Abstract: For 3 yr we studied the reproductive responses of desert rodents in the Owens Valley of eastern California (average annual precipitation 14 cm): four nocturnal heteromyids–the kangaroo rats Dipodomys microps and D. merriami and the pocket mice Perognathus formosus and P. longimembris–and one diurnal sciurid, the antelope ground squirrel, Ammospermophilus leucurus. Reproductive status was assessed by autopsies of adults trapped at approximately monthly intervals. Reproduction differed conspicuously among the five species. Our analysis illustrates effects of body size, phylogenetic association, and adaptation to the desert environment upon reproductive performances and associated life—history parameters. Most breeding occurs in late winter and early spring. Winter rains cause a series of pulses in vegetation growth and an attendant increase in availability of water in food plants, which contribute to rodent reproduction. Among the four heteromyids, onset of breeding is sequential according to body size, with the largest first. Pocket mice hibernate (P. formosus typically 3½ mo, P. longimembris 6½ mo), which restricts their breeding season compared to that of Dipodomys, but breeding normally begins following hibernation. The males of all species precede females in reproductive readiness; sperm production begins 1@2 to 2 mo before mating begins. Some male D. merriami remain spermatogenic throughout the year, and the mating season of this species is the longest (typically 2½ mo) and most variable of any of the species. D. merriami typically produces only two young, which are weaned in just less than 3 wk. It can breed repeatedly under favorable conditions and is the only species in which we observed reproductive maturity of both male and female juveniles in the season of birth. D. merriami has the highest annual reproductive potential of any of the five species studied. D. microps, although larger than D. merriami and sharing similar traits of small litter size and rapid growth, has a more restricted mating season (typically 1½ mo), but its breeding success generally exceeds that of D. merriami. The diet of saltbush leaves consumed by D. microps is atypical within this generally granivorous rodent family. Saltbush is a perennial shrub with highly predictable spring growth of leaves that are used by lactating mothers and developing young. Consequently the breeding season of D. microps is less variable and shorter than that of D. merriami. D. microps typically produces one litter per year and the juveniles typically do not mature sexually in the season of their birth. Due to their small size, seasonal dormancy, and restricted reproductive season, pocket mice are more prone to reproductive failure than are Dipodomys. We observed a complete reproductive failure in both species of Perognathus in a year when winter—spring temperature was below average and precipitation only 47% normal. Perognathus have larger litter size (°5 young) than Dipodomys. Consequently, the total annual reproductive potential of Perognathus is close to that of Dipodomys. The relative energy investment and attendant risks for production of a given litter are considerably greater in Perognathus than in Dipodomys, particularly in P. longimembris, which is at the lower limit of body size in rodents. Nonetheless both species of Perognathus have the potential for breeding twice inan unusually favorable year. The pattern of reproduction of the marmotine sciurid A. leucurus contrasts sharply with that of heteromyids. It breeds only once a year, at a fixed time and with a mating season that lasts only 2 wk. Litter size is larger (average 8 or 9) and more variable (range 5—14) than that of any of the heteromyids. Growth and development are slow: 8 wk to weaningin contrast to 〈 3 wk in the heteromyids. The breeding strategy of A. leucurus differs quantitatively from that of other marmotine rodents; it remains homeothermic throughout the year, never hibernates, and has one of the largest litter sizes and slowest growth rates of any sciurid. We suggest that Dipodomys and Ammospermophilus can reproduce when annual precipitation is as low as one standard deviation below the mean, whereas Perognathus generally require more precipitation for successful reproduction due to the short available time between the end of hibernation and the onset of summer drought. The five species differ in year—to—year flexibility in response to variable food supply. Dipodomys show little variation in litter size, but adjust frequency of reproduction and time for maturation of young. Perognathus have a greater range of potential litter size and less potential for adjusting frequency of litters. For A. leucurus frequency is fixed but litter size is highly adjustable; at early stages (ovulation, fertilization, implantation) it can be high, and later adjusted downward during gestation by resorption of of embryos and during lactation by nestling attrition. Dipodomys and Ammospermophilus represent alternative strategies of reproductive timing and flexibility, and each responds to environmental seasonality and unpredictability on a different time scale. D. merriami is a "pulse matcher"; it responds directly and quickly to the occurrence of pulses of food production. A leucurus is a "pulse averager"; its slow and prolonged reproductive effort is programmed to match the historical probability distribution of precipitation and food production. Despite the diversity of reproductive tactics in the five species, they are all long—lived and maintain relativley stable population levels, which is critical not only to reproduction itself but to survival between successive breeding opportunities in a desert environment.

Abstract: Rates of oxygen consumption during hover-feeding of wild, unrestrained, adult male Anna hummingbirds (Calypte anna) were measured with an artificial outdoor feeder converted into a respirometer mask. A computer sampled changes in O2 concentration in air drawn through the mask, automatically detecting the presence of a hummingbird from a drop in the O2 concentration, and photoelectrically timing the duration over which the feeder functioned as a mask. Birds coming to the feeder were weighed on a trapeze perch suspended from a force transducer. Feeding bouts consisted of sallies which carried the head in and out of the feeding mask about once a second. The volume of O2 consumed per feeding sally was linearly related to the length of the sally. The energy cost of hover-feeding in five hummingbirds, mean mass 4·6 g, was 41·5 ± 6·3 ml O2g−1 h−1.