altitudes, where warm weather comes on much later and more abruptly, these breeding dates would be correspondingly later and the breeding season still more restricted.
While males and females occur in practically equal numbers, mating seems to be promiscuous; there is no permanent pairing off.
The number of young per litter, as ascertained from counts of embryos, varies from 4 to 11. The average, from very extensive records kept by the United States Public Health Service (McCoy, 1912, p. 1070), may be inferred to be very close to 7.2. The same records serve further to show that there is some variation in size of litter from month to month. The average for February is 6.9; for March, 7.3; for April, 7.5; for May, 6.8. The tendency seems to be toward slightly larger litters in April, which is beyond the date of maximum number of pregnant females (see fig. 12). Number of mammæ (nipples), which is usually six pairs in this species, occasionally but five, is no criterion for number of young per litter. All the evidence at hand indicates that each female raises but one litter each year. A female ground squirrel was taken on the University campus at Berkeley, on March 13, 1918, which contained eleven embryos each of which measured three-fourths of an inch long. Eight of these were contained in one branch of the uterus and three in the other. Another female taken at the same time contained eight embryos, each of which measured five-eighths of an inch long. Five of these were in one branch of the uterus and three in the other.
On April 6, 1918, G. R. Stewart and the junior author dug out two female ground squirrels which had been previously gassed in their burrows. One of these females was found in a nest with four small young which we took to be about ten days old, since their eyes were not yet open. These baby squirrels averaged 170 millimeters or 6¾ inches in length; a typical one weighed 61 grams, or a little over 2 ounces. They were well covered with hair, which already showed on the back the characteristic dappled pattern of the adult squirrel. The tail, however, was nearly round and showed little sign of the fringe of hairs along the sides. Their stomach contents showed no sign of their having eaten green vegetation or anything else than milk. Data from other sources indicate that the young are not completely weaned until they are at least half grown. The other female secured in an adjoining burrow not over ten feet distant was found to contain seven small embryos each of which measured three-eighths of an inch in length. These embryos could not well have reached full development short of two or three weeks, so we have a variation of nearly a month in time of birth at one locality.
Cases such as those just given are thought to be exceptional and may serve in part to explain the occurrence of late litters such as have been the basis of the claim that this animal has two litters a season. Litters of young squirrels which sometimes appear very late in the season are, too, likely to be merely the result of efforts to replace first litters of young which have met an untimely death. Thus two litters might be born in one season, though only one raised.
Shaw (1916, p. 4) gives 24 or 25 days as the period of gestation in the Columbian Ground Squirrel (Citellus columbianus) in the region about Pullman, Washington. The period of gestation of the California
27
