The number, size, and survival of bear cubs emerging from winter dens depend on maternal condition prior to entering the den. We hypothesized that delayed implantation provides flexibility in timing of birth such that pregnant females are able to track environmental or body conditions long after conception to optimize reproductive output in a changing environment. We tested the hypotheses that causative links between maternal condition and size of newly emerging brown bear (Ursus arctos) cubs were females in superior condition give birth earlier and, thereby, lactate longer in the den than females in poorer condition; and females in superior condition produce more milk or higher quality milk, which accelerates cub growth relative to females in poorer condition. No brown bear with a body fat content ≤ 20% produced cubs even though breeding occurred. Brown bears that were fat gave birth earlier than those that were lean. Cubs nursing from fat mothers grew faster than those nursing from lean mothers. The combination of an earlier birth date and faster growth by cubs produced from fat mothers increased mass of brown bear and polar bear (U. maritimus) twins at den emergence by 330– 360 g for each unit increase in percent maternal body fat content when entering hibernation.
Understanding the ecology and evolution of mammalian reproduction has been a major focus of modern biology (Charnov 1982; Sadleir 1969; Sandell 1990; Trivers and Willard 1973). Newborn sex ratios, adult body size, mating tactics and investment, mating and birthing times, and parental offspring investment are selectable traits (e.g., Kovach and Powell 2003; Powell and King 1997; Sandell 1989). One important aspect of reproduction in many carnivores is delayed implantation (Sandell 1990). For example, north temperate bears mate in spring to early summer, the fertilized eggs or blastocysts implant late fall to early winter, and cubs are born in early to midwinter (Ramsay and Stirling 1988). For the next 3-5 months while the female continues to hibernate, milk produced from maternal reserves is the only nourishment for cubs.
Sandell (1990) proposed that delayed implantation evolved when female fitness increased by prolonging the interval between mating and birthing beyond a normal gestation length to maximize female choice or male competition during mating. He did not consider that the timing of implantation and parturition might also vary to optimize female fitness. Several studies in other mammals have shown that food availability and female condition influence pregnancy length and date of birth in species with and without delayed implantation (Ben-David 1998; Berger 1992; Bowyer 1991; Cook et al. 2004; Rachlow and Bowyer 1991; Woodroffe 1995). Although species without delayed implantation are limited in the extent to which pregnancy length can be modified (Sandell 1990), we hypothesized that species with delayed implantation may have greater flexibility such that varying the timing of implantation and birth may be an important part of their reproductive strategy (Woodroffe 1995; Zhang et al. 2009). This may be particularly true in species that give birth and begin nursing during a prolonged fast.
Reproductive success in North American bears in a wide variety of ecosystems depends on seasonal food resources that 541 enable females to accumulate fat (Alt 1989; Atkinson and Ramsay 1995; Derocher and Stirling 1998; McLellan 2011; Ramsay and Stirling 1988; Schwartz and Franzmann 1991; Schwartz et al. 2006). For example, only 14% of adult female American black bears (Ursus americanus) in south-central Alaska without older, accompanying offspring during the preceding fall were observed with spring cubs when their fall body fat content averaged 19% ±3% (X̄ ±SD) but 67% of such females had cubs when their fall body fat content increased to 35% ±5% (Belant et al. 2006). Similarly, the body fat content of the leanest polar bear (Ursus maritimus) that produced cubs in western Hudson Bay was about 20%, although most had body fat contents > 40% (Atkinson and Ramsay 1995). Unfortunately, the causative links between food resources, maternal condition (i.e., body fat content) prior to hibernation, and cub production and survival in the spring are unknown.
Two competing but not exclusive hypotheses may account for the link between maternal condition and cub production, size, and survival. These hypotheses include females in superior condition give birth earlier and, thereby, lactate longer in the den than do females in poorer condition (i.e., cubs born to fat mothers are older at den emergence than those born to lean mothers); and females in superior condition produce more milk or higher quality milk that accelerates cub growth relative to the cubs produced by leaner females (i.e., cubs born to fat mothers are better nourished than those born to lean mothers). The latter hypothesis assumes that female condition, and not cub demand, is the primary determinant of variation in milk production.
A variable birth date and milk production capability as modulated by female condition would enable females to track environmental or body conditions long after conception (Ferguson et al. 1996). Such flexibility would allow females to optimize reproductive output in a changing environment. Thus, females in superior condition could produce larger young at den emergence that maximized their fitness, but females in poorer condition could either forgo implantation and thereby not produce cubs during a particular year or produce offspring at a reduced cost by shortening lactation or reducing the intensity of lactation while denned. The trade-off or cost to the 2nd alternative is that smaller offspring emerging from the den may be less competitive with offspring of the same or differing species in better condition and thereby have a reduced chance of survival (Atkinson and Ramsay 1995; Dahle et al. 2006; Derocher and Stirling 1996; Mattson et al. 2005; Ramsay and Stirling 1988). For example, small polar bear cubs (<6 kg) in western Hudson Bay had less than a 20% chance of survival during their 1st year as compared to large cubs (≥22 kg) that had as much as an 80% chance of survival (Derocher and Stirling 1996).
If female reserves rather than cub demand determine variation in the amount or quality of the milk during early lactation, the birth of additional cubs beyond the norm (i.e., triplets versus twins) might not result in completely compensatory increases in milk production. Although twins are most common among North American brown bears (Ursus arctos; 2.0 ±0.2 cubs/litter in 13 interior populations and 2.3 ±0.3 cubs/litter in 6 coastal, salmon-feeding populations), triplets or larger litters can occur (Hilderbrand et al. 1999a; McLellan 1994). Thus, triplets should grow more slowly and be smaller at den emergence than twins when maternal reserves are limited. Any lack of a compensatory increase in milk production when nursing triplets as compared to twins is unlikely to be due to limited mammary gland capacity because not all 6 mammary glands on brown bears produce milk during early lactation (i.e., there is surplus capacity), and because milk production increases 4-fold between hibernation and later in the summer at peak production (Farley and Robbins 1995).
Limited milk production by leaner females also may increase competition between littermates, reduce survival within a litter, and thereby allow females to focus limited resources on the most robust offspring (Dahle et al. 2006; Ramsay and Stirling 1988). However, failure to sense maternal reserves correctly and to modulate reproductive investment during hibernation could result in death of the cubs and prolong the interbirth interval as females reaccumulate sufficient reserves to initiate the reproductive process (Hissa et al. 1998; Iverson et al. 2001; McDonald 1998; Moschos et al. 2002; Spady et al. 2009). Interbirth intervals in brown bears range from 2 to 5 years and, thus, any delay could markedly reduce maternal fitness (Dahle and Swenson 2003; McLellan 1994). Consequently, these links between maternal reserves and cub production should be highly selected traits that should be apparent in studies of both captive and wild bears.
Materials and Methods
Animals, diets, and housing for captive bears.—We randomly assigned 6 captive, adult (6- to 20-year-old) female brown bears at the Washington State University Bear Research, Education, and Conservation Center to mass and body fat targets prior to hibernation. We then varied the amount of food that each bear was fed from the beginning of August to the middle of October, 2000–2011, to achieve those targets. Targeted body fat contents for each female ranged from 17% to 40%, which covered much of the range observed in free-ranging brown bears (Belant et al. 2006; Hilderbrand et al. 1998, 1999b, 2000; Robbins et al. 2004).
We fed apples and a commercial chow containing 21% protein to each bear twice daily at 0800 and 1600 h. Bears were released daily into a 0.9-ha irrigated exercise yard for 8-16 h/day where they freely grazed white clover (Trifolium repens) and grasses (Poa pratensis and Bromus gracilis). Females bred naturally in May with 1 or 2 resident males. During the course of the study, a total of 4 males were used, although only 2 were available during any 1 breeding season. Mate choice was determined by each female as all females were released into the exercise yard with the 2 resident males.
Hibernation or the cessation of feeding and foraging started in late October when all bears were closed into their individual dens and runs, and hibernation ended mid- to late March when 542 feeding resumed. We chose these times because appetites and, therefore, the amount of food offered by us had dropped significantly by late October. When brown bears are deprived of food during the active season, they show very obvious signs of discomfort (i.e., growling, roaring, and extreme aggression). However, when food was removed in late October, the females exhibited silence and indifference. Similarly, bears had become increasingly active by mid- to late March when daytime temperatures were noticeably warmer and vegetation was starting to grow. Thus, both October and March behaviors indicated that the bears were voluntarily going into and coming out of hibernation.
Females hibernated singly in individual 3 × 3-m concrete dens and were given a bale of straw to form a bed. Dens were connected to 3 × 6-m outside, concrete-floored runs. The doors between the dens and outside runs were left open so each den was illuminated by natural light. Bears were never closed into their dens. Thus, bears could always exit and enter the den, but they could not go into the large outside exercise yard during hibernation. The overall den area was closed to human entry during hibernation to minimize disturbance except during infrequent entries for weighing cubs.
Animals used in this study were cared for in accordance with current animal care and use guidelines approved by the American Society of Mammalogists (Sikes et al. 2011). This project was approved by the Washington State University Institutional Animal Care and Use Committee under protocol 03054–003.
Measuring maternal body fat content, birth date, and growth of captive cubs.—At the start of hibernation, we measured body mass of the adult females using electronic scales weighing to the nearest 0.5 kg. Body composition was estimated using bioelectrical impedance analysis (model 101 A; RJL Systems, Detroit, Michigan—Farley and Robbins 1994; Hilderbrand et al. 1998). All bears were fasted for 16 h before measurements to ensure empty gastrointestinal tracts (Felicetti et al. 2003).
We determined birth dates by installing in each den a camera (MGB600 high-resolution black and white with QLAV2 2.6- 6 mm lens in V28CC housing; Silent Witness, Surrey, British Columbi, Canada), 3 infrared illuminators (model SWIR-24 volt AC; Silent Witness), and a microphone (Verifact Microphone, model A connected to AP-4 Audio Base Station; Louroe Electronics, Van Nuys, California). The cameras and microphones were connected to a computer and video screen (X240 16 Channel High Definition Digital Recorder with 500- GB hard drive; Open Eye, Spokane, Washington) distant to the den that permitted storage and replay of previously recorded images and sound. Behaviors associated with birth included characteristic, prolonged licking of the vulva and surrounding area by the female in a sitting, upright posture and ultimately small cub screams followed by lateral recumbency of the mother for lactation. Newborn cubs were highly vocal and screamed loudly in response to any movement by the mother. Thus, the time of birth could be identified within minutes.
We weighed cubs opportunistically to the nearest 0.01 kg throughout the first 3 months of life with a final mass determined at 90 days. Adult females had been trained previously with voice commands to leave the den and go into their outside runs. The door between the inside den and outside run was closed as mothers exited the den, and cubs were removed, sexed, identified, and their masses were determined. Cubs were quickly returned to the den, the den door was opened such that the mother could return to comfort her cubs, and all people left the area. We did not measure birth masses because of concerns that disturbance at that time might cause cub abandonment. Newborn brown bear cubs weigh approximately 650 g (Farley and Robbins 1995; Quick 1969). Thus, although variation in size at birth could account for gram-sized differences in masses at den emergence, larger kilogram-sized differences must occur after birth and be caused by differences in milk intake.
During the latter years of the study, we also measured masses of cubs on April 1 when they began exiting the den and following their mother into the outside runs. Cubs had not consumed solid food by either 1 April or 90 days. Thus, growth rates of cubs during the study depended solely on milk intake. We measured and compared growth characteristics of 9 litters of twin cubs containing 11 males and 7 females and 2 litters of triplets containing 5 males and 1 female. Two litters of twins were produced by each of 3 females, 3 litters of twins were produced by 1 female, and 1 set of triplets was produced by each of 2 females.
Measuring maternal condition and size of wild cubs.—We do not know of any published data on wild brown bears that would permit a direct comparison to any of the values generated with the captive bears. Wild brown bear cubs and their mothers are not handled during hibernation in North America, and the same females are rarely captured at the beginning and end of hibernation. Thus, a direct comparison of results between captive and wild brown bears is not possible currently. However, limited data relating maternal condition during the preceding fall and cub size in March does exist for polar bears living in northeastern Manitoba, that is, Hudson Bay (Atkinson and Ramsay 1995). Polar bears evolved about 152,000 years ago from an ancestral brown bear (Lindqvist et al. 2010), and female polar bears hibernate when producing cubs (Ramsay and Stirling 1988). Thus, our use of polar bear data collected by Atkinson and Ramsay (1995) provides the opportunity to determine if the relationship between maternal condition and spring cub size is characteristic of north temperate bears in general.
Statistical analyses.—Birth date, cub growth, and mass relative to maternal condition were evaluated with regression analyses. Analysis of variance was used to test for differences in mean masses of male and female cubs (SAS Institute IncL 1998). Means are reported ±1 SD.
Adult females at the start of hibernation weighed from 123 to 249 kg and ranged in body fat content from 18% to 38%. Birth dates ranged from 3 to 20 January. The relationship 543 between maternal body fat content and birth date was curvilinear with birth date increasingly delayed below a body fat content of approximately 30% (Fig. 1). Because the structural size or lean body mass of individual bears ranged from 98 to 168 kg, the correlation between birth date and maternal mass was weaker (R2 = 0.47) than that between birth date and maternal body fat content (R2 = 0.74; Fig. 1). No female with a body fat content < 20% produced cubs even though we observed multiple matings during the preceding spring.
Cub mass of twins and triplets increased curvilinearly as lactation progressed (Fig. 2). Cub mass and litter mass of twins at either 90 days of age (Fig. 3) or on 1 April (Fig. 4) were highest when maternal body fat content was high at the start of hibernation. At 90 days, male cubs in sets of twins (10.2 ±1.2 kg) did not differ in mass from female cubs (9.7 ±1.4 kg; F1,17 = 0.72, P = 0.41). However, each unit increase in percent maternal body fat content at the start of hibernation increased cub mass of twins at 90 days by an average of 170 g/cub. Cub size within each litter was very similar (0.3 ±0.3 kg) except for a set of cubs born to the leanest female (Fig. 3). The difference in cub size within that litter at 90 days was 2.2 kg, or a 28% difference between littermates. Cub size at 90 days in 2 sets of triplets averaged 55% ±7% of twin cubs raised by mothers of the same body fat content (26.7% and 30.3%; Fig. 2). Total litter mass of the triplets at 90 days was 83% of twin litters.
The combination of an earlier birth and increased rate of mass gain in cubs produced by fatter females led to increasing differences in cub mass between litters as lactation progressed. Twin brown bear cubs nursing from the fattest females were as much as 80% larger when exiting the den as those nursing from the leanest females. Mass at den emergence of captive brown bear and wild polar bear cubs raised as twins increased from 330 to 360 g for each unit increase in percent maternal body fat content at the start of hibernation (Fig. 4).
Brown bears were able to vary both the birth date and growth rate of their cubs in response to maternal fat stores. Although birth dates of wild brown bears have not been measured, estimated birth dates of 149 litters of wild American black bears in Virginia ranged over a 39-day period 544 from 26 December to 4 February (Bridges et al. 2011). The extent of the variation in birth dates of black bears is similar to the range of estimated implantation dates (37 days) that occurred in European badgers (Meles meles), which also have delayed implantation (Woodroffe 1995). The fattest females of both the captive brown bears and wild badgers gave birth or implanted earliest, which allowed their offspring the most time to grow before leaving the den (bears) or experiencing low food availability during the summer (badgers—Woodroffe 1995). Larger bear cubs would likely be less susceptible to infanticide (Ben-David et al. 2004) and better able to follow their mother after den emergence (Amstrup et al. 2000). This is particularly important for polar bears, which inhabit constantly shifting and drifting sea ice and often must swim (Durner et al. 2011). Thus, delayed implantation provides the opportunity to optimize female fitness both at the time of breeding as well as many months later during implantation, gestation, birth, and early lactation (Woodroffe 1995; present study).
Bears are known for their prodigious appetites prior to hibernation. Daily intakes by brown bears can be as high as 34% ±6% of body mass for fruits and 15% of body mass for salmon (Hilderbrand et al. 1999b; Welch et al. 1997). The attractiveness of foods that can be used for fattening, including natural foods (e.g., fruits and salmon for brown bears or seals for polar bears) and improperly stored or discarded human or pet foods, garbage, or livestock carrion, can be appreciated because small differences in female body fat content lead to large differences in cub production, size, and potentially survival following den emergence (Derocher and Stirling 1996, 1998). Although there are selective pressures that limit fat accumulation in many prey species due largely to the increased risk of predation (Lima 1986), most adult bears probably do not experience an increased predation risk associated with fatness and, therefore, have almost no upper limit to fat accumulation. In many species, leptin, an adipocyte- secreted protein, plays an important role in regulating reproduction and lactation and could be particularly important in hibernators as a means to sense available fat stores and, thereby, modulate reproductive investment (Moschos et al. 2002).
The consequences of differing birth dates, neonatal nutrition, and subsequent cub size at den emergence may be multigen- erational because growth and fat storage characteristics of adult offspring can be influenced by both nutrition of their mother and reserves prior to parturition and their own nutrition after birth (Mortensen et al. 2010). Although the current study demonstrated that maternal fat stores effected cub birth date and growth during hibernation, the masses of wild polar bear yearlings following fat mothers (35–45% body fat) were one- third larger (approximately 40 kg) than those following lean mothers (10–20% body fat—K. D. Rode and E. V. Regehr, United States Fish and Wildlife Service, pers. comm.). This relationship between maternal fatness and offspring size that extends into the 2nd year of life may reflect maternal dominance, more efficient hunting strategies, use of more- productive environments, or other processes that boost both maternal and offspring nutrition. Such differences during growth of young bears may have long-term, population-level consequences in that most successful reproduction may be by energetically efficient, large males and fat females and their offspring (Dahle et al. 2006; Kovach and Powell 2003; McLellan2011).
Funding was provided the Interagency Grizzly Bear Committee, United States Fish and Wildlife Service, and the Raili Korkka Brown Bear and Nutritional Ecology Endowments at Washington State University. We thank J. P. Whiteman, K. D. Rode, R. A. Powell, and A. E. Derocher for thoughtful comments on the manuscript.