16 BULLETIN 1091, U. S. DEPARTMENT OF AGRICULTURE. {buI^No^I, 
is maturing the animal's activities appear to be concentrated on this 
work. During September, 1919, when a good crop of grass seed was 
ripening following the summer rains, a kangaroo rat under observa- 
tion made repeated round trips to the harvest field of grass heads. 
Each outward trip occupied from 1 to 1^ minutes, while the unload- 
ing trip into the burrow took only 15 to 20 seconds. 
One individual in a laboratory cage, which had not yet been given 
a nest box, busied itself in broad daylight in carrying its grain supply 
into the darkest corner of the cage. When a nest box is supplied the 
individual will retreat into its dark shelter, and will only come forth 
after darkness has fallen unless forcibly ejected, but will store the 
food supplied. 
In another case an animal escaped while being handled, and sought 
refuge behind a built-in laboratory table, where it could not be re- 
covered without tearing out the table. For four days and nights it 
had the run of the laboratory. On the first night of its freedom it 
found and entered a burlap bag of grass seed that had been taken 
from a mound. A trail of seed and chaff next morning showed that 
it had been busily engaged in making its new quarters comfortable 
with bedding and food. After four nights of freedom it was cap- 
tured alive in a trap, and later it was found that it had moved from 
the corner behind the table to the space beneath a near-by drawer, 
where it had stored about 2 quarts of the grass seed and a handful of 
the oatmeal used for trap bait. 
BREEDING HABITS. 
Observations on breeding habits have consisted mainly in taking 
records from the females trapped at all seasons of the year through- 
out the course of the investigation, and from examinations made dur- 
ing poisoning operations, and yet from this source the number of 
pregnant females taken or of young discovered is disappointingly 
small. The records indicate a breeding period of considerable length, 
extending from January to August, inclusive. It is possible that the 
length of the period may be increased by a second litter from the 
earliest breeding females in summer, but the large percentage of 
nonpregnant or nonbreeding animals which occurs throughout the 
season would indicate a wide variation in the time of breeding of 
different individuals. 
Trapping in February and March for the purpose of securing 
greater numbers of female specimens, begun with the idea that these 
months were most likely to be the breeding months, has invariably 
yielded an unsatisfactory number of nonbreeding specimens and 
males. Unfortunately, the numbers of females secured in some 
months were not sufficient to be significant if worked out in per- 
