Karyotype evolution in the grammicus complex
caveat
possible phyletic sequences
the probable phyletic sequence
Clarkii group chromosomes and origin of Em karyotypes
Sex chromosome evolution
Clearly the extraordinary karyotypic variability and cryptic speciation in the grammicus complex will provide material for testing chromosomal speciation hypotheses and other ideas reputing to explain the association between chromosomal diversity and prolific speciation. However, to be used effectively as a test, the sequence of karyotypic evolution in grammicus should be derived as accurately and as independently as possible from any hypothesis it will be used to test. On the other hand, if the molecular mechanism(s) of Robertsonian mutations were well understood, the cytological evidence might still be a powerful constraint on the phylogenetic interpretation independent of the hypotheses to be tested. Unfortunately, this understanding of Robertsonian change does not exist.
Thus, Cole (1970, 1971a), accepting the postulated translocation mechanisms of White (1954) and Matthey (1949), which predict that centric fissions should be much less probable than centric fusions, reconstructed the chromosomal phylogeny of the spinosus group of Sceloporus on the assumption that its chromosomal evolution was unidirectional from primitively high chromosome numbers to low numbers by centric fusions. On the other hand, Todd (pers. comm.) suggested that chromosomal evolution in Sceloporus, and particularly within grammicus, can readily be interpreted by his hypothesis of the simultaneous fissioning of all metacentrics by "centric misdivision" (Todd, 1970), followed by the sorting out of the fissions as polymorphisms through the populations of the species. Yet again, White, while now accepting the comparative data which suggest that centric fissioning (= "centric dissociation"--White, 1969) is not uncommon as a mode of Robertsonian mutation, dismisses Todd's thesis of the simultaneous fissioning of all metacentrics because this is "equivalent to a belief in miracles, which has no place in science" (White, in press, p. 501). And, to show that the possibility of simultaneous fissioning cannot be discounted out of hand, I can suggest at least two "mechanisms" which might produce just White's "miracle:"
Certainly such mutator genes are known which affect other chromosome loci (Ives, 1950). Todd's (1970) thesis rests on the a priori assumption of a mechanism for "centric misdivision," while White's (1973) counter argument seems to be based on equally a priori assumptions from his mechanism for "centric dissociation." It seems to me most unwise to use any of these coined terms "with a very precise implication as to the mechanism," as White (1969) suggested; when, in plain fact, the mechanisms which generate the mutations are far from being demonstrated and are much farther from being understood in the details of their biochemical and molecular modes of action.
Instead, I consider that at this stage in our understanding of the molecular mechanics of Robertsonian mutation (e.g., see Brinkley and Stubblefield, 1970; Comings and Okada, 1970) it will be far better in the analysis of grammicus phylogeny which follows to use the terms fission and fusion as simple descriptions of what can be determined from nature about the direction and sequence of the Robertsonian mutations.
I have shown above that karyotypic variability per se in the grammicus complex cannot be used safely as an independent guide to sequence(s) of karyotypic derivation. Also, since neither morphological nor biochemical data are yet available which could contribute to this analysis, discussion of the phyletic sequences must depend largely on what is known about ecology and biogeography for the various chromosomal races of the complex. Although these limited data are not conclusive, they suggest the tentative rejection of the more complex, and intuitively unlikely, of the possible phyletic sequences.
The apparent possibilities fall in four categories:
Figs. 7a and 7b illustrate the approximate geographical ranges of the karyotypically distinctive populations of the grammicus complex, and Tables 2, 3, 4, 5, 6, 7, and 8 summarize their ecological distributions in these ranges. A logical and consistent simple fissioning sequence can be developed which readily accounts for this biogeographic distribution.
Judging by their present range and ecological diversity, the Standard grammicus must have been widespread on the Mexican Plateau before their karyotypic evolution began. Then, sometime during the history of Standard, the FIS-5 mutation became fixed on its northwestern periphery. However, since the F5 population probably has always been separated from other chromosomally derived populations, either by intervening standard populations or by the hottest part of the Chihuahuan Desert; and since it presumably was derived independently from the other populations, there is no information which would allow an estimate of F5's age with respect to the other derived populations. Therefore, no profit will derive from considering this population further in the present report.
The derivation of P1 is least certain. Two possibilities were suggested by Hall and Selander (in press):
The reconstitution of a metacentric-6 would not be an intrinsically unlikely mutation if fusions between chromosomes of greatly different lengths (e.g., fusion with an acrocentric-6 element and a micro chromosome) are especially disadvantageous; since in the F6 karyotype, the acrocentric-6 elements could then fuse successfully only with one another. Either model is geographically plausible.
Of the remaining derived populations, F6 seems the oldest. Its present disjunct range in dry areas, where local populations are found only in comparatively mesic local habitats, and the restriction of its continuous distribution only to more mesic parts of the Sierra Volcanica Transversal together suggest that F6 achieved a maximum distribution during a Pleistocene pluvial period. Then, F6 probably ranged throughout the entire Sierra Volcanica Transversal-Sierra Madre Oriental system.
The next largest distribution (assuming that the coastal and plateau populations have a single ancestry) is that of the F5+6 population, It seems likely (though not yet confirmed) that the F5+6 population bisects F6 in the area of northern Guanajuato and Hidalgo, where F5+6 inhabit areas almost as mesic as those inhabited by F6. From this possible center, F5+6 has spread into much more xeric areas / in San Luis Potosi and along the coast. This range suggests a possible spread coinciding with increasing aridity in the post-Pleistocene period. An alternative hypothesis would derive the coastal and plateau populations independently from an F6 ancestry. However, the ecological derivations would remain the same, with both F5+6 independently spreading into more xeric habitats.
FMl and FM2 also occupy xeric habitats, which extend south of the F5+6 range, with FMl occupying the geographically intermediate region. The apparently very small ranges of these populations suggest that they may be phylogenetically the most recently derived. The fact that they have not been found anywhere except in close association with human agriculture suggests that they may have evolved contemporaneously with the cultivation of primitive crops in the area of central Mexico. The area between the northern FMl and southern F5+6 has not been sufficiently sampled to rule out the possibility that populations karyotypically intermediate between F5+6 and FMl may be found there. Furthermore, if the FM populations did evolve in close association with agriculture, the more primitive intermediate populations may have been eliminated by man-induced habitat modifications. It is also quite possible that these populations did not grow large before giving rise to the later derivatives.
These known distributions are clearly compatible with the simple fissioning sequence, as geographical and ecological derivations can be traced in the same sequence as the karyotypes; whereas, on the other hand, it seems unlikely that independent simultaneous fissioning events would involve just these chromosomes and would be arrayed geographically in just such a fashion that this clear linear sequence could be traced.
Superficially, the observed distributions also seem compatible with the Todd model. This would assume that the original fissioning occurred in a single event, and the fissioned chromosomes then diffused outward as polymorphisms from a center until they became either fixed in local populations or lost. By this model, all of the fissions were either fixed or retained as polymorphisms in the FM2 population, with increasingly fewer of the fissions having reached or having been retained by the more peripheral populations to the north. However, if the actual events did follow this model, it seems strange that the fissions have spread only to the north. Furthermore, it is awkward that the probable "central" FM2 population seems to have differentiated so much more recently than either of the more "peripheral" F6 or F5+6 populations. Finally, it seems even more strange that the polymorphic populations (P1, FMl, and FM2) occupy such small parts of the total range of grammicus, if we are to believe that at one time polymorphisms for all of the macrochromosomes were spreading throughout the range of grammicus now occupied by the chromosomally derived races.
Therefore, the simplest phyletic model, that of a direct sequence of fissions from Standard through F6 to FM2, seems most in accord with the available evidence, although until more is known about the biochemical and morphological relationships of the various grammicus complex populations, the Todd model cannot be completely discounted. Except for the possibility that the P1 population is derived from F6 by a centric fusion, there is no evidence supporting either independent multiple fissioning or more complex patterns of derivation. And, except for the single derivation branches leading to F5 and P1, and a possible single derivation branch to one of the F5+6 populations, all of the remaining chromosomal races can be traced in a single linear sequence of derivation, which can still be traced through a sequence of parapatric contacts.
Although evidence from the crevice-using Sceloporus is sufficient to determine the primitive macrochromosome morphology and to suggest sequences of macrochromosomal fissioning away from it in the grammicus complex, to fully interpret the phylogenetic significance of the Em chromosomal condition found in asper and the megalepidurus group, the karyology of certain non-crevice-using Sceloporus must also be considered. The crevice-using species all belong to Smith's (1939) large-sized, large-scaled division of the genus; most species of which have been karyotyped either by Cole or myself. Besides the crevice-users, only two other species in this division, S. clarkii and melanorhinus (= the clarkii group), have female karyotypes with 20 microchromosomes. In fact, the common microchromosomal pattern in clarkii group females is indistinguishable from that of the standard karyotype crevice-using females, although the male microchromosomal patterns in the clarkii group species differ from one another and from standard (Figs. 8 and 9). Also, both clarkii group species differ from the standard karyotype by being fixed for fissions of macrochromosome pairs 1,3,4, and 5. (Cole, 1970, and Lowe et al., 1967, postulated that this pattern is primitive relative to the standard macrochromosomal pattern, but this idea is indefensible for the same reasons that the FM2 cannot be considered primitive in the grammicus radiation.) However, here we are concerned with the similarities of the microchromosomal patterns between the clarkii group species and the species with the standard karyotypes, which suggest that all may have a fairly close common ancestry within the large-scaled Sceloporus division.
One aspect of the microchromosomal variation in the clarkii group is especially interesting. Cole (1970) reports that both clarkii and melanorhinus are polymorphic for an enlarged microchromosome found in his KB or 20+1+19 patterns. I find this enlarged micro-chromosome indistinguishable from the Em chromosome fixed in asper and the megalepidurus group (Figs. 8, 9). Cole quite reasonably suggests that this Em condition is derived and may be due to a tandem duplication resulting from unequal crossing over. Cole found the Em chromosome in clarkii sampled from two areas (Santa Cruz Co., Arizona, and near La Concha, Sinaloa) which represent two of three subspecies and the opposite extremes of clarkii's geographic range. In the Santa Cruz Co. sample, the Em chromosome was present at a frequency of 0.25 (N=51); and in the total of two individuals collected at the Sinaloa locality, one was heterozygous for the mutation. The remaining clarkii karyotyped by Cole (12 lizards from 7 other localities and 16 from an 8th) did not show the Em chromosome. Of 12 clarkii I have karyotyped, including eight from the Playa Escondido area of Mazatlan, Sinaloa, none conclusively showed the Em chromosome, although two individuals might have been homozygous for the mutation based only on diakinesis figures (see Fig. 10. Many of my clarkii preparations were made during my 1966 expedition and did not benefit from a colchicine pretreatment, and hence have few if any serviceable mitoses).
Of seven melanorhinus karyotyped by Cole (1970), two of three individuals from one locality near Acapuico, Guerrero, were heterozygous for the Em mutation, while the third individual from that locality lacked it, as did the remaining four from two other localities. Of the six melanorhinus I have karyotyped, only one from Rio Maria Basio, west of Manzanillo, Colima, was heterozygous Em; while all of the remaining specimens, representing a second locality near Manzanillo and two localities near San Bias, Nayarit, lacked the Em chromosome.
It seems most unlikely that the two clarkii group species, the two megalepidurus group species, and asper would have independently evolved these very similar appearing Em chromosomes. On the other hand, particularly since the other microchromosomal similarities between these species also suggest that they are closely related, the Em mutation probably originated in the common ancestor of these five species. If so, this common ancestor must have been the progenitor of both the clarkii group and all of the crevice-using Sceloporus as well. The survival of the Em chromosome as a polymorphism in both clarkii and melanorhinus is surprising enough, but its presence in the crevice-using assemblage as well indicates that it must have survived as a polymorphism in the clarkii group from before it and the crevice-users diverged from one another. Furthermore, the polymorphism must also have been retained for a while as such by the primitive crevice-users to explain its fixation in two rather different sections of this radiation.
Analysis of the sex chromosomes of the clarkii group and the crevice-users provides additional evidence on their relationships.
As described above, all crevice-users are characterized by identical X1X2Y sex trivalents in males, as originally noted by Cole et al. (1967) for jarrovii and poinsettii and Axtell and Axtell (1971) for another jarrovii population. On the other hand, Lowe et al. (1967) and Cole (1970) did not identify sex chromosomes in the clarkii group karyotypes. However, my studies have revealed distinguishable sex chromosomal heteromorphisms in both species of the clarkii group. This is most obvious in melanorhinus and will be discussed first.
The four females of the seven melanorhinus karyotyped by Cole (1970) had 2n=40 karyotypes identical to the common clarkii karyotype, which has two pairs of metacentric macrochromosomes (2 and 6 of the standard karyotype), eight pairs of acrocentric macrochromosomes, and 20 microchromosomes. However, Cole's three males had only 39 chromosomes in their karyotypes: one micro was "missing" and there were five instead of the usual four metacentric macros, with the anomalous metacentric in the pair 6 size range. This replaced one of the acrocentrics in the female karyotype. Cole called this 39 chromosome pattern the KC karyotype and compared it with the similar but not sex correlated 40 chromosome KC karyotype of S. clarkii. However, he also noted that the apparent sex correlation in melanorhinus was "suggestive."
Unfortunately Cole's material did not provide meiotic figures to clarify the situation. Four of the six melanorhinus I have karyotyped were males, and each of these also had the 2n=39 KC karyotype as described by Cole (Fig. 8c), while my two females had the 2n=40 normal pattern. (Actually, as can be seen in Fig. 8c, the anomalous "metacentric" chromosome is more submetacentric than the pair 6 chromosomes, and Cole's indication that it was metacentric is due to a poor pairing of the small metacentrics, as can be seen by inspection of his Fig. 4c.) Additionally, one of my males (incidentally, the individual heterozygous for the Em mutation) yielded several diakinesis figures, all of which showed clear trivalents involving the anomalous metacentric (Fig. 9d). The strict sex correlation of this KC karyotype (all 7 males karyotyped had it, while all 6 females lacked it) and the trivalent formation in diakinesis both clearly indicate that this metacentric must be the Y of an X1X2Y sex chromosome system--but clearly the Y here is different from that found in the crevice-users (Figs. 8 and 9).
Presumably this metacentric Y was generated by the centric fusion of one of the 5b acrocentrics and a more primitive Y chromosome, which seems to have been a comparatively large acrocentric micro-chromosome (Fig. 8). As a result of the fusion, the second 5b macro-chromosome thereby became linked with sex as an X2 chromosome. The X1 or the more primitive X, is in the size range of the second or third smallest pair of microchromosomes, as it is in the crevice-users, and it is probably homologous with the X of these species.
Furthermore, in comparisons of the karyotypes of melanorhinus and the crevice-users (Fig. 8), it can be seen that the short arm of the melanorhinus Y is comparable in size to the long arm of the crevice-user Y. Therefore, these arms are probably homologous and approximate the more primitive acrocentric Y from which the biarmed Y chromosomes of the two taxa were independently derived by centric fusions, respectively with a macro chromosome in melanorhinus and with a microchromosome in the crevice-users.
As noted above, no sex chromosomal heteromorphism was seen in clarkii by Lowe et al. (1967) or Cole (1970). However, the close relationship between the clarkii and melanorhinus karyotypes in other respects, and the size relationships in melanorhinus between its X, and the short arm of its biarmed Y (presumably homologous to the whole of a more primitive Y in clarkii), suggest that a sex chromosomal heteromorphism should exist in clarkii as well. Since male clarkii have an even 2n, this heteromorphism would most likely be of the XY type. Unfortunately none of my preparations from male clarkii yielded useful mitotic figures, but the predicted XY heteromorphism is detectable in the diakinesis figures. In most cells, one of the larger microchromosomal bivalents is distinctly unequal (Fig. 9c). (Cole's, 1970, Fig. 3a also demonstrates this heteromorphic pairing in diakinesis: the unequal bivalent can be seen near the periphery of this figure at about 4:30 o'clock.) Fig. 10 shows additional arrays of the microchromosomes from representative diakinesis figures of several individuals, which again demonstrate the unequal bivalents. Although I cannot confirm the morphological relationships of the components of the heteromorphic pair with mitotic preparations from my material, the size relationships of these chromosomes with one another and with the other microchromosomes are consistent with the idea that the Y chromosome is a large acrocentric micro and the X is a smaller micro, equivalent to the X. of the crevice-users and melanorhinus, as would be expected. Furthermore, in Cole's (1970) Fig. 1c, illustrating a male karyotype of clarkii, it appears that the second chromosome in the microchromosomal array is improperly paired. This is a large acrocentric and may be the Y. However, Cole paired it with a submetacentric micro, which more likely pairs with the fourth micro in his karyotype.
So, to recapitulate, clarkii probably retains the primitive sex chromosomal condition in the clarkii-crevice-user radiation; i.e., an XY condition, where the Y is a large acrocentric microchromosome equivalent in size to pair 9 or 10 in the standard karyotype; and where the X is a smaller acrocentric, equivalent in size to pair 11 or 12 in the standard karyotype. The melanorhinus-type Y would then be formed by a fusion between the clarkii-type Y and one of the pair 5b acrocentric macro chromosomes. The standard-type Y would then be formed by a fusion of the clarkii-type Y and another large acrocentric or sub-acrocentric microautosome, probably homologous to pair 7 in the clarkii group karyotypes (Figs. 8a, b).
In comparison with the sex chromosomes of the primitive 2n=34 sceloporine karyotype (e.g., Pennock et al., 1969; Cole 1971a, 1971b), even the sex chromosomes in clarkii are derived. The X chromosome may be homologous to this more primitive sceloporine X, but the sceloporine Y is minute, being by far the smallest micro in the 2n=34 karyotype. With the present data it is unclear how the large Y and 18 microautosomes of the clarkii karyotype derive from the minute Y and 20 microautosomes of the primitive 2n=34 karyotype, particularly if their X chromosomes are homologous.
The frequency with which complex sex chromosomal systems have evolved independently within Sceloporus and in other iguanid genera suggests that it may be selectively advantageous for them to increase the amount of sex linkage. Many iguanids show no cytologically distinct sex chromosomal heteromorphism. This is presumably the primitive condition (Gorman et al., 1967). Several Anolis species (Hall et al., unpub.) and most of the sceloporines show an XY heteromorphism involving the smallest microbivalent. So, whether the primitive sex chromosomes are cryptic or not, they are very small microchromosomes. This cytological evidence and the facts that sex is frequently indeterminant or determined by only a single "gene" in fish and amphibia (Ohno, 1967), and that at least one lizard (Chevalier, 1969; Gorman, in press), snakes (Ohno, 1967; Gorman, in press), and birds (Ohno, 1967) are female heterogametic all suggest that sex determination in reptiles and lizards was based primitively on only very slight amounts of heterogamety and that the differentiation of distinctive sex chromosomes has occurred within the reptilia in several independent lines. In Sceloporus several lines also show increases in the chromosomal linkage of sex, as indicated by distinctive heteromorphisms;
The increases in some of these lines must have been at least partially independent. Also, in Anolis sex linkage has increased independently in at least three lines (Gorman, in press). Finally, the genus Polychrus also shows at least one line with increased sex linkage. Many iguanid species show pronounced sexual dimorphisms, and it is interesting to speculate that the increased sex linkage is adaptive in that it simplifies developmental control of the dimorphic characters. But, at least the sceloporine species with increased sex linkage are not conspicuously more dimorphic than are those that have the primitively small amount of sex linkage. However, this is a possible correlation that deserves more attention than I have given it to date.
