76 
JOUENAL OF MAMMALOGY 
desert deer-mice are nearly or quite lacking in pigment, while those of 
mice from the more humid coastal regions vary from purplish to nearly 
black. Again, the fur of the ventral surface of the body is whiter, in 
desert forms, owing to the relatively greater length of the terminal 
unpigmented zone of the individual hairs. Akin to this is the fact that 
the ventral white area of the pelage tends to extend higher upon the 
sides of the body, a circumstance which would seem to render the ani- 
mal more, rather than less conspicuous. 
(4) The dorsal white stripes of the skunk are broader in the desert 
races.’ It has rarely been contended that the skunk owes its peculiar 
fur pattern to the need for concealment. The class of facts cited in 
the last two paragraphs seems to indicate that depigmentation, rather 
than concealing coloration per se, is the thing which results from life 
in arid regions. 
(5) I shall offer evidence below which tends to show that the need 
for concealing coloration on the part of these rodents has been greatly 
overestimated. 
The general correlation between depth of pigmentation and atmos- 
pheric humidity is, of course, a widely recognized fact. Not only 
may we contrast the pale desert races with their dark relatives from 
the humid coast belt of the northwest, but various intermediate sta- 
tions may be chosen, whose birds and mammals display intermediate 
shades in their feathers and fur.^^ This principle is not, of course, one 
of universal application, the correlation in respect to color-tone being 
very imperfect. Nevertheless, the phenomenon is so widespread and 
so well known that I need not even cite specific cases in the present 
discussion. 
Now the hypothesis of conceahng coloration through natural selec- 
tion might doubtless be advanced with a certain degree of plausibility 
to cover this entire situation. It could be pointed out that regions 
with more humid climates (at least those having a higher rainfall) have 
darker (appearing) soils than the less humid ones, and that therefore 
the effective correlation may be between pigmentation and soil color, 
rather than between pigmentation and humidity per se. But the fact 
that outside of the desert regions the ground, in a state of nature, is 
commonly covered by vegetation renders such an argument very 
unconvincing. Green grass is doubtless somewhat darker than dry 
^ See Grinnell (Univ. of California Pub. Zool., vol. 12, p. 257). 
Instances of this tendency among mice are discussed in papers by the 
present writer, in the American Naturalist (April-May, 1918,) and the Journal of 
Experimental Zoology (April 5, 1920). 
