SCIURIDJE— sciurus. GG9 



or subspecific forms are permanently white below. The large number of 

 synonyms that have arisen in consequence of this variability in color, together 

 with the variability itself, render the recognition and characterization of the 

 species exceedingly difficult. The abundance of the material accessible for 

 the study of the species occurring north of Mexico renders this part of the 

 work comparatively easy, and demonstrates clearly the large amount of both 

 individual and geographical variation one must be prepared to recognize, 

 frequently at least, in the different specific representatives of this perplexing 

 group. Some of the Mexican species seem to be fully as variable in colora- 

 tion as any that are met with in the United States, so that color alone becomes 

 a wholly unsafe guide for the determination of the species. On the other 

 hand, Sciurus fossor, of the west coast of the United States, is as constant- 

 in its coloration as any Mammal with which I am acquainted, and the Brazilian 

 form of Sciurus cestuans, although widely distributed, presents only a small 

 range of variation in either color or size. The large Squirrel of Western 

 Brazil and the neighboring region to the westward (S. variabilis), though 

 subject to considerable variation in color, is far less variable than some of the 

 Mexican and North American ones. Three, at least, of the Central American 

 forms also vary less than many of the others : these are Sciurus gerrardi, 

 S. tephrogaster, and S. cestuans var. rufoniger, which range southward into 

 the northern States of South America. 



The material at my command, though large in amount (probably con- 

 siderably exceeding that ever before collectively examined by any previous 

 investigator), is quite insufficient for a satisfactory study of the tropical forms, 

 and the results arrived at are considered as open to future revision. Among 

 the numerous (some thirty or more) nominal species of authors, I have been 

 able to recognize only nine or ten that seem to me valid, with two additional 

 subspecies. These latter may possibly be entitled to specific rank, but I 

 think that, on the whole, the number of species will, by future investigation, 

 be still further reduced, rather than increased, and that I have erred in recog- 

 nizing too many species rather than too few. In the majority of instances, 

 I have been able to make the collocation of the nominal species with a good 

 degree of confidence ; in other cases, of course, only with much doubt ; while 

 two or three names I have been wholly unable to identify. One is doubtfully 

 American, and one or two I have provisionally adopted as possibly valid 

 species are yet unknown to me from specimens. 



In regard to the geographical distribution of the species, it is worthy 



