1915.] Allen, Review of the South American Sciuride. 167 
are not sharply defined, as with the posterior border of the nasals in mam- 
mals. In the work here presented all measurements were taken with 
dividers in a straight line between the most extreme points of the parts 
measured. In the case of the nasal bones in squirrels, it is difficult to 
measure the same specimen twice alike; and the difference of half a milli- 
meter in the length of the nasals will very essentially affect the result in 
computing, e. g., the ratio of the length of the nasals to the interorbital 
breadth. In such cases there is, further, the element of the personal equa- 
tion whenever the work of two or more investigators is subjected to com- 
parison. In the case of averages, such slight aberrations are likely to offset 
each other, and hence the only trustworthy ratios are those based on a 
normal determined from averages. 
Pelage and coloration. 
Squirrels which live in temperate and boreal zones have always soft, 
thick fur, with abundant underfur, the thickness of the coat being corre- 
lated with climatic conditions; the coat is hence longer and thicker in winter 
than in summer. A similar correlation would be expected in squirrels 
inhabiting tropical and subtropical zones. It not only prevails similarly 
under tropical conditions, but here environment over-rides genetic relation- 
ship. In tropical lowlands the pelage is shorter and thinner, usually with- 
out underfur, and often more or less coarse and hispid, in contrast with that 
of species of the same genus, or even with that of subspecies of the same 
species, living at altitudes of 6000 to 12,000 feet in nearby mountain ranges. 
This is true not only of the pygmy squirrels of the genus Microsciurus, but 
of various members of Kchinosciurus (as mentioned repeatedly by Nelson 
(J. c., p. 21 and passim) of Mexico and Central America, and in Microsciurus, 
Mesosciurus, and Guerlinguetus in South America. The character of the 
pelage has therefore little taxonomic significance except among conspecific 
subspecies. | 
The case is quite different in the matter of coloration, which, as has 
often been observed, is frequently a trustworthy index of genetic relation- 
ship. The Tamiasciurus group of North America has a distinctive style 
common to the group as a whole, but it is not so strongly marked as in the 
ground squirrels of the genera Tamias and Eutamias. What would be 
called a distinct ‘pattern’ of coloration is absent in most tree squirrels, 
where, in some groups, individual variation in color runs riot, as 1s excep- 
tionally exemplified in the Echinosciurus group of Mexico and Central 
America. On the other hand, certain groups, as Microsciwrus, Guerlin- 
guetus, Urosciurus, and a numerous group of species and subspecies in Meso- 
