By definition the genetic system of a species includes all intrinsic aspects of its biology which serve to generate its genetic variability and to regulate its evolutionary responses to selection. Hence, variations in genetic system parameters may have rather profound regulatory effects on evolutionary processes in the lineages they characterize. And since genetic system parameters are themselves genetically determined, these may also show interesting patterns of evolutionary change. Unfortunately, because genetic systems do not fossilize, their past actions and history can be reconstructed only from their present manifestations. However, given independent data on the phylogeny of a radiation (derived through paleontology, phenetics, and various comparative studies) and the details of variations in genetic system parameters of its present species, one can deduce and construct a conceptual model for the genetic system of an ancestral species and its subsequent evolutionary derivations. The model should logically account for the observed phylogenetic relations and present genetic system parameters in the species derived from this ancestry. If the deduced model for the genetic system and its evolution is predictive and general enough in its application, its validity can then be tested against other radiations: those which show similar phylogenetic patterns should reveal predicted genetic system parameters when examined, and vice versa. The more frequently such correlations are observed, the w more faith one may have in the validity of the model. Such a comparative and model building approach will be used here to elucidate possible roles of genetic system parameters in the speciation process.
All responsible evolutionists now accept the idea that if a species is divided geographically by an absolute barrier to gene flow, the separated populations in their independent evolutionary responses will eventually differentiate enough to become reproductively isolated if they should come together again (Mayr, 1963). This generally slow process of allopatric speciation presumably requires no specializations of a lineage's genetic system beyond those involved in the evolutionary maintenance of adaptive genotypes, and will serve as a "control" to help identify the possible functions of variable genetic system parameters associated with more rapid modes of speciation. Bush (1975), in a comparative review, distinguished three potentially faster kinds of speciation besides the slow allopatric mode he termed "la." These are: "1b," speciation by allopatric "founder" populations (Mayr, 1970); "II," parapatric speciation (usually "stasipatric," involving chromosomal differentiations [White, 1968, 1969, 1973]); and "III," sympatric speciation (usually involving a change in host-specificity by parasites). Here I will apply the comparative and model building approach to genetic system parameters involved in parapatric speciation.
Workers as diverse as Goldschmidt (1940) and Mayr (1969) have noted that the fixation of chromosomal rearrangements of kinds usually not found as intrapopulation polymorphisms between species of a lineage correlates with the apparently rapid evolution of isolation by these species without any obvious history of their geographic isolation. Comparative studies of all vertebrates by Wilson, et al. (1975) and detailed studies at the familial level in mammals (Todd, 1970, in canids) and in lizards (Paull, et al., 1976, in iguanids) have documented what Goldschmidt (1940) noted: that such non-allopatric "chromosomal" speciation is alsoassociated with an apparently rapid phyletic evolution of chromosomally variable lineages in comparison to chromosomally conservative and allo-patrically speciating lineages. In fact, this correlation is pronounced enough in some groups that it led Goldschmidt (1940) and Todd (1970) to present bizarre mechanisms to account for it: the "systemic mutation" and "hopeful monster" of Goldschmidt and Todd's "karyotypic fissioning" and "genetic potentiation").
More realistic attempts to explain functions of chromosomal differentiation and other aspects of genetic systems in rapid speciation have taken into account the apparent concentration of this kind of chromosomal variability in lineages which have intrinsically subdivided populations, and therefore effectively small deme sizes, because of limited vagility or other behavioral characteristics which achieve the same effect (White, 1968, 1969, 1973; Mayr, 1969, 1970; Arnason, 1972; Bush, 1975; Wilson, et al., 1975). All of these models agree that genetic drift (random sampling error) in effectively small local populations will occasionally allow a chromosomal rearrangement to become fixed in a deme which significantly reduces the reproductive fitness of heterozygotes for it by increasing their meiotic malassortment. All models assume that the rearrangement does not reduce the fitness of homozygous carriers for it with respect to homozygous ancestral types. Also, excepting White (White, et al., 1967; White, 1968, 1969), most authors believe that sufficient isolation to allow such chance fixations can occur only on a species' periphery where a differentiating population would be temporarily isolated and could expand into previously unoccupied areas (Spurway, 1953; Spurway and Callan, 1960; Wallace, 1959; Key, 1968; Mayr, 1969, 1970; Patton, 1969; Bush, 1975). Many also seem to believe that the rearrangement must be pushed to fixation by meiotic drive (White, 1973) or that it must confer some immediate selective advantage in a developmental sense to aid the fixation and to allow the mutant population to expand away from its parent species' range into previously unoccupied territory (Bush, 1975; Wilson, et al., 1975). Not only must a rare rearrangement be fixed by a rare chance event, but this doubly rare rearrangement must at the same time confer a significant developmental advantage to the individuals carrying it.1 Of course, completion of the speciation involves more than just fixing the chromosomal difference.
Once a deme becomes chromosomally differentiated from its ancestral species, the reduced reproductive fitness of chromosomally heterozygous hybrids is assumed to serve in further contacts with the parental species to favor the completion . of reproductive isolation. Either one or both of the following mechanisms are supposed to be involved in this process:
The reduced fitness cannot by itself ever serve as a complete barrier to the transmission of genes from one chromosomal homozygote to the other, because this would require the heterozygote to be completely sterile, which would of course lead to elimination of the rearrangement in its first generation of heterozygosity. On the other hand, short of postulating Goldschmidtiam hopeful monsters or complete allopatric isolation (where the chromosomal rearrangement would have no obvious function), clearly the rearrangement must somehow work in situations where parental and derived populations remain in contact, to allow complete reproductive isolation to be evolved more rapidly than it could be achieved with' the otherwise necessary allopatric isolation. Superficially, selection against hybridization seems to offer an easy mechanism to achieve this result (e.g. see Bush, 1975).
However, as I will show below, when genetic system parameters involved in chromosomal differentiation, hybridization, speciation, and rapid phyletic evolution are compared in actual and theoretical detail with similar parameters of allopatrically speciating lineages, the speciation model summarized above is seen to need many revisions and modifications. Quite unexpected effects deriving from limited vagility and the reduced fitness of chromosomal heterozygotes combine to acheive a seemingly complete block to gene flow between the respectively homozygous populations without either complete hybrid sterility or any need for premating isolation. In fact, quite paradoxically to the intuitive idea that hybridization (specifically parapatric hybridization-Woodruff, 1973; Bush, 1975) will favor the evolution of premating isolation, the evidence suggests that the hybridization will serve actually to delay the evolution of premating isolation. The study also suggests modifications and clarifications to other aspects of the chromosomal speciation model and provides an explanation for and testable predictions about the frequently rapid phyletic change associated with chromosomal speciation.
[CHROMOSOMAL VARIATION AND EVOLUTION IN SCELOPORINE IGUANID LIZARDS]