On the Origins of New Forms of Life

3: Karyotypes: Stasis and Variation

(Continued from the previous page)

An organism with a distinctive karyotype (i.e., with a distinctive set of chromosomes) will often be treated as a distinct species. Another characteristic that commonly results in two populations being treated as separate species -- or separate subspecies for that matter -- is morphological distinctness (physical differences that can be detected among specimens), a fact already discussed in the section on the definition of species. Any set of physical traits may be involved, but most commonly such traits are ones detectable by visual inspection. Populations distinguished on the basis of morphological traits are so common and are mentioned so often in subsequent discussion that they will be designated by a special name: somasets. Types distinguished on such a basis will be termed somatypes. There will also be a special name for populations and types distinguished on the basis of karyotypes, but first a few technicalities.

Cells and Chromosomes. The cell is to biology what the atom is to chemistry. All living things other than viruses are composed of cells. A cell is a membrane-bounded compartment, usually microscopic, filled with a complex liquid called cytoplasm. Within the cytoplasm are various structures with specific functions. Broadly speaking, all cellular organisms can be divided into two categories, bacteria and eukaryotes (viruses, which are not cellular, constitute an additional major category). Most eukaryotes are multicellular (animals, plants, and most fungi). There are also many single-celled eukaryotes. Each eukaryotic cell has a set of linear chromosomes enclosed in a nuclear membrane. A bacterial chromosome is single, circular, and not so enclosed.

A chromosome is a complex structure with two main components: (1) a chain of millions of purine and pyrimidine molecules known as deoxyribonucleic acid (DNA) and (2) an intricate structural framework that supports and manipulates that chain. It is the DNA that contains the hereditary information directing the development of an organism. Along the DNA chain present in each chromosome are distinct, relatively small, regions, known as genes, which govern the various traits of an organism. They are actual segments of the chain. Each gene occurs at a particular location on a chromosome, known as the gene's locus (pl. loci). A single human chromosome may contain thousands of loci affecting thousands of different traits.

Karyotypes. The chromosomes of a eukaryote can usually be sorted into types on the basis of length and other physical characteristics, as well as their genetic content. For example, the largest human chromosome is called Chromosome 1. Each human being has two copies of this chromosome, and two copies of each other type of human chromosome, except for the sex chromosomes. A man has a single copy of each sex chromosome, one X chromosome and one Y chromosome. A woman has two X chromosomes, but no Y chromosome.

Although the same set of chromosomes is normally present in an each cell of an individual multicellular organism, eukaryotes treated as different species, even closely related ones, often have different sets. Often, when a chromosome from one such form is aligned with an otherwise identical chromosome from another, certain loci do not match. For example, a locus present in one chromosome may be absent or inverted in the other. Likewise, two otherwise identical sets of loci may occupy two different positions on the two chromosomes. Commonly, too, the loci present in a single chromosome will be shared out in blocks into two or more separate chromosomes in another organism. There may also be differences in the number of chromosomes present. Disparities of all these kinds, where the chromosomes have been restructured relative to each other, are called structural differences.

When viewed from this structural perspective, the set of chromosomes characteristic of a particular type of organism is called its karyotype, an important term in stabilization theory. As defined under stabilization theory, a karyotype is a set of chromosomes in which a particular set of loci is distributed onto particular chromosomes in a particular order and relative orientation (the DNA segments between the loci also are distributed in a particular order and orientation). Similarly, a chromosome pair is here defined as two chromosomes in the same cell that have the same set of loci, distributed in the same order and relative orientation (again, the regions between the loci are assumed to be distributed in the same particular order and orientation). With respect to the genetic information that it contains, the same locus can differ between the two members of a chromosome pair. For example, a locus for eye color on one chromosome of a pair might contain a gene for blue eyes, but contain a gene for brown eyes on the other. Such variant genes, segments of DNA differing in molecular composition but occurring at the same locus on different chromosomes, are known as alleles.

The sorts of structural differences just described distinguish, for example, the human karyotype from that of a chimpanzee (Pan troglodytes). Humans and chimpanzees do not have the same number of chromosomes and there are also differences in the structure of the individual chromosomes. The Y chromosome differs markedly in size in humans and chimpanzees. An obvious structural difference is that the equivalent of human Chromosome 2 exists as two separate chromosomes (2A and 2B) in the chimpanzee. Moreover, various other human chromosomes cannot be aligned intact with those of a chimpanzee. For example, there are regions on human chromosomes 1 and 18 that are inverted relative to the same regions on the equivalent chimpanzee chromosome. Two large inversions also distinguish a human Y chromosome from that of a chimpanzee. There are many other structural differences differentiating these karyotypes. In general, the karyotypes of more distantly related organisms are more extensively rearranged relative to each other.

Chromotypes and Chromosets. Since many populations treated as separate species differ with respect to karyotype, many hybrids have chromosomes that do not exactly match in pairs. This mismatching is a result of the normal process of sexual reproduction, where a parent typically passes only one chromosome of each of the types present in its karyotype to its offspring. Since the other parent does the same, an F₁ hybrid receives a pair of a given type only when that same type is found in both its parents. Any type of chromosome found in only one parent will have no match in the hybrid. Individuals with such unmatched chromosomes, ones not occurring in pairs, are known as structural heterozygotes.

Such mismatches disrupt the production of gametes and reduce the fertility of the affected individual. Because of this reduction in fertility, populations with distinct karyotypes have often been treated as separate species — their hybrids are relatively infertile, sometimes markedly so. A second characteristic, then, the presence of two distinct karyotypes, often causes two populations to be treated as separate species. As White (1973: 338) clearly states:

In subsequent discussion, a set of individuals sharing the same karyotype will be termed a chromoset; a type of organism having a particular karyotype will be called a chromotype. Although it is true many populations treated as separate species are distinct chromosets, many treated as different subspecies are too. In a broad range of eukaryotes, populations of both these kinds are known (see Table 3.1).

Such differences often are so small that they are not easily observed, and yet they are present. Many such minor deletions, duplications, and inversions differentiate the karyotypes of humans and chimpanzees. Therefore, to the many pairs of chromotypes exhibiting obvious structural differences, we must add many more pairs that differ with respect to such "cryptic" structural rearrangements, ones that are not immediately apparent under the microscope, but that are detectable by other, more discriminating techniques. Even small differences of this sort can have a marked adverse effect on the fertility of hybrids.

While such karyotypic differences between closely related forms are indeed very common, it should be mentioned that there are examples of closely related fruit flies in the genus Drosophila that differ morphologically, but that appear not to differ with respect to karyotype. This fact receives much attention because geneticists so often use fruit flies in their experiments. There are also cases of the same kind known among mosquitoes. As Sumner (2003: 197) points out, the existence of such exceptions does not change the fact that "karyotypes usually differ between organisms, even closely related ones."

The individuals composing a single somaset can often be broken down into two or more distinct chromosets (recall that a somaset is a population distinguished on the basis of morphological traits). For example, Mus musculus (house mouse), though treated as a single species, has distinct chromosets (with distinct geographic ranges), each with its own particular karyotype. Each of the chromosets within the single somaset Mus musculus constitutes a reproductively stable population because all members of a particular chromoset have the same karyotype. But when the members of different chromosets come into contact and interbreed, they produce structurally heterozygous offspring of reduced fertility. This phenomenon is seen not only in mice, but also in a broad range of organisms. So in this case the group of organisms treated as a species (M. musculus) is equivalent to the somaset. Here, taxonomists do not treat the chromosets as separate species.

On the other hand, chromosets of the same somaset are often treated as different species because their karyotypes differ, especially when the structural differences in question severely affect the fertility of hybrids. Reeves's Muntjac (Muntiacus reevesi) and the Indian Muntjac (M. muntjac) look the same, and so form a single somaset. Nonetheless, they are usually classified as different species (here, the two chromosets are treated as a separate species). Matings between these two types of muntjacs produce hybrids of low fertility. Chromosets of a single somaset typically have distinct geographic ranges that come into contact only along their margins, if they come into contact at all.

Note that distinct karyotypes defining distinct chromotypes need not differ at the genetic level. Each can contain the same genes in a rearranged state. For this reason, the different karyotypes all specify the development of the same somatype. For example, various chromosets of Mus musculus can be genetically indistinguishable. The same is true of the distinct chromosets composing the Rattus sordidus complex.

In the upcoming discussion of stabilization theory, the vagueness of the word species and the consequent difficulties of its application in a theoretical context are avoided through the use of chromoset, chromotype, somaset, and somatype, words more specific than species. These names allow populations to be discussed as groups with a particular characteristic — say, by a particular morphology, without any implication being made that they have some other characteristic, such as reproductive isolation. Table 3.2 provides some examples of how these terms can be applied.

With this approach, interbreeding is no longer the major issue it is within the context of neo-Darwinian theory because it is of interest only in itself as a phenomenon, not as a criterion used in defining categories. A somaset is a somaset whether it interbreeds with other somasets or not. Likewise, a chromoset is a chromoset whether or not it interbreeds with other chromosets.

This change in terminology also overcomes various practical difficulties. For example, the most common definition of the word species (i.e., Mayr's) cannot be applied in the case of the many types of organisms not engaging in sex, but the names chromoset, chromotype, somaset, and somatype can. Populations that have been treated as species have often been either sets of individual organisms having the same karyotype (i.e., chromosets), or sets of such sets having the same somatic form.

Sources of Variation

Natural selection cannot act without variation. If all individuals are the same, there are no individual differences among which to select. So evaluation of a mechanism's potential to produce variation is essential in assessing its evolutionary significance. In stabilization theory all genetic variation is assigned to three broad categories according to the source from which the variation is derived:

1. point mutations;
2. meiotic recombination;
3. chromosomal mutations.

The nature of each of these three sources of variation will be explained in the next three sections.

Point Mutations. The word mutation is used to refer both

1. to a heritable change in an organism and
2. to a change in the genetic material (DNA).

Although its meaning is usually a little more specific, point mutation will here be defined as any local change in DNA that does not produce a new karyotype. Point mutations usually have no discernible effect on the development of an organism (that is, they alter the DNA, but the alteration does not change the organism's traits). When they do, most affect a single trait. For example, in the fruit fly Drosophila melanogaster, white-eye mutants that are otherwise normal result from a point mutation affecting a single gene.

Meiotic Recombination. In eukaryotes that reproduce sexually there is an alternation of generations, involving two cyclic stages, haploid and diploid, each of which produces the other. In the haploid stage each cell of the organism contains one autosome of each type (an autosome is any chromosome other than a sex chromosome). In diploid cells the autosomes are present in pairs. For example, in the human life cycle, a multicellular diploid organism alternates with single-celled, haploid gametes (i.e., sperm and egg). In organisms having a single pair of sex chromosomes, each haploid cell contains one sex chromosome, but each diploid cell contains two. In humans, and many other types of animals, the diploid stage is the one producing a multicellular organism visible to the naked eye. But in many plants the haploid phase is more prominent. Many other eukaryotes are unicellular and microscopic throughout both phases (e.g., yeast). In organisms that undergo this alternating cycle, the process of cell division producing the haploid stage from the diploid is called meiosis. During meiosis the two members of each chromosome pair come into direct contact, exchange DNA, separate from each other, and pass into separate cells.

During meiosis, when two chromosomes belonging to the same pair exchange DNA, they trade alleles at some loci. For example, at the eye-color locus there might be an exchange of blue- and brown-eye alleles. As a result, the chromosome that originally specified blue eyes would be altered to specify brown eyes. The offspring individual receiving this chromosome, then, would have a different combination of traits than if the exchange had not occurred. For example, she would have brown eyes and, say, brown hair (instead of blue eyes and brown hair). For this reason, variation of this sort is called recombination. There is a recombination of traits (in this case, brown eyes with brown hair, instead of blue eyes with brown hair). Once these exchanges have occurred during the initial stage of meiosis, the two members of each chromosome pair break away from each other and pass into separate gametes without undergoing any further alteration.

In exchanges of this sort, between the two members of a chromosome pair, no new loci are added or deleted. Nor is the relative order or orientation of the loci altered in any way. Such exchanges only trade alleles between equivalent loci. All the variation is allelic. No structural variation (additions and deletions of loci, or rearrangements of loci) is produced. There is only variation of preexisting alleles at preexisting loci that occur in a preexisting order on preexisting chromosomes.

Therefore, if the variation at a particular locus is considered for a particular diploid chromoset with a karyotype in which all autosomes are paired, only a certain finite number of different variants can be produced by meiosis. For example, if variation at the eye color locus were considered for the chromoset as a whole, there might be an allele for blue eyes present at the eye-color locus on some chromosomes of some individuals in the chromoset, an allele for brown eyes present on other chromosomes, and one for green eyes on still others, but there might be none for yellow or orange eyes.

Under such circumstances, no matter how many exchanges occurred, meiosis would never produce an individual with orange or yellow eyes. The only ways those colors could arise would be either

1. through the creation of a new allele on some chromosome within the population by some process other than meiotic recombination (i.e., through the occurrence of a point mutation), or
2. through the introduction of a preexisting allele from outside the chromoset.

Similarly simple reasoning leads to the conclusion that meiotic recombination can have only limited effect on a trait controlled by genes at multiple loci (to see this reasoning click here). This idea is nothing new, but it has been glossed over in neo-Darwinian theory. Long ago, even before the nature of point mutation was known, geneticist E. B. Babcock (1918: 120) commented that

So Babcock is saying that a karyotype imposes a limit on variation. This fact is clear. In defining a particular set of loci, a karyotype constitutes a stable domain within which point mutations and exchanges of alleles can occur. Such changes at the various loci of a karyotype can occur without changing the structure of the karyotype. But viewed within the context of the chromoset it defines, a karyotype is normally stable. A karyotype, and its corresponding chromoset, can therefore be thought of as a genetic context. Under stabilization theory

1. chromosets are the context within which the Mendelian processes described in neo-Darwinian theory occur;
2. new karyotypes are created by other, non-Mendelian processes.

Since a chromoset is a population of fully intrafertile individuals, interbreeding is unhindered and there is selection for favorable alleles at each locus. The karyotype defines the population in which such allelic (Mendelian) selection can occur. Mendel's Laws form the basis for virtually all the evolutionary mechanisms normally discussed in neo-Darwinian theory. But they apply only to meiosis involving chromosome pairs. They say nothing about the meiotic behavior of unpaired chromosomes. Nor do they tell us anything about the production of new types of organisms by means of non-Mendelian processes (such as most of the stabilization processes to be discussed in Section 4). They do not explain how new chromosets arise.

Chromosomal Mutations. Suppose two hybridizing somasets belong to the same chromoset. Then in their hybrids the sort of genetic variation resulting from meiosis will be limited to the kind just described (meiotic recombination). For, when individuals with the same karyotype mate, the structure of the karyotype remains stable. There is little or no tendency for the number of chromosomes to change from one generation to the next. Chromosomes are not subjected to forces that rip them apart, rearrange them, and reassort them into new karyotypes.

In hybridization between chromosets, however, an additional type of genetic variation occurs due to chromosomal mutations. When individuals from distinct chromosets mate, some or all of the chromosomes of the resulting (F₁) hybrid will be either unpaired or inexactly paired. Two chromosomes are perfectly paired if and only if they have the same loci in the same order and orientation (the regions between those loci must also lie in the same order and orientation). Chromosomes entirely lacking a match do not undergo a regular distribution into gametes as paired chromosomes do. No cellular mechanism exists to deal with unpaired chromosomes. So they pass into gametes at random. In consequence, different gametes end up with different chromosome complements. Some contain one chromosome of a given type; others contain none of that type. When chromosomes are partially matched, the affected chromosomes undergo breakage and rearrangment during meiosis. Such partially paired chromosomes (i.e., ones in which some subregions match, but others do not) exchange nonequivalent loci during meiosis so that both their genetic content and their overall structure are altered. In some cases, where partial matches exist between three or more chromosomes, they join together to form a chain, an event that can also lead to erratic assortment, breakage, and restructuring. Even attempts on the part of the cell to repair the damage can lead to alterations when broken fragments are attached to new chromosomes (translocations). Note that Mendel's Laws, which apply to the meiotic behavior of paired chromosomes, do not apply to the sorts of processes producing chromosomal mutations.

There are additional types of chromosomal mutations, yet to be discussed, that do not result from hybridization. However, chromosomal mutations are more common when hybridization does occur. Whereas point mutations affect individual genes without changing the structure of the chromosome on which the affected gene resides, chromosomal mutations involve restructuring, deletion, and/or duplication of chromosomes, as well as their reassortment into new sets. They also involve the deletion, duplication, and/or reordering of loci (and/or segments of DNA between loci) within a chromosome or set of chromosomes. In short, as defined here, a chromosomal mutation is any mutation producing a new karyotype.

Chromosomal mutations are of three general types, all of which are commonly induced by hybridization between distinct chromosets:

1. Alteration of the structure of individual chromosomes (chromosomal rearrangement), which may involve reversal in the orientation of a portion of a chromosome (inversion) or a transfer of part of chromosome to another location on the same chromosome or on some other chromosome (translocation); It can also involve the deletion or duplication of a portion of a chromosome.
2. Deletion or duplication of entire chromosomes (aneuploidy), or duplication of entire sets of chromosomes (polyploidy);
3. The combination, in a single organism, of chromosomes previously found only separately in two distinct chromosets (chromosomal reassortment).

Chromosomal mutations are sometimes called "gross" mutations because, by affecting entire chromosomes, sets of chromosomes, or large blocks of genes within chromosomes, they recombine and/or duplicate and/or delete hundreds, or even thousands, of genes at a time. On the other hand, the effect of a point mutation, even when detectable, is typically limited to a particular trait or, at most, to a set of related traits. A chromosomal mutation typically affects many traits because it involves many loci. Admittedly, some point mutations have more obvious effects than others. Likewise, the effects of some chromosomal mutations are more limited than those of others. However, in general, chromosomal mutations have effects so large that they are qualitatively distinct from those of point mutation. The effect is of a different order of magnitude.

Even in the absence of point mutation, the reassortment, duplication, and deletion of multiple large blocks of genetic material occurring with chromosomal mutations can have major developmental effects. The simplest demonstration of this fact is seen in F₁ hybrids, which are often markedly different from their parents even though the individual chromosomes are passed unaltered from parent to offspring. Consider how different the common mule is from either of its parents. Here the change is brought about simply by combining in a single organism (i.e., the mule) the unaltered chromosomes of two different organisms (i.e., horse and ass). It results solely from interactions of genes present in new combinations or in different dosages. Thus, even by itself, the reassortment of unaltered chromosomes into new karyotypes — without point mutation and chromosomal restructuring — can be sufficient to bring about the production of new somatypes. In interchromoset crosses producing partially fertile hybrids, the numerous additional chromosomal mutations that occur in later hybrid generations can produce a broad spectrum of morphological variability.

Due to the great amplification of chromosomal mutations in interchromoset hybridization, meiosis in such hybrids (structural heterozygotes) is often severely disrupted. For this reason, interchromoset hybrids produce many more inviable gametes than do organisms with fully paired karyotypes. Many of the gametes do not contain the necessary genetic information to make them viable. Obviously, the production of fewer viable gametes will result in reduced fertility. However, the viable gametes they do produce are far more variable in genetic content than are those produced by ordinary meiosis in an individual with a fully paired karyotype. The chromosomes present in the two parents are present in the gametes in various combinations that could not occur in either parental type. In fact, in later-generation hybrids new chromosomes, not present in either of the parents participating in the initial cross, are often present. These have been built up out of blocks of genes present only on separate chromosomes in the parents (or present on the same chromosome in different relative order). Therefore, in later generations, karyotypes vary greatly from one hybrid individual to another with respect to genetic content and level of chromosome pairing. Thus, interchromoset hybridization produces individuals with combinations of genes, and with restructured chromosomes, that could never arise from intrachromoset matings. Traits therefore vary far more among such individuals than among individuals produced by intrachromoset matings. Under such circumstances, meiosis becomes a far more potent source of variation.

On the other hand, in the case of intrachromoset matings, point mutation is the only well-characterized source of variation other than meiotic recombination. But point mutations are extremely rare because the process that duplicates DNA, the hereditary material, is remarkably reliable. Among eukaryotes, point mutation rates are very low. Consider, for example, the number of gametes that bear a mutation in a particular gene. The frequency of such gametes varies from one type of organism, and from one type of gene, to another, but the highest rate given by Dobzhansky et al. (1977: 69) was one gamete in ten thousand. The low end of the range given by the same authors was about one in a billion. Moreover, a point mutation, by definition, affects a very local region of a particular chromosome.

For example, the DNA polymer chain contained in a single chromosome might contain on the order of 100,000,000 linked units, known as base pairs (each is a purine molecule paired with a pyrimidine). Gregory et al. (2006) say human Chromosome 1 contains around 286 million base pairs and 3,141 genes A single point mutation in Chromosome 1 would typically affect only one or a few of those 286 million pairs and no more than one gene.

What is more, most point mutations have no effect on development unless they happen to occur in a gene (the apparently functionless regions between genes make up the majority of the DNA in the typical eukaryotic chromosome). Even mutations that do occur in genes often have no effect on development. When they do, they are very rarely advantageous and only one or a few traits are affected. Many point mutations that do have an effect are detrimental, or even lethal, a fact long recognized. For all these reasons, point mutations are not a plausible source of rapid, major evolutionary change. Thus, when all matings are within a single chromoset, stabilization theory assumes that point mutations are a relatively insignificant source of new variation.

Certain types of chromosomal mutations (e.g., aneuploidy, polyploidy), though they occur at increased rates in hybrids, can also arise in the progeny of non-hybrid individuals. Individuals with Down's syndrome, a form of aneuploidy, are a familiar example. People exhibiting this syndrome have a normal human karyotype except that they have three copies of chromosome 21, a condition known as "trisomy 21." Note that, as was the case with the common mule, no chromosomes are rearranged in the case of Down's syndrome, but numerous traits are nevertheless affected. The multiple changes result from an increased dosage of the many different genes on Chromosome 21.

Because of the mechanisms just discussed (and because of certain additional mechanisms to be discussed in Section 4), variability is a characteristic trait of later-generation hybrids. High levels of variability within a natural population therefore are usually an indication the population is of hybrid origin. Variability is also an indication of fertility (McCarthy 2006), since (1) F₁ hybrids must be at least partially fertile if later generations are to occur, and (2) variation produced by hybridization is typically seen only in later generations. Thus, variability itself is a factor that can aid not only in identifying hybrid populations, but also in predicting the fertility of hybrids themselves.

Stasis

Even in cultures preceding the advent of science it was believed hybridization would cause distinct populations to blend and become homogeneous (see discussion in Section I). This idea is widespread even today. Experience with actual hybrid zones, however, has shown that hybridizing somasets usually remain distinct. Typically each has a broad geographic range in which there is relatively little morphological variation, but the region between the hybridizing forms contains a variable population of morphologically transitional hybrids. In this section we will look at some examples of stable hybrid zones and then consider the genetic basis of this stability.

Stability of Hybrid Zones. There are many well-documented examples of hybridizing populations remaining stably distinct. For example, Collar (1997: 421) notes that Buffon's Macaw (Ara ambigua) and the Military Macaw (A. militaris) are "sometimes treated as conspecific, but in spite of evidence of interbreeding the characters of the two forms are consistently different over their respective ranges."

Indeed, differences between hybridizing somasets can remain stable over long periods of time — many such hybridizing pairs have shown no significant tendency to blend, even after decades of observation. For example, a stable hybrid zone exists in southeastern Queensland, Australia, between two mammalian somasets, Herbert's Rock Wallaby (Petrogale herberti) and the Brush-tailed Rock Wallaby (P. penicillata). These are also distinct chromosets. Structural heterozygotes are present within the zone, but not elsewhere. And yet, the two are not blending and losing their distinctive morphologies. In the United States the sunflowers Helianthus annuus and H. petiolaris form hybrid populations (i.e., large variable hybrid populations) in many areas, but remain distinct outside regions of hybridization. Also in the United States, the butterflies Collias eurytheme and C. philodice produce a full range of hybrids in some localities, but show no evidence of fusing elsewhere. Fooden (1997: 228) evaluated monkeys in a hybrid zone between Macaca mulatta (Rhesus Macaque) and Macaca fascicularis (Long-tailed Macaque) in southeast Asia. Across the zone a rapid morphological transition occurs from one somatype to the other. And yet outside the zone they remain uniform and pure.

Some zones are surprisingly stable. Perhaps, the best-known hybrid zone involving a pair of North American birds is between two woodpeckers, the Yellow-shafted Flicker (Colaptes auratus) and the Red-shafted Flicker (C. cafer). It extends some 3,000 km from New Mexico and Texas to southeastern Alaska and contains huge numbers of hybrids partially fertile in both sexes. Every conceivable variant between C. auratus and C. cafer exists within the zone. Nevertheless, hybrids very rarely occur outside the zone. There is no indication the parents are going to blend and lose their distinctive traits. The two have remained distinct in all geographic regions outside the zone of contact. An expert on this zone, W. S. Moore (1995: 5) says it is "at least 4000-7000" years old. Obviously, then, this means the situation is extremely stable — No merging of the two parental populations has occurred, though intense hybridization has been taking place at least since the time when the pyramids were built at Giza.

Two wood warblers were once treated as separate species, Audubon's Warbler (Dendroica auduboni) and the Myrtle Warbler (D. coronata). They are quite different in appearance. However, when these two small songbirds were found to hybridize, the American Ornithologists' Union decided they should both be called "Yellow-rumped Warbler" and said the single binomial Dendroica coronata should apply to both. Intense hybridization occurs between these birds from southern Alaska to southwestern Alberta. Throughout this large region virtually all birds are hybrid. Nevertheless, ornithologist Robert Zink (1995: 703) estimates at current rates of hybridization it would take at least "3,200,000 generations (likely over 6,000,000 years) for the fusion of these two taxa to include 3,000 km (only a part of the total range [of these two birds])." This is an underestimate, he says, and yet it "greatly exceeds the estimated time for the duration of a passerine [i.e., songbird] species in the fossil record (0.5 to 1.0 million years; Brodkorb 1971)." We can therefore condense Zink's comments to a single sentence: These two somasets will always remain distinct.

Mayr (1982: 284) refers to "a case of two species of California oaks (Quercus), hybrids of which are known from the Pliocene to the present, and yet where the two species have retained their essential integrity." The Pliocene Epoch ended some 1.6 million years ago. Surely, if Mayr's oaks were going to merge, they would have done so by now. He goes on to say, "the genetics of such situations is not understood at all, for it seems as if some part of the genotype of the two species is not affected by the hybridization. The two species, in such a case, seem to remain 'reproductively isolated,' in the sense that they do not fuse into a single population."

Why are Hybrid Zones Stable? Thus, observation shows hybridizing populations usually remain distinct despite extensive interbreeding. When Mayr was writing, twenty-five years ago, it is true "the genetics of such situations" was not understood. But various factors have since been identified that keep hybridizing somasets from blending over time. Two types of explanations are generally offered. One is cast in terms of environmental factors, while the other is in terms of dispersal and selection against hybrids. The usual formulation of explanations of the former type is the environmental gradient model. This attributes the maintenance of hybrid zones to the differing habitat requirements of the two parental somatypes. For example, Good et al. (2000) argue that adaptation to distinct environments maintains differences between Glaucous-winged and Western gulls (Larus glaucescens and L. occidentalis). Fritsche and Kaltz (2000) make a similar case for the hybrid zone between two plants commonly used in herbal medicine, Prunella grandiflora (Large Self-heal) and P. vulgaris (Common Self-heal).

Nevertheless, the location of a hybrid zone often seems not to depend on environmental conditions. For example, in the northeast corner of the Sinai Peninsula is a narrow hybrid zone between two chromosets usually treated as races of Acomys cahirinus (Cairo Spiny Mouse). Wahrman and Goitein (1972: 235) say it seems

A hybrid zone between two birds, the Black and Painted francolins (Francolinus francolinus and F. pictus), extends across India from the Arabian Sea to the Bay of Bengal. That between the Rock Pigeon (Columba livia) and Hill Pigeon (C. rupestris) begins in northern India and ends in southern Siberia. Each of these avian zones passes through such a wide variety of environments it seems unlikely zone maintenance is related to habitat. In point of fact, hybrid zones often occur where there seems to be no significant change in the environment. For example, regarding the shrew hybrid zone mentioned above, Benedict (1999a: 135) notes that it follows an irregular course and is not associated with any particular type of soil or vegetation. Indeed, the region of Nebraska in question is monotonously uniform with respect to most environmental factors.

Such situations, where the role of the environment is at best obscure, are explained as tension zones. Tension zones occur when dispersal of hybrids from the zone is balanced by influx of parental individuals. The mechanism is simple. When parental types have a reproductive advantage versus the hybrids, they keep the zone narrow by moving into it at higher rates than hybrids move out. This bias in dispersal keeps the genetic influence of hybrids from spreading outside the zone. In general, the larger the selective disadvantage against hybrids, the narrower a tension zone will be. The more mobile the participating organisms, the wider it will be. One probable example of a tension zone is that between the Black-capped Chickadee (Parus atricapillus) and the Carolina Chickadee (P. carolinensis). It extends across the eastern United States from New Jersey to Kansas through a variety of environments. Bronson et al. (2003) monitored the reproductive success of mated pairs within this zone. Unmixed parental pairs of either parental type produced more than twice as many fledglings per nest as did hybrid pairs.

Much debate has focused on whether environmetal gradients or tension zones are more important in maintaining zone stability. The two, however, are not mutually exclusive. A tension zone can exist in a region of transitional habitat. Thus, Delport et al. (2004) argue both habitat characteristics and a balance of dispersal and selection seem to play a role in maintaining a hybrid zone between two large south African birds, the Damaraland and Red-billed hornbills (Tockus damarensis and T. erythrorhynchus). Regarding a hybrid zone between the Cherry Stone Clam (Mercenaria mercenaria) and Southern Quahog (M. campechiensis), Bert and Arnold (1995) came to similar conclusions.

Karyotypes and the Maintenance of Hybrid Zones. Distinct chromosets break up spatially into separate populations that hybridize where they come into contact. Spatial segregation on either side of such zones occurs because

1. populations that are mixed with respect to karyotype produce many infertile offspring, and so tend to shrink; and
2. chromosets, which are pure with respect to karyotype, produce relatively fertile offspring and tend to expand. Expansion brings the chromosets into contact, but infertility of structural heterozygotes resulting from that contact limits overlap of the chromosets.

With time, hybrid populations come to occupy relatively narrow interface regions (hybrid zones) between more extensive regions occupied by pure, karyotypically uniform, fully fertile parental chromosets. The population dynamics of such situations fit the tension zone model, but presumably, each of the parental chromosets would tend to occupy those regions where they had a reproductive advantage (as in environmental gradient models). In fact, examples are known of chromosets breaking up along an environmental gradient.

Thus, in those situations where hybridizing populations differ in karyotype, chromosomal models provide an explicit explanation of how populations remain morphologically distinct despite ongoing interbreeding. The chromosets on either side of such zones can, and often do, have different genes. The distinct genetic content of their karyotypes program the development of different organisms with different morphologies. The genetic uniformity (and hence morphological uniformity) within each of the two hybridizing chromosets is maintained by selection for karyotypic uniformity within each of the two chromosets. For, in an otherwise uniform population, there is selection against any aberrant individual with a differing karyotype. In matings with karyotypically normal individuals any such individual produces structurally heterozygous offspring of low fertility (see above).

Hybrid zones are widely viewed as sources of "gene flow" (i.e., as causing genetic mixture of the participating populations). However, as we have just seen, a tension zone between chromosets may serve as a buffer, actually preventing gene flow. There is gene flow into the zone, but not between the two parental populations. Although the two are interbreeding, they are, in a real sense, reproductively isolated. In spatial computer simulations, hybridizing chromosets can remain distinct indefinitely (McCarthy et al. 1995). In narrow hybrid zones, gene flow is prevented by the reduced ability of hybrids to survive and reproduce, and in wide ones it is prevented by distance itself.

Conclusion. Hybridization is associated with two seemingly contradictory phenomena, variation and stasis. However, this contradiction is illusory. In fact, the variation produced by hybridization is limited to the hybrid populations produced by such interbreeding. Morphological stasis is a feature of the interbreeding parental forms, which retain their distinctive characteristics outside the zone of contact. Cases are known where hybridizing populations have remained distinct for thousands or even millions of years. True, hybrid zones are highly stable in the sense that they can continue to exist for eons of time. Yet, in general, within most such zones, the hybrids themselves are extremely variable.

Stabilization theory distinguishes three types of variation. At one level, there is the slow accumulation of point mutations. This process is ongoing and is not a result of meiosis. It produces new alleles. At another level there is intrachromoset variation resulting from meiotic recombination involving paired chromosomes. Allelic variation of this type is described and predicted by the rules of Mendelian genetics. Intrachromoset meiosis can produce numerous changes in a single generation. But change from this source is inherently limited because it involves the recombination of preexisting traits and alleles. At a third level there are chromosomal mutations. These are often, but not exclusively the result of meiotic reassortment and restructuring in structural heterozygotes produced by interchromoset matings. Chromosomal mutations bring about radical, rapid changes in morphology. They create new sets of loci and new karyotypes. They can duplicate, delete, and recombine thousands of genes in a single generation. Chromosomal mutations brought about by interchromoset matings not only produce new karyotypes, but also combine, in a single organism, traits that were previously found only separately in one or the other parent. Moreover, they produce a wide variety of different hybrid individuals with different combinations of such traits. Variation of this type is non-Mendelian.

It is the first two of these three levels of variation that are emphasized in modern evolutionary discussion (neo-Darwinian theory). The third level, which is the focus of stabilization theory, receives relatively little attention, apparently because the theoretical framework underpinning contemporary evolutionary discussion is based on Mendel's Laws, which apply only to intrachromoset matings. Such rules say nothing about processes that bring about karyotypic change. Such changes fall outside their scope. In the next section we will consider in more detail how new types of organisms arise via stabilization processes and the chromosomal mutations associated with them. This discussion will be illustrated by numerous examples.