From the Eastern United States. 57 



more sharply bicolorcd tail and paler, grayish yellow eolor, with- 

 out traee of the russet usually seen in americanus, and much less 

 distinct dorsal stripe. 



The differences in color characterizing these two animals are 

 rather difficult to- describe, but nevertheless they are of such a 

 kind as to appeal immediately to the eye, especially when speci- 

 mens in the flesh are examined. In many adults of S. canadensis 

 the color of the dorsal surface is nearly homogeneous yellowish- 

 brown or grayish-brown throughout, with merely the faintest 

 possible trace of darkening in the mid-dorsal region. There is 

 usually .an indication of a very narrow yellowish line separating 

 the color of the sides from the white of the belly. This is apt 

 to be more distinct in the region of the cheeks and neck. The 

 white ventral surface has frequently a soiled yellowish cast, which 

 is oftenest met with in mid-summer. The pencil is usually 

 white, and this color frequently involves the whole tip of the 

 tail, sometimes for a distance of 30 mm., a feature very rarely 

 'seen in the shorter-tailed S.. americanus. Grayish examples of 

 amerieanus are sometimes met with among specimens, taken in 

 the summer, but with the exception of these very few of the 

 smaller race approach-in color even the brightest individuals of 

 S. canadensis. In the gray phase Sitomys americanus canadensis 

 bears a somewhat close resemblance to S. americanus arcticas 

 (Mearns), the type of which in the Museum of Comparative 

 Zoology at Cambridge, Mass., I have examined. The 'former 

 may, however, be at once distinguished by its much longer tail, 

 proportionally longer than in americanus, instead of propor- 

 tionally somewhat shorter, as is the case with arcticus. 



So far as I can see, Sitomys americanus canadensis shows no 

 cranial or dental characters to separate it from its near allies. 



As in all members of the genus, there is here considerable 

 variation in actual size as well as in proportions. This variation 

 for each form (americanus and canadensis) proves to be much less 

 than recent writers have generally accredited to " Hesperomys 

 leucapus." Both Allen (Bull. M. C. Z., i, 1869, 227, 228) and Coues 

 (Monog. N. Am. Rod., 1877, 53) allow a large range of variability 

 in the ratio of tail vertebras to total length. Nevertheless, this 

 character proves to be sufficiently constant to be of considerable 

 diagnostic value. Mr. Allen says (1. c., pp. 227-228) : "But the 

 most variable character consists in the relative length * * * 

 of the caudal vertebra?. About one-fifth of the Massachusetts 



