ranges, and migratory routes of both Yellow and Black-throated Blue Warblers, we posit that a vagrant Black-throated Blue Warbler occurring far west of its expected range may have bred with a Yellow Warbler in the northern United States or southern Canada (i.e., Montana, Alberta, or Saskatchewan, all three of which have late June records of singing Black-throated Blue Warbler; Cornell Lab of Ornithology 2021), and that the hybrid offspring followed a Yellow Warbler–like fall migration route through New Mexico en route to its nonbreeding grounds. While it is possible that the hybridization event occurred locally within New Mexico, the Black-throated Blue Warbler's higher-latitude, northerly breeding range suggests that a northern origin is more likely. Other possible breeding localities for this pair certainly exist, but without genetic or feather samples we could not confirm the site of origin for the UNM hybrid.
We note that Yellow Warbler and Black-throated Blue Warbler are 2 species with a relatively high number of documented hybrid pairings, especially compared to other species in the genus Setophaga (Table 1). In total, we found 5 instances of Yellow Warbler hybrids (3 intrageneric, 2 intergeneric) and 4 instances of Black-throated Blue Warbler hybrids (3 intrageneric, 1 intergeneric). Although this does not speak to the probability of hybridization between these 2 species, it may suggest that they have a higher propensity for hybridization than related warbler species. Propensity for hybridization and predictability of intrageneric and intergeneric hybridization within Parulidae are promising subjects for future work.
Boundaries of reproductive isolation among warblers
Species boundaries between and among warblers appear porous, with increasingly documented rates of hybridization between distantly related taxa (McCarthy 2006, Willis et al. 2014). This is detectable in genomic data, particularly with excess introgression of the carotenoid-related gene BCO2 (Baiz et al. 2020). The fact that many hybrid zones occur between sister pairs (Krosby and Rohwer 2010, Toews et al. 2016, Brelsford et al. 2017), together with the likely introgression of the gene BCO2 between distantly related species, suggests that some hybrids—even rare combinations—may provide fodder for adaptive introgression. Simulations have confirmed that a few migrants are sufficient to provide novel adaptive alleles to populations (Galloway et al. 2020), and these events of unexpected gene flow between non-sister taxa may have helped fuel this colorful radiation. The importance of gene flow in rapid radiations has received increasing attention in recent years (Meier et al. 2017, Marques et al. 2019, Gillespie et al. 2020). Future modeling and bioinformatic developments may elucidate whether adaptive introgression events are more likely to be fueled by hybrid zones or rare hybrids. The imperfect reproductive isolation between species with dramatic plumage and vocal differences, as in the example of Yellow Warbler × Black-throated Blue Warbler, remains a curious pattern whose importance (or lack thereof) we have just begun to unveil.
In conclusion, we uncovered instances of hybridization in >93% of North American parulid warbler species and we found consistent patterns in carotenoid and wing bar inheritance between parental species pairs, consistent with past genomic studies of warbler pigmentation. We identified a striking cross of a Yellow × Black-throated Blue Warbler, whose identification was supported by evidence of plumage pattern inheritance without genetic markers. Our results may provide a framework for hybrid identification based on phenotype and they suggest several promising lines of future research on rare hybrids: (1) models and bioinformatic assessment of the prediction that rare hybrids may have helped fuel dramatic color patterns seen in Parulidae (as in Galloway et al. 2020); (2) spectrographic and biochemical analysis of pigmentation concealing patterns (i.e., black eumelanins concealing rufous pheomelanin, as we hypothesized could be the case for the Black-throated Blue Warbler); and (3) quantification of hybridization propensity within Parulidae and heterogeneity of plumage pattern inheritance. These topics have the potential to deepen our understanding of evolutionary patterns, rapid radiations, and polygenic inheritance.
We thank M. Baumann, A. Jaramillo, M. Retter, the Macaulay Library at the Cornell Lab of Ornithology, and members of the Facebook Advanced Bird ID and Hybrid Birds of North America groups. Funding was provided by the University of New Mexico Graduate and Professional Student Association (JLW).
† Indicates joint first authorship