mechanisms

The Population Genetics Approach To Speciation

If you survey the various modes and models of speciation that have been proposed over the years, you should come away from such a survey with a vague feeling that you still don't know very much about speciation. The reason is that, in spite of all the time and effort put into speciation theory (especially by Ernst Mayr; e. g., Mayr, 1970), we still have only a vague set of scenarios (see review by Bush, 1975). These scenarios say little about the form or extent of genetic differentiation between populations to be expected during and after speciation. This is compounded by the fact that studies of interspecific genetic differentiation have found few, if any, trends.

The heart of the problem is that speciation theory has developed in partial isolation from theoretical population genetics. In fact, with the exception of the relationship between the founder effect and genetic drift, it is almost as if the two areas of evolutionary biology have no relationship at all. Part of the reason for this is undoubtedly historical; speciation theory having been developed by field biologists and taxonomists, while population genetics was developed by experimentalists and theoreticians. The second factor is that a survey of reproductive isolating mechanisms reveals such a diversity of phenomena which keep species distinct, that one could conclude that there are no general processes in speciation.

Starting with two papers published in 1980 (Templeton, 1980a, b; 1981; 1982b), Alan Templeton of Washington University (St. Louis) has attempted to rectify this situation. He has effectively bridged the gap between classical speciation theory and population genetics, with the result that we now have a more coherent, and predictive, classification of speciation mechanisms.

Templeton begins by defining several categories of speciation. They are organized as follows:

Transilience Modes

These are speciation mechanisms in which the isolating barriers depend upon genetic differences in which the intermediate stages are unstable (i. e., unstable equilibria). All of these modes require some process either in addition or in opposition to natural selection.

A. Genetic Transilience: is based on a strongly epistatic system with a few major genes. This results in multiple adaptive peaks (see Wright's Shifting-Balance Process), and speciation is brought about by a non-selective process causing a peak shift. The requirements for a genetic transilience are fairly stringent. It not only requires a genetic system like that described above, but also depends upon reorganization of the genome via inbreeding to a degree that some level of genetic variability must be maintained. These requirements are so restrictive that, for example, the majority of founder events do not lead to speciation via a genetic transilience.

B. Chromosomal Transilience: occurs when variant chromosomal arrangements, characterized by reduced fitness in the heterozygotes (underdominance), become fixed in different populations. Species which are likely to undergo this process would have subdivided populations with relatively high levels of inbreeding.

C. Hybrid Maintenance: Hybridization leads to offspring that perpetuate themselves as an independent lineage. The most common mechanism would be through hybridization accompanied by polyploidy. Another possibility would be apomixis or parthenogenesis

D. Hybrid Recombination: this occurs when the genetic instabilities induced by hybridization result in the formation of a novel, viable genetic arrangement that is incompatible with either of the parental genomes.

Divergence Modes

In divergence modes speciation occurs in a more continuous fashion, with natural selection being the main driving force.

A. Adaptive Divergence: This occurs when the population is subdivided by some extrinsic barrier. Subsequent micro- evolutionary changes cause the accumulation of enough genetic differences to bring about reproductive isolation. The extrinsic barriers need not be purely geographical, and Templeton (1980a) warns against equating this category with the classical allopatric speciation scenario.

B. Clinal Divergence: This occurs when populations are arranged along some geographical (and selective) cline, and there is sufficient isolation by distance to render the homogenizing effects of migration a negligible factor. This model of speciation has been developed in detail by Endler (1973, 1977), who described it as parapatric speciation.

C. Habitat Divergence: This mode involves a species inhabiting a heterogeneous environment. Microhabitat specific selective pressures, accompanied by habitat selection and/or assortative mating result in speciation without isolation by distance.

It is important to note that these mechanisms are not mutually exclusive; some speciation events may involve more than one mechanism. Templeton offers the example of a founder event causing a genetic transilience, which is coupled by the fixation of a chromosomal variant (Chromosomal Transilience), and these are followed by an Adaptive Divergence associated with selective pressures in the new habitat.

The probability of these modes of speciation depends on the ancestral population structure and the type of separation associated with the process of speciation. These comparisons are made for some of the categories in the following figure taken from Templeton (1980). In addition, the figure shows which of these mechanisms may be found involved with the same speciation event.

The relative likelihood of various forms of speciation according to Templeton (1980a):

Type of Split	Adaptive Divergence	Genetic Transilience	Chromosomal Transilience	Clinal Divergence
Founder Event	most likely if original population is subdivided	most likely if original population is panmictic	most likely if original population is subdivided	very unlikely
Large Subdivision	most likely if original population is subdivided	very unlikely	most likely if original population is subdivided	very unlikely
None	very unlikely	very unlikely	most likely if original population is subdivided	most likely if original population is subdivided

Note: "subdivided": population broken up into demes linked by migration; "panmictic": population a single randomly mating aggregate.

Summary:

Adaptive Divergence: Will occur throughout a range of population structures (from subdivided to panmictic), but will most likely occur if the original population is subdivided. This form of speciation requires some form of separation between populations.

Genetic Transilience: Will occur only in founder event situations, and is most likely where the ancestral population is panmictic.

Chromosomal Transilience: Because of the underdominant effects of most chromosomal variants, this will most likely occur via a founder event in which the founders come from an already subdivided population.

Clinal Divergence: By definition, this form occurs when there is no separation among populations, but it is most likely to occur in populations which are already subdivided into semi-isolated demes.