AMERICAN BATS — HANDLEY 145 



3. No other two species of the subgenus were known to be sympatric. 



4. Relatively few specimens have been available for study. Until recently, 

 both species were represented only in the U.S. National Museum. The Texas 

 Cooperative Wildlife Research Unit has had an example of each species since 1942; 

 the Museum of Zoology, University of Michigan, has had both since 1946 (skins 

 and skulls of P. townsendii; P. mexicanus in alcohol) ; and the Museum of Natural 

 History, University of Kansas, has had both since 1948. 



5. With the exception of the type series of P. mexicanus (preserved as dry skins 

 and skulls), all early collections of Mexican big-eared bats in the National Museum 

 were preserved in alcohol, and in a random selection of specimens for removal 

 of skulls, only P. mexicanus happened to be chosen. The presence of P. town- 

 sendii, thus, was not detected. 



When Mexican material was assembled from several museums for 

 this study, variation was found to be slight in some series, great in 

 others. The entire sample segregated into two groups (figs. 16, 17), 

 each of which included both sexes and various ages. The amount of 

 variation far exceeded that observed in samples from other parts of 

 the range of the subgenus Corynorhinus (fig. 17). With one exception, 

 all specimens could be referred without hesitation to one group or 

 the other on the basis of a dozen diagnostic characters. Each speci- 

 men had most of the characters of its group, i.e., any specimen with 

 dark coloration also had small tragi and few cross-ribs on the inter- 

 femoral membrane, etc. (fig. 16). 



The two forms have distinct although broadly overlapping geo- 

 graphic ranges (fig. 18). The range of P. mexicanus — the cooler, 

 moister, higher elevations of the Sierra Madre Occidental, the 

 transverse volcanic belt of central Mexico, and the Sierra Madre 

 Oriental (Goldman and Moore, 1946) — is strikingly similar to the 

 ranges of the dark colored races of the harvest mouse, Reithrodon- 

 tomys megalotis, outHned by Hooper (1952, p. 51) and to that of the 

 white-footed mouse, Peromyscus difficilis (Osgood, 1909, p. 179). 

 The range of the other form, P. townsendii, is in part complementary. 

 It inhabits the lower elevations of the arid plateau and desert ranges 

 of north-central Mexico north of the transverse volcanic belt, and 

 the arid valleys of Jahsco, Morelos, and Oaxaca south of the trans- 

 verse volcanic belt. In addition, it occurs together with P. mexicanus 

 in the southern extremity of the Sierra Madre Occidental (Sierra de 

 Valparaiso, 8,200 feet) and in the transverse volcanic belt (Santa 

 Rosa, 9,500 feet, Convento de Acolman, and Lago Texcoco, 7,500 

 feet. However, it should be noted that the latter locahties are rela- 

 tively arid, despite their high elevation (Davis, 1944, p. 371; Gold- 

 man, 1951, p. 146). 



This pattern of distribution might exist without intergradation 

 between the two forms if they were: (1) Subspecies with seasonal 

 migrations; (2) ecologically isolated subspecies; or (3) distinct, 

 partially sympatric, species. 



