f sheaths. This
necessitates the cutting of the stems into suitable lengths for
carrying, and the stored material appears to be merely cut sections of
the stems. Close examination, however, discloses the heads within, and
shows that as in other instances seed storage is the end sought. These
pieces are packed beautifully parallel like so many matches, and vary
from a minimum length of 20 to a maximum of 37 millimeters, averaging
about 30. Count of 2 grams of the above _Sporobolus_ material shows that
there are 125 separate cut sections per gram, or a total of
approximately 680,000 pieces in this one lot of storage, indicating a
remarkable activity on the part of the individual rat (Pl. VIII, Fig.
1).

[Illustration: PLATE VII. FIG. 1.--DEN EXCAVATED ON RANGE RESERVE.

Vertical section through Den No. 1, of Table 1 (p. 20), showing the
complex system of burrows, some of them plugged with closely packed
storage (outlined in white), the depth of the den, and the widened
chambers centrally located.]

[Illustration: PLATE VII. FIG. 2.--CONTENT OF EXCAVATED DEN.

Storage content of Den No. 1 (Fig. 1, above), showing the separate
species of plants listed in Table 1. The rod is 1 meter long. The large
pile on the left is composed of seed-laden heads of crowfoot grama
(_Bouteloua rothrockii_), the large pile on the right consists of heads
of six-weeks grama (_Bouteloua aristidoides_), the pile of heads in the
center is desert plantain (_Plantago ignota_), and the smallest heap is
composed of buckwheat-bush seeds (_Eriogonum polycladon_).]

The number of lots of storage (24) studied in detail, extending as it
does over a period of three years with seasons of varying growth
conditions, is not sufficient to permit the construction of a curve
showing increase and decrease in quantity of stored material with
growing seasons and intervals between; but the results indicate a very
decided increase during the autumn storing season, and continuing large
well into the winter, since some outside material can still be obtained
until midwinter. From about February to April a decrease may be noted,
followed, if the spring growth of annuals be good, by a slight increase;
and we can very nearly predict the general character of the increases
and decreases by the precipitation and consequent growth conditions.

TABLE 2.--_Quantity of storage per den correlated with time of year and
growth conditions of preceding season (chiefly from Unit