Because r_{ max} represents the genetically fixed, physiologically determined maximum possible rate of increase, data on earliest possible age of female reproduction, highest possible birth rate of female young, and longest possible female reproductive life span were used for a, b, and n, respectively. Calculated values, therefore, represent physiologically possible, not ecologically possible, intrinsic rates of increase (Hennemann, 1983, 1984; Hayssen, 1984; McNab, 1984b). Values of n were derived from longevity records for captive animals, and as these were all large values of similar duration (14-16 years), they had very little effect on r_{ max}. All species considered have one litter per year, and because their sex ratios at birth are about 50:50, variation in b was due to differences in litter size. Therefore, age of first reproduction and litter size had the greatest effect on r_{ max}. Intrinsic rate of increase scales to body mass (Fenchel, 1974), and we removed this effect by comparing each calculated r_{ max} with the value expected (r_{ maxe}) on the basis of body mass (Hennemann, 1983).
Dimensionless numbers for each of the four variables used in calculating composite scores were derived as follows. Ratios of measured to predicted values were used for basal metabolism (H_{ br}) and minimum wet thermal conductance (C_{ mwr}). Thermoregulatory ability at low temperatures is closely related to the ratio H_{ br}/C_{ mwr} (McNab, 1966). This ratio was used, therefore, to gauge each species' cold tolerance. For D_{ d} we used the ratio of food categories actually used by a species to the total number of food categories taken by all species tested (D_{ dr}). The ratio of calculated to expected intrinsic rates of natural increase was used to derive r_{ maxr}. Composite scores were calculated as
Composite score = [(H_{ br}/C_{ mwr}) + D_{ dr} + r_{ maxr}]/3 Eq. 3
The correlation between number of climates these species occupy and their composite scores was tested by linear regression.
According to monthly live-trapping records, the body mass of free-ranging female raccoons increased from 3.6 ±0.6 kg during summer to 5.6 ±0.8 kg in early winter, and the mass of free-ranging males increased from 4.0 ±0.5 to 6.7 ±0.9 kg during the same interval. These seasonal changes in body mass were due to fluctuations in the amount of body fat and represent a mechanism for storing energy during fall for use in winter. In summer, captive and trapped male and captive female raccoons had the same body mass (4.73 ±0.61, 4.41 ±0.70, and 4.67 ±0.88 kg, respectively, Table 2). Mass of captive females did not change between seasons, whereas captive males were heavier in winter than summer (p<0.005; Table 2). This seasonal change in mass of our captive males was of a much smaller magnitude (0.6 kg) than that observed for wild males (2.7 kg). During winter, captive males (5.34 ±1.39 kg) were heavier than captive females (4.49 ±0.98 kg; p<0.005; Table 2). Thus, our captive animals maintained a body mass throughout the year that was intermediate to the range of values found for wild raccoons in the same area.
