SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY · NUMBER 542

Metabolic Adaptation to Climate
and Distribution of the Raccoon
Procyon lotor and Other Procyonidae

John N. Mugaas, John Seidensticker,
and Kathleen P. Mahlke-Johnson

SMITHSONIAN INSTITUTION PRESS
Washington, D.C.
1993

ABSTRACT

Mugaas, J. N., J. Seidensticker, and K. Mahlke-Johnson. Metabolic Adaptation to Climate and Distribution of the Raccoon Procyon lotor and Other Procyonidae. Smithsonian Contributions to Zoology, number 542, 34 pages, 8 figures, 12 tables, 1993.—Although the family Procyonidae is largely a Neotropical group, the North American raccoon, Procyon lotor, is more versatile in its use of climate, and it is found in nearly every habitat from Panama to 60°N in Canada. We hypothesized that most contemporary procyonids have remained in tropic and subtropic climates because they have retained the metabolic characteristics of their warm-adapted ancestors, whereas Procyon lotor evolved a different set of adaptations that have enabled it to generalize its use of habitats and climates. To test this hypothesis we compared Procyon lotor with several other procyonids (Bassariscus astutus, Nasua nasua, Nasua narica, Procyon cancrivorus, and Potos flavus) with respect to (1) basal metabolic rate (Ḣ_b), (2) minimum wet thermal conductance (C_mw), (3) diversity of diet (D_d), (4) intrinsic rate of natural increase (r_max), and, where possible, (5) capacity for evaporative cooling (E_c). We measured basal and thermoregulatory metabolism, evaporative water loss, and body temperature of both sexes of Procyon lotor from north central Virginia, in summer and winter. Metabolic data for other procyonids were from literature, as were dietary and reproductive data for all species.

Procyon lotor differed from other procyonids in all five variables. (1) Procyon lotor's mass specific Ḣ_b (0.46 mL O₂·g^-1·h^-1) was 1.45 to 1.86 times greater than values for other procyonids. (2) Because of its annual molt, Procyon lotor's C_mw was about 49% higher in summer than winter, 0.0256 and 0.0172 mL O₂·g^-1·h^-1·°C^-1, respectively. The ratio of measured to predicted C_mw for Procyon lotor in winter (1.15) was similar to values calculated for Potos flavus (1.02) and Procyon cancrivorus (1.25). Values for other procyonids were higher than this, but less than the value for Procyon lotor (1.76) in summer. On a mass specific basis, Bassariscus astutus had the lowest C_mw with a ratio of 0.85. (3) Procyon lotor utilized three times as many food categories as Procyon cancrivorus, Nasua nasua, and Bassariscus astutus; about two times as many as Nasua narica; and nine times as many as Potos flavus. (4) Intrinsic rate of natural increase correlated positively with Ḣ_b. Procyon lotor had the highest r_max (2.52 of expected) and Potos flavus the lowest (0.48 of expected). The other procyonids examined also had low Ḣ_b, but their r_max's were higher than predicted (1.11-1.32 of expected). Early age of first female reproduction, fairly large litter size, long life span, high-quality diet, and, in one case, female social organization all compensated for low Ḣ_b and elevated r_max. (5) Although data on the capacity for evaporative cooling were incomplete, this variable appeared to be best developed in Procyon lotor and Bassariscus astutus, the two species that have been most successful at including temperate climates in their distributions.

These five variables are functionally interrelated, and have co-evolved in each species to form a unique adaptive unit that regulates body temperature and energy balance throughout each annual cycle. The first four variables were converted into normalized dimensionless numbers, which were used to derive a composite score that represented each species' adaptive unit. Procyon lotor had the highest composite score (1.47) and Potos flavus the lowest (0.39). Scores for the other procyonids were intermediate to these extremes (0.64-0.79). There was a positive correlation between the number of climates a species occupies and the magnitude of its composite score. Linear regression of this relationship indicated that 89% of the variance in climatic distribution was attributed to the composite scores. Differences in metabolic adaptation, therefore, have played a role in delimiting climatic distribution of these species.

It was clear that Procyon lotor differed from the other procyonids with respect to thermoregulatory ability, diet, and reproductive potential. These differences have enabled it to become a highly successful climate generalist, and its evolution of an Ḣ_b that is higher than the procyonid norm appears to be the cornerstone of its success.

Official publication date is handstamped in a limited number of initial copies and is recorded in the Institution's annual report, Smithsonian Year. Series cover design: The coral Montastrea cavernosa (Linnaeus).

Library of Congress Cataloging-in-Publication Data
Mugaas, John N.
Metabolic adaptation to climate and distribution of the raccoon Procyon lotor and other Procyonidae / John N. Mugaas, John Seidensticker, and Kathleen P. Mahlke-Johnson.
p. cm.—(Smithsonian contributions to zoology; no. 542)
Includes bibliographical references (p. )
1. Raccoons-Metabolism-Climatic factors. 2. Procyonidae-Metabolism-Climatic factors. 3. Raccoons-Geographical distribution. 4. Procyonidae-Geographical distribution. I. Seidensticker, John. II. Mahlke-Johnson, Kathleen. III. Title. IV. Series.
QL1.S54 no. 542 [QL737.C26] 591 s-dc20 [599.74´443´04542] 93-3119

The paper used in this publication meets the minimum requirements of the American National Standard for Permanence of Paper for Printed Library Materials z39.48—1984.

Contents

Metabolic Adaptation to Climate
and Distribution of the Raccoon
Procyon lotor and Other Procyonidae

John N. Mugaas, John Seidensticker,
and Kathleen P. Mahlke-Johnson

John N. Mugaas, Department of Physiology, Division of Functional Biology, West Virginia School of Osteopathic Medicine, Lewisburg, West Virginia 24901. John Seidensticker and Kathleen P. Mahlke-Johnson, National Zoological Park, Smithsonian Institution, Washington, D.C. 20008.

Introduction

Defining the Problem

Procyonid Origins

The major carnivore radiations took place about 40 million years before present (MYBP) in the late Eocene and early Oligocene (Ewer, 1973:363; Wayne et al., 1989). Between 30 and 40 MYBP, a progenitor split into the ursid and procyonid lineages, which evolved into present-day bears, pandas, and raccoons (Wayne et al., 1989). The taxonomic relatedness of pandas to bears and raccoons has been tested extensively and a number of authors have summarized current thinking on the problem (Martin, 1989; Wayne et al., 1989; Wozencraft, 1989a, 1989b; Decker and Wozencraft, 1991). Davis (1964:322-327) and others (Leone and Wiens, 1956; Todd and Pressman, 1968; Sarich, 1976; O'Brien et al., 1985) place the giant panda, Ailuropoda melanoleuca, with the ursids. The taxonomic status of the red panda, Ailurus fulgens, appears to be less certain. Some current investigations align the red panda with bears (Segall, 1943; Todd and Pressman, 1968; Hunt, 1974; Ginsburg, 1982; Wozencraft, 1984:56-110; 1989a), whereas others place them intermediate to procyonids and bears (Wurster and Benirschke, 1968; Sarich, 1976; O'Brien et al., 1985), or in close relationship to the giant panda (Tagle et al., 1986).

The procyonid radiation took place in North America and produced forms that were mostly arboreal and omnivorous (Eisenberg, 1981:122; Martin, 1989). The center of this diversification occurred in Middle America (Baskin, 1982; Webb, 1985b) during the Miocene (Darlington, 1963:367; Webb, 1985b). Fossil procyonids from the late Miocene are represented in Florida, California, Texas, Nebraska, Kansas, and South Dakota (Baskin, 1982; Martin, 1989) and include such genera as Bassariscus, Arctonasua, Cyonasua, Paranasua, Nasua, and Procyon (Baskin, 1982; Webb, 1985b). During the Miocene procyonids underwent a modest radiation within tropical and subtropical climates of North America's central and middle latitudes. Cyonasua, which has close affinities to Arctonasua (Baskin, 1982), appears in tropical South America in the late Miocene and immigrated there either by rafting across the Bolivar Trough or by island-hopping through the Antilles archipelagoes (Marshall et al., 1982; Marshall, 1988). Thus, procyonids were found on both continents prior to formation of the Panamanian land bridge (Darlington, 1963:367, 395; Marshall et al., 1982; Marshall, 1988). Origins of Bassaricyon and Potos are obscure but probably occurred in tropical rainforests of Middle America (Baskin, 1982; Webb, 1985b). A subsequent Pleistocene dispersal carried several modern genera ([Table 1]) across the Panamanian land bridge into South America (Webb, 1985b). Bassariscus and Bassaricyon represent the most primitive genera in Procyoninae and Potosinae subfamilies, respectively ([Table 1]; Wozencraft, 1989a; Decker and Wozencraft, 1991).

In the early Tertiary, mid-latitudes of North America were much warmer than they are now, but not fully tropical, and temperate deciduous forests, associated with strongly seasonal climates, occurred only in the far north (Barghoorn, 1953; Colbert, 1953; Darlington, 1963:589, 590). Major climatic deteriorations, with their attendant cooling of northern continents, occurred during the Eo-Oligocene transition, in the middle Miocene, at the end of the Miocene, and at about 3 MYBP (late Pliocene). This last deterioration corresponds with closure of the Panamanian isthmus (Berggren, 1982; Webb, 1985a). Climatic deterioration went on at an accelerating rate during the late Tertiary, with glacial conditions developing at the poles by the mid-Pliocene (Barghoorn, 1953). Therefore, throughout the Tertiary, as continents cooled, northern climate zones moved toward the tropics (Barghoorn, 1953; Colbert, 1953; Darlington, 1963:589, 590, 594, 595; Webb, 1985a).

Table 1.—Classification of recent Procyonidae after Wozencraft (1989a) and Decker and Wozencraft (1991). Information in parenthesis indicates general geographic distribution (modified from Kortlucke and Ramirez-Pulido (1982) and Poglayen-Neuwall (1975)): S.A. = South America; C.A. = Central America; M. = Mexico; U.S. = United States; C. = Canada. Lower case letters preceding geographic areas signify north (n), south (s), and west (w).

Order Carnivora Bowdich, 1821
Suborder Caniformia Kretzoi, 1945
Family Procyonidae Gray, 1825
Subfamily Potosinae Trouessart, 1904
Genus Potos E. Geoffroy and G. Cuvier, 1795
P. flavus (S.A., C.A., M.)
Genus Bassaricyon Allen, 1876
B. alleni[A] (S.A.)
B. beddardi[A] (S.A.)
B. gabbii[A] (nS.A., C.A.)
B. lasius[A] (C.A.)
B. pauli[A] (C.A.)
Subfamily Procyoninae Gray, 1825
Genus Bassariscus Coues, 1887
B. astutus (M., wU.S.)
B. sumichrasti (C.A., M.)
Genus Nasua Storr, 1780
N. narica[B] (nS.A., C.A., M., swU.S.)
N. nasua[B] (S.A., sC.A.)
Genus Nasuella Hollister, 1915
N. olivacea (S.A.)
Genus Procyon Storr, 1780
P. cancrivorus (S.A., sC.A.)
P. gloveralleni[C] (Barbados)
P. insularis[C] (Maria Madre Is., Maria Magdalene Is.)
P. lotor[C] (C.A., M., U.S., sC.)
P. maynardi[C] (Bahamas, New Providence Is.)
P. minor[C] (Guadeloupe Is.)
P. pygmaeus[C] (M., Quintana Roo, Cozumel Is.)

[A] The several named forms of Bassaricyon are a single species, Bassaricyon gabbii (Wozencraft, 1989a).
[B] These are considered conspecific in some current taxonomies (Kortlucke and Ramirez-Pulido, 1982); however, the scheme followed here maintains them as separate species (Decker, 1991).
[C] Several named forms of Procyon are a single species, Procyon lotor (Wozencraft, 1989a).

During the late Miocene, late Pliocene, and Pleistocene, the Bering land bridge between North America and Asia formed periodically, offering an avenue for dispersal between northern continents (Darlington, 1963:366; Webb, 1985a). However, by the late Tertiary, northern continents had cooled to the extent that climate, with its attendant sharply defined vegetative zones, became the major factor limiting dispersal by this route (Darlington, 1963:366; Webb, 1985a). Those Holarctic mammals that did cross the Bering land bridge in the late Tertiary were "cold-adapted" species associated with relatively cool, but not alpine, climates (Darlington, 1963:366; Ewer, 1973:369). Among carnivores this included some canids, ursids, mustelids, and felids (Darlington, 1963:393-395, 397; Webb, 1985a). Procyonids, however, did not cross the Bering land bridge into Asia, and Ewer (1973:369) ascribes this to their being an "essentially tropical group." Miocene radiation of procyonids occurred at a time when two of the four major climatic deteriorations (middle and late Miocene) were taking place (Webb, 1985a, 1985b). These deteriorations had the effect of cooling the middle latitudes to the extent that temperate forest forms began to appear in mid-latitude floras, along with a rapid influx of herbaceous plants (Barghoorn, 1953). The procyonid radiation did not penetrate beyond these climatically changing middle latitudes, which implies that these animals were "warm-adapted," and were, therefore, physiologically excluded from reaching the Bering land bridge. Today, three of the six genera and over half of the 18 species that comprise Procyonidae ([Table 1]; Wozencraft, 1989b) remain confined to tropical regions of North and South America (Hall and Kelson, 1959:878-897; Poglayen-Neuwall, 1975; Kortlucke and Ramirez-Pulido, 1982; Nowak and Paradiso, 1983:977-985).

Typical Procyonids

McNab (1988a) contends that basal metabolism is a highly plastic character in evolution, and he has amply shown that ecologically uniform species are more apt to share common metabolic rates than taxonomically allied species from drastically different environments (McNab, 1984a, 1986a, 1986b, 1988a). Procyonids represent a taxonomically allied group that shared a common ecological situation for millions of years; consequently, members of this family might be expected to show some uniformity in their Ḣ_b. Basal and thermoregulatory metabolism of several procyonids have been measured: kinkajou, Potos flavus (Müller and Kulzer, 1977; McNab, 1978a; Müller and Rost, 1983), coatis, Nasua nasua (Chevillard-Hugot et al., 1980; Mugaas et al., in prep.), and Nasua narica (Scholander et al., 1950c; Mugaas et al., in prep.), ringtail, Bassariscus astutus (Chevalier, 1985), and crab-eating raccoon, Procyon cancrivorus (Scholander et al., 1950c). In general, these species have Ḣ_b's that are 40%-80% of the values predicted for them by the Kleiber (1961:206) equation. Lower than predicted Ḣ_b is viewed as an energy-saving adaptation for procyonids living in relatively stable tropical climates (Müller and Kulzer, 1977; Chevillard-Hugot et al., 1980; Müller and Rost, 1983). This implies that lower than predicted Ḣ_b is a general procyonid condition and that it represents a characteristic that evolved in response to the family's long association with tropical and subtropical forest environments.

The Atypical Procyonid

Although most procyonids are found in only tropical to subtropical climates, the North American raccoon, Procyon lotor, ([Figure 1]) has a much broader distribution that extends from tropical Panama (8°N) to southern Canada. In Alberta, Canada, its range reaches the edge of the Hudsonian Life Zone at 60°N (for distribution maps see Hall and Kelson, 1959:878-897, and Poglayen-Neuwall, 1975). Range extensions and an increase in numbers have been noted in Canada and in parts of the United States since the 19th century (Lotze and Anderson, 1979; Kaufmann, 1982; Nowak and Paradiso, 1983:977-985). Thus, Procyon lotor is more complex ecologically than other procyonids, particularly when one takes into account its highly generalized food habits (Hamilton, 1936; Stuewer, 1943; Stains, 1956:39-51; Greenwood, 1981) and the wide range of habitat types (forest, prairie, desert, mountain, coastal marsh, freshwater marsh) and climates (tropical to north temperate) in which it is successful (Whitney and Underwood, 1952:1; Hall and Kelson, 1959:885; Lotze and Anderson, 1979; Kaufmann, 1982). On this basis it is clear that Procyon lotor has deviated from the typical procyonid portrait and has become the consummate generalist of the Procyonidae.

Figure 1.—North American raccoon, Procyon lotor.

The Hypothesis

Our general hypothesis was that whereas most contemporary procyonids have retained the metabolic characteristics of their warm-adapted ancestors, Procyon lotor possesses a different set of adaptations, which either evolved as characteristics unique to this species or were acquired from its ancestral stock. In either case, its unique adaptations have given Procyon lotor the physiological flexibility to generalize its use of habitats and climates and expand its geographic distribution to a much greater extent than other procyonids.

Hypothesis Testing

We tested our hypothesis by comparing Procyon lotor with several other procyonids (Bassariscus astutus, Nasua nasua, Nasua narica, Procyon cancrivorus, and Potos flavus) on the basis of their (1) basal metabolic rate (Ḣ_b), (2) minimum wet thermal conductance (C_mw), (3) diversity of diet (D_d), (4) intrinsic rate of natural increase (r_max), and, when data were available, (5) capacity for evaporative cooling (E_c). In a genetic sense each one of these variables is a complex adaptive characteristic, expression of which is determined by the interaction of several genes (Prosser, 1986:110-165). Experience has shown that a given species will express each one of these variables in a specific manner that is relevant to its mass, physiology, behavior, and environmental circumstance. Thus, different expressions of these variables may represent specific climatic adaptations (Prosser, 1986:16) that have been selected-for by evolutionary process. Because these variables are interrelated with respect to regulation of body temperature and energy balance, they have co-evolved in each species to form an adaptive unit. For each species, measured and calculated values for the first four variables were converted into dimensionless numbers and used to derive a composite score that represented its adaptive unit. Climatic distributions of these species were then compared relative to their composite scores.

Adaptive Significance of the Variables

Basal Metabolic Rate and Intrinsic Rate of Natural Increase

Basal metabolic rate represents the minimum energy required by an animal to maintain basic homeostasis (Lusk, 1917:141; Kleiber, 1932, 1961:251; Benedict, 1938; Brody, 1945:59; Robbins, 1983:105-111). For mammals, Ḣ_b appears to be determined by complex interactions between their body size (Kleiber, 1932, 1961:206; Benedict, 1938; Brody, 1945:368-374; Hemmingsen, 1960:15-36; McNab, 1983b; Calder, 1987), the climate in which they live (Scholander et al., 1950c; McNab and Morrison, 1963; Hulbert and Dawson, 1974; Shkolnik and Schmidt-Nielsen, 1976; McNab, 1979a; Vogel, 1980), their food habits (McNab, 1978a, 1978b, 1980a, 1983a, 1984a, 1986a, 1986b, 1988a, 1989), and their circadian period (Aschoff and Pohl, 1970; Prothero, 1984). Some species have higher mass-specific Ḣ_b than others, and this variation appears to be tied to ecological circumstances rather than taxonomic affinities (McNab, 1988a, 1989). Basal metabolic rate is important ecologically because it serves as a measure of a species' minimum "obligatory" energy requirement, and under many circumstances, it represents the largest energy demand associated with a daily energy budget (King, 1974:38-55; McNab, 1980a; Mugaas and King, 1981:37-40). Recently it also has been implicated as a permissive factor with respect to r_max of mammals (Hennemann, 1983; Lillegraven et al., 1987; Nicoll and Thompson, 1987; Thompson, 1987) via its direct effect on their rates of development and fecundity (McNab, 1980a, 1983a, 1986b; Hennemann, 1983; Schmitz and Lavigne, 1984; Glazier, 1985a, 1985b). The implication of this latter point is that those species with higher Ḣ_b's also have faster rates of development and greater fecundity and hence enjoy the competitive advantage of a higher r_max. Basal metabolism is, therefore, "a highly plastic character in the course of evolution" (McNab, 1988a:25) that has a profound influence on each species' life history.

Minimum Thermal Conductance

Whole-body resistance to passive heat transfer is equal to tissue resistance plus coat resistance. Within limits, these resistances can be altered; tissue resistance can be varied by changes in blood flow, whereas coat resistance can be changed by piloerection, molt, and behavior. When whole-body resistance is maximized (maximum tissue and coat resistances), passive heat transfer is minimized. The inverse of resistance is conductance; therefore, maximum whole-body resistance is the inverse of minimum thermal conductance (C_m). Minimum thermal conductance is readily derived from metabolic chamber data, and it is commonly used to describe an animal's capacity to minimize passive heat transfer. Minimum thermal conductance interacts with Ḣ_b and body mass to set the maximum temperature differential a mammal can maintain without increasing its basal level of heat production. The low temperature in this differential is the lower critical temperature (T_lc).

Mass-specific C_m for mammals is negatively correlated with body mass (McNab and Morrison, 1963; Herreid and Kessel, 1967; McNab, 1970, 1979b; Bradley and Deavers, 1980; Aschoff, 1981), and for any given mass its magnitude is 52% higher during the active, rather than the inactive, phase of the daily cycle (Aschoff, 1981). However, some mammals have C_m's that are higher or lower than would be predicted for them on the basis of body mass and circadian phase. Seasonal variation in C_m (higher values during summer than winter) has been reported for many northern mammals that experience large annual variations in air temperature (Scholander et al., 1950a; Irving et al., 1955; Hart, 1956, 1957; Irving, 1972:165). Some tropical mammals with very thin fur coats, and others with nearly hairless bodies, have high C_m's (McNab, 1984a), as do burrowing mammals (McNab, 1966, 1979b, 1984a) and the kit fox, Vulpes macrotis (Golightly and Ohmart, 1983). Some small mammals with low basal metabolic rates tend to have lower than predicted C_m's: small marsupials (McNab, 1978a), heteromyid rodents (McNab, 1979a), several ant eaters (McNab, 1984a), the arctic hare, Lepus arcticus (Wang et al., 1973), the ringtail, Bassariscus astutus (Chevalier, 1985), and the fennec, Fennecus zerda (Noll-Banholzer, 1979). Thus, in spite of its mass dependence, C_m also has been modified during the course of evolution by selective factors in the environment and by the animal's own metabolic characteristics.

Capacity for Evaporative Cooling

Latent heat loss occurs as a result of evaporation from the respiratory tract and through the skin, and except under conditions of heat stress, it "is a liability in thermal and osmotic homeostasis" (Calder and King, 1974:302). E_c, defined as the ratio of evaporative heat lost to metabolic heat produced, can be used to quantify thermoregulatory effectiveness of evaporative cooling and to make comparisons of heat tolerance between species. Thermoregulatory effectiveness of latent heat loss is not just a function of the rate of evaporative water loss but also of the rate of metabolic heat production (Lasiewski and Seymour, 1972). For example, a low metabolic rate minimizes endogenous heat load and thus conserves water, whereas the opposite is true of high metabolic rates (Lasiewski and Seymour, 1972). Some mammals that live in arid regions have evolved low metabolic rates and thus capitalize on this relationship to reduce their thermoregulatory water requirement (McNab and Morrison, 1963; McNab, 1966; MacMillen and Lee, 1970; Noll-Banholzer, 1979). What is evident, therefore, is that an animal's capacity for increasing latent heat loss must evolve together with its Ḣ_b and C_m in response to specific environmental demands.

Diet

McNab (1986a, 1988a, 1989) demonstrated that, for mammals, departures of Ḣ_b from the Kleiber (1961:206) "norm" are highly correlated with diet and independent of phylogenetic relationships. McNab's analysis indicates that for mammals that feed on invertebrates, those species with body mass less than 100 g have Ḣ_b's that are equal to or greater than values predicted by the Kleiber equation, whereas those with body mass greater than 100 g have metabolic rates that are lower than predicted. Grazers, vertebrate eaters, nut eaters, and terrestrial frugivores also have Ḣ_b's that are equal to or greater than predicted, whereas insectivorous bats, arboreal folivores, arboreal frugivores, and terrestrial folivores all have rates that are lower than predicted. McNab (1986a) found animals with mixed diets harder to categorize, but in general he predicted that their Ḣ_b's would be related to (1) a food item that is constantly available throughout the year, (2) a food item that is most available during the worst conditions of the year, or (3) a mix of foods available during the worst time of the year. Although these correlations do not establish cause and effect between food habits and Ḣ_b, McNab's analysis does make it clear that the relationship between these variables has very real consequences for an animal's physiology, ecology, and evolution.

Experimental Design and Summary

In this investigation we measured basal and thermoregulatory metabolism, evaporative water loss, and body temperature of raccoons from north central Virginia. Measurements were conducted on both sexes in summer and winter to determine how season and sex influenced these variables. We then compared the data for this widely distributed generalist with data from literature for its ecologically more restricted relatives. Dietary data for all species were taken from literature, as were reproductive data for calculation of r_max.

Our analysis demonstrated clear differences between Procyon lotor and other procyonids with respect to Ḣ_b, C_mw, D_d, and r_max. The composite score calculated from these variables for Procyon lotor was much higher than those derived for other species, and there was a positive correlation between the number of climates a species occupies and the magnitude of its composite score. Data on evaporative water loss, although not complete for all species, suggested that tropical and subtropical procyonids have less capacity for evaporative cooling than Procyon lotor or Bassariscus astutus. It was clear, therefore, that with respect to its thermal physiology, Procyon lotor differed markedly from other procyonids, and we contend that these differences have allowed this species to become a highly successful climate generalist and to expand its distribution into many different habitats and climates. Our analysis also suggested that the cornerstone of Procyon lotor's success as a climate generalist is its Ḣ_b, which is higher than the procyonid norm.

Acknowledgments

The authors would like to thank John Eisenberg and Devra Kleiman for their support and encouragement throughout the study. This investigation was supported by research grants from the West Virginia School of Osteopathic Medicine (WVSOM), and Friends of the National Zoo (FONZ). Logistic support was provided by the National Zoological Park's Conservation and Research Center (CRC), and the departments of Mammalogy and Zoological Research. Our ability to conduct physiological research at CRC was made possible by the thoughtful support and encouragement provided by Chris Wemmer. His excellent staff at CRC, especially Jack Williams, Junior Allison, and Red McDaniel, were very helpful in providing hospitality and logistical support to the senior author and his family during their various visits to the Center. The assistance of several people at the National Zoo also is gratefully acknowledged: Mitch Bush and Lyndsay Phillips not only provided veterinary support throughout the investigation, but also performed surgical procedures required to implant temperature-sensitive radio transmitters in several raccoons; Olav Oftedal made his laboratory available to us at various times and loaned us equipment to use at CRC; Miles Roberts and his staff provided care for our captive raccoons in the Department of Zoological Research during various parts of the investigation. Greg Sanders and Ken Halama, supported by FONZ assistantships, cared for our captive raccoons at CRC, provided assistance in the laboratory whenever needed, and were an invaluable source of aid. Their friendship and help is gratefully acknowledged. Ellen Broudy and Andy Meyer, supported by WVSOM and a student work study grant, respectively, provided assistance in the laboratory. David Brown, John Eisenberg, Mary Etta Hight, Brian McNab, Steve Thompson, and W. Chris Wozencraft critically reviewed various phases of the manuscript and provided many helpful suggestions. We deeply appreciate the work of Jean B. McConville, whose beneficial editorial suggestions helped us improve several early versions of the manuscript. We also gratefully acknowledge Diane M. Tyler, our editor at the Smithsonian Institution Press, whose expertise helped us mold the manuscript into its final form. Jill Mellon and Sriyanie Miththalapa, supported by FONZ traineeships, assisted in measuring the daily cycle of body temperature in raccoons. The Virginia Commission of Game and Inland Fisheries gave us permission to use wild-caught raccoons in this project.

Materials and Methods

Live-trapping

Raccoons were caught from May 1980 through December 1984 on a trapping grid of 30 to 35 stations (one or two "live traps" per station) that covered about one-third of the National Zoological Park's Conservation and Research Center (CRC) near Front Royal, Virginia (Seidensticker et al., 1988; Hallett et al., 1991). Animals were trapped during 10 consecutive days each month, and in this five-year interval 407 raccoons were captured and marked with tattoos and ear tags. All captured animals were individualized with respect to age, reproductive status, physical condition, parasite load, and mass and body dimensions. These data characterized the structure and dynamics of the raccoon population at CRC and provided information on the annual cycle of fattening for raccoons in north central Virginia.

Animals used for metabolic measurements were captured at CRC about 1.5 km south of the trapping grid and thus were genetically representative of the area. Six males were captured and measured during the summer of 1983. These animals were kept isolated for a week before being measured and were released later that summer at the site of their capture. The other seven animals used in our study were from the collection of the National Zoological Park and all of them had their origins at CRC.

Metabolic Studies

Basal and Thermoregulatory Metabolism

Metabolic measurements, conducted at CRC, were carried out on eight males during July and August 1983, on four females and three males from November 1983 through March 1984, and on four females during June and July 1984.

Raccoons were housed throughout the study such that they were constantly exposed to a natural cycle of temperature and photoperiod. Weather records for the Front Royal area indicate that average temperatures are around -0.5°C in January and 23.3­°C in July (Crockett, 1972). Light:dark (L:D) periods for the latitude of CRC (48°55'N; United States Department of the Interior Geological Survey, 1972), calculated from duration of daylight tables (List, 1971:506-512), were 14.9:9.1 and 9.4:14.6 hours L:D for summer and winter solstices, respectively, and 12.2:11.8 hours L:D for vernal and autumnal equinoxes.

Our animals were fed a measured amount of food daily, and they usually ate most of what was provided. Occasionally these animals would eat very little or none of their ration, and on some days they would eat all that was given to them. We fed them either feline diet (ground horse meat) or canned mackerel (Star-kist®[1]) along with high-protein dog chow (Purina®). When available, fresh fruit also was added to their diet. Water was always provided ad libitum.

[1] The use of product brand names in this publication is not intended as an endorsement of the products by the Smithsonian Institution.

Measurements were conducted during the raccoons' daily inactive period (sunrise to sunset) in both summer and winter. Oxygen consumption was measured in a flow-through metabolism chamber at 5°C intervals from -10°C to 35°C. Animals were held at each temperature until the lowest rate of oxygen consumption had been obtained and maintained for at least 15 minutes. During each determination, oxygen consumption was monitored for 30 minutes to one hour beyond a suspected minimum value to see if an even lower reading could be obtained. Raccoons attained minimum levels of oxygen consumption more quickly at warm (>10°C) than at cold temperatures. Depending on the temperature, therefore, each measurement took from two to five hours to complete. On days when two measurements could be completed, the second trial was always at a temperature 10°C warmer than the first.

The metabolism chamber was constructed from galvanized sheet metal (77.5 × 45.5 × 51.0 cm = 180 liters) and was painted black inside. Within the chamber, the animal was held in a cage (71 × 39 × 33 cm) constructed from turkey wire that also was painted black. This cage prevented the raccoons from coming into contact with the walls of the chamber, yet it was large enough to allow them to stand and freely move about. The bottom of the cage was 11 cm above the chamber floor, which was covered to a depth of one cm with mineral oil to trap urine and feces.

During measurements, the metabolism chamber was placed in a controlled-temperature cabinet (modified Montgomery Ward model 8969 freezer). Air temperature (T_a) in the metabolism chamber was regulated with a Yellow Springs Instrument model 74 temperature controller. T_a was controlled to ± 1.0°C at temperatures below freezing, and to ± 0.5°C at temperatures above freezing. The chamber air and wall temperatures were recorded continuously (Linseis model LS-64 recorder) during each experiment, and, except during temperature changes, they were always within 0.5°C of each other.

Columns of Drierite® and Ascarite® removed water vapor and carbon dioxide, respectively, from air entering and leaving the chamber. Dry carbon-dioxide-free room air was pumped into the chamber (Gilman model 13152 pressure/vacuum pump) at a rate of 3.0 L/min (Gilmont model K3203-20 flow meter). Downstream from the chemical absorbents, an aliquot (0.1 L/min) of dry carbon-dioxide-free air was drawn off the chamber exhaust line and analyzed for oxygen content (Applied Electrochemistry model S-3A oxygen analyzer, model 22M analysis cell, and model R-1 flow control). All gas values were corrected to standard temperature and pressure for dry gas. Oxygen consumption was calculated from the difference in oxygen content between inlet and outlet air using Eq. 8 of Depocas and Hart (1957).

Each raccoon was fasted for at least 12 hours before oxygen consumption measurements began. At the start and end of each metabolic trial the animal was weighed to the nearest 10 g (Doctors Infant Scale, Detecto Scales, Inc., Brooklyn, N.Y., U.S.A.). The body mass used in calculating minimum oxygen consumption and evaporative water loss was estimated from timed extrapolations of the difference between starting and ending weights, and the time at which these variables were measured.

Evaporative Water Loss

During metabolic measurements at temperatures above freezing, evaporative water loss was determined gravimetrically. Upstream from the chemical columns, an aliquot of air (0.1 L/min) was drawn off the exhaust line and diverted for a timed interval through a series of preweighed (0.1 mg)

-tubes containing Drierite®. The aliquot then passed through a second series of

-tubes containing Ascarite® before entering the oxygen analysis system. Evaporative water loss was calculated using [Eq. 1]

where Ė is evaporative water loss (mg·g^-1·h^-1), m_w is mass of water collected (mg), .V_e is rate of air flow into the chamber (3.0 L/min), .V_a is the rate of air flow through the

-tubes (0.1 L/min), t is length of the timed interval (h), and m is the estimated mass of the raccoon at the time of sampling (g).

Body Temperature

Veterinarians at the National Zoological Park surgically implanted calibrated temperature-sensitive radio transmitters (Telonics, Inc., Mesa, AZ, U.S.A.) into abdominal cavities of two female and two male raccoons. Transmitter pulse periods were monitored with a digital processor (Telonics TDP-2) coupled to a receiver (Telonics TR-2-164/166). During some metabolic measurements, body temperatures of these animals were recorded to the nearest 0.1°C at 30-minute intervals. The daily cycle of body temperature of these raccoons also was measured once a month.

Calibrations

Calorimeter

At the conclusion of these experiments, the accuracy of our calorimetry apparatus was tested by burning an ethanol lamp in the metabolism chamber. During these tests a CO₂ analyzer was incorporated into the system (Beckman, LB-2). Results demonstrated that we measured 84% of the oxygen consumed by the lamp as well as 84% of the water and CO₂ it produced; standard deviation = ± 2.6, ± 5.0, and ± 3.6, respectively (n = 27). Average respiratory quotient (RQ) calculated from these data was O.657 ± 0.008 (n = 27), which is 99.5% of that predicted (0.66). McNab (1988b) reports that the accuracy of open-flow indirect calorimetry systems, such as ours, depends on the rate of air flow through the animal chamber. If flow rates are too low, there is inadequate mixing of air within the chamber, and the rate of oxygen consumption, as calculated from the difference in oxygen content of air flowing into and out of the chamber (Depocas and Hart, 1957), is underestimated. At some critical rate of air flow, which is unique to each combination of chamber and animal, this situation changes such that measured rates of oxygen consumption become independent of any further increase in flow rate (McNab, 1988b). In recent tests of our system, where we burned the ethanol lamp at a variety of chamber flow rates, the efficiency of measurement increased linearly as flow rate increased, and the critical rate of air flow was about 6.7 L/min. This appeared to explain why a flow rate of 3.0 L/min underestimated oxygen consumption of the ethanol lamp.

Our earlier tests of the efficiency of our system indicated that although we underestimated actual oxygen consumption of the ethanol lamp, we did so with a fair degree of precision; probably because flow rates were closely controlled. During our metabolic measurements, chamber flow rates also were closely controlled at 3.0 L/min, and we believe, therefore, that these measurements also were carried out with a high degree of precision. Consequently, all measured values of oxygen consumption and water production were considered to be 84% of their actual value and were adjusted to 100% before being included in this report.

Body Temperature Transmitters

The calibration of all temperature-sensitive radio transmitters drifted over time. Transmitters were calibrated before they were surgically implanted and again after they were removed from the animals. Although the drift of each transmitter was unique, it was also linear (S. Tomkiewicz, Telonics, Inc., pers. com.). All body temperature measurements were corrected from timed extrapolations of the difference between starting and ending calibrations.

Statistical Methods

Values of oxygen consumption, evaporative water loss, and body temperature were plotted as a function of chamber air temperature. Linear regressions of oxygen consumption at temperatures below the thermoneutral zone (T_n), and evaporative water loss at temperatures above freezing, were determined with the SAS (1982) GLM procedure. Lower critical temperature (T_lc) was determined graphically from intersection of the line representing Ḣ_b and the regression line representing oxygen consumption below T_n. Slopes and intercepts of regression lines, as well as other mean values, were compared with t-tests (Statistical Analysis System, 1982; Ott, 1984:138-175). Unless indicated otherwise, data are expressed as mean ± standard deviation (s.d.).

Estimating Intrinsic Rate of Natural Increase

We employed the method first described by Cole (1954) to calculate r_max:

where a is potential age of females first producing young, b is potential annual birth rate of female young, and n is potential age of females producing their final young. After life-history data were substituted into [Eq. 2], r_max was determined by trial and error substitution (Hennemann, 1983).

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).

Comparison of Adaptive Units

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

The correlation between number of climates these species occupy and their composite scores was tested by linear regression.

Results

Body Mass

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.

Table 2.—Body mass in kg and basal metabolism (mL O₂·kg^-0.75·h^-1) of Procyon lotor in summer and winter (s.d. = standard deviation and n = number of observations).

Basal Metabolic Rate

Within thermoneutrality, Ḣ_b (mL O₂·g^-1·h^-1) was 0.54 ± 0.09 for trapped males in summer, 0.46 ± 0.07 for captive males in summer, 0.42 ± 0.07 for captive females in summer, 0.47 ± 0.06 for captive males in winter, and 0.46 ± 0.10 for captive females in winter ([Figures 2],[ 3]). Ratios of these measured values to those predicted by the Kleiber (1932, 1961:206) equation are 1.28, 1.12, 1.02, 1.17, and 1.09, respectively. To minimize the effect of body size (Mellen, 1963) and to facilitate comparisons between sexes and seasons and between captive and trapped animals, basal metabolism also was calculated as a function of metabolic body size (mL O₂·kg^-0.75·h^-1; [Table 2]). Based on this analysis, trapped summer males had a higher basal metabolism than captive males (p<0.025) or females (p<0.005) in either season ([Table 2]). There was no difference in basal metabolism between captive males and females in either summer or winter, and there was no seasonal difference in their basal metabolic rates ([Table 2]).

Minimum Thermal Conductance

Minimum wet and dry thermal conductances were calculated using Eqs. 4 and 5

where C_mw is wet and C_md is dry conductance (mL O₂·g^-1·h^-1·°C^-1); Ḣ_r is the lowest resting metabolic rate measured at each temperature (mL O₂·g^-1·h^-1); Ė_eq is oxygen equivalent for heat lost by evaporation [Ė_eq = mL O₂·g^-1·h^-1 = Ė·λ/γ, where Ė is evaporative water loss (mg·g^-1·h^-1), λ is heat of vaporization for water (2.43 J/mg), and γ is heat equivalent for oxygen (20.097 J/mL)]; T_b is body temperature (°C); and T_a is chamber air temperature (°C). Only data from animals equipped with temperature-sensitive radio transmitters were used for these calculations.

Table 3.—Minimum wet and dry thermal conductances (mL O₂·g^-1·h^-1·°C^-1) of Procyon lotor in summer and winter. Means of values were calculated from equations 3 and 4 (s.d. = standard deviation and n = number of observations).

Figure 2.—Relationship between oxygen consumption and chamber air temperature for raccoons in summer: captive females, open circles; captive males, closed circles; trapped males, open squares. Sloping lines represent regressions of oxygen consumption on chamber air temperature, and horizontal lines, basal metabolism.

Figure 3.—Relationship between oxygen consumption and chamber air temperature for raccoons in winter: captive females, open circles; captive males, closed circles. Solid sloping line represents regression of oxygen consumption on chamber air temperature for males and females, and the horizontal line, basal metabolism for males and females.

C_mw was calculated for each season from metabolic measurements made at all air temperatures below T_lc ([Table 3]). Because evaporative water loss was not measured at temperatures below freezing, C_md was calculated only from metabolic determinations made at air temperatures between T_lc and 0°C. There was no difference between males and females in summer for either C_mw or C_md (mL O₂·g^-1·h^-1·°C^-1). Data for each sex were combined to give a summer average of 0.0256 ± 0.0028 for C_mw, and 0.0246 ± 0.0019 for C_md ([Table 3]). These summer conductances were 49% higher (p<0.005) than those calculated for winter females (0.0172 ± 0.0023, and 0.0161 ± 0.0027 for C_mw and C_md, respectively; [Table 3]). C_mw and C_md were not different from each other in either summer or winter, which indicated that in both seasons evaporative water loss contributed very little to heat dissipation at temperatures below T_n. Comparisons of thermal conductances calculated on the basis of metabolic body size (Mellen, 1963) gave the same results.

Evaporative Water Loss

Evaporative water loss increased as chamber temperature increased in both summer and winter ([Figures 4],[ 5]). In summer, the pattern of increase was different for females and males. Polynomial regressions for trapped and captive males produced equations that describe a concave relationship between T_a and evaporative water loss, whereas the equation for females describes a sigmoid curve ([Table 4]; [Figure 4]). For females, water loss increased rapidly at temperatures above 25°C ([Figure 4]). The intercepts and coefficients of the X, X², and X³ terms of the polynomial regression equations ([Table 4]) were compared (t-tests) to determine if they differed from each other. The coefficients in the equation for trapped males differed from those for captive females in the X² (p<0.05) and X³ (p<0.025) terms. The intercept and coefficients of the equation for captive males, however, were not different from those for either captive females or trapped males. Although this lack of difference is understandable in the case of trapped males, where the shape of the two curves is similar (concave), it is not so clear for the sigmoid curve of captive females ([Figure 4]). Perhaps the lack of difference in this case is simply due to the small number of observations available for captive males (n = 10; [Table 4]). Nonetheless, in summer at 35°C, both captive and trapped males relied less on evaporative cooling than did captive females ([Figure 4]).

In winter, males and females had similar rates of evaporative water loss across the full range of temperatures tested ([Figure 5]). Therefore, data for both sexes were combined. The intercept and coefficients of this equation ([Table 4]) did not differ from those for summer females, but they did differ from those in the regression for trapped males in the X² (p<0.05) and X³ (p<0.025) terms. As was the case for females in summer, rates of water loss for winter animals increased most rapidly at temperatures above 25°C ([Figure 5]).

Figure 4.—Relationship between evaporative water loss and chamber air temperature for raccoons in summer: captive females, open circles; captive males, closed circles; trapped males, open squares. Lines represent polynomial regressions of evaporative water loss on chamber air temperature.

Figure 5.—Relationship between evaporative water loss and chamber air temperature for raccoons in winter: captive females, open circles; captive males, closed circles. Lines represent polynomial regressions of evaporative water loss on chamber air temperature.

Table 4.—Polynomial regression equations describing evaporative water loss (mg·g^-1·h^-1) of Procyon lotor in summer and winter (X = chamber temperature (°C), Y = evaporative water loss, n = number of observations, R² = coefficient of determination, and SEE = standard error of estimate).

Thermoregulation at Low Temperatures

Body Temperature

Body temperatures in [Figure 6] are those recorded during metabolic measurements from animals equipped with surgically implanted, temperature-sensitive radio transmitters. Each point was recorded during the lowest level of oxygen consumption at each T_a. In both summer and winter, T_b's were lowest during metabolic measurements at T_a's around T_lc. At T_a's below T_lc, T_b's increased ([Figure 6]), which is an unusual response. Under similar conditions, other procyonids either maintain a nearly constant T_b or allow it to fall slightly (Müller and Kulzer, 1977; Chevillard-Hugot et al., 1980; Müller and Rost, 1983; Chevalier, 1985). For our raccoons, confinement in the metabolism chamber at low temperatures must have stimulated a greater than necessary increase in metabolic rate such that heat production exceeded heat loss, which caused T_b to become elevated.

Figure 6.—Relationship between body temperature and chamber air temperature in summer (panel A), and winter (panel B): captive females, open circles and solid lines; captive males, solid circles and dashed lines. Solid vertical lines represent lower critical temperatures.

Table 5.—Regression equations describing oxygen consumption (mL O₂·g^-1·h^-1) of Procyon lotor at temperatures below their lower critical temperature (I = x-intercept (°C), n = number of observations, R² = coefficient of determination, SEE = standard error of estimate for the y-intercept (a) and slope (b), X = chamber temperature (°C), and Y = oxygen consumption).

Summer

During summer, T_lc for male raccoons was 20°C, whereas for females it was 25°C ([Figure 2]). Regression equations calculated to describe oxygen consumption at T_a's below T_lc are presented in [Table 5]. For three groups of summer animals, slopes of regressions are identical. This indicates that minimum conductances of these three groups were equivalent. Intercepts of these equations are different, which suggests a difference in metabolic cost of thermoregulation between these groups ([Figure 2]); captive males had a lower intercept than either trapped males (p<0.005) or captive females (p<0.05), but there was no difference in intercepts of captive females and trapped males. These regression equations, therefore, also were derived using values of oxygen consumption expressed in terms of metabolic body mass (Mellen, 1963). Relationships between intercepts of these equations are different than those for regressions in [Table 5]. Intercept for females was intermediate to, and not different from, those of the two groups of males. However, captive males still had a lower intercept than trapped males (p<0.025). Thus, in summer, thermoregulatory metabolism was less expensive for captive than for trapped males, and in spite of a 5°C difference in their T_lc's ([Figure 2]), captive males and females had similar thermoregulatory costs.

Regression lines for three groups of animals in summer extrapolate to zero metabolism at values equivalent to, or greater than, normal T_b; 38.8°C for trapped males, 37.6°C for captive males, and 41.1°C for captive females ([Table 5]). Thus, all three groups had minimized thermal conductance at T_a's below T_lc (Scholander et al., 1950b; McNab, 1980b). Minimum wet thermal conductance calculated for raccoons in summer with [Eq. 4] ([Table 3]) is numerically similar to these "slope" values ([Table 5]), and it was, therefore, considered to be the best estimate of C_mw for Procyon lotor during that season (0.0256 mL O₂·g^-1·h^-1·°C^-1).

Winter

During winter T_lc for both sexes decreased to 11°C ([Figure 3]). Regression equations of thermoregulatory metabolism for males and females in winter are not different from each other in either slope or intercept. These data, therefore, were combined into a single equation ([Table 5]). Slope and intercept of this equation are both lower (p<0.005 and p<0.05, respectively) than those for summer animals ([Table 5]). Identical results were obtained from comparisons using regressions derived from oxygen consumption expressed in terms of metabolic body mass (Mellen, 1963). Thermoregulatory costs at any temperature below 20°C were lower for winter than summer animals ([Figures 2],[ 3]).

Table 6.—Regression equations describing oxygen consumption (mL O₂·g^-1·h^-1) of Procyon lotor at temperatures below their lower critical temperature in winter (A = females with radio transmitters, B = females without radio transmitters, C = males, I = x-intercept (°C), n = number of observations, R² = coefficient of determination, X = chamber temperature (°C), and Y = oxygen consumption).

Figure 7.—Relationship between body temperature and time of day at various months of the year: captive females, open circles; captive males, closed circles. Vertical cross-hatched areas represent civil twilight.

The regression line for Procyon lotor in winter ([Table 5]) extrapolates to zero metabolism at 35.2°C, which is below normal T_b ([Figures 6],[ 7]). This suggests that not all raccoons measured in winter minimized thermoregulatory metabolism or conductances at T_a's below T_lc (Scholander et al., 1950b; McNab, 1980b). To assess this possibility, data for these animals were divided into three groups: (A) females with radio transmitters, (B) females without radio transmitters, and (C) males ([Table 6]). Regression equations of metabolism below T_lc were derived for each group, and based on extrapolated T_b's at zero metabolism, only the two females with implanted radio transmitters (group A) minimized thermoregulatory metabolism and conductance. Had animals in groups B and C also minimized their thermal conductances, while retaining their measured metabolic rates, their rates of heat production would have been disproportionately higher than their rates of heat loss. Equation 4 predicts that under these conditions their body temperatures would have been elevated to 42.0°C and 40.4°C, respectively. Thus, in order to avoid such a large increase in body temperature, animals in groups B and C increased their thermal conductances in preference to lowering their metabolic rates. The regression equation of thermoregulatory metabolism for all winter animals ([Table 5]), therefore, overestimates minimum metabolic cost of temperature regulation below T_lc, and its slope underestimates C_mw. Consequently, the best estimate of C_mw for Procyon lotor in winter is the value calculated for group A animals with [Eq. 4] (0.0172 mL O₂·g^-1·h^-1·°C^-1; [Table 3]), and the minimum cost of thermoregulatory metabolism at any T_a below T_lc is best estimated by substituting this value into [Eq. 4] and solving for Ḣ_r.

Thermoregulation at High Temperatures

Body Temperature

In both summer and winter, T_b's increased during metabolic measurements at T_a's above T_lc ([Figure 6]). This response also was seen during metabolic measurements conducted on other procyonids (Müller and Kulzer, 1977; Chevillard-Hugot et al., 1980; Müller and Rost, 1983; Chevalier, 1985).

Summer

During summer our data suggested that the upper critical temperature (T_uc) was higher than 35°C. The lowest rates of oxygen consumption at T_a = 35°C occurred after 1.5 to 2.5 hours of exposure to that temperature. Prolonged exposure to this temperature in summer did not make animals restless, and their rate of oxygen consumption was very stable throughout each measurement. Body temperature responses at T_a = 35°C were recorded from two males and two females that had implanted radio transmitters. With the exception of one male, T_b's were maintained near 38°C ([Figure 6]). The one exception (a male) maintained its T_b at 39.3°C. At T_a = 35°C, summer males had rates of evaporative water loss that were lower than those of summer females ([Figure 4]). At this temperature, males dissipated 35% ± 6% and females 56% ± 18% of their metabolic heat via evaporative water loss. Thus, at T_a = 35°C, males must have utilized modes of heat transfer other than evaporative cooling (convective and conductive heat transfer) to a greater extent than females.

Winter

Body temperature, evaporative water loss, and metabolic data indicated that, in winter, T_uc was very close to 35°C. In winter, the lowest level of oxygen consumption was recorded during the first hour after the chamber had reached T_a = 35°C. Unlike summer, animals became restless after the first hour at 35°C, at which point their oxygen consumption increased and showed a high degree of variability. Body temperature responses at 35°C were recorded from both females that had implanted radio transmitters. In one case, T_b rose from 37.9°C at the end of the first hour to 40.5°C by the end of the second hour, and as it did not show signs of leveling off, we terminated the experiment. We exposed that same animal to T_a = 35°C one other time during winter. In that instance, its T_b rose to 40.0°C during the first 30 minutes and was maintained at that level for three hours with no apparent distress. The other female elevated its T_b from 37.3°C to 39.0°C during the second hour at T_a = 35°C and maintained its T_b at that level for two hours. Thus, during winter, prolonged exposure to T_a = 35°C stimulated more of an increase in T_b than it did in summer. During winter, both males and females increased evaporative water loss at T_a = 35°C ([Figure 5]) but only to the extent that they dissipated 35% ± 10% of their metabolic heat production. Thus, even in winter, convective and conductive heat transfers were still the most important modes of heat loss at this temperature.

Daily Cycle of Body Temperature

The daily cycle of raccoon T_b's during summer and winter are presented in [Figure 7]. In general, T_b's showed a marked circadian cycle in phase with photoperiod. T_b's rose above 38°C for several hours each night but remained below 38°C during daytime. During summer, with the exception of one female whose record was not typical ([Figure 7]), T_b's rose above 38°C shortly after sunset, whereas in winter T_b's did not rise above 38°C until several hours after sunset. Once T_b was elevated it usually remained so until just before or after sunrise ([Figure 7]). During summer, T_b was above 38°C for 85% or more of the time between sunset and sunrise (87% for the female with the typical body temperature pattern, and 85% and 98% for males), whereas in winter it was elevated for only 47%-78% of the time between sunset and sunrise (47% and 61% for females, and 67% and 78% for males). During night, T_b would oscillate between 38°C and about 39°C, such that two peak values occurred. These peak values presumably corresponded to two periods of heightened nighttime activity. During summer, one of these peaks occurred before and the other after 24:00 hours, whereas in winter both peaks occurred after 24:00 hours. With the exception of one female in winter ([Figure 7]), the lowest T_b of the day for both sexes was near 37°C, and this typically occurred during daytime ([Figure 7]).

Discussion

Basal Metabolic Rate

Background

Basal metabolism represents the minimum energy required by a mammal to maintain endothermy and basic homeostasis (Lusk, 1917:141; Kleiber, 1932, 1961:251; Benedict, 1938:191-215; Brody, 1945:59; Robbins, 1983:105-111). Mammals with lower than predicted Ḣ_b maintain endothermy and enjoy its attendant advantages at a discount, whereas others, with rates that are higher than predicted, pay a premium (Calder, 1987). Such variation in Ḣ_b appears to be tied to ecological circumstances rather than taxonomic affinities (Vogel, 1980; McNab, 1986a, 1988a, 1989), and depending on environmental conditions, each rate provides an individual with various advantages and limitations. During the course of evolution, therefore, each species' Ḣ_b evolves to provide it with the best match between its energy requirements for continuous endothermy, its food supply, and the thermal characteristics of its environment.

Captive versus Wild Raccoons

Male raccoons trapped in summer had higher Ḣ_b's than our captive animals in any season ([Table 2]). The higher rate of metabolism of these trapped males could have been due to the stress of captivity or to the fact that "wild" animals actually may have higher metabolic rates than those that have adjusted to captivity. If the latter is true, then our data for captive animals underestimated the actual energy cost of maintenance metabolism for Procyon lotor in the wild. At present, we have no way of determining which of these alternatives is true.

Seasonal Metabolism of Raccoons

In some temperate-zone mammals, Ḣ_b is elevated in winter, which presumably increases their "cold-hardiness." Conversely, lower summer metabolism is considered to be a mechanism that reduces the potential for heat stress. Such seasonal variation in Ḣ_b has been found in several species: collard peccary, Tayassu tajacu (Zervanos, 1975); antelope jackrabbit, Lepus alleni (Hinds, 1977); desert cottontail, Sylvilagus audubonii (Hinds, 1973); and, perhaps, cold-acclimatized rat, Rattus norvegicus (Hart and Heroux, 1963). Unlike these species, our captive raccoons showed no seasonal variation in Ḣ_b ([Table 2]). Instead, raccoons achieved "cold-hardiness" in winter and reduced their potential for heat stress in summer with a large seasonal change in thermal conductance ([Table 3]).

Table 7.—Metabolic characteristics of several procyonid species.

[a] Meas is measured basal metabolism (mL O₂·g^-1·h^-1). H_br is the ratio of measured to predicted basal metabolism where the predicted value is calculated from Ḣ_b = 3.42·m^-.25 (Kleiber, 1932, 1961:206) and m is body mass in grams.
[] Meas is measured minimum thermal conductance (mL O₂·g^-1·h^-1·°C^-1). C_mwr is the ratio of measured to predicted minimum thermal conductance where the predicted value is calculated from C_m = 1.0·m^-0.5 (McNab and Morrison, 1963; Herreid and Kessel, 1967), and m is body mass in grams.
[c] T_b is body temperature during the active (α) and rest (ρ) phases of the daily cycle (°C).
[d] T_n is the thermoneutral zone as defined by the lower (T_lc) and upper (T_uc) critical temperatures (°C).
[e] Conductance calculated as the slope of the line describing oxygen consumption at temperatures below the lower critical temperature.
[f] Conductance calculated from C_mw = Ḣ_r/(T_b - T_a), where Ḣ_r is resting metabolic rate at temperatures below T_lc, and other symbols are as described elsewhere.
[g] Inactive-phase thermal conductance: estimated from Scholander et al. (1950b), assuming that active-phase thermal conductance is 52% higher than values determined during the inactive phase (Aschoff, 1981).

Comparison of Procyon lotor with Other Procyonids

Procyon lotor has a much higher mass-specific Ḣ_b than other procyonids ([Table 7]). To quantify the magnitude of this difference, we compared the measured value for Procyon lotor with one calculated for it from a mass-specific least-squares regression equation ([Eq. 6]; R² = 0.78) derived from data for those procyonids with lower than predicted Ḣ_b: Potos flavus, Procyon cancrivorus, Nasua nasua, Nasua narica, and Bassariscus astutus ([Table 7]).

Ḣ_b in Eq. 6 is basal metabolism (mL O₂·g^-1·h^-1) and m is body mass (g). Measured values of Ḣ_b for Procyon lotor were 1.45 to 1.86 times greater than those predicted for it by Eq. 6 ([Table 8]).

Table 8.—Basal metabolism (mL O₂·g^-1·h^-1) of Procyon lotor as predicted by [Eq. 6] (Ḣ_b = 2.39·m^-0.25). Body masses, used to calculate predicted values, and measured values were taken from [Table 7].

Influence of Diet on Basal Metabolism

Background.—With respect to Ḣ_b, McNab (1986a:1) maintains that "the influence of climate is confounded with the influence of food habits," and that departures from the Kleiber (1961) "norm" are best correlated with diet. Although this does appear to be the case for diet specialists, the analysis is not so clear-cut for omnivorous species (McNab, 1986a). His analysis also indicates that an animal's "behavior" (i.e., whether it is terrestrial, arboreal, subterranean, aquatic, etc.), secondarily modifies the influence of food habits on Ḣ_b. For example, terrestrial frugivores have Ḣ_b's that are very near predicted values, whereas arboreal frugivores have rates that are much lower than predicted (McNab, 1986a).

Table 9.—Food habits of some Procyonids. References for foods were as follows: Potos flavus, Procyon cancrivorus, and Nasua nasua taken from Bisbal (1986); Nasua narica taken from Kaufmann (1962:182-198); Bassariscus astutus taken from Martin et al. (1951), Taylor (1954), Wood (1954), Toweill and Teer (1977), and Trapp (1978); Procyon lotor taken from Hamilton (1936), Stuewer (1943:218-220), Stains (1956:39-51), and Greenwood (1981). Symbols represent either qualitative (#) or quantitative (+, †) assessments of feeding habits: # indicates that the animal was observed eating the food; + and † represent volume and frequency, respectively, of food utilization. No attempt was made to account for seasonal variation in the use of these foods.

Food Habits of Procyonids.—Food habits of six procyonids for which metabolic data are available are presented in [Table 9]. All six species clearly have mixed diets. Compared to other species, Procyon lotor is highly catholic in its diet, taking food from almost twice as many categories as Nasua narica, three times as many as Procyon cancrivorus, Nasua nasua, and Bassariscus astutus, and nine times as many as Potos flavus.

For those species for which food habit data are quantified, we used Eisenberg's (1981:247-251) substrate/feeding matrix method, where "substrate" is analogous to McNab's (1986a) "behavior," to construct the following feeding categories that are based on the major food groups utilized by each species ([Table 9]).

1. Potos flavus: (1) arboreal/frugivore, insectivore.
2. Procyon cancrivorus: (1) semiaquatic/crustacivore, molluscivore, insectivore, piscivore, carnivore.
3. Nasua nasua: (1) terrestrial/insectivore, arachnidivore, carnivore, frugivore.
4. Bassariscus astutus: (1) terrestrial/carnivore, insectivore, frugivore.
5. Procyon lotor: (1) terrestrial/carnivore, granivore, frugivore, insectivore; and (2) semiaquatic/crustacivore, molluscivore, insectivore, piscivore, carnivore.

Food Habits and Basal Metabolism.—The most important foods in the diet of Procyon lotor are vertebrates, nuts, seeds, and fruits ([Table 9]). These are the same foods that are eaten by those dietary specialists that have Ḣ_b's equivalent to, or higher than, values predicted for them by the Kleiber equation (McNab, 1986a). The most important foods in the diets of Potos flavus, Procyon cancrivorus, and Nasua nasua are invertebrates and fruit ([Table 9]), and these foods are eaten by dietary specialists that have lower than predicted Ḣ_b's (McNab, 1986a). Major foods in the diet of Bassariscus astutus are terrestrial vertebrates, insects, and fruit ([Table 9]). Dietary specialists that eat terrestrial vertebrates have higher than predicted Ḣ_b's, whereas those that feed on insects have Ḣ_b's that are lower than predicted (McNab, 1986a). Year-round utilization of vertebrates by Bassariscus astutus suggests that it also should have a metabolic rate that is equivalent to or higher than predicted, rather than lower (McNab, 1986a). However, perhaps year-round inclusion of insects in its diet (Martin et al., 1951; Taylor, 1954; Wood, 1954; Toweill and Teer, 1977; Trapp, 1978), plus water-and energy-conserving advantages of a low metabolic rate, each exert a stronger selective influence on Ḣ_b than do vertebrates in its diet.

Summary.—The basal metabolic rate of these procyonids does appear to be influenced by diet. But, it is apparent from this family's evolutionary history and tropical origins that climate also has had a profound influence on its member's metabolism. The history of the family and the data presented here ([Table 7]) suggest that lower than predicted Ḣ_b is a feature that evolved very early as the primary metabolic adjustment to a tropical climate. From this perspective, it could be argued that climate would have been the major selective force determining Ḣ_b, whereas food habits would have had a secondary influence.