Black (1937:195) and Cockrum (1952:180-181) reported stomach contents of P. b. attwateri from Cherokee County containing acorn pulp, seeds, and insects. Analysis of 38 stomachs of the brush mouse (Table 2) show acorns to be the most commonly used food in winter and spring. Seed coats were only rarely found, and insects were absent. Two captive females preferred acorns. Live beetles and grasshoppers of numerous kinds were decapitated and their inner parts eaten. Seeds (wheat, corn, and oats) were also eaten. Inasmuch as acorns appear to be the chief food, it is not surprising that the brush mouse is usually found on cliffs that support stands of blackjack oak (Quercus marilandica). Other oaks are present, but I have no evidence that the brush mouse eats their acorns. A. Metcalf told me that he observed in December, 1960, a released brush mouse interrupt its movement toward a hole in a cliff-face along Cedar Creek, Cowley County, in order to pick up an acorn (judged to be from the blackjack oak) in daylight. The mouse carried the acorn into the hole in the cliff. I have observed that captive brush mice eat acorns of the blackjack oak but not some other kinds of acorns.
Behavior
The chief differences observed between the brush mouse and other species of the genus Peromyscus in Kansas can be summarized as follows: the brush mouse is a superior and more cautious climber; seldom jumps from high places when under stress; is capable of finding its way better in partial darkness; has a stronger preference for acorns; and sometimes buries or hides seeds or acorns. These are all behavioral adaptations that seem in harmony with its mode of life.
Buck, Tolman, and Tolman (1925) showed the balancing function of the tail in Mus musculus. Climbers (for example, squirrels) often possess long, well-haired tails. It is reasonable to suggest (as did Hall, 1955:134) that the long, tufted tail is an adaptation for a scansorial existence. Little observation is necessary to observe how such a tail is used in balancing. Furthermore, it is used as a prop when the mouse is climbing a vertical surface. Dalquest (1955:144) mentioned tree-climbing in P. boylii from San Luis Potosí, México. It may occur in P. b. attwateri or in P. b. cansensis also, but there is no evidence as yet to prove it.
The brush mouse can seldom be induced to jump from heights of two feet or more. Rather it tends to scamper downward or to remain in place. It often swings itself over an edge, holding to it by its hind feet, and sometimes to it lightly with its tail, and reduces a short jump by almost the length of its body. Such caution seems to be an adaptation in a mouse that lives as a climber.
Many animals of cavernous habitats have small eyes (see Dobzhansky, 1951:284). Some nocturnal animals (for example, owls) have large eyes. The brush mouse has large, protuberant eyes; it lives in the deep crevices and fissures of the cliffs on which it is found, but it is not strictly a cave-dwelling animal. Perhaps large eyes aid the brush mouse in performing activities in the partial darkness of a deep crevice or hole in a cliff. Brush mice experimentally placed in what appeared to be total darkness fed, built houses of cotton, and ran and climbed in the usual manner.
On several occasions the captive brush mice hid surplus seeds and on other occasions hid acorns by burying them and sometimes by placing them in a small jar. The mice never carried the surplus food into their house.
Black (1937:195) has claimed that the brush mouse builds a nest similar to that of the nest of the pack rat, Neotoma floridana. Hall (1955:134) doubts this to be the case. Dalquest (1953:144) described a nest of P. boylii from San Luis Potosí as seven inches in diameter, made of leaves, and found in a hollow tree. Drake (1958:110) noted that P. b. rowleyi lives in holes and crevices in rocky bluffs in Durango, México. I have found this to be the case for P. b. attwateri, as did A. Metcalf (unpublished) for P. b. cansensis. Nests of sticks and leaves were taken apart by Metcalf, and all sign indicated only the presence of the pack rat. I have observed that there are no such houses on the cliffs along Shoal Creek, Cherokee County, and that no pack rats have been obtained from there (pack rats have not been reported from Cherokee County). Blair (1938) found two brush mice (P. b. attwateri) in the house of a pack rat in Oklahoma. Nests of the brush mice that occur in Kansas have not been found.
A lactating, pregnant female (KU 81833) of P. b. attwateri, containing three embryos, was obtained on December 24, 1959, and shows that this subspecies breeds in winter. Accumulated records for the subspecies indicates year-round breeding (see Cockrum, 1952:181). Another female obtained on March 27, 1960, was probably lactating.
Pregnant females of P. b. cansensis (KU 84892, 84895, and 84890) were obtained from the type locality on April 1-2, 1961, containing 3, 4, and 5 embryos respectively. This indicates, perhaps, increased breeding in spring; five was the highest number of embryos found in brush mice in Kansas.