The "standard" or S karyotype
The "enlarged micro" or Em karyotype
Karyotypic diversity of Sceloporus grammicus
Introductory comments
S or standard grammicus
P1 or polymorphic-1 grammicus
F6 or fission-6 grammicus
F5 or fission-5 grammicus
F5+6 or fissions 5+6 grammicus
FM or multiple fissions grammicus
Contact zones, hybridization, and reproductive isolation in grammicus
All species reported here except those belonging to the megalepidurus group, S. asper, the clarkii group, and the karyotypically derived populations of the grammicus complex have indistinguishable karyotypes, which will be termed the "standard" or S pattern (Figs, 2, 3, 4a, b, c, d, 5a, 6b, and 8a). This is the karyotype described by Cole et al. (1967) for S. jarrovii (see also Axtell and Axtell, 1971) and poinsettii, although my arrangement of the chromosomes differs somewhat from theirs. Females with the standard karyotype have 32 chromosomes, including 12 macrochromosomes and 20 microchromosomes; while males have only 19 microchromosomes due to a sex chromosomal heteromorphism of the x1x2y type. Ordered in sequence from largest to smallest, macrochromosome pairs 3 and 4 are almost exactly metacentric and are approximately identical in size. These cannot be reliably distinguished from one another, although in most mitotic spreads the remaining macrochromosomes can be. Pairs 1, 5, and 6 are all slightly submetacentric but conspicuously differ in size from one another and from pairs 3 and 4. Pair 2 is only slightly smaller than 1, but it is conspicuously submetacentric, with the long arms about 1.5 times the length of the short. In particularly good preparations, minute satellites can be seen at the tips of the long arms of pair 2. These are presumably homologous with the satellites reported for the similar chromosomes of other Sceloporus (Lowe et al., 1967; Cole and Lowe, 1968; Cole, 1970, 1971a, 1971b, 1972; Jackson and Hunsaker, 1970), some other iguanids (Jackson and Hunsaker, 1970), and possibly even with those reported in primitive teiids (Gorman, 1970). In diakinesis arrays (Figs. 9a, b) four size classes may reliably be determined among the macrochromosomal bivalents. Pairs 1 and 2 are close enough in size that their identification is equivocal, 3 and 4 are indistinguishable from one another, but are distinctly smaller than I and 2, and 5 and 6 differ enough from each other and from the larger bivalents to be readily identifiable.
The microchromosomal morphology of the S karyotype is less certain because of preparation artifacts and the fact that the sizes of the smaller ones approach the limits of optical resolution. Yet, a careful examination of the clearest preparations from several species reveals an apparently consistent pattern of microchromosomal morphology (Figs. 2, 3, 4a, b) . Although the microchromosomal structure of the remaining species assumed to have the S karyotype cannot be seen as well, their microchromosomes can be arranged in the same size sequence, and I have no reason to believe that they differ from the clearer preparations.
In males the y chromosome is the largest microchromosome and it is submetacentric (short arm more than half as long as the long arm). The unpaired x1 (the original x chromosomeósee discussion below), designated #8 in the male karyotype (Fig. 2), is acrocentric and is about the size of the second or more probably third smallest pair of acrocentric to subacrocentric microchromosomes. However, exactly which of the individual microchromosomes in this size range is unpaired cannot be determined with certainty in any mitotic spread. The x^ chromosome, designated #7 in the male karyotype (Fig. 2), is clearly one of three largest subacrocentric microchromosomes, and is probably slightly larger and slightly more conspicuously subacrocentric than the other two, which are designated pair 9. These sex chromosomes form a clear trivalent in male diakinesis figures (Figs. 9a, b; 10). Besides the sex chromosomes, the microautosomes then include 5 pairs of acrocentric to subacrocentric chromosomes and 3 pairs of submetacentric to metacentric chromosomes. The acro's are arrayed in decreasing order of size from pairs 9 through 13. Pair 9 is subacro, 10 acro, 11 subacro, and 12 and 13 probably acro. Pair 14 is submetacentric and probably between pairs 9 and 10 in size. Pairs 15 and 16 are probably metacentric and somewhat smaller than 14.
All karyotyped species of the torquatus group (Figs. 2, 3) appear to have this microchromosomal pattern, as do all species and populations of the grammicus group except the FM2 population of the grammicus complex (see below).
In species with the S karyotype, the y chromosome is distinctly larger than the three large subacrocentric microchromosomes which include the x2 chromosome (#7 and pair 9). Three species: megalepidurus (Figs. 4f, 9e) and pictus (Figs. 4e, 10h), which form part of the present megalepidurus group (Smith and Lynch, 1967), and as per (Figs. 4g, 8d), currently placed in the formosus group (Smith and Taylor, 1950), all have karyotypes indistinguishable from S except that one of the larger microchromosome pairs (probably 9) is replaced by a much larger pair of subacrocentric chromosomes which are intermediate in size between the y and pair 6. These enlarged microchromosomes are similar to or slightly shorter than the long arms of pair 6. The chromosome pattern distinguished by these comparatively enlarged microchromosomes is hereby designated the Em karyotype, and will be discussed further below. Other karyotyped species in the megalepidurus and formosus groups (as these are presently defined) are karyotypically quite different from the Em pattern, indicating that these groups probably are not natural as they stand (Hall, in prep., and see below).
By far the greatest karyotypic vari ation found among the species characterized by the microchromosomal pattern of the S karyotype occurs between and within populations now placed in Sceloporus grammicus. Chromosome numbers of the approximately 1300 grammicus studied range from the 2n=31 of standard males to a maximum of 2n=46 in some females of the FM2 population. Fig. 5 clearly shows that most of this variation is Robertsonian and results from centric fissions.
Table 1 lists the number of species of each lizard family known to have karyotypes with 12 approximately metacentric macro- chromosomes (see second to last column). Most of these species also show a relatively constant pattern of macrochromosome size and centromere placement. Thus pairs 1 and 2 are usually similar in size with 1 nearly metacentric and 2 clearly submetacentric and sometimes satellited on its long arms. Pairs 3 and 4 are usually indistinguishable--both are nearly exactly metacentric and are distinctly but not greatly smaller than 2. Pair 6 is usually rather small and slightly submetacentric, and pair 5 is generally intermediate in size between 4 and 6 but usually closer to 4 and also slightly submetacentric. Fig. 6 compares the agamid, Agama caucasica (6a), and a standard grammicus (6b) to demonstrate this point. More importantly, all 12 karyotyped species of live-bearing and crevice-using Sceloporus also have exactly the standard grammicus macro chromosomal pattern. It would patently be absurd to suppose that all of the species with 12 metacentric macrochromosomes listed in Table 1 (and in the present report) independently evolved their similar macro chromosomal patterns from ancestors with 46 or 48 chromosome karyotypes. And it would be even more implausible to suppose that this supposedly primitive karyotype survives in Sceloporus only as one end of a polymorphic series in a small fraction of one species (FM2 grammicus) and in one other species (S. merriami--Cole, 1971a) which is very distantly related to grammicus.
It is much more reasonable to suppose that the many species with 12 metacentric macrochromosomes retain a primitive pattern. A complete fissioning of the standard 12 metacentrics would always give 24 similar appearing acrocentric chromosomes, whatever the sequence of the fissioning; but, on the other hand, the various acrocentrics in a 46 or 48 chromosome karyotype could be fused in numerous combinations, and only a few would produce the size and arm ratios observed in the standard karyotype. Furthermore, whatever opinions have been held in the past about the improbability of centric fissioning as a mode of karyotypic evolution (e.g., Matthey, 1949; White, 1954, 1957; Cole, 1970, 1971a); Webster et al. (1972) have clearly documented a case of centric fissioning in the Iguanidae.
The conclusion is thus inescapable that the female 2n=32, male 2n=31 standard karyotype is primitive in the grammicus radiation, and that karyotypes with higher numbers of macrochromosomes must have been derived from the standard karyotype by centric fissioning. These derived karyotypes in grammicus are described in order of the increasing amount of their fissioning. Ecological and geographic distributions of the various grammicus populations will be analyzed in detail in Hall and Alvarez (in prep.), and are summarized here in Tables 2 through 8.
Collection localities are mapped in Hall and Selander (in press) and in more detail in Hall and Alvarez (in prep.).
The grammicus populations characterized by the S karyotype (Figs. 4b, 5a, 6b) are themselves designated S, or standard, and have the widest distribution of any of the karyo- typically distinct grammicus populations. These S populations range geographically from northern Coahuila, discontinuously, to southern Oaxaca (Fig. 7a, 7b, and see also Figs, 2 and 3 in Hall and Selander, in press), and ecologically from the upper edges of the Chihuahuan Desert in southern Coahuila and Zacatecas to mountain rain forest in Oaxaca and Veracruz (Table 2). Most of the karyotypically derived populations of grammicus are found across the middle of the range defined by standard grammicus and appear to replace them geographically, thereby separating northern and southern populations of standard grammicus. Also, though geographically close to the continuous southern population of standard grammicus, intervening populations of chromosomally derived forms appear to separate the standard population on the floor of the Valley of Mexico from the standard populations to the south and east of the Valley.
Populations given the P1 designation are polymorphic for a fission of chromosome 1 (the FIS-1 mutation,. Fig. 5b) and are found only above 3200 m elevation on the three mountains, Tialoc, Ixtaccihuati, and Popocatepetl, that form the eastern divide of the Valley of Mexico (Table 4). The P1 population appears to be completely surrounded by the population of F6 grammicus (described next, below) which encircles the mountains at intermediate elevations (Figs. 7a, 7b), and see also Fig. 3 in Hall and Selander, in press).
Most individuals in the P1 population have S karyotypes, but some are characterized by heterozygosity for the FIS-1 mutation, and three individuals were found to be homozygous for the fission. The frequency of this mutation is 0.103 in a total of 301 individuals from throughout the P1 range above the F6 grammicus. The FIS-1 mutation was found in all areas where reasonably large samples were taken (18 was the largest number of individuals sampled from any 1 km area which did not include at least one FIS-1 heterozygote). There are hints that the frequency of the fission may exhibit geographic or microgeographic variation within the range of P1, but present sample sizes and localities are poorly chosen for exploring this question. Any treatment of this possibility must depend on further collecting and should be designed to distinguish between differences in frequency due to drift in local breeding populations vs systematic differences due to selection along geographical or ecological gradients. However, it is clear that the FIS-1 polymorphism is well established and widespread in the P1 population.
Clearly, the FIS-1 polymorphism is not selectively disadvantageous, or it would not have spread over the approximately 2,500 km2 range of P1. I have not yet thoroughly studied meiotic assortment in FIS-1 heterozygotes, but a check of 500 second meiotic metaphase cells from one Individual showed a 0.022 frequency of malassortment products from the fission trivalent (3 spreads showed duplications and 8 showed deficiencies for arms of the fissioned chromosome 1). No cells were scored that were aneuploid for only one other macro- chromosome. If these counts represent an actual rate of malassortment and are not an artifact of preparation, and they are typical for all heterozygotes, and if the aneuploid sperm function in fertilization, then there should be a 0.022 selective disadvantage for the heterozygous condition in males. Of course, I have no information on meiotic assortment in female meiosis. If the fission was genetically neutral, and the single specimen examined was typical, one would think that the presumed anti-heterozygote selection should have been enough to eliminate it from the population soon after its origin. Four counter possibilities are offered:
No attempt has yet been made to test any of these questions experimentally.
Populations designated F6 are all characterized by fixation of a centric fission of chromosome pair 6 in the standard karyotype (the FIS-6 mutation. Fig. 5c). F6 populations have a much wider distribution than does the P1 population. As indicated by my collections, F6 are continuously distributed in the humid forests of the central section of the Sierra Volcanica Transversal from western Michoacan to the eastern side of the Valley of Mexico (Figs. 7a, 7b; see also Figs. 2 and 3 in Hall and Selander, in press). On the east side of the Valley of Mexico, F6 occupy the humid forest belt between the S population found on the Valley floor and east and south of the Valley below about 2400 m, and the Pi population which is found above about 3200 m along the eastern divide of the Valley. An apparently disjunct population of F6 was found on the Nevado de Colima in Jalisco, to the west of the continuous range; while to the north, similarly disjunct populations occur along the intermediate eastern slopes of the Sierra Madre Oriental in San Luis Potosi, on a high peak above the town of Marcela in Tamaulipas, and near springs in three deep canyons in north central Nuevo Leon. All of these F6 populations inhabit what seem to be the most humid forest associations in their respective areas (Table 3). Oaks seem to be an important component of most forests where F6 were collected.
Aside from the obvious hybrids discussed below and by Hall and Selander (in press), only two FIS-6 heterozygotes have been karyo- typed. Both were from the north side of Cerro la Malinche in south- eastern Tiaxcala, but were separated by 12 km distance and 900 m in elevation. Both heterozygotes came from collections that contained respectively 7 and 8 karyotypically standard individuals. A third sample of 5 standard individuals was taken at an intermediate locality and elevation. Three possible explanations for these heterozygotes are offered:
More sampling and/or biochemical and morphological studies of the heterozygotes will be required before any of these possibilities can be assessed. Meiotic assortment has not yet been studied in any of the F6 hybrids or in these heterozygotes.
The fission-5 karyotype is standard except for fixation of a centric fission of chromosome pair 5 (the FIS-5 mutation. Fig. 5d). Ten individuals from two localities in western Chihuahua, separated by about 30 km, had this karyotype (Fig. 7a, see also Fig. 2, Hall and Selander, in press). The F5 lizards were found living on fallen logs in the pinon-juniper woodland association (Table 5). Presumably these samples are representative of a population of unknown extent on the northwestern border of the distribution of grammicus.
The F5+6 karyotype is characterized by fixations of both the FIS-5 and FIS-6 mutations (Fig. 5e), Two widespread and possibly connected populations are characterized by this karyotypic pattern (Fig. 7a, see also Fig. 2, Hall and Selander, in press).
One is a high elevation population, centered on the dry plateau lands of San Luis Potosi, Guanajuato, Queretaro, and northern Hidalgo, Ecologically this population ranges from the upper Chihuahuan Desert, where the lizards inhabit Yucca, Agave, Opuntia, and occasionally mesquite trees, to comparatively dry oak and pine woodlands at higher elevations and in the southeastern parts of its range, where the lizards are found on logs and trees (Table 6).
The second, a coastal population, was found sporadically from the lower Rio Grande Valley of Texas and Mexico south along the coastal hills and foothills of the Sierra Madre Oriental. I do not know if these sampled populations are now connected, as the area has been too poorly collected to tell. These lizards were found only on large mesquite trees along arroyos or, in one case, on scrub oak at the top of the coastal Sierra San Carlos in Tamaulipas (Table 6).
It also seems possible that the coastal and plateau populations either are now or were in the past united through connections in the area of southern Tamaulipas, northern Veracruz, and northern Hidalgo. A few grammicus in museum collections were taken from these areas, but I do not yet have any karyotypic data from them, nor have I determined the validity of the localities.
An almost certainly introduced F5+6 population is found within the city of Kingsville, Texas. Its ancestors were probably brought in on "ironwood" logs esteemed by the local ranchers for fence posts, which were presumably cut in the once extensive wooded areas along the flood plain of the lower Rio Grande River. Grammicus were seen on these mesquite-like trees in the Santa Ana Wildlife Refuge, Alamo, Texas, which is one of the few areas where the trees still survive in any number. Since grammicus from near Rio Grande City, upstream, were F5+6, I assume that the lizards from the Alamo locality also belong to this population, although I was not able to karyotype any.
The populations given this designation are cytogenetically by far the most complex of any lizard material I have examined; and most unfortunately, the average quality of the available preparations is poorer than that of most of the other forms treated here, even without considering material from animals which were dead for several hours before being processed. Furthermore, laboratory analyses of this material are not yet complete. However, some tentative generalizations about the distribution and cytogenetics of these populations can be made at this time.
The FM designation is given to karyotypes in which most or even all of the macrochromosomes have fissioned and to the populations characterized by these karyotypes. FM grammicus have been collected only in an area approximately 100 km square straddling the Hidalgo-Mexico state line (Fig. 7a, 7b, and see Figs. 2 and 3 in Hall and Selander, in press), and all of the collection localities were either from semi- cultivated areas where the lizards lived on Opuntia, Agave, or occasionally on Yucca or trees and/or from edificarian habitats (Table 7 and Table 8). Except for unquestioned hybrids and backcrosses with the Standard grammicus in the hybrid zone described below, all FM individuals are homozygous for fissions of macro chromosome pairs 2 (FIS-2), 5 (FIS-5), 6 (FIS-6), and one of the two pairs 3 and 4, which cannot be reliably distinguished in the S karyotype (the fissioned pair is arbitrarily designated 3 and the mutation FIS-3). Given the limitations of the available material, it appears that the FM grammicus may actually be subdivided into two cytogenetically and geographically distinguishable populations.
Ten grammicus from the western and two northernmost areas where FM individuals were collected, besides being homozygous for the four fissions, FIS-2, FIS-3, FIS-5, and FIS-6, were polymorphic for fissions of the two remaining pairs of metacentric macrochromosomes: FIS-1, ~ 50% frequency and FIS-4, ~ 10% frequency (Fig. 5f) . Although I have not been able to count the microchromosomes of all of these individuals with certainty, where I am confident of the counts, all had the standard 19 or 20 micros, depending on their sex. Lizards from these northern and western localities are designated FMl to distinguish them from the bulk of the FM specimens, described below and designated FM2, which all seem to show an even higher state of chromosomal fissioning.
All of the grammicus falling into the FM2 category were collected in an area about 55 x 20 km2 between Pachuca (state of Hidalgo) and San Juan Teotihuacan (state of Mexico). These FM2 differ most importantly from FMl and all other crevice-using Sceloporus by having an extra pair of microchromosomes, presumably generated by centric fissioning of one of the metacentric micros (the FIS-m mutation) (Fig. 5g). Other differences of FM2 from FMl are a higher frequency of FIS-4 (~50% frequency, rather than ~10%), and the complete or nearly complete fixation of FIS-1 (frequency 50% in FMl). Whether the metacentric-1 condition is present at a low frequency in the FM2 population or completely absent will be impossible to determine without better material than I presently have available.
The arm ratios of pairs 1 and 4 in the Standard karyotype are similar, and when only one or two metacentric chromosomes are present in a somewhat fuzzy mitotic spread from FM2, it is difficult to decide to which of the two pairs the metacentric(s) should be assigned. However, as nearly as can be determined, the majority of FM2 with only one metacentric were heterozygous for the metacentric of Standard pair 4, although a few of the metacentrics might have been large enough to fall in the pair 1 size range. Among the individuals with two metacentrics, two reproductively mature males each showed two macrochromosomal trivalents in diakinesis spreads, a clear indication that their two metacentrics were non-homologous. However, both individuals came from near the contact zone with Standard grammicus, and in one of them both trivalents seemed to fit in the 3-4 size class, suggesting that this might be a backcross individual (see discussion on hybridization below). The second doubly heterozygous male, although one of its metacentrics is almost certainly in the pair 1 size range, showed one or two extra microchromosomes beyond the extra pair typical of all FM2 in mitotic spreads and one or two extra "bivalents" in diakinesis spreads. This lizard can also be reasonably interpreted as a backcross, with the extra microchromosome(s) resulting from meiotic malassortment in the hybrid parent (it should be remembered that Standard grammicus have 16 microautosomes while the FM2 have 18). Finally, two FM2 females which could not be definitely identified as backcrosses showed two metacentrics in the pair 3-4 size range and one in the pair 1 size range. However, both did come from near the contact zone so that the possibility that they are backcrosses cannot be discounted. Hopefully, the electrophoretic analyses of these specimens being carried out by Sheldon Guttman will eventually allow them to be precisely allocated.
We may then tentatively conclude that there are actually two cytogenetically distinct FM populations: FMl, which has the standard grammicus microchromosomal pattern (19 in the male, 20 in the female), a polymorphic FIS-1 condition with the metacentric present at about a 0.5 frequency, and a low frequency (0.1) polymorphism for FIS-4, and therefore with 2n's in the female usually ranging between 40 and 42 but possibly going as low as 38; and FM2, with 21 microchromosomes in the male, 22 in the female, a fixed or nearly fixed FIS-1 condition, and a polymorphic FIS-4 present at about a 0.5 frequency, and therefore with 2n's in the female usually between 44 and 46. Further reports on the cytogenetics of the FM populations will be deferred until more collections and better preparations are available.
The six or seven cytogenetically distinctive "races" of grammicus show a mosaic distribution in Mexico (Fig. 7a, 7b) see also Figs. 2 and 3 in Hall and Selander, in press) . Some form of grammicus can be found, at least in scattered local populations, almost anywhere in Mexico above 1000 m elevation north of the Isthmus of Tehuantepec where their preferred escape cover, wood or plant crevices can be found. On the plateau they seem to be completely excluded only from the hottest deserts; and, on the other hand, some grammicus populations extend almost to sea level to the east of the plateau. Furthermore, grammicus do not always seem to be strictly limited by the absence of their preferred cover, since I have occasionally found them using rock crevices. Therefore, because of their wide distribution on the Mexican plateau, most of the chromosomal races must at least occasionally contact neighboring populations, if they are not in fact constantly in contact. Yet, despite these probable contacts, I have found neither overlaps nor wide zones of polymorphism between karyotypically distinctive populations. (Although it is barely possible that the FMl population represents the latter situation.)
For example, in the northern half of Mexico, based on karyotyped specimens, two distinctive races have been found within about 50 km of one another in several areas: F5+6 and S, in central Nuevo Leon and also near Ciudad Zacatecas; F6 and S, at the southern border of Nuevo Leon and Tamaulipas; and F6 and F5+6, in eastern San Luis Potosi. Similarly, F5+6 and FMl, and FMl and FM2 collections are separated by no more than about 70 km. In none of these areas is there evidence for geographic or ecological barriers which would separate the respective populations. There simply has not been an opportunity to do the necessary collecting in the intervening areas to establish the nature of the contacts. Also, although the intervening terrain is some of the most inaccessible to collectors in all of North America, there seem to be no vegetational, physiographic, or climatic barriers separating the F5 grammicus of Chihuahua from the S population of Durango. The failure to find any indication of gradual transition from one cytogenetic system to another in any of these cases must be significant.
However, in several areas of the Valley of Mexico (see Fig. 3, Hall and Selander, in press), intensive collecting has pinpointed geographic contacts between three sets of karyotypically distinct populations:
Lizards heterozygous for the chromosomal conditions fixed between their respective "pure" populations were recovered from each of these contacts and are presumed to represent hybrids between the pure populations. In each of these contacts (except the incompletely examined S x F6) hybrids were found in belts of parapatric contact no more than 500 meters wide. This width does not exceed a reasonable dispersal distance for single individuals. The hybridization is therefore parapatric according to the nomenclature of Woodruff (MS).
Hybridization between P1 and F6 populations has been analyzed in Hall and Selander (in press) and will be summarized here. One fixed chromosomal difference (FIS-6) and two fixed isozyme differences (LDH-2 and GOT-1) were used as genetic markers to determine the ancestry of 153 individuals collected from a transect through the zone of parapatric hybridization on Cerro Potrero near the town of Rio Frio (see Fig. 3b, Hall and Selander, in press). In the sample from this transect there were 13 presumptive F1 hybrids, as indicated by their heterozygosity for all three genetic markers. Such F1 hybrids were apparently quite fertile, as shown by the many presumptive backcross individuals, heterozygous for only one or two of the genetic markers. Twenty-seven of these were backcrosses with F6 and another 29 were P1 backcrosses. Yet, there was no evidence of either an F2 generation or of introgression into F6 populations beyond the first generation of backcross; and, if there was any introgression into P1 beyond the first generation backcross, it was very slight and the evidence for it equivocal at best. The apparent deficiency of certain marker combinations in the back-crosses suggested that some recombination products survived poorly, and the lack of evidence for any second or later generation introgression suggested that those backcrosses surviving to adulthood were effectively sterile. Hall and Selander (in press) therefore concluded, notwithstanding the unquestionable evidence for both hybridization and back-crossing, that the P1 and F6 populations were genetically isolated from one another, and were therefore good biological species--at least with respect to one another, even though individuals of the two populations did not appear to behaviorally recognize this fact. This conclusion was supported by the electrophoretic evidence that other aspects of the genetic systems of the two species were more different in samples collected only two km apart on either side of the zone of hybridization than they were between two FIS-6 samples collected some 500 km apart.
A similar but less complete analysis is possible for the contact between Standard and FM2 in the valley of San Juan Teotihuacan, using as genetic markers the four (or five) macrochromosomal differences fixed between the two populations. Seven-presumptive F1 hybrids were found in some 150 karyotyped lizards from the mapped area of the Teotihuacan Archeological Zone (Millon, 1970). An eighth F1 hybrid was found in a contact area about 12 km E of the Archeological Zone. Again, as in the F6 x P1 hybridizations, F1 hybrids were not completely sterile, as revealed by the recovery of unquestionable backcrosses: one to Standard and three to FM2. Subject to the limitations in the data discussed above (pp. 50, 68), four more individuals might also be back-crosses to FM2 on the basis of being double heterozygotes or because they have three metacentric macrochromosomes. By comparison with the F6 x Pi hybridization, it would seem that there are fewer hybrids here and that these hybrids apparently have a much lower fertility in backcrossing. However, this might be artifact from differences between the population structures of the grammicus in the two areas and our sampling of it. In the Rio Frio area, lizards are concentrated on more-or-less randomly dispersed logs and dead trees, while in the Archeological Zone they are concentrated along the gridwork formed by the ancient walls of the old city (Millon, 1970).
Details of the microgeography of hybridization between S and FM2 will be reported when biochemical studies of the collections are completed by Sheldon Guttman [2003 note: unfortunately the biochemical studies were not completed]; however, some general conclusions from the preliminary mapping of the collection localities of the specimens and our karyotypic determinations of them can be given here. Intermediate scale mapping of the two populations indicates that they meet along an approximately east-west line following the valley of San Juan Teotihuacan (see Fig. 3a, Hall and Selander, in press). Within the Archeological Zone, as mapped by Millon (1970), our most intensive collecting was done in his quadrangles N2W1, N3W1, and N4W1 west of the "Street of the Dead" (see Fig. 1 in Millon, 1970). Apparently, in the area between the "Pyramid of the Moon" and the "Explorations of 1917," the "Street of the Dead" relatively effectively impedes local dispersal, as pure FM2 are found on the east side. (Three FM2 were taken from an ancient wall leading away from the "Puma Mural" and one F, hybrid was taken in the "Compound of the Four-Temple Complexes.") Most of the hybrids were found in N2W1 and N3W1 along a line running nearly NNW from the region of the "Explorations of 1917." The pure FM2 population was more-or-less restricted to the wedge-shaped area of lava and tufaceous soil figured by Mooser (1968) between this line and the "Street of the Dead." Standard grammicus are then found to the west of this line and to the east of the "Street of the Dead" mostly on alluvial soil. In this area of rather favorable habitat for grammicus, the hybrid zone was no wider than about 300 m.
As in the case of hybridization between P1 and F6 grammicus, though the hybridization between FM2 and S definitely involves backcrossing, there is no evidence from chromosomal markers for genetic exchanges beyond the first generation backcross. Therefore, by the criteria used by Hall and Selander (in press) to show that Pi and F6 are good biological species with respect to one another, FM2 and S are also good species-at least with respect to one another. By extension, if the contacts between the remaining chromosomal races prove to be similar to those described above (and there is no reason to think otherwise), then these other cytogenetically distinctive populations of grammicus are also probably good species with respect to one another.
During the present study, several karyotypically novel grammicus were found, including four with unquestionably triploid karyotypes. Three were morphological males: respectively, a Standard from southern Coahuila, a P1 from near Rio Frio, and an FM2 from the Teotihuacan Archeological Zone, These and other novelties will be described in future works, but the fourth triploid lizard defies easy categorization and deserves special mention here, as it may have a bearing on possible consequences of hybridization between FM2 and Standard. This animal is a morphological female with two standard chromosome sets plus one FM2 set and was collected in or near the zone of hybridization between FM2 and Standard.
With respect to this unique triploid female, it is worth note that six of eight F1 hybrids from the valley of San Juan Teotihuacan were female. The sample is too small for the female bias to be statistically significant, but these observations suggest the interesting speculation that the triploid individual resulted from fertilization of a parthenogenetically developing unreduced hybrid ovum. Hybrid origins have been suggested for most parthenogenetically reproducing animal populations (Maslin, 1968; Asturov, 1969; Schultz, 1969), and in almost every case where diploid parthenogenones are sympatric with closely related bisexual species, sterile polyploid hybrids and/or polyploid parthenogenetic species have been found (previous refs., Neaves, 1971; Lowe et al., 1970), indicating that parthenogenetically developing ova may easily be fertilized. Furthermore, parthenogenetic populations are known or are suspected to occur in the Agamidae and Chamaeleontidae (Hall, 1970), the two families most closely related to the Iguanidae. It is therefore not completely unreasonable that an incipient case of parthenogenesis was found in the S x FM2 hybrid zone. However, other explanations for the triploid female cannot be discounted, as triploidy has clearly occurred in grammicus in cases where a hybrid origin is not suspected. But, conversely, if these four cases of triploidy occur randomly in my sample of 1300 grammicus, the probability that one of them would be included in a sample of nine hybrids, such as that taken from the Teotihuacan hybrid zone, is about 0.03.
This situation suggests the further speculation that incipient parthenogenetic clones induced by hybridization may be not uncommon in natural zones of hybridization. However, if as is the case in both of the grammicus hybrid zones studied, the population density of males is high enough and the reproductive isolation low enough that such parthenogenetic clones would be bred into genetic difficulties within a few generations by increases of polyploidy because of the parthenogenones( inability to prevent males from fertilizing already developing ova. If this speculation pertains, it may be possible to find and isolate cases of such incipient parthenogenesis under conditions of controlled hybridization as Schultz (1973) has done in an analogous situation.