Potential Mechanisms of Phenotypic Divergence in Body Size Between Newfoundland and Mainland Black Bear Populations

Color profile: Generic CMYK printer profile Composite Default screen

1650

Potential mechanisms of phenotypic divergence in body size between Newfoundland and mainland black bear populations

Shane P. Mahoney, John A. Virgl, and Kim Mawhinney

Abstract: Phenotypic variation in body size and degree of sexual size dimorphism of North American black bears (Ursus americanus) was quantified for populations from New Brunswick, Quebec, Ontario, Maine, Alaska, and the island of Newfoundland. Based on a model of island biogeography developed by Case, we predicted that body size should be larger in Newfoundland bears than in mainland populations. The presence of few large predators and minimal competition between herbivore prey on Newfoundland allow an appropriate test of the model (i.e., food availability for bears may differ between populations on the mainland and in Newfoundland). In addition, sexual-selection theory predicts that the coevolution of polygyny and large size will be coupled with an increase in sexual size dimorphism. Therefore, we also predicted that among the six populations, male body mass should scale hyperallometrically with female body mass (i.e., slope > 1). Analysis of deterministic growth curves indicated that bears from Newfoundland attained greater asymptotic body size than populations on the mainland, which supports our first prediction. On average, the relative difference in asymptotic body mass between females from the island and mainland populations was 55%, while the relative difference between males was 37%. However, we found that sexual size dimorphism did not increase disproportionately with body mass among the six populations, which refuted our second prediction. We discuss a range of abiotic and biotic selection pressures possibly responsible for larger body size in Newfoundland bears. We suggest that the ability to exploit seasonally abundant and spatially dispersed dietary protein by female and male black bears on the island has been and is still a primary environmental factor selecting for large body size in Newfoundland bears. Although the relationship between sexual size dimorphism and body size is tenuous (slope ≤ 1), it does suggest that (an)other adaptive mechanism(s), opposing sexual selection for extreme male size, explain(s) a large amount of the variation in sexual size dimorphism among black bear populations.

Résumé : La variation phénotypique de la taille et de l’importance du1660 dimorphisme sexuel de la taille a été quantifiée chez des populations nord-américaines d’Ours noirs (Ursus americanus) du Nouveau-Brunswick, du Québec, de l’Ontario, du Maine, de l’Alaska et de Terre-Neuve. D’après un modèle de biogéographie insulaire élaboré par Case, nous avons prédit que la taille des ours de Terre-Neuve devait être supérieure à celle des ours des populations continentales. La présence limitée de prédateurs de grande taille et la compétition minimale entre les proies herbivores à Terre-Neuve sont des conditions appropriées pour tester le modèle (i.e., la disponibilité de la nourriture peut être diffé- rente chez les populations insulaires et les populations continentales). De plus, la théorie de la sélection sexuelle prédit que la coévolution de la polygynie et d’une grande taille devait s’accompagner d’une augmentation de l’importance du dimorphisme sexuel de la taille. Nous avons donc prédit en outre que, chez les six populations, la masse corporelle des mâles devait être hyperallométrique par rapport à la masse des femelles (i.e., pente > 1). L’analyse des courbes de croissance déterministes indique que les ours de Terre-Neuve atteignent une taille asymptotique supérieure à celle des ours des populations continentales, ce qui vérifie notre première prédiction. En moyenne, la différence relative entre la masse asymptotique des femelles insulaires et celle des femelles des populations continentales a été évaluée à 55 % et la différence relative entre les mâles, à 37 %. Cependant, le dimorphisme sexuel de la taille n’a pas augmenté de façon disproportionnée en fonction de la masse corporelle chez les six populations étudiées, ce qui infirme notre deuxième prédiction. Nous examinons une série de pressions de sélection possibles, abiotiques aussi bien que biotiques, qui pour- raient être responsables de la taille plus grande des ours de Terre-Neuve. Nous croyons que la capacité des ours mâles et femelles d’exploiter des sources saisonnières abondantes et éparses de protéines alimentaires dans l’île a été et demeure le facteur environnemental déterminant de la sélection en faveur d’une grande taille chez les ours de Terre- Neuve. Bien que la relation entre le dimorphisme sexuel de la taille et la taille elle-même soit ténue (pente ≤ 1), elle

Received August 16, 2000. Accepted July 31, 2001. Published on the NRC Research Press Web site at http://cjz.nrc.ca on September 7, 2001. S.P. Mahoney. Wildlife Division, Department of Forest Resources and Agrifoods, P.O. Box 8700, Building 810, St. John’s, NF A1B 4J6, Canada John A. Virgl.1 Ecological Developmental and Statistical Analysis, 222 Haight Place, Saskatoon, SK S7H 4W2, Canada. Kim Mawhinney. Parks Canada, 1869 Upper Water Street, Halifax, NS B3J 1S9, Canada. 1Corresponding author (e-mail: [email protected]).

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:14 AM Color profile: Generic CMYK printer profile Composite Default screen

Mahoney et al. 1651

indique tout de même qu’un ou plusieurs autres mécanismes évolutifs qui s’opposent à la sélection sexuelle favorisant le gigantisme des mâles expliquent une grande partie de la variation du dimorphisme sexuel de la taille chez les populations d’ours noirs.

[Traduit par la Rédaction] breeding system (Ralls 1977). For example, in monogamous Introduction Mahoney et al. species, individuals are typically small to medium-sized, male Within mammalian species, body size typically varies across parental investment can be high, and there is little or no sexual latitudinal and longitudinal gradients (McNab 1971; Ralls size dimorphism. Conversely, extreme polygyny is associated and Harvey 1985; Geist 1987; Brown 1995). Explanations with large body size, minimal male parental investment, and for the observed latitudinal pattern in body size include ad- a high degree of sexual size dimorphism. Comparative studies aptations for temperature, primary productivity, seasonal un- that regress male body mass on female body mass among predictability of food resources, and prey size. While the primate species have generally shown that male body mass classic correlation between increase in body size and decrease scales hyperallometrically (i.e., slope > 1; Fairbairn and in temperature (Bergmann’s rule) is controversial (Geist 1987), Preziosi 1994) with female body mass (Clutton-Brock et al. a number of studies do support this hypothesis (McNab 1971; 1977; Leutenegger 1978). Sexual selection for large body Burnett 1983; Owen 1989; Quin et al. 1996). Alternatively, size in males, and the associated advantage in terms of increased Rosenzweig (1968) demonstrated that size in mammalian mating opportunities, appears to be the primary mechanism carnivores was explained more by primary productivity than that explains the variation in sexual size dimorphism be- by temperature, and suggested that highly productive envi- tween monogamous and polygynous species (Fisher 1958; ronments should select for larger body size. Boyce (1979) Clutton-Brock et al. 1977; Ralls 1977; Leutenegger 1978). linked increasing latitude with primary productivity and sea- Therefore, an increase in body size for polygynous species sonality, and predicted that variability in food resources should be correlated with an increase in the degree of sexual should select for longer fasting endurance, which is posi- size dimorphism. However, environmental factors, such as tively correlated with body size. Finally, given the relation- spatial and temporal variation in availability of high-quality ship between maximum prey size and predator body size food resources, availability of receptive females, and length (Schoener 1969; Vézina 1985), an increase in prey size with of the mating period, acting on both female and male body latitude may also be coupled with an increase in predator size can constrain sexual selection for increasing size in size (Ralls and Harvey 1985). All of these environmental males (Fisher 1958; Clutton-Brock et al. 1977; Ralls 1977). factors, operating through evolutionary time and space, have We investigated geographic variation in body size and sex- likely contributed to the patterns of body-size variation in ual size dimorphism in North American black bears (Ursus species (Gould 1996). americanus) from five populations on the mainland and the Studies have also shown a link between body-size varia- population on the island of Newfoundland. Bears on New- tion and biogeographical isolation, with insular populations foundland have coexisted with wolves (Canis lupus) and car- often being larger or smaller than mainland populations ibou (Rangifer tarandus) since the end of the Wisconsin ice (Foster 1964; Case 1978). While lagomorphs, ungulates, foxes, age, except during the last 80 years, when wolves have been raccoons, and snakes tend to be relatively smaller on islands, extirpated from the island (Dodds 1983). Moose (Alces alces), other groups such as cricetid rodents, bears, and iguanid liz- which were introduced around the turn of the century (Pimlott ards display an increase in size relative to mainland popula- 1953), are the only other large prey of Newfoundland bears. tions. Based on optimal body size theory and predator–prey Relative to the number of different large predators and herbi- dynamics, Case (1978) developed a model to predict the di- vores on the mainland, the presence of few large predators rectional shift in body size among ecotype groups inhabiting and minimal competition between herbivore prey on New- island and mainland environments. His primary assumptions foundland meet the assumptions of the island biogeographical were that, in general, islands should (i) contain fewer preda- model described above (i.e., Case 1978) and allow an appro- tors on consumers, (ii) contain fewer consumer competitors, priate test of the model. Thus, based on the model developed and (iii) have a more benign climate. Given these conditions, by Case (1978) and the territorial behaviour of black bears, the increase in average food availability should be associated we first predicted that female and male black bears from with selection for increase in body size, provided that body Newfoundland would be larger than mainland populations. size is not overly constrained by other physical or biotic factors Furthermore, because black bears are polygynous and ex- (Case 1978). Although there were exceptions, Case found a hibit almost no male parental investment (except gamete loose correlation between island gigantism and territorial contribution), we used sexual-selection theory to predict that species and between island dwarfism and nonterritorial species. sexual size dimorphism should increase hyperallometrically Like the proposed latitudinal responses presented above, with body size among populations. island–mainland shifts (or phenotypic divergence) in body size likely reflect current and (or) historical differences in Materials and methods abiotic and biotic environmental selection pressures (Case and Schwaner 1993; Brown 1995; Abrams 1996). Animal collection and morphometrics Across mammalian species there is a good correlation be- Morphometric data were collected from black bear populations tween the proportional difference in female and male sizes, inhabiting six geographic areas in North America: Alaska, Ontario, or the degree of sexual size dimorphism, and the type of Quebec, New Brunswick, Newfoundland, and Maine. More

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:14 AM Color profile: Generic CMYK printer profile Composite Default screen

1652 Can. J. Zool. Vol. 79, 2001

Table 1. Capture methods and periods for black bears from six geographically separate populations in North America. Capture method Capture period Newfoundland Hunting/livetrapping 1985–1993 (May–November) New Brunswick Hunting 1989–1991 (May–June, October) Ontario Livetrapping 1989–1991 (May–August) Quebec Hunting 1983–1987 (June–July) Alaska Hunting/livetrapping 1977–1985 (January–December) Maine Hunting/livetrapping 1979–1991 (January–December)

specifically, the Alaska measurements were obtained from bears and the residual sum of squares (SSres) and coefficient of determi- captured within the areas of Moose Pens and Fingers Lake, and nation (r2) were also calculated. Ontario data were collected from individuals captured in the Chapleau area (northern Ontario). All bears from Quebec were captured within Statistical analyses the La Vérendrye Wildlife Reserve (southwestern Quebec) and indi- To assess the fit of each growth model to the data, a one-way viduals from Maine were captured within the counties of Spectacle analysis of variance was used to examine SS for body mass, and Pond, Stacyville, and Bradford. Studies involved a mixture of cap- res the shapes of the curves were visually assessed (Zullinger et al. ture methods and capture periods (Table 1). We acknowledge that 1984). Following selection of the model, total body length and differences in capture methods may have influenced the results to chest girth were then fitted to the same equation. Parameter esti- some degree, but we assume that whether hunting or trapping was mates among populations were judged to differ significantly if the used is incidental to the observed patterns of growth and adult size standard errors did not overlap. among populations. For descriptive purposes, means and standard deviations of neck Measured variables included body mass, total body length, chest girth were also calculated among age-classes (i.e., 0, 1, 2,…, ≥10). girth, and neck girth (except for Alaska and New Brunswick bears). Each age-class included bears captured within a calender year (i.e., Body mass was measured to the nearest 1 kg. Total body length, or January–December). For example, young of the year were placed contour length, was measured from the tip of the nose to the tip of in age-class 0. Because of the limited neck-girth values for bears the tail. Chest girth, or heart girth, was recorded as the circumfer- older than 10 years, all individuals ≥10 years of age were pooled ence of the body immediately posterior to the shoulders. Neck for each sex. girth represents the circumference of the neck directly behind the jaw. All length and girth measurements were recorded to the near- The relationship between sexual size dimorphism and body mass est 1 cm with a flexible steel tape. among populations was analyzed by least squares regression of log10(male body mass) on log10(female body mass), where male and female masses were asymptotic estimates obtained from growth Age determination and categorization curves. This represents an appropriate test of the allometric rela- Age of bears at first capture, except young of the year, was de- tionship between male and female body masses (Clutton-Brock et termined from an extracted premolar and by counting cementum al. 1977; Leutenegger 1978; Fairbairn and Preziosi 1994). Corre- annuli (Willey 1974). Age of recaptured bears was determined by lating the ratio of male to female body masses with average body back-dating to the initial capture date. For all bears age was ad- mass is not recommended, owing to the mathematical interdepen- justed to the month of capture. For example, assuming that the par- dence between dependent and independent variables (see Fairbairn turition period is in January (Alt 1983), a 2-year-old individual and Preziosi 1994). We specifically tested the hypothesis that the captured in August was recorded as being approximately 2.67 (2 + relationship between male and female body masses among popu- 8/12 months) years of age. Similarly, young of the year captured in lations would be hyperallometric (i.e., slope > 1). All statistical April were 0.33 years old. We assumed that this adjustment would analyses were performed with the SAS statistical package for minimize biases associated with gross changes in nutrient-reserve microcomputers (version 6.0; SAS Institute Inc. 1990). mass (fat and protein) that occur in black bears between hiberna- tion periods. Results Postnatal growth Geometric growth was analyzed using the following three growth models: Selection of a growth model Analysis of variance indicated that the moderate amounts –K(t–I) 3 [1] M(t)=A·[1 – 1/3 e ] of unexplained variance (SSres) associated with each growth equation for female and male black bears were very similar [2] M(t)=A·e–e–K(t–I) among models (F[2,33] = 0.01, P > 0.50). However, the shape [3] M(t)=A·[e–K(t–I) +1]–1 of the fitted curve varied enough among the models to generate differences in parameter estimates. For example, the where M(t) is body mass, total body length, or chest girth at age t, –1 von Bertalanffy model produced the largest estimates of as- A is asymptotic mass or size, K is the growth-rate constant (year ), ymptotic mass, while the logistic model generated the high- and I is age (years) at the inflection point. Equations 1, 2, and 3 est estimates of growth rate (see also Zullinger et al. 1984). represent the von Bertalanffy, Gompertz, and logistic growth curves, respectively, (Zullinger et al. 1984). The Gompertz model typically produced parameter estimates Data for individuals from each geographic area were fitted to that fell between estimates for the von Bertalanffy and logis- growth models, separately for each sex, using the Gauss–Newton tic equations. We decided to use the middle values generated method (PROC NLIN; SAS Institute Inc. 1990). Parameter esti- by the Gompertz model to estimate asymptotic body sizes mates (mean ± 1 SE) were generated after convergence was met, and growth rates among black bear populations.

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:15 AM Color profile: Generic CMYK printer profile Composite Default screen

Mahoney et al. 1653

Table 2. Estimates (mean ± 1 SE) of sex-specific asymptotic body mass (A, kg), growth-rate constant (K, year–1), and age at the inflection point (I, years) for black bears from six geographically separate populations in North America.

Females Males AK IAK I Newfoundland 101.1 ± 9.6 0.302 ± 0.096 1.9 ± 0.4 178.6 ± 9.7 0.313 ± 0.045 3.2 ± 0.3 New Brunswick 67.1 ± 2.4 0.514 ± 0.104 1.1 ± 0.3 145.5 ± 11.8 0.305 ± 0.050 2.7 ± 0.3 Ontario 67.3 ± 3.1 0.431 ± 0.081 2.1 ± 0.3 147.9 ± 8.6 0.253 ± 0.036 3.9 ± 0.3 Quebec 54.8 ± 1.4 0.593 ± 0.071 1.9 ± 0.1 118.2 ± 8.2 0.337 ± 0.048 3.2 ± 0.3 Alaska 68.4 ± 1.8 0.603 ± 0.063 1.6 ± 0.1 129.0 ± 4.5 0.418 ± 0.042 2.6 ± 0.1 Maine 65.8 ± 1.0 0.526 ± 0.021 2.1 ± 0.1 116.2 ± 2.4 0.522 ± 0.021 2.4 ± 0.1 Mean 70.7 ± 6.4 0.494 ± 0.045 1.8 ± 0.2 139.3 ± 9.5 0.356 ± 0.038 3.0 ± 0.2 Note: Parameters were estimated from the Gompertz model.

Phenotypic variation in body size and growth rate exhibited the highest growth rates, while males from the Estimates of asymptotic body mass indicated that female other populations, including Newfoundland, reached asymp- and male black bears from Newfoundland were significantly totic mass at 14–16 years of age and had lower growth rates. larger than individuals from mainland populations, which Except for populations from Newfoundland and Maine, the supports our first prediction (Table 2). The average difference growth rate of male bears was significantly lower than that in asymptotic mass between female bears from the island of females (Table 2). (101 kg) and mainland populations (65 kg) was approximately Analysis of other correlates of body size generated similar 55%, while the difference between males from the island patterns, although the results were more equivocal. Bears (179 kg) and mainland populations (131 kg) was approxi- from Newfoundland did not consistently exhibit the extreme mately 37%. estimates of asymptotic size, but they were invariably larger Among mainland populations, asymptotic estimates of body than the average for all populations. For example, asymp- mass were less variable for females than for males (Table 2). totic total body length was greatest for female bears from For example, female bears from New Brunswick, Ontario, Newfoundland and Ontario, followed by Alaska (Fig. 2). Alaska, and Maine did not differ in mass, but females from Total body lengths of females from New Brunswick, Quebec, these four populations were significantly larger than females and Maine were similar. Male bears from Ontario attained from Quebec. In contrast, asymptotic masses of male bears the largest total body length, followed by Newfoundland and from Ontario and New Brunswick were similar, and males Alaska (Table 3). Total body length for males from New from these two populations were significantly larger than Brunswick was greater than that for males from Maine but males from Quebec, Alaska, and Maine. In addition, males similar to that for males from Quebec. Asymptotic chest from Alaska were larger than males from Maine but similar girth in females from Newfoundland was significantly larger in size to those from Quebec (Table 2). than in females from mainland populations (Table 4). Among For all populations, age at the inflection point for females mainland populations, chest girth in females was largest for (1.8 ± 0.2 years; mean ± 1 SE) was less than for males (3.0 ± bears from Alaska, followed by those from Ontario, New 0.2 years), and the variation in inflection point among popu- Brunswick, Maine, and Quebec. Chest girth in males from lations for each sex was moderate (Table 2). For example, Newfoundland, New Brunswick, Ontario, and Alaska was age at the inflection point ranged from 1.1 years for females larger than in males from Quebec and Maine (Table 4). Mean from New Brunswick to 2.1 years for females from Ontario neck girth for bears older than 7 years of age, particularly and Maine. For males, age at the inflection point ranged females, was also largest for the population from Newfound- from 2.4 years for Maine bears to 3.9 years for Ontario land, followed by Ontario, Quebec, and Maine (Table 5). As bears. Age at the inflection point for Newfoundland females with body mass, the growth-rate constants obtained for total body length and chest girth were significantly higher for fe- and males was 1.9 and 3.2 years, respectively. males than for males, except in populations from Newfound- Variation in asymptotic body mass and age at the inflection land and Maine (Tables 3 and 4). point were correlated with sex-specific and among-population differences in growth rate. For mainland populations, females reached asymptotic mass at approximately 7–8 years Sexual size dimorphism of age (Fig. 1). In contrast, females from Newfoundland did Using the asymptotic values obtained from each popula- not reach asymptotic mass until they were 12 years of age. tion, a regression of male body mass on female body mass Thus, although females from Newfoundland had a signifi- generated a significant relationship (F[1,5] = 10.86, P = 0.03, cantly slower growth rate than females from mainland popu- r2 = 0.73). However, the slope (0.69) of the relationship was lations, they spent more time growing and attained a greater not statistically different from 1 (F = 2.18, P = 0.21), indicat- asymptotic size (Table 2). Relative to females, age at asymp- ing that male body mass did not increase hyperallometrically totic mass and growth rate for males were more variable with female body mass (Fig. 3). Although our estimate of among populations (Fig. 1, Table 2). Males from Alaska and the allometric exponent (slope) must be tempered by our low Maine attained asymptotic mass at 10–11 years of age and sample size, it does suggest a trend towards a decrease in

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:16 AM Color profile: Generic CMYK printer profile Composite Default screen

1654 Can. J. Zool. Vol. 79, 2001

Fig. 1. Geometric growth in body mass for female (᭡) and male (᭝) black bears (Ursus americanus) from different populations in North America. Predicted curves were calculated from the Gompertz model.

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:18 AM Color profile: Generic CMYK printer profile Composite Default screen

Mahoney et al. 1655

Fig. 2. Geometric growth in total body length for female (᭡) and male (᭝) black bears from different populations in North America. Predicted curves were calculated from the Gompertz model.

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:20 AM Color profile: Generic CMYK printer profile Composite Default screen

1656 Can. J. Zool. Vol. 79, 2001

Table 3. Estimates (mean ± 1 SE) of sex-specific asymptotic total body length (A, cm) and growth-rate constant (K, year–1) for black bears from six geographically separate populations in North America.

Females Males AK AK Newfoundland 164.9 ± 3.5 0.479 ± 0.088 183.2 ± 3.0 0.560 ± 0.060 New Brunswick 145.6 ± 1.3 0.688 ± 0.101 170.3 ± 2.5 0.496 ± 0.051 Ontario 162.0 ± 2.0 0.487 ± 0.046 192.7 ± 4.2 0.335 ± 0.048 Quebec 146.1 ± 1.7 0.874 ± 0.118 168.9 ± 3.4 0.479 ± 0.048 Alaska 157.6 ± 1.8 0.653 ± 0.074 181.1 ± 2.4 0.490 ± 0.047 Maine 147.5 ± 0.7 0.901 ± 0.022 164.4 ± 1.1 0.906 ± 0.024 Mean 154.0 ± 3.5 0.676 ± 0.073 176.7 ± 4.4 0.542 ± 0.075 Note: Parameters were estimated from the Gompertz model.

Table 4. Estimates (mean ± 1 SE) of sex-specific asymptotic chest girth (A, cm) and growth-rate constant (K, year–1) for black bears from six geographically separate populations in North America.

Females Males AK A K Newfoundland 98.1 ± 3.5 0.448 ± 0.104 119.7 ± 3.4 0.343 ± 0.045 New Brunswick 82.8 ± 1.4 0.549 ± 0.105 115.0 ± 4.4 0.279 ± 0.047 Ontario 83.6 ± 1.5 0.453 ± 0.054 117.0 ± 5.2 0.252 ± 0.048 Quebec 76.9 ± 1.9 0.581 ± 0.132 99.5 ± 3.4 0.382 ± 0.049 Alaska 86.8 ± 1.4 0.657 ± 0.099 116.3 ± 3.5 0.311 ± 0.046 Maine 80.8 ± 0.6 0.632 ± 0.023 97.3 ± 1.1 0.609 ± 0.022 Mean 84.9 ± 3.0 0.550 ± 0.034 110.4 ± 3.9 0.364 ± 0.050 Note: Parameters were estimated from the Gompertz model.

sexual size dimorphism with increase in body size among sures that may be correlated with variation in body size populations. among populations, our explanations remain hypothetical. One possible factor is the difference in spatial and temporal varia- Discussion tion in food resources, particularly the availability of dietary protein (Schoener 1969; Case 1978; Case and Schwaner 1993). Adult female and male black bears from the island of New- In this context, the island of Newfoundland may represent a foundland were determined to be significantly larger than relatively unique foraging environment for North American bears from mainland populations. On average, the difference black bears. For example, caribou and black bears have co- in asymptotic body mass between females from the island existed on the island since the end of the Wisconsin ice age, and mainland populations was 55%, while the difference be- while moose were twice introduced around the turn of the tween males was 37%. Although bears from Newfoundland 20th century (circa 1878 and 1904; Pimlott 1953). Thirteen generally exhibited lower growth rates than mainland popu- subpopulations of caribou inhabit Newfoundland and moose lations, females and, to a lesser degree, males spent more have extensively colonized the island (the current estimate is time growing and attained asymptotic mass later than indi- 150 000) (Mahoney 20002). In accordance with Case’s viduals from mainland populations. Less correlative measures model (1978), minimal interspecific competition between of body size, such as total body length and chest girth, gen- these herbivores creates the potential for an abundance of erated more equivocal results, but asymptotes for Newfound- caribou and moose calves that can provide a rich source of land bears were invariably among the largest. Therefore, the protein for bears during the spring and early summer results supported our prediction that Newfoundland black (Mahoney 1986; Mahoney et al. 1990; Schwartz and bears would be larger than bears from the five mainland pop- Franzmann 1991; Ballard 1992). Although hard mast crops are ulations. unavailable on the island, early-successional forest stands and There are a number of potential environmental selection wetland and “tundra” heath habitats can support an abun- pressures, which are not mutually exclusive, that could ex- dance of soft mast during the late summer and autumn plain the observed phenotypic divergence in body size be- (Damman 1993). Furthermore, millions of capelin (Mallotus tween Newfoundland and mainland bears. However, without villosus) spawn on accessible beaches in early summer and explicit statistical testing of the environmental selection pres- are fed upon by bears, as are the large numbers of spawning

2 S.P. Mahoney. 2000. A synthesis and interpretation of the biology of woodland caribou on the island of Newfoundland. Vols. 1–15.

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:20 AM Color profile: Generic CMYK printer profile Composite Default screen

Mahoney et al. 1657

Atlantic salmon (Salmo salar) that provide North American bear populations with high dietary energy and protein during the late summer and autumn (Welch et al. 1997; Jacoby et al. 1999). In a recent study of brown bears, Ferguson and McLoughlin (2000) also linked higher primary productivity and the presence of anadromous fishes to larger mass of females in coastal bear populations relative to interior populations. Mainland black bear populations also have access to a range of food items, but differences in the temporal continu- ity of dietary protein prior to each growth diapause (i.e., hi- bernation) may cause mainland populations to be relatively more protein-limited than Newfoundland bears. If the average seasonal availability of food is greater for Newfoundland bears than for mainland populations, then being able to maximize the energy and protein acquired from plants, caribou and moose calves, capelin, and salmon likely requires that individuals travel efficiently from one locally abundant food source to another. Although no work has been done on body size and travel efficiency in bears, allometric patterns predict that body size will be generally positively correlated with cursorial development (e.g., larger lungs and chest girth and longer limbs), which would enable individuals to move more efficiently over greater foraging distances (Calder 1984; Schmidt-Nielsen 1994). Furthermore, while black bears are omnivorous, they are also predatory, and for solitary terrestrial carnivores there is a strong correlation between maximum prey size and predator body size (Schoener 1969; Vézina 1985). Thus, larger body size in Newfound- land black bears may be a phenotypic response allowing them to take advantage of abundant but spatially dispersed dietary protein, and conferring the ability to efficiently kill large prey. The shift in body size between Newfoundland and mainland black bear populations may also be linked to current and (or) historic differences in the strength of interspecific competition for prey (Abrams 1996; Losos 1996; McPeek 1996). During the past 80 years, black bears have been the exclusive large carnivore on the island. Wolves were extirpated around 1920, and coyotes (Canis latrans) have been present on the island only since 1985 (Larivière and Crête 1993). Again, in accordance with one of the assumptions of Case’s model (Case 1978), limited competition with other carnivores for large prey during the past 80 years or so may also have influenced body size in Newfoundland black bears. Whether or not phenotypic divergence in body size between Newfoundland and mainland black bear populations

is genetically linked remains to be determined. The New- female and male black bears of different age-classes from Newfoundland, Ontario, Quebec, and Maine. foundland black bear is currently recognized as a distinct subspecies (Ursus americanus hamiltoni), based on differences in skull size and shape (Cameron 1956). More recently, Paetkau and Strobeck (1996) sequenced eight haplotypes

from black bear populations across North America but could n±1SD)of find no strong phylogenetic split between the Newfoundland and mainland populations. Still, heritability of body morphology is typically high (approximately 0.50; Roff 1992), and the larger size of Newfoundland bears could be partially

explained by genetic drift in a small isolated founder popula- Newfoundland Ontario Quebec Maine

tion (Paetkau and Strobeck 1996; Slatkin 1996; Lynch et al. Neck girths (mea 1997). During postglacial colonization of the island, a neu- Numbers in parentheses are sample sizes; nd, no data. tral or nonadaptive mutation in the gene complex regulating Note: 10 59.5 ± 2.4 (4) 72.6 ± 10.7 (10) 51.6 ± 5.2 (8) 69.4 ± 8.8 (16) 47.4 ± 4.0 (5) 70.3 ± 6.5 (3) 46.7 ± 2.9 (38) 63.7 ± 9.9 (12) ≥ body size may have provided the Newfoundland population Table 5. Age (years)0 Females123 26.34 ± 4.6 (2) 36.1 Males5 ± 5.5 (6) 41.56 ± 7.3 (10) 51.2 28.97 ± ± 16.2 4.4 (5) 46.4 (6) 40.08 ± ± 2.9 48.2 5.6 (5) nd ± (13)9 4.1 (21) 48.7 48.6 ± ± 4.2 Females 23.9 2.1 (11) ± (5) 50.0 nd 55.0 4.2 ± ± (4) 1.4 42.5 0.9 (2) 59.3 (1) (4) ± 13.8 39.6 (2) nd ± 60.3 4.8 (1) (7) 22.0 ± 74.0 4.0 (1) 48.6 (3) Males 76.5 ± ± 6.8 59.3 2.1 (4) ± (2) 6.0 44.1 (4) ± 42.0 5.2 ± (12) 3.3 19.0 (5) 38.0 (1) ± 50.6 3.6 ± 51.6 51.7 (4) 5.7 ± ± (8) 47.1 8.1 1.8 ± (11) 42.2 50.5 (7) 74.5 3.9 ± ± (1) (7) 3.7 3.9 36.6 (11) (7) Females ± 4.4 (9) 52.0 30.0 ± ± 42.0 62.3 14.7 3.5 ± ± (11) (7) nd 42.8 2.4 6.2 20.9 ± (7) 66.0 (6) ± 4.1 ± 1.2 (14) 3.9 (5) 39.1 (6) ± 44.6 4.9 ± (19) 7.0 29.9 48.2 (8) ± 49.2 ± Males 38.7 3.0 ± 2.6 nd ± (9) 2.6 (3) 18.6 4.4 (5) ± (66) 4.2 46.3 35.3 (199) ± ± 1.5 4.1 (3) 55.9 (80) ± 28.8 73.4 7.9 ± ± 41.0 (4) 47.4 19.2 3.4 6.9 ± ± ± (122) (5) 3.0 6.2 45.6 5.4 (45) (95) ± (252) 64.0 3.4 42.7 ± (10) ± 5.7 Females 5.0 (2) 43.0 (103) 31.3 ± ± 3.9 4.8 54.7 (37) 48.0 (145) ± ± 59.2 60.9 6.1 3.2 ± ± (46) (5) 7.5 14.8 (3) (8) 44.7 ± 2.9 (23) 58.9 ± 4.7 (15) 42.4 64.5 ± Males 46.6 ± 4.3 ± 0.7 (51) 3.7 (2) (22) 58.8 ± 5.9 (13) 56.1 44.7 ± 61.8 ± 5.7 ± 3.1 (38) 3.7 (11) (6) 62.0 ± 5.9 (4)

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:21 AM Color profile: Generic CMYK printer profile Composite Default screen

1658 Can. J. Zool. Vol. 79, 2001

Fig. 3. Allometric relationship between male and female body masses for black bears from six geographically separate populations in North America. The slope of the relationship was greater than 0 (P = 0.03) but not significantly less than 1 (P = 0.21).

with the ability to respond to differences in protein availability, variation in adult body mass in black bears, then we would prey size, and levels of interspecific competition. Alterna- expect that an increase in adult size would be associated tively, variation in body size among phylogenetically similar with an increase in sexual size dimorphism (Fisher 1958). In black bear populations may largely reflect the norm of reaction other words, an increase in female mass should be accompa- to these different environmental selection pressures (Roff nied by a disproportionate increase in male mass, which 1992; Stearns 1992). We do not know the exact genetic and would generate an allometric slope greater than 1 (i.e., environmental factors that led to the phenotypic divergence hyperallometry; Fairbairn and Preziosi 1994). Therefore, we in body size between the island and mainland populations, predicted that the allometric relationship between male body but we believe that the ability of bears to efficiently exploit mass and female body mass in black bears should have an seasonally abundant and spatially dispersed dietary protein exponent greater than 1. Although tenuous, our results did on Newfoundland has been and is still a primary factor. not support this prediction, and we found a trend towards As in other polygynous species, adult male black bears decreasing sexual size dimorphism and increasing body size are larger than adult females. For body mass, age at the in- among black bear populations (Fig. 3). flection point in females occurred near weaning (i.e., mean In a study on western bobcats (Lynx rufus), Dobson and 1.8 years), while geometric growth in males did not slow Wigginton (1996) also found a tendency for female body down until they were approximately 3 years of age. Thus, size to increase disproportionately with male body size (i.e., although the growth rate in male bears was less than or slope = 0.85), but as in our results the exponent was not sig- equal to that in females, males spent more time growing and nificantly less than 1. The allometric patterns exhibited by achieved greater asymptotic body mass. Theoretically, the both black bears and bobcats suggest that (an)other factor(s), larger size of male black bears could have evolved from besides sexual selection, explain(s) a large amount of the intrasexual competition for females and (or) the relaxation variation in male body size and the associated sexual size of competition for food between females and males (i.e., dimorphism within these species. Because the evolution competitive character displacement; Selander 1972; Clutton- of polygyny and large size appear to be quite inseparable Brock et al. 1977; Shine 1989). For black bears, however, (Leutenegger 1978), we hypothesize that similar allometric the spatial and temporal availability of animal protein, and patterns to those for black bears and bobcats can be expected phenology and dispersion of nutritious plant items, are ex- within other species where the extent of polygyny is tempered pected to be similar for both sexes. This should minimize the by local environmental conditions. Factors that constrain potential for strong niche separation, although whether or polygyny, such as the feasibility of males monopolizing ei- not female and male bears differ in their propensity to prey ther food resources or females (which depends on the length on ungulates remains unclear. We believe that sexual size di- of the mating period and the spatial and temporal dispersion morphism in bears has mostly evolved through intrasexual of food or receptive females), will oppose sexual selection selection in males for increased competitive abilities and for extreme size in males (Clutton-Brock and Harvey 1978). larger home ranges, which is associated with a polygynous Furthermore, these opposing processes may scale differently mating system (see also Bunnel and Tait 1981). among hierarchical levels (Wiens 1989), and population-level If sexual selection is the key factor explaining most of the responses to varying local environmental-selection pressures

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:21 AM Color profile: Generic CMYK printer profile Composite Default screen

Mahoney et al. 1659

may show different allometric patterns than species-level Dodds, D. 1983. Terrestrial mammals. In Biogeography and ecol- associations (Ralls 1977; Leutenegger 1978; Fairbairn and ogy of the island of Newfoundland. Edited by G.R. South. Dr. Preziosi 1994; McLoughlin and Ferguson 2000). W. Junk Publishers, The Hague, the Netherlands. pp. 509–550. Fairbairn, D.J., and Preziosi, R.F. 1994. Sexual selection and the evolution of allometry for sexual size dimorphism in the water Acknowledgments strider, Aquarius remigis. Am. Nat. 144: 101–118. Our greatest appreciation is extended to the following people Ferguson, S.H., and McLoughlin, P.D. 2000. Effect of energy avail- and government agencies who provided body-morphology ability, seasonality, and geographic range on brown bear life history. Ecography, 23: 193–200. data: D. Larsen, T. McCarthy, C. Schwartz, and T. Wettin Fisher, R. 1958. The genetical theory of natural selection. Dover, with the Department of Fish and Game, Alaska; K. Craig New York. with the Fish and Wildlife Branch, New Brunswick; M. Obbard Foster, J.B. 1964. Evolution of mammals on islands. Nature (Lond.), with the Wildlife Research Section, Ontario; H. Jolicoeur 202: 234–235. with the Government of Quebec; L. Berchielli with the Wild- Geist, V. 1987. Bergmann’s rule is invalid. Can. J. Zool. 65: 1035– life Research Center, New York; and C. McLaughin with the 1038. Department of Inland Fisheries and Wildlife, Maine. We are Gould, S.J. 1996. Full house: the spread of excellence from Plato heavily indebted to Mr. H. Abbott, who for nearly 10 years to Darwin. Three Rivers Press, New York. gathered and assembled morphometric data on captured ani- Jacoby, M.E., Hilderbrand, G.V., Servheen, C., Schwartz, C.C., mals in Newfoundland. S. Larivière provided technical assis- Arthur, S.M., et al. 1999. Trophic relations of brown and black tance during the production of the manuscript. We benefited bears in several western North American ecosystems. J. Wildl. from the constructive comments of B. McLaren and two Manag. 63: 921–929. anonymous reviewers. Funding for this project was provided Larivière, S., and Crête, M. 1993. The size of eastern coyotes: a by the provincial government of Newfoundland and Labra- comment. J. Mammal. 74: 1072–1074. dor, which has continued its commitment to the development Leutenegger, W. 1978. Scaling of sexual dimorphism in body size of wildlife science. and breeding system in primates. Nature (Lond.), 272: 610–611. Losos, J.B. 1996. Phylogenetic perspectives on community ecology. References Ecology, 77: 1344–1354. Lynch, J.M., Whelan, R., Fituri, A.I., and Hayden, T.J. 1997. Cranio- Abrams, P.A. 1996. Evolution and the consequences of species intro- metric variation in the Eurasian badger, Meles meles. J. Zool. ductions and deletions. Ecology, 77: 1321–1328. (Lond.), 242: 31–44. Alt, G.L. 1983. Timing of parturition in black bears (Ursus ameri- Mahoney, S.P. 1986. The black bear in Newfoundland. Osprey, 16: canus) in northeastern Pennsylvania. J. Mammal. 64: 305–307. 10–21. Ballard, W.M. 1992. Bear predation on moose: a review of recent Mahoney, S.P., Abbott, H., Russell, L.H., and Porter, B.R. 1990. North American studies and their management implications. Alces, Woodland caribou calf mortality in insular Newfoundland. In 28: 162–176. Transactions of the 19th Congress of the International Union Boyce, M.S. 1979. Seasonality and patterns of natural selection for of Game Biologists, Trondheim, Norway, 1989. Edited by S. life histories. Am. Nat. 114: 569–583. Myrberget. Norwegian Institute for Nature Research, Trondheim. Brown, J.H. 1995. Macroecology. University of Chicago Press, pp. 592–599. Chicago. McLoughlin, P.D., and Ferguson, S.H. 2000. A hierarchical pattern Bunnel, F.L., and Tait, D.E.N. 1981. Population dynamics of bears— of limiting factors helps explain variation in home range size. implications. In Dynamics of large mammal populations. Edited Ecoscience, 7: 123–130 by C.W. Fowler and T.D. Smith. John Wiley and Sons, New McNab, B.K. 1971. On the ecological significance of Bergmann’s York. pp. 75–98. rule. Ecology, 52: 845–854. Burnett, C.D. 1983. Geographic and climatic correlates of morpho- McPeek, M.A. 1996. Linking local species interactions to rates of logical variation in Eptesicus fuscus. J. Mammal. 64: 437–444. speciation in communities. Ecology, 77: 1355–1366. Calder, W.A. 1984. Size, function, and life history. Harvard Uni- Owen, J.G. 1989. Population and geographic variation of Peromyscus versity Press, Cambridge, Mass. leucopus in relation to climatic factors. J. Mammal. 40: 98–109. Cameron, A.W. 1956. A new black bear from Newfoundland. Paetkau, D., and Strobeck, C. 1996. Mitochondrial DNA and the J. Mammal. 37: 538–540. phylogeography of Newfoundland black bears. Can. J. Zool. 74: Case, T.J. 1978. A general explanation for insular body size trends 192–196. in terrestrial vertebrates. Ecology, 59: 1–18. Pimlott, D.H. 1953. Newfoundland moose. Trans. N. Am. Wildl. Case, T.J., and Schwaner, T.D. 1993. Island/mainland body size Nat. Resour. Conf. 18: 563–581. differences in Australian varanid lizards. Oecologia, 94: 102–109. Quin, D.G., Smith, A.P., and Norton, T.W. 1996. Eco-geographic Clutton-Brock, T.H., and Harvey, P.H. 1978. Mammals, resources, variation in size and sexual dimorphism in sugar gliders and and reproductive strategies. Nature (Lond.), 273: 191–195. squirrel gliders (Marsupialia: Petauridae). Austr. J. Zool. 44: Clutton-Brock, T.H., Harvey, P.H., and Rudder, B. 1977. Sexual di- 19–45. morphism, socionomic sex ratio and body weight in primates. Ralls, K. 1977. Sexual dimorphism in mammals: avian models and Nature (Lond.), 269: 797–800. unanswered questions. Am. Nat. 111: 917–938. Damman, A.W.H. 1993. An ecological subdivision of the island of Ralls, K., and Harvey, P.H. 1985. Geographic variation in size and Newfoundland. Monogr. Biol. 48: 163–206. sexual dimorphism of North American weasels. Biol. J. Linn. Dobson, F.S., and Wigginton, J.D. 1996. Environmental influences Soc. 25: 119–167. on the sexual dimorphism in body size of western bobcats. Roff, D.A. 1992. The evolution of life histories: theory and analy- Oecologia, 108: 610–616. sis. Chapman and Hall, New York.

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:22 AM Color profile: Generic CMYK printer profile Composite Default screen

1660 Can. J. Zool. Vol. 79, 2001

Rosenzweig, M.L. 1968. The strategy of body size mammalian Slatkin, M. 1996. In defense of founder-flush theories of speciation. carnivores. Am. Midl. Nat. 80: 299–315. Am. Nat. 147: 493–505. SAS Institute Inc. 1990. SAS user’s guide: statistics, version 6.0. Stearns, S.C. 1992. The evolution of life histories. Oxford Univer- SAS Institute Inc., Cary, N.C. sity Press, Oxford. Schmidt-Nielsen, K. 1994. How animals work. Cambridge Univer- Vézina, A.F. 1985. Empirical relationships between predator and sity Press, Cambridge, Mass. prey size among terrestrial vertebrate predators. Oecologia, 67: Schoener, T.W. 1969. Models of optimal size for solitary predators. 555–565. Am. Nat. 103: 277–313. Welch, C.A., Keay, J., Kendall, K.C., and Robbins, C.T. 1997. Schwartz, C.C., and Franzmann, A.W. 1991. Interrelationship of Constraints on frugivory by bears. Ecology, 78: 1105–1119. black bears to moose and forest succession in the northern Wiens, J.A. 1989. Spatial scaling in ecology. Funct. Ecol. 3: 385– coniferous forest. Wildl. Monogr. No. 113. 397. Selander, R.K. 1972. Sexual selection and dimorphism in birds. In Willey, C.H. 1974. Aging black bears from first premolar tooth Sexual selection and the descent of man. Edited by B.G. Campbell. sections. J. Wildl. Manag. 38: 97–100. Aldine, Chicago. pp. 180–230. Zullinger, E.M., Ricklefs, R.E., Redford, K.H., and Mace, G.M. Shine, R. 1989. Ecological causes for the evolution of sexual di- 1984. Fitting sigmoidal equations to mammalian growth curves. morphism: a review of the evidence. Q. Rev. Biol. 64: 419–461. J. Mammal. 65: 607–636.

J:\cjz\cjz79\cjz-09\Z01-122.vp Thursday, August 30, 2001 11:43:22 AM