of winter pelage remaining on the rump and likewise does not appear to have been actively molting. KU 72085, obtained on October 13, 1956, 4 mi. E Barada, Richardson Co., Nebraska, is in the process of molting from summer to winter pelage, but has tufts of old winter pelage on the rump.

Geographic Variation

Geographic variation, both in color of pelage and in external and cranial dimensions, is less in R. megalotis in the region studied than in most other cricetine species that occur there. Nevertheless, meaningful variation is present. The assumption that variation in R. megalotis paralleled in degree that of other species, Peromyscus maniculatus for example, led to untenable taxonomic conclusions by some previous workers.

Color of Pelage

Color of pelage is remarkably uniform, considering the geographic extent of the area involved, over most of the northern part of the central grasslands. Perhaps this uniformity results partly from the predilection of the western harvest mouse for grassy habitats, for in most areas on the Great Plains the species is restricted to riparian communities, principally along river systems, where soils, cover, and other conditions approximate those of corresponding habitats farther to the east to a much greater degree than do conditions in upland habitats. Differential selective pressure, therefore, theoretically would be less between eastern and western populations of R. megalotis than in an upland-inhabiting species. In any event, specimens from western Nebraska, Wyoming, northern Colorado, and adjacent areas average only slightly paler dorsally than specimens in corresponding pelages from the eastern parts of Nebraska and Kansas, and many individuals from the two areas can be matched almost exactly.

To the southwest, on the other hand, a trend toward paler (pale brownish, less blackish) upper parts is apparent. Specimens from southwestern Kansas and adjacent parts of Colorado and Oklahoma average slightly paler in comparable pelages than specimens from northeastern Kansas and eastern Nebraska, but most specimens from farther southwest, in northern New Mexico and southwestern Colorado, are discernibly, although not markedly, paler than mice from northern and eastern populations.

A "pectoral spot," fairly common in some populations of R. megalotis east of the Missouri River (see Hoffmeister and Warnock, 1955:162-163), is present in only a small percentage of the specimens we have studied, and when present is usually only faintly developed.

External and Cranial Size

Fig. 1. Geographic variation in five measurements of Reithrodontomys megalotis on the central Great Plains. The size of each sample is given, along with total length, length of tail expressed as a percentage of the head and body, length of ear, greatest length of skull, and length of rostrum. The approximate distribution of the species in the region shown and the approximate boundary between the subspecies R. m. aztecus and R. m. dychei also are indicated.

As seen in Figure 1, the tail and especially the ear are longer in mice from New Mexico and adjacent areas than in specimens from northern localities. The ear, only slightly variable in size in the northern part of the region, is markedly longer in the southwest, averaging more than 2 mm. longer in specimens from New Mexico and adjacent southwestern Colorado than in specimens from Nebraska and eastern Kansas; specimens in a zone from central Colorado through southwestern Kansas and adjacent Oklahoma generally have ears of a size between the two extremes. As concerns the tail we note a slight trend toward increasing length (best expressed as percentage of length of body) from north to south throughout the central plains, but in general the trend is more pronounced southwestwardly. Variation in length of tail and length of ear, therefore, appear to be in accord with Allen's Rule. Length of body and length of hind foot seem not to vary significantly in specimens we have studied.

The skulls of specimens examined differed only slightly, except that the rostrum is significantly longer and relatively, if not actually, narrower in specimens from the south and southwest than in mice from the rest of the region under study. The rostrum is longest (average 7.7 mm.) in specimens from the vicinity of the type locality of R. m. aztecus, but is relatively long (7.5-7.6 mm.) in populations from as far north as northeastern Colorado and southwestern Nebraska. An average greater occipitonasal length (greatest length of skull) in specimens from the south and southwest results mostly from the longer rostrum.

Recognition of two subspecies of R. megalotis on the central Great Plains seems justified on the basis of the geographic variation discussed above. One subspecies, for which the name R. m. aztecus is applicable, occurs in the southwest and is characterized by the culmination of trends in the region studied to paler upper parts, longer tail, longer ear, and longer, relatively narrower rostrum—characters that appear at least partly independent of each other as concerns gradation toward the smaller, darker-colored populations to the northward. The latter, while exhibiting some differences in color (slightly paler westwardly) and length of tail (shorter northwardly), stand more or less as a unit in contrast to the mice from the southwest, and represent, in our judgment, a single subspecies, R. m. dychei. The area of intergradation between the two subspecies is relatively broad, considering all the characters mentioned, and assignment of some intergrades is admittedly difficult.

Reithrodontomys megalotis aztecus J. A. Allen

Reithrodontomys aztecus J. A. Allen, Bull. Amer. Mus. Nat. Hist., 5:79, April 28, 1893 (type locality, La Plata, San Juan Co., New Mexico).

Reithrodontomys megalotis aztecus, A. H. Howell, N. Amer. Fauna, 36:30, June 5, 1914.

Reithrodontomys megalotis caryi A. H. Howell, Jour. Mamm., 16:143, May 15, 1935 (type locality, Medano Ranch, 15 mi. NE Mosca, Alamosa Co., Colorado).

Distribution.—Western and southern Colorado, southeastern Utah, northeastern Arizona and northern New Mexico, east to the panhandles of Texas and Oklahoma and to southwestern Kansas.

Remarks.—For comparisons with Reithrodontomys megalotis dychei, geographically adjacent to the northeast, see account of that subspecies.

When Howell (1935:143) named Reithrodontomys megalotis caryi from the San Luis Valley of Colorado he compared it directly only with R. m. megalotis from southern New Mexico and northern Chihuahua. Few adults were available to Howell from the San Luis Valley, accounting for the