MORPHOh/IETRICS, EVOLUTION, AND CYTOTAXONOMY OF MAINLAND BATS OF THE GENUS MACROTUS

Davis, Breizt L . , and Robert J. Baker (Dept. Biology, and T h e Mliseum, Texas Tech Uniuersity, Lubbock, Texas) 1974. Morplzometrics, Euoltition, and Cytotaronomy of Mainlatld Bats of the Gentis hlacrotus (Chiroptera: Pl~~llloston~ati~lae). Slyst. Zool. 23:2639.-Chromosoma1 data reveal the existence of two mainland species of Macrotus that are parapatric in distribution: -Individuals can be identified to species by both chromoso~~~al features and cranial n~orphology. A stepwise multiple discriminant analysis and canonical variate analysis show tllat the two species are morphonletrically divergent. The distribution of these two taxa is proof that parapatry occurs in species with lugh vagility. If cl~romosomal data had not been available or if chronlosomal divergence had not accompanied speciation in Macrotus, this unique pattern of distribution probably would not have been detected. Allopatric, stasipatric, and centrifugal speciation are considered in light of the presently available data for Macrotus. [Macrottrs; morphometrics; cytotaxonomy.] Geographic variation in chron~osome The primary reason for studying the cynumber is rather uncommon among bats togenetics and morphometrics of Macrotus of the family Phyllostomatidae. Of the 77 was to determine the mechanism of evospecies of phyllostomatids thus far karyolution and the degree of hybridization betyped (Baker, 1973), four species show tween the two cytotypes. Robertsonian geographic variation-Uroderma bilobatum variation is the most likely mechanism by (Baker et al., 1972), Macrotus zoaterhousii which a chromosomal change of this nature (Nelson-Rees, et al., 1968), and Micronyccould occur (Nelson-Rees et al., 1972). If teris hirsuta and Vanapyressa pusilla (Baker centric fusions or fissions are the mechanism et al., 1973). by which the cytotypes evolved, loss or Chromosomal variation in Macrotus, gain of genetic material probably would which previously was thought to have two have occurred in the heterochromatic rechromosomal races, is of the Robertsonian gion (Jackson, 1971) and a cross between type (the fundamental number remains the two cytotypes should be successful, at constant but the diploid number varies). least on a chromosomal basis of homology. The cytotype with a diploid number (2N) Because of the relative ease with which of 46 and a fundamental number ( F N ) of specimens of Macrotus can be obtained and 60 has been reported as having a geographic because the species is chromosomally varirange from Alamos, Sonora, south through able, detailed studies were initiated. Morelos, and Guerrero (Nelson-Rees et Several possibilities are obvious from the al., 1968). The 2N =40, FN = 60 cytotype outset. First, the two cytotypes could be had a known distribution from Carbo, freely interbreeding. If this were the case, Sonora, north to Arizona and California. the genetics of the system could be examAlamos (the most northern published record ined by using chromosome number, morfor the 2N = 46 cytotype) is 335 kilometers phology, and meiotic studies. Second, the south of Carbo (the most southern reported two could be sympatric and produce no locality of the 2N = 40 cytotype). Varihybrids. Third, the two could have alloation in diploid number, as understood at patric distributions. the outset of this study, was confined to the Our studies have revealed a parapatric subspecies M. z(;. californicus (Anderson, distribution (as discussed for mammals by 1968, 1969; Nelson-Rees, et al., 1968). Vaughan, 1967) with no phenetic interMORPHOMETRICS OF MACROTUS FIG.1.-Skull of Macrotus showing cranial measurements used in statistical studies. Names of measurements are given in the text. mediates. Baker ( 1 9 7 0 ~ ) pointed out that ". . . studies of zones where two chromosomal races are contiguous could give considerable information on gene flow, dispersal potential, and isolating mechanisms in these populations." We have collected specimens for karyotypic analysis and, based on these and on specimens from museums, have evaluated the karyotypic and morphometric characters as potential population markers. In addition, this report evaluates the systematic status of mainland Mncrotus. MATERIALS AND METHODS Bats were trapped in nlistnets set over water or taken,from caves or abandoned mines. Karotypes were prepared using a modification of the in vivo bone marrow techniques described by Patton (1967) and Baker ( 1970b). The modifications were: a 5-hour in vivo culture after injection of from .20 to .25 milliliters of 0.04% vinblastine sulfate (Velban of Eli Lilly); a 25minute incubation of the bone marrow in 0.9% sodium citrate solution; and a total time in the first fixative of 25 minutes ( a 10-minute fixation before disruption of the cell button, followed by 15 minutes of additional fixation). The remainder of the procedure was as described by Raker (1970b). At least five somatic spreads were examined for each specimen. Metacentric, submetacentric, subtelocentric, acrocentric, and fundamental number ( F N ) were used as defined by Patton ( 1967). Cranial measurements were taken with dial calipers, calibrated in twentieths of millimeters, and rounded to the nearest tenth of a millimeter. The number before each measurement identifies that character at various places in the text and tables. Cranial measurements (Figure 1) were: (1) greatest length of skull, AR; ( 2 ) condylobasal length, CB; ( 3 ) occipitonasal length, AD; ( 4 ) mastoid breadth, EF; ( 5 ) braincase breadth, GH; ( 6 ) braincase depth, IJ; ( 7 ) interorbital breadth, KL; (8) posterior zygomatic breadth, k1N; ( 9 ) postpalatal length, OP; (10) maxillary toothrow, QR; (11) width at M2, ST; and (12) canine breadth, UV. External measurements used were (13) length of forearm and (14) length of the third metacarpal. Data were arranged into five procedures, labeled A thru E in text and tables. Procedure A (sexes pooled) consisted of all specimens and examined for intergroup divergence without regard for secondary sexual dimorphism. Procedures I? (males only) and C (females only) examined for intergroup divergence and eliminated the effects of secondary sexual dimorphism. Specimens in each of procedures A (sexes pooled), R (males), and C (females) were grouped for statistical analyses. Group 1 was all specimens with a karyotype of 2N = 40 and all specimens north of Quiriego, Sonora, which were not karyotyped (localities 1-16 and the 2N =40 component of locality 17). See specimens examined and Figure 2 for population identification. Not all of the Quiriego specimens were karyotyped. Certain of these specimens were grouped with the 2N = 40 cytotype based on morphometric analyses carried out prior to those reported here. These specimens were consistently classified with the 2N = 40 cytotype and were treated as such. Group 2 consisted of the northern Sinaloan locality 18. Group 3 included all 2N = 46 specimens and those from Quiriego that were (as in group 1) morphometrically similar to the 2N = 46 cytotype (localities 19-21 and the 2N = 46 portion of 17). Group 4 included the southern Sinaloan localities (22-24). Group 5 was the Jaliscan samples (localities 25-30). Group 6 included specimens from the Mexican states of Morelos, Puebla, Guerrero, and Oaxaca (localities 31-38). The last two procedures, D and E, were designed to examine geographic variation within each cytotype. Procedure D analyzed the 2N =40 cytotype, sexes pooled. Specimens for Procedure D were grouped as follows (letters identify that group in the figures and text): ( a ) localities 1 and 2; ( b ) localities 3-5; ( c ) localities 6-8; ( d ) localities 9-14; ( e ) locality 15; ( f ) locality 16; ( g ) locality 17 (2N = 40 race) ; and ( h ) locality 18. Specimens of Procedure E, analyzing the 2N = 46 cytotype, were divided into 7 groups as follows: ( i ) locality 17 (2N =46); ( j ) localities 19-21; ( k ) localities 22-24; (1) localities 25-30; ( m ) localities 33-35; ( n ) localities 31-32; and ( 0 ) localities 36-38. See specimens examined and Figure 2 for population localities. Statistical methodology has been described in detail by Baker et al. (1972). Univariate analysis of variance (ANOVA) was used to assess intergroup morphometric divergence. The effects of secondary sexual dimorphism was assessed by an ANOVA for a single population. I t has been show11 in mormoopid bats (Smith, 1972), in Desmodus personal communication Alberto SYSTEMATIC ZOOLOGY FIG. 2.-Localities from which specimens of Maci40tus were studied. Numbers identify that locality in the list of specimens examined. Cadena, and eluded to by Atchley (1971) for Chironomus Diptera: Chironomidae, that the amount of secondary sexual dimorphism may vary from population to population. Sample size prohibited multivariate analysis of sexual variation. For these rea5-ons results of the ANOVA for secondary sexual dimorphism is not presented. Canonical variate analysis aided in assessing the degree of multivariate divergence. A standardized canonical variate coefficient was computed by multiplying the canonical variate coefficients for a particular character by the pooled standard deviation. The standardized coefficient aids in indicating which characters account for the variation found within each canonical variate. The variables with a high coefficient are those that make the greatest contribution to the discrimination of the groups along that axis. The Mahalanobis generalized distance statistic (qp)was calculated for each pair of groups in each analysis and was also used to generate a probabilistic classification matrix to indicate the amount of phenetic overlap among groups. A stepwise discriminant analysis was performed, resulting in a listing of characters according to their discriminatory MORPHOMETRICS OF MACROTUS 29