Hybridization is a phenomenon of basic relevance in biology (McCarthy 2006), and zoologists have often considered it a sporadic event (Mayr 1963). Hybridization, the mating between two different species that results in offspring, might be considered a rare phenomenon because it is geographically restricted and because hybrids represent between 0.1 and 5% of populations (Mallet 2005, Arnold 1997, 2006). Nevertheless, it has been suggested that about 10% of all animal species hybridize naturally (Mallet 2005). With the advance of molecular techniques, the number of reports on natural hybridization and hybrid speciation has increased in the past decades (Brelsford 2011).

Hybridization can occur in a contact zone, there is, in areas where the ranges of two populations overlap; and those can become a hybrid zone, where hybridization occurs regularly (McCarthy 2006). A hybrid can be defined as an individual mixing the physical traits of individuals of two different but not necessarily related taxa. Hybrids can express the traits of their parents in three ways: combination, when a trait resembles only one of the parents; intermediacy, when the traits seem to express a mix between the parents; and heterotic, when the hybrid presents a trait well outside of the range of parental variation (McCarthy 2006).

McCarthy (2006) listed all reported cases of hybridization in the family Corvidae. In this family, natural hybridization is rare (Pitelka et al. 1956). Even so, one of the most complex case of hybridization involves the interbreeding between Corvus corone corone Linnaeus, 1758 and Corvus corone cornix Linnaeus, 1758 with an extensive hybrid zone (Picozzi 1976, Rolando 1993, Haas et al. 2009, Wolf et al. 2010, Brodin et al. 2013).

New World jays do not hybridize as often or as extensively as Old-World ones. Some records mention occasional hybridization involving congeneric species, for instance Cyanocorax morio (Wagler, 1829) and Cyanocorax formosus (Swainson, 1827) (Pitelka et al. 1956), two morphologically distinct species, whereas others involve hybridization between sister species, for example C. formosus and Cyanocorax colliei (Vigors, 1829) (McCarthy 2006, Anjos and de Juana 2018) and Cyanocitta cristata (Linnaeus, 1758) and Cyanocitta stelleri (Gmelin, 1788) (Williams and Wheat 1971). Hybridization in captivity has also been observed, comprising almost half of the known cases, more often when there is no conspecific individual to pair up with (Hardy and Wheat 1982), for instance between Cyanocorax yncas (Boddaert, 1783) and C. cristata (Pulich and Dellinger, 1981).

The Neotropical jays belonging to the genus Cyanocorax Boie, 1826 comprise a very distinct group of New World corvids. Cyanocorax chrysops (Vieillot, 1818), the Plush-crested Jay, is a polytypic species widely distributed in South America, ranging from northern Brazil to northern Argentina, with four subspecies currently accepted: C. c. chrysops (Vieillot, 1818), C. C. diesingii Pelzeln, 1856, C. C. insperatus Pinto and Camargo, 1961 and C. C. tucumanus Cabanis, 1883. This species is found in various types of lowland forests and temperate rainforest, also inhabiting patches of forest in open areas, riparian forests and even disturbed areas. Cyanocorax cyanopogon (Wied, 1821), the White-naped Jay, in turn, inhabits mostly areas of ‘caatinga’ and ‘cerrado’, but it is also seen in secondary woodland, riparian forests, and in the borders of tropical deciduous forests. It is Brazilian endemic, distributed from the state of Maranhão to the state of Minas Gerais, also occurring in southeastern Pará, eastern Mato Grosso and Goiás (Madge and Burn 1994, Anjos et al. 2009, 2019a, b). Regarding the preferences of habitat occupancy, C. chrysops prefers to occupy the understory and the middle level, exhibiting some variation in habitat use (Uejima et al. 2012). On the contrary, the unique information available about the preferences of C. cyanopogon is the use of the ground when foraging (Barros et al. 2014). The information about the reproductive behavior of both species is very limited (Uejima et al. 2012).

Cyanocorax chrysops and C. cyanopogon are considered sister species (Bonaccorso et al. 2010) and due their morphological similarity they have been previously treated as conspecific (Pinto 1954, Pinto and Camargo 1961, Blake and Vaurie 1962). In addition, Pinto (1954) described a taxon based on a single specimen from Alagoas, Cyanocorax chrysops interpositus, considering it as an intermediate form between the two species. Later, Hardy (1969) questioned the validity of C. c. interpositus, arguing that the specimens assigned to this subspecies were actually molting, worn, and subadults of C. cyanopogon, discarding its intermediate position both in geographic and genetic sense. There are no reliable records or specimens confirming the existence of intermediate individuals between the two so far, but we found two museum specimens showing intermediate plumage patterns between C. chrysops and C. cyanopogon from two different sites in Brazil, which we identified as hybrids. Thus, our objectives were to describe these hybrid specimens and discuss the possible causes behind the interbreeding and the existence of contact zones.

We analyzed 80 specimens of C. chrysops and 137 of C. cyanopogon housed at Museu de Zoologia da Universidade de São Paulo (MZUSP) and Museu Paraense Emílio Goeldi (MPEG) (see Supplementary Material S1) in order to describe the plumage pattern of each hybrid and then compare it with the plumage of their parents. We analyzed plumage coloration following Smithe (1975) and Munsell (1994) color catalogues (hereafter S and M, respectively). In addition, we also analyzed records of both species on the Brazilian online database WikiAves (http://wikiaves.com.br) aiming to delimit a more precise range limit of both taxa, trying to identify putative contact zone areas.

We found two specimens, one from Três Lagoas, state of Mato Grosso do Sul (MZUSP 64191), and a second from Comendador Gomes, state of Minas Gerais (MZUSP 103009), showing intermediate plumage characters between C. cyanopogon and the nominate form C. c. chrysops. The most remarkable plumage characters exhibited by these particular specimens were: (1) back and wings dark grayish brown, (2) tail indigo blue or dark grayish brown + indigo blue and (3) nape light bluish gray/dull violaceous blue or white/lavender blue (Table 1). The main difference between the two is that the specimen from Três Lagoas has yellowish underparts whereas the specimen from Comendador Gomes shows white underparts, with each one being more similar to one of its parental species, resembling C. c. chrysops and C. cyanopogon, respectively. Following the classification given by McCarthy (2006), the hybrids showed both, intermediate traits (characters 1 and 3 in both hybrids and character 2 in the hybrid 2) and a combined trait (character 2 in the hybrid 1).

There are specimens of both species from Três Lagoas, C. c. chrysops (MZUSP 64192) and C. cyanopogon (MZUSP 73779) (Figs 1–3) in addition to records from WikiAves database; WA12379 (C. c. chrysops) and WA 3230616 (C. cyanopogon). The existence of these specimens and records demonstrates that parental forms come into contact in this region, one of the prerogatives to designate a true hybrid according to McCarthy (2006). Otherwise, in Comendador Gomes, a place near the São Paulo border, there are no specimens or records of both species rather than the hybrid.

Identifying the causes of hybridization requires understanding how ecology, demography and phenotype influence mate choice in a particular species (Willis 2013). Many factors could be involved in hybridization, but environment and behavior are known to be the main drivers causing this phenomenon (McCarthy 2006, Randler 2006); however, a close genetic relationship is considered an important factor leading to it (Randler 2006). In fact, it is been argued that species with small genetic distances are more prone to hybridize (Gholamhosseini et al. 2013), and sister taxa tend to interbreed more than non-sister taxa due their similar morphology (Randler 2002, 2004, 2006). There is no doubt about the close relationship between C. chrysops and C. cyanopogon based on morphology and molecular studies (Bonaccorso et al. 2010). Not only morphological similarity, but similarity in mating signals might be also a significant factor inducing interspecific pairing and hybridization (Gholamhosseini et al. 2013) and even the voice repertoire of the two seems to be very similar (Anjos 2019a), which would facilitate recognition of mating signals. The case of C. chrysops and C. cyanopogon match all these statements, revealing that multiple and related factors are involved and could be used to explain why they hybridize.

McCarthy (2006) pointed to breeding range overlap between C. chrysops and C. cyanopogon in southern Pará, but no specimen was mentioned or was found by us to substantiate his statement. Willis (1992), in turn, believed that the two taxa meet in western São Paulo, while Madge and Burn (1994) asserted that the two species meet in eastern Mato Grosso. Our findings agree with Willis (1992) since one of the specimens was collected at the border between the states of Mato Grosso do Sul and São Paulo (Fig. 4). Moreover, it is possible to identify two conspicuous areas where the ranges of both species overlap: southern Minas Gerais and eastern Mato Grosso do Sul, the very same areas where we found the hybrid specimens. However, eastern Mato Grosso and Pará could be potential contact areas due the proximity of the range limits of both taxa, especially for C. C. diesingii, which shows an easternmost record very close to the westernmost records of C. cyanopogon, and because C. C. diesingii inhabits patches of open drier vegetation (Pacheco and Olmos 2005, Lees et al. 2008, Anjos et al. 2009, Whittaker 2009, Olmos et al. 2011, Santos et al. 2011), environments very similar to those inhabited by C. cyanopogon.

We are not able to ascertain if there is a stable hybrid zone, and there is no information about an area where individuals of C. chrysops and C. cyanopogon interbreed regularly. Possible reasons for the paucity of records of intermediate individuals could come from the great similarity of the parental species, which are often erroneously identified: individuals of C. chrysops are misidentified as C. cyanopogon and vice-versa. When the parental taxa are similar, the phenotype of the offspring produced by hybridization might be also more similar to one of the parents, making it difficult to identify visually as intermediate so they could just have remained unnoticed. On the other hand, the lack of intermediate individuals could just reflect the rarity of the interbreeding, as hybridization between other species of Cyanocorax is rare. For instance, C. chrysops is sympatric with other congeneric species, C. cyanomelas (Vieillot, 1818) and C. caeruleus(Vieillot, 1818) in part of its range (Goodwin 1976, Madge and Burn 1994) and there is no evidence of hybridism between these species, which have a dissimilar morphology when compared to C. chrysops, also demonstrating that sympatry by itself may not be a factor leading to hybridization. Thus, without any other clear indication and based on data we have, we suppose that hybridization between the two taxa occurs only occasionally in contact zones.

Hybridization is more common in areas where two related species meet and one of them is rarer. Thus, restricted mate choice drives the interbreeding (Hubbs 1955, Mayr 1967, Short 1969, Gillespie 1985, Grant and Grant 1997, Wirtz 1999, 2000). In the absence of conspecifics, individuals have to “choose” between mating heterospecifically or not mating at all (Randler 2006), a hypothesis known as “making the best of a bad job” (Baker 1996). The occurrence of C. cyanopogon in Mato Grosso do Sul is rare (Godoi et al. 2013), and there are only three municipalities where this species was recorded in WikiAves database. Otherwise, there are several records of C. chrysops and museum specimens from the same state, proving that it is the commonest species there. On the other hand, in southwestern Minas Gerais both species seem to be uncommon, as C. cyanopogon is considered rare in northern São Paulo (Anjos 2019a), and C. chrysops is known by a handful of records in the same region.

Deforestation and habitat modification affect the range of the species, and those inhabiting non-forest environments are expanding their distribution throughout Brazil, for instance Patagioenas picazuro (Temminck, 1813) (Willis and Oniki 1987), Ramphastos toco Statius Müller, 1776 (Sick 1997), Rhynchotus rufescens (Temminck, 1815) (Willis and Oniki 2002), and Cyanocorax cristatellus (Temminck, 1823) (Lopes 2008). One of the effects of range expansions is that it increases the possibility of taxa coming into secondary contact due the lack of a natural barrier. For instance, an invading species, initially rare locally, might mate heterospecifically because conspecific partners are at low frequency (Rheindt and Edwards 2011, Duckworth and Semenov 2017). The same occurs in transitional zones, which often show disparities in the relative abundances of species (Jansson et al. 2007, Lepais et al. 2009, Larson et al. 2013). Cyanocorax cyanopogon is expanding its range southward due the habitat modification, reaching the state of Espírito Santo (Anjos 2019a). The range of C. chrysops, in turn, does not seem to be expanding. This species is not very vulnerable to habitat fragmentation, being present in urban areas in southern and southeastern Brazil (Anjos 2019b). Cyanocorax cyanopogon and C. chrysops have a very diverse diet (Madge and Burn 1994, Anjos et al. 2009, Barros et al. 2014) and exhibit plasticity and differences in habitat use (Uejima et al. 2012, Barros et al. 2014). Therefore, they probably do not compete for resources when co-occurring, so that these features also contribute to the ability of individuals to persist in disturbed or transitional areas.

The two sites where the hybrids were collected are transitional areas from drier vegetation formations (Cerrado, southwestern Minas Gerais and eastern Mato Grosso do Sul) to areas of humid forests (western São Paulo), and this may explain the differences in abundance of both species, easing the possibility of hybridization. Moreover, with the advance of deforestation, we detected putative new contact areas. Both species were recorded on some urban surroundings in the cities of Araxá (WA 1333137 and WA 745949), Belo Horizonte (WA 2006717 and WA 47372) in Minas Gerais, and Meridiano (WA 1702233 and WA 775584) in São Paulo state, showing that C. chrysops and C. cyanopogon can also meet in very modified environments and new events of hybridization are likely to occur.

In summary, the best predictors to explain the occurrence of hybrids between C. chrysops and C. cyanopogon are those from morphological and behavior similarity coupled with a putative rarity of conspecific pairs. Our findings highlight the need to conduct further studies in areas of range overlap to investigate the rate and frequency of hybridization and its consequences for the genetic integrity of the species involved.

We thank Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) (grant 2013/22026-1). LFS is supported by Conselho Nacional de Desenvolvimento Cientifico e Tecnológico research productivity fellowship (process #302291/2015-6). We are thankful to Alexandre Aleixo (MPEG) for the specimens’ loan and Gustavo Bravo for assisting with the procedures associated with the loan. Thiago Costa and two anonymous reviewers helped us to improve an earlier version of this manuscript.