The typical crows (Corvus) are mostly all-dark, blackish birds, but there are several examples of contrastingly two-toned plumage, the pale tone varying from grey to pure white. In the Eurasian crow complex, formed from west to east by black Corvus corone corone, black-and -grey Corvus corone cornix and black Corvus corone orientalis, and in the African complex of black-and-white Corvus albus and black Corvus edithae, the existence of hybrid zones shows the phylogenetic closeness of the morphs within each complex. Recent molecular studies have confirmed these relationships and shown C. corone (orientalis) to be close to black-and-white Corvus pectoralis, a southeastern Asian morph (Haring et al. 2012), and in Wallacea, black Corvus unicolor is close to black-and-white Corvus typicus (Jønsson et al. 2012), although hybrids have never been reported in these two cases. All the two-toned patterns produce a superficially similar pied appearance, with a wide pale area centred on the anterior part of the trunk and contrasting with the glossy black of the rest of the bird, including the face. However, the colour of the feather bases varies regionally; those of Eurasian birds are grey, those of African birds are white, and those of Wallacean birds are dark-grey, in my opinion a first suggestion of the different origins of these pattern alternations. Based on their phylogenetic trees, Haring et al. (2012) proposed that pied plumage appeared several times independently during crow evolution and that reversals to all-dark colour have also occurred. In support of the existence of such to-and-fro changes (each time constrained by intervening evolution), as far as it is true that ontogeny recapitulates phylogeny, I would add that Corvus dauuricus is usually pied in juvenile plumage, virtually all-dark in the first winter (although the two-toned pattern is still evident in close view), and later pied again (Cramp & Perrins 1994). Repeated gain and loss of colour patterns during phylogeny may depend on simple threshold mechanisms in birds (Price & Pavelka 1996). It is also interesting that a substantial gene flow appears across the Eurasian crow complex, with the possible addition of C. pectoralis (Haring et al. 2012). Therefore, the hybrid 'zones' would actually be much wider, and perhaps less stable, than might be suggested by phenotypic differences. The lack of genetic structuring may even indicate a genetic continuum across the whole Eurasian crow complex, with no 'pure' population. This would in turn mean that the observed 'hybrids' are only phenotypic intermediates between more stable and widespread 'pure' morphs within one species. As the low level of genetic variation weakens any explanation based on the adaptation of the different morphs to different habitats, their geographical distribution is enigmatic. Moreover, there is no indication of past genetic isolation in the Eurasian crow complex (references in Haring et al. 2012). This weakens the traditional, long-held explanation for the origin of different morphs: that their ancestors would have become isolated in different refugia during glacial periods (allopatric speciation). It is likely that two-toned and all-dark morphs have alternated during crow evolution and are maintained in phenotypically separate populations by a few genes, difficult to detect by researchers and ecologically unimportant in their allelic differences. This is not to say that colour patterns have no adaptive value: there is an obvious enhancement of the visual effect of plumage displays (described, for example, in Goodwin 1976 and Cramp & Perrins 1994) for the pied morphs compared with their all-dark counterparts; the latter might counterbalance this advantage through melanin-dependent benefits, such as increased feather resistance to deterioration (Goldstein et al. 2004) or an 'honest' signal of better antioxidant and immune systems (Moreno & Møller 2006). To have substantial gene flow across a phenotypically narrow and stable hybrid zone, most of the adaptive alleles must be independent of colour patterns, so that alleles from one parental population spread into the other through individuals that crows perceive by colour as completely belonging to the latter, although actually they are hybrids. When it was only a tentative explanation for narrow C. corone corone × C. corone cornix hybrid zones, assortative mating was mainly seen as arising from differential habitat selection (Saino & Villa 1992) rather than from the flocks being kept separate, even in the absence of ecological segregation, by selective aggression (Rolando & Laiolo 1994). However, based on preliminary field experiments for flock reactions to dummy birds of different morphs, Londei et al. (1994) outlined a testable model to explain the persistence of narrow hybrid zones: on both sides of the hybrid zone, a representative of the minority phenotype would be perceived to be more unusual the greater its distance from the median line of the zone, i.e. the line of the most frequent contact of different phenotypes. Increasing avoidance of the minority colour pattern would result. I re-proposed the same possible mechanism, integrated by imprinting, for the little-studied C. albus × C. edithae hybrid zone (Londei 2008). In addition to imprinting, social experience with non-relatives may contribute to assortative mating in birds (Freeberg 1996). Moreover, the rather overlooked finding that C. corone corone drive away C. corone cornix and hybrids from winter flocks in which the all-black birds are numerically dominant (Saino & Scatizzi 1991) suggests that something more than sexual selection is involved in the segregating mechanism, which might work all the year round through social interactions and result in marginalization of the minority phenotypes. Social selection, with many examples in birds (Price 2008), encompasses both sexual and non-sexual contexts, which often have no clear boundary (Lyon & Montgomerie 2012). Thus, several learning stages might be responsible for, and social selection the consequence of, the intolerance of unusual-looking conspecifics in crows. These birds seem sensitive to 'incongruent' stimuli in their cross-modal recognition of group members, based on familiarity (Kondo et al. 2012). Whether intolerance comes only from the members of the more frequent phenotype, or also from the minority birds, with previous 'imprinting' resulting in self-marginalization, is open to question and requires further work to produce mathematical, constraining representations of the model. To stress the importance of the local proportions of alternative phenotypes both at the time when the preference is learnt and afterwards, when different-looking conspecifics interact, I provisionally name this the reject-the-unusual model of segregation. This model is worth considering in the understanding of hybrid zones. Sexual imprinting has recently been emphasized as a likely cause of their limits, based on simulations through artificial neural networks: offered stylized representations of pure and hybrid phenotypes, the network learned the pure phenotypes better and faster (Brodin & Haas 2006) and the resulting homotypic preferences were able to maintain a stable hybrid zone (Brodin & Haas 2009). Field experiments might fill the gap between such suggestive network simulations and experiments on the cognitive abilities of captive crows. For example, 'hybrid' dummies might first be tested for their 'incongruence' (at different levels of detectability and novelty) with group-housed crows and then presented to free-living flocks at increasing distances from the median line of a hybrid zone, with intolerance being expected to increase. Simulations suggested to Brodin and Haas (2006, 2009) that 'speciation by perception' might work as parapatric as well as sympatric speciation. They proposed sexual imprinting in young individuals as the most likely factor in natural situations. The reject-the-unusual model of segregation leads to all of the learning mechanisms for the perception of 'incongruent' stimuli that may result in social selection. In the case of birds bearing a novel trait, marginalization might induce them to congregate in sufficient numbers in a suitable space to form a new population. They might colonize an area not yet reached by the original morph during a slow expansion (parapatric speciation), or a within-range but less preferred habitat already occupied by less competitive individuals of the original morph (sympatric speciation). In the latter case, the birds of the new morph would prevail, as their marginalization would only be due to their colour, virtually a neutral character in terms of non-social requirements. Crow preference for clear-cut colour patterns, as suggested by simulations (Brodin & Haas 2006) as well as by the mere aspect of several morphs, would contribute to make the new morph uniform across its range. A genetically determined mate preference has still been seen as a valid alternative to a learnt one in maintaining a hybrid zone narrow (Haas et al. 2010). However, as it would only require a new trait to be learnt and not a new trait to be genetically linked to a preference for it, a learnt preference, with effects on social selection, may in my opinion be a more likely explanation for the frequent alternation of colour patterns in crow evolution and their geographical distribution. Although socially selected traits often seem to be shaped by ecological factors (Price 2008), a new trait may become established with little cost through a learnt preference (Witte et al. 2000) and, in the case of crows, the small genetic difference between different morphs (Haas et al. 2009), and even greater genetic variation within than between morphs (Haring et al. 2012), suggests that learning-dependent social selection may have little ecological constraint. Trevor Price encouraged me to have my viewpoint published. Elisabeth Haring and an anonymous reviewer offered useful comments on a first draft of the manuscript. Paul F. Donald helped me to express concepts in readable English.
No takes yet. Share an insight, caveat, or question.
Tiziano Londei (2013) studied this question.
Synapse has enriched 2 closely related papers on similar clinical questions. Consider them for comparative context: