From the point of view of a speciation geneticist, the study system Steeves et al. (2010) examine is fascinating. Habitat modification has facilitated the colonization of poaka (also known as the pied stilt; Himantopus himantopus leucocephalus), a generalist wading bird native to Australia, into New Zealand, where it has come into secondary contact with endemic kakī (also known as the black stilt; Himantopus novaezelandiae), a braided river specialist (Figs 1, 2 and 3). The ensuing hybridization between the two stilts provides a natural laboratory to help understand both the genetics of speciation and the evolutionary consequences of biological invasions that are accompanied by hybridization (Hewitt 1988). But because kakī are one of the world’s most endangered birds — only 98 individuals were tabulated in 2010, up from approximately 23 in 1981 — the study system is exceedingly fragile (Bird et al. 2009).
In plants, hybridization has been a major evolutionary force in the creation of new species through, for example, allopolyploidy (Grant 1961; Arnold 1997; Rieseberg 1997; Soltis & Soltis 2000; Vellend et al. 2007). In contrast, speciation via hybridization is thought to occur relatively rarely in animals (Mallet 2007). From a theoretical perspective, the most likely effects of hybridization in animals include everything from relatively low level amounts of introgression in which species boundaries are more or less maintained (Fitzpatrick et al. 2010), to the formation of a hybrid swarm (a population composed of a diversity of recombinants) (Harrison 1993). Its genome disintegrated, the formation of a hybrid swarm spells the genetic extinction of a hybridizing species if it has no other parental populations away from the swarm. Linking this back to Homo sapiens, human-induced disturbance is considered the most frequent underlying cause of extinction by hybridization (Seehausen et al. 2008), and, unfortunately, the literature is now replete with examples of this phenomenon (Rhymer & Simberloff 1996).
Anecdotally, kakī have been long thought by many biologists to already be genetically extinct, the few remnant individuals simply representing black-plumaged recombinant phenotypes in a hybrid swarm (Fig. 1). This is a reasonable assumption given the theoretical outcomes of biological invasions that include hybridization in animals (Fig. 4). Allendorf et al. (2001) identified three general outcomes of anthropogenically induced hybridization, with the caveat that outcomes in real biological situations exhibit a continuum: (i) hybridization without introgression; (ii) widespread introgression; and (iii) complete admixture. Hybridization without introgression usually occurs if hybridization events are very rare, with any fertile F1s subject to the same stochastic processes that influence the fitness of any individual in a population, or, if hybridization events are common, because F1s are sterile or suffer greatly reduced fitness. In the stilts, hybridization has been occurring for at least 50 years (approximately 15 generations), and breeding data, plus the presence of backcrossed individuals in the population, indicate F1s are fertile and reproducing. The second possible outcome, widespread introgression, denotes genetic admixture in and around the point of hybridization, but with the presence of geographically isolated or distant reservoir parental populations that are unaffected by hybridization. That the entire kakī distribution is found in the Upper Waitaki Basin negates the gene reservoir concept.
Sadly, for kakī, the most probable expected outcome is extinction of the species through the formation of a hybrid swarm, genetic disintegration of its genome and genetic assimilation into poaka (Fig. 4). In the context of a biological invasion, a hybrid swarm can form rapidly as invading alleles diffuse from the point of contact into the native species (Kot et al. 1996; Pialek & Barton 1997; Huxel 1999; Perry et al. 2001; Blum et al. 2010). The swarm is characterized, at least initially, by relatively high levels of nuclear and cytonuclear disequilibria because of the tendency for the two co-adapted genomes to have a selective advantage over hybrids (Nei & Li 1973; Szymura & Barton 1991; Arnold 1993; Barton & Gale 1993). If the swarm has a continued influx of invading migrants carrying parental, unrecombined genomes into the swarm, it is expected to decay as a function of migration rate, recombination rate, effective population size and the fitness of hybrids relative to parentals. With no source of parental native genomes and a steady stream of parental invasive genomes migrating into the swarm, genetic drift can lead to the total genetic loss of the native species unless some neutral or selectively advantageous alleles persist as assimilated new variation in the invading species’ genome (Barton 1979; Harrison 1990). There are many cases in which the level of genetic variation in a hybridizing invasive species increases, and in some instances this variation can be adaptive (Lee 2002; Lambrinos 2004; Richards 2005).
Contrary to the expected effects of hybridization on kakī, Steeves et al. (2010) found little evidence of cytoplasmic introgression, and the nuclear introgression uncovered was in the direction from endangered kakī into poaka. In other words, based on genetic variation at the mitochondrion and eight microsatellite loci, endangered kakī are persisting as a distinct evolutionary unit despite multiple generations of ‘successful’ hybridization with poaka. From a species genetic perspective, it is remarkable that species boundaries of these closely related taxa have been maintained in the face of hybridization.
Here, the discrepancy between the model-based expectation and empirical reality has its explanation in the biological details. The lack of cytoplasmic introgression between poaka and kakī can be explained in part by the dominance theory of Haldane’s rule (Haldane 1922), which states that when hybrids show reduced fitness, the effects are stronger in the heterogametic sex. That females are the heterogametic sex in birds and mitochondrial inheritance is matrilineal explain the limited mitochondrial introgression in part, especially since female hybrid stilts are known to suffer reduced fitness. Ironically, another probable factor in kakī survival is small population size. This contributed to a transient male-biased sex ratio of kakī during the late 1900s, so that most matings were between male kakī and female poaka or hybrids, combinations that are known to produce offspring of reduced fitness. The small population size also led to an exceptionally high mortality rate of stilts during the period in which the two species have been in contact and hybridizing. Thus, few of the offspring during the period of hybridization contributed to the current genetic pool.
The history of kakī conservation management is emblematic of the challenges that hybridization presents. In the case of kakī, hybridization was first ignored so as to focus efforts entirely on the reproductive success of pure-breeding pairs. An allozyme study indicating that dark hybrids were genetically intermediate led to the equal inclusion of dark hybrids in conservation efforts, but a subsequent study that suggested mixed pairs suffer from reduced fitness led to a renewed focus on pure-breeding kakī. Despite this tumultuous management history, the results from the Steeves et al. study and the increased number of kakī suggest the effort spent on its conservation is working. Still, it is premature to sound the victory bell. Molecular techniques used today by most conservation geneticists, including those employed by Steeves et al., allow one to sample only a fraction of the genome. Thus, poaka alleles that have introgressed into kakī may have gone undetected. Addressing that question will require a more comprehensive analysis of the genomic architecture of these hybridizing taxa. One of the great hopes of next-generation sequencing technologies is that genome-scale sequencing scans will be possible for studies of nonmodel taxa, including those on the brink of extinction (Lerner & Fleischer 2010).