A New Species from a Hybrid Population

We have seen that hybrids can form a new species in a single step if they reproduce asexually or if they are polyploid. In such cases, the new species is reproductively isolated from its parents and gains an advantage through heterosis. Surprisingly, new species have sometimes arisen from a hybrid population of diploid, sexually reproducing organisms and, therefore, without the immediate reproductive isolation caused by asexuality or polyploidy.

The best understood example involves hybridization between two species of sunflowers, Helianthus annuus and H. petiolaris (Rieseberg et al. 1995a,b, 1996; Fig. WN22.4A). These grow on different soils (heavy clays vs. dry sandy soil, respectively), but their ranges overlap over much of the central and western United States (Fig. WN22.4B). At least three hybrid species have arisen independently and are each associated with a distinct habitat: H. anomalus and H. deserticola are restricted to the Great Basin desert of the southwestern United States, whereas H. paradoxus is endemic to brackish marshes in western Texas. These three species were originally thought to have a hybrid origin because they each carry a mixture of characteristics, some typical of H. annuus and others of H. petiolaris. This hypothesis was confirmed by genetic markers.

First-generation hybrids between H. annuus and H. petiolaris, and between these and hybrid species, are almost sterile, with low pollen fertility and seed set. This is partly due to multiple chromosomal rearrangements, which cause severe meiotic irregularities in heterozygotes (Box 12.2). At least ten rearrangements distinguish the parent species, and various combinations of these are found in homozygous form in the hybrid species. Remarkably, seven further rearrangements are fixed in H. anomalus. These arose by mutation in the hybrid population and contributed to its emerging reproductive isolation. The formation of these hybrid species has been mimicked by artificial crosses between H. annuus and H. petiolaris (Rieseberg et al. 1996). These separate hybrid lineages were propagated for five generations and then scored for 197 genetic markers. As expected, there was almost no introgression in regions that had different chromosome arrangements, because in those locations recombination is suppressed (see Fig. WN22.4C). Even in regions with the same gene order, however, introgression was much more variable than expected. Most regions had fewer H. petiolaris markers than expected, but some had more (see Fig. WN22.4C).

These patterns were very similar between the three replicates. Because the sunflowers were reared in the greenhouse with little mortality, they reflect selection for increased pollen fertility (a form of sexual selection). (Regions of chromosomes carrying alleles that reduce fertility in a new genetic background introgress less, and vice versa.) Indeed, over five generations, fertility rose from less than 4% to more than 90%. Remarkably, the pattern of introgression seen in these experiments was similar to that in the natural hybrid species, H. anomalus (Fig. WN22.4C). Thus, the selection for fertile hybrids that was measured in these greenhouse experiments had also contributed to the origin of a hybrid species in nature.

Within H. anomalus, in regions that do not differ in chromosome arrangement from either parent species, blocks of genome inherited from one or the other parent are quite large (Fig. WN22.4D). This implies that hybrid species were established quite quickly—within 20–40 generations—so that recombination did not have time to break up these blocks (Ungerer et al. 1998). In cases such as this, a fit hybrid genotype is rapidly fixed by selection and is then partially reproductively isolated from its progenitor species by chromosomal rearrangements and other kinds of incompatibility. The process is easiest to envisage if it occurs in a temporarily isolated population (i.e., in allopatry). When the new genotype comes back into contact with one or both parental species, prezygotic isolation may be reinforced, as discussed below. If the new genotype becomes adapted to a distinct ecological niche, as in Helianthus, then it may survive and spread over a wider area.

We have described hybrid speciation in some detail, because it is relatively well understood, rather than because it is common (for reviews, see Rieseberg 1997). Polyploidy and parthenogenesis aside, there are fewer than a dozen well-documented examples. Indeed, if this mode of speciation were widespread, it would be impossible to classify species into a tree-like phylogeny (Fig. WN22.5): Species would typically descend from several parent lineages and not just one.