Male-biased sexual size dimorphism is typical of polygynous mammals, where the degree of dimorphism in body mass is related to male intrasexual competition and the degree of polygyny. However, the importance of body mass in monogamous mammals is largely unknown. We investigated the effect of body mass on life-history parameters and territory size in the red fox (Vulpes vulpes), a socially monogamous canid with slight sexual dimorphism. Increased body size in males appeared to confer an advantage in territory acquisition and defense contests because heavier males held larger territories and exerted a greater boundary pressure on smaller neighbors. Heavier male foxes invested more effort in searching for extrapair matings by moving over a wider area and farther from their territories, leading to greater reproductive success. Males that sired cubs outside their own social group appeared to be heavier than males that only sired cubs within their social group or that were cuckolded, but our results should be treated with caution because sample sizes were small. Territory size, boundary pressure, and paternity success were not related to age of males. In comparison, body mass of females was not related to territory size, probability of breeding, litter size, or cub mass. Only age affected probability of breeding in females: younger females reproduced significantly less than did older females, although we did not measure individual nutritional status. Thus, body mass had a significant effect on life-history traits and territory size in a socially monogamous species comparable to that reported in polygynous males, even in the absence of large size dimorphism.
Body mass is a key life-history component of many mammalian species. For example, increased body mass of males may confer an advantage in competition for resources such as mates or territories. Similarly, in females, greater body mass can be associated with larger litter size, greater neonate weights, and greater neonate survival (e.g., Festa-Bianchet et al. 1998; Moehlman and Hofer 1997). In polygynous species, male-biased sexual size dimorphism is typically a consequence of intrasexual competition between males for access to sexually receptive females (Andersson 1994; Kleiman 1977). Most mammals are polygynous, with moderate to small male-biased size dimorphism (Rails 1977); only approximately 3% of species are considered monogamous. Within these monogamous species sexual size dimorphism is low or absent (Kleiman 1977).
Genetic monogamy implies a single mate, generally low variance in reproductive success (but see, e.g., Thomas and Coulson 1988), and, consequently, limited opportunity for sexual selection. Hence, the evolutionary selective pressures on monogamous males are considered less pronounced than in polygynous males (Wiegmann and Nguyen 2006). However, even within monogamous species, males have to compete to acquire mates, territories, or both. Moreover, monogamous males should not forego extrapair reproductive opportunities where these occur (Trivers 1972), such that increased physical size also may be beneficial within such systems. Yet, although widespread in studies of polygynous species (e.g., Loison et.al. 1999; Weckerly 1998), studies examining the importance of body mass on territory size or reproductive success in monogamous species are largely absent.
Species in the family Canidae are considered predominantly monogamous (Kleiman 1977), although both social and genetic polygyny have been reported (Baker et al. 1998, 2004; Kitchen et al. 2006; Macdonald 1979; Zabel and Taggart 1989). The red fox (Vulpes vulpes) is the most widely distributed extant canid and occupies a diverse range of habitats (Baker and Harris 2004). Throughout this range, red foxes exhibit marked variation in territory size and group size (Cavallini 1996a), being found both as monogamous pairs and as small groups consisting of a dominant pair and subordinate animals; typically, but not exclusively, subordinates are nondispersing female offspring from previous years (Baker et al. 1998; Macdonald 1979; von Schantz 1981,1984a, 1984b). Subordinates were not thought to reproduce (Hartley et al. 1994; Macdonald 1979) although, in years of high resource availability, they may rear their own cubs successfully, indicating a switch from social monogamy to social polygyny (von Schantz 1984b; Zabel and Taggart 1989). Furthermore, male foxes may engage in extraterritorial movements during the mating period and genetic analyses have shown that these lead to successful extrapair paternities (Baker et al. 2004; White et al. 1996). Therefore, red foxes provide an excellent opportunity to study the effect of body mass in a monogamous species as, despite having only slight sexual dimorphism, there are clear indications that male intrasexual competition occurs, particularly where resource availability is high.
We used data from a long-term study on urban red foxes in Bristol (Baker et al. 1998, 2000, 2004; Baker and Harris 2004) to quantify the effects of body mass on individual life-history traits and territory size. In this population, males were on average 15.6% heavier than females, a degree of sexual size dimorphism typical of red foxes in Europe and Asia, where males are on average 17.4% heavier than females. In particular, we investigated the effect of body mass on territory size and reproductive success of male and female foxes. Given the high resource availability associated with urban environments, and given that food appears not to be limiting in our population (Ansell 2004), we hypothesized that territory size does not reflect resource availability and we therefore predicted that territory size would be independent of body mass of both males and females. We also hypothesized that body mass confers a competitive advantage to males; we predicted that heavier males would exercise greater boundary pressure on neighbors, and would move farther and faster, and that this advantage would lead to greater reproductive success, that is, heavier males would sire more litters. Similarly, for females we hypothesized that body mass is a predictor of reproductive success, and hence predicted that heavier females would have larger litters and heavier cubs.
Materials and Methods
Study site and capture methods.—During 1978–1989, the study site encompassed the entire city of Bristol, United Kingdom (144 km2), and was used to collect data on demographic processes. To focus on more detailed behavior, between 1990 and 1994, the study site was reduced to ∼1.5 km2 in the northwestern suburbs of Bristol. This area was then expanded to ∼15 km2 after a dramatic reduction in population density due to a sarcoptic mange (Sarcoptes scabiei) epizootic from 1994 to 1996 (Baker et al. 2000); the disease was then endemic from 1996 to 2004, and densities postmange (range 4.0–5.5 adults/km2) were significantly lower than premange (7.8–25.8 adults/km2—Soulsbury et al. 2007a). Foxes were captured by netting from den sites or in baited box traps placed in residential gardens (Baker et al. 2001). Animal capture and handling followed guidelines approved by the American Society of Mammalogists (Gannon et al. 2007). All individuals were weighed, sex was determined, age was estimated from tooth wear (Harris 1978), and individuals were then ear-tagged (Rototags; Dalton Supplies Ltd., Nettlebed, Henley-on-Thames, Oxfordshire, United Kingdom). Foxes were classified as cubs (<6 months), subadults (6–12 months), or adults (>12 months—Harris and Trewhella 1988). Date of birth of all animals was assumed to be 1 April (Harris and Smith 1987). Throughout the study, cadavers from tagged and untagged individuals were recovered. Cause of death, mass, nutritional status (only during 1988–1991), and, for females, the presence, size, and number of fetuses were recorded.
Radiotracking methods.—Full-grown individuals were fitted with commercial transmitters (Biotrack Ltd., Wareham, Dorset, United Kingdom) or transmitters manufactured at the University of Bristol. Radiocollars weighed approximately 120 g, that is, 1.6–2.5% of the mass of males and 1.7–2.5% of the mass of females, with battery life span lasting an average of 1.5 years (range 6 months–4years). Animals were radiotracked between 2000 and 0400 h Greenwich Mean Time on foot using a 3-element hand-held yagi antenna and CE12 receiver (Custom Electronics of Urbana Inc., Nokomis, Florida). Locations were taken every 5 min; foxes were assigned to a 25 × 25-m grid cell using a 1:1,250 scale map. Most fixes were taken from a distance of <50 m, and locational data could easily be related to specific gardens, and often to a particular part of a garden, by taking bearings from nearby roads. The high density of roads on the study area, combined with the fact that the majority of gardens were <625 m2 (often much less), meant that fixes were both quicker to obtain and far more accurate than is possible in other types of environments. Radiotracking was done on a seasonal basis, with tracking nights spread as evenly as possible throughout each 3-month season for each fox. Seasons were defined as: spring, March–May; summer, June–August; autumn, September–November; winter, December–February. Analyses were performed in TRACKER 1.1 (Camponotus AB, Solna, Sweden).
All foxes sighted while radiotracking were recorded; the location, time, and duration of observations, sex, status, and activities of the foxes seen were noted. Behaviors were defined as nonaggressive, aggressive, or unknown. Nonaggressive encounters included grooming, play, and close spatial associations in which none of the foxes involved exhibited aggression toward one another. Aggressive encounters included physical fighting, agonistic vocal behavior without physical contact, and spatial associations with some degree of aggression. Radio-collared females were located daily during the immediate postnatal period when activity by females is restricted to the breeding den site. This behavior and additional behaviors from males and females, such as babysitting and provisioning, were used to determine behavioral linkage to cubs.
Seasonal influence on body mass.—The mass of adult male and female foxes varies seasonally because of changes in the pattern of fat deposition and utilization (Saunders et al. 1993) but not all animals were captured in every season. Consequently, we determined the degree to which mass of adults at capture varied seasonally and calculated conversion factors to standardize body mass for those animals included in the analyses. A backward stepwise general linear model was carried out separately for males and females to determine whether season (fixed factor), age in months (covariate; log transformed for males, reciprocal transformed for females), and individual (random factor) significantly affected mass at capture for adult males and females captured in 2 or more seasons during 1990–2004. Pregnant females and foxes infected with sarcoptic mange were excluded from analyses. In this and all subsequent statistical analyses, we tested for and met any prerequisite test assumptions. Where analyses indicated that season significantly affected body mass of adults, the ratio of mass in spring, summer, and winter was calculated relative to mass in autumn; these ratios were then used to estimate the projected autumn mass for those adults captured in other seasons (hereafter standardized mass). Autumn was chosen as the reference because it was the season in which the majority of foxes were radiotracked.
Measurements of territory size.—In both the pre- and postmange periods, individual foxes were tracked until a clear asymptote was achieved in each season (after Harris et al. 1990). Typically this was achieved between 200 and 250 active fixes, but foxes were tracked for longer. The mean (± SE) number of active fixes collected per season was 370 ± 30 (males) and 321 ± 39 (females), representing 4–10 nights of tracking. Seasonal radiotracking indicated high concordance between territory boundaries of males and females (X̄ = 77.8% ± 21.8% SE); thus we were not misrepresenting territory size (Iossa 2005).
Territory size was defined as the 95% harmonic mean isopleth area in autumn (hereafter, male or female territory size), this being considered most representative of normal territory size because movement patterns of adults during spring and summer are heavily influenced by cub-rearing activities (Baker et al. 1998), and during winter males enlarge their home ranges while seeking extrapair copulations (hereafter, winter-range size—White et al. 1996). Despite not representing an actual territory expansion, winter-range size reflects the area searched for extrapair copulations. Territory size could not be compared directly to body mass, because the sarcoptic mange epizootic from 1994 to 1996 caused a sharp decline in population density and a concurrent increase in territory size (Baker et al. 2000; Soulsbury et al. 2007a). Because territory sizes in the pre- and postmange periods were markedly different, the mean size of all territories of males in each year (hereafter, annual male territory size) was included as a covariate in multiple linear regression analyses of the relationship between standardized mass (independent variable), age (months, independent variable), and male territory size (dependent variable) and male winter-range size (95% harmonic mean isopleth; dependent variable). Foxes were only included once in each model.
Territory sizes of females were similarly examined. Again, territories could not be compared pre- and postmange, so the mean of all territories of females in each year (hereafter, annual female territory size) was included as a covariate in multiple linear regression analyses of the relationship between mass (independent variable), age (months, independent variable), and female territory size (dependent variable).
Boundary pressure.—Because a small male surrounded by large individuals may potentially be expected to hold a small territory because of asymmetrical competitive pressures (e.g., Pétrie 1984), the influence of the body mass of neighboring territory holders on the size of an individual's territory was investigated using an index of relative boundary pressure (P), which was calculated by comparing the ratio of body masses between each focal male and its neighbors relative to the length of shared borders. It was important to include shared border length between foxes, because territories of males were bordered on average by 3 other males (range 2–6 males) with shared boundary lengths ranging from 240 to 1,140 m (X̄ = 565 m ± 59 SE), and so the relative pressure exerted (positive or negative) varied. We did not have data for all neighboring foxes in all circumstances; therefore, P was calculated for the total boundary length for which data on neighboring individuals were available. Only those individuals whose boundary pressure could be estimated for >50% of their territory perimeter or where >2 of their neighbors' ranges and body masses were known, or both, were included in the analyses. Specifically:
where Ri is the body mass ratio between the focal fox and his ith neighbor, bi is the length of the border shared with the ith neighbor, and T is the total boundary length with all n neighbors. The relationship between boundary pressure (dependent variable) and annual male territory size (covariate) was analyzed using multiple linear regression. However, it must be noted that the boundary pressure data points included in this regression were not completely independent because the pressure exerted by fox a upon fox b was dependent on the pressure exerted by b upon a. However, each total known boundary pressure on fox a was composed of the pressure exerted by fox b and several (in this study, 2–6) other neighbors. We carried out multiple linear regression between territory size (dependent variable), annual male territory size (covariate), and boundary pressure (independent variable). We also carried out a 2nd analysis, using the mean asymmetry in body mass of neighbors (dependent variable) without incorporating boundary length into the calculation.
Male movements in winter.—To further investigate male reproductive tactics in winter, we examined whether heavier males moved farther or faster per night. Both distance travelled and speed of movements were compared to body mass of individual males using Pearson's correlation coefficient. We also used a general linear model to examine the linear distance travelled from the male's territory (dependent variable), with male body mass (independent variable) and annual female territory size (covariate). Additionally, for 5 males with adequate spatial knowledge, we examined how many female territories they contacted.
Genetic analyses.—From 1992 onward, skin ejected during ear-tagging was stored at — 20°C in 100% ethanol in 2-ml plastic eppendorf tubes; 10 polymorphic microsatellites were amplified using the procedures described in Soulsbury et al. (2007b). Two problems were encountered when assigning parentage: significant levels of allelic dropout at 2 loci (Soulsbury et al. 2007b), and assigning parentage among closely related individuals (Olsen et al. 2001). Parentage assignments of offspring genotypes were initially analyzed with the likelihood-based method for paternity inference in CER VUS 2.0 (Marshall et al. 1998). However, the exclusionary power (Chakraborty et al. 1988) assumes that parents are not related; if this assumption is not met, the resolving power is overestimated (Olsen et al. 2001). To circumvent these problems, candidate parents were screened using CERVUS to remove all individuals without positive log-likelihood ratio scores to the offspring. These remaining candidate parents were then assigned parentage using a decision matrix (Fig. 1); no cubs were assigned parentage where heterozygote mismatches occurred. Allele frequency was calculated using only the 286 foxes included in the analyses. Using the previously defined behaviors, adults could be linked to cubs, but maternity or paternity was only assigned using genetics.
Male reproductive success.—Patterns of paternity were considered for 3 cohorts of cubs (2002–2004) for which we had sampled the majority of putative fathers in the population (59%). In total, 286 individuals were genotyped (60% of the population). The number of candidate parents varied each year: 120 (2002), 136 (2003), and 169 (2004). Many of these tagged individuals were caught once as cubs or subadults and subsequently disappeared; however, because the longest period between tagging and retrapping was 3.5 years, they could not be excluded as candidate parents because they might have bred undetected. Individual reproductive success was measured as the number of litters because it was not always possible to capture every cub and, postmange, there was no evidence of mixed-paternity litters (Iossa 2005), unlike premange (Baker et al. 2004). Even for litters where no father was sampled, analyses indicated 1 paternal genotype. The relationship between the number of litters sired and standardized mass of males and age of males was analyzed using a general linear model; 1 male was present in 2 cohorts, so for this individual mean number of litters sired was used. Finally, the relationship between cuckolding behavior (extrapair versus within-pair sire) and body mass of the sire was analyzed using a Mann-Whitney U-test with P-value from the exact distribution.
Female reproductive success.—Female reproductive success was assessed in 2 ways: 1st we examined the relationship between age (independent variable), body mass (independent variable), and number of fetuses found in cadavers (dependent variable) using a general linear model. Secondly, for 3 cohorts of cubs (2002-2004) with maternity assigned genetically, we assessed the relationship between maternal age (independent variable), maternal body mass (independent variable), and presence–absence of breeding using general linear mixed models with a binomial error structure, (glmmPQL function), and minimum litter size using general linear mixed models with a binomial error structure with a Gaussian error structure. Dam was included in the model as a random variable to account for multiple breeding events. Minimum litter size was the minimum number of cubs as assessed by behavioral observations and genetic parentage; in cases where we believe we captured all the cubs in a litter, observed litter size equaled genetic parentage.
The relationship between mass of cubs (dependent variable) and maternal age, maternal body mass, minimum litter size, and sex of cubs (independent variables) was examined using general linear mixed models with a Gaussian error structure. Individual was included as a random variable; because it was not possible to capture all cubs at the same age, age of cubs (days since 1 April) was included as a covariate to control for age at capture. Sample sizes vary between analyses because not all measures were available for all foxes in all years. During 1988–1991, cadavers of adult females were examined for the extent of kidney fat reserves, scored on a scale of 0 (no fat) to 5 (kidney completely covered in fat). Kidneys that had been damaged at death and females that may have died of an infection were excluded, because this may lower kidney fat reserves. However, we cannot exclude the possibility that some females that were killed, for example, by road traffic accidents, also may have been suffering some infection. To determine whether there was any relationship between body mass and nutritional status, we also considered the relationship between age (independent variable), body mass (independent variable), and their interaction to explain variance in kidney fat scores (dependent variable) using a general linear model. Finally, to assess whether nutritional status affected reproductive performance, we examined the relationship between kidney fat scores of females and the number of fetuses using a Spearman correlation coefficient test. This and all other analyses were performed using SPSS 12.0 for Windows (SPSS Inc., Chicago, Illinois) or R 2.5.0 (The R Foundation for Statistical Computing—www.r-project.org). All results are shown as mean ± SE unless stated otherwise.
Seasonal influence on body mass.—Thirty-four adult male and 27 female foxes were captured for ≥2 seasons during 1978–2004 (n = 69 and 58 captures, respectively; Table 1). Age did not significantly affect body mass of males (backward stepwise general linear model: F = 1.53, d.f. — 1, 31, P = 0.226) and was removed from the model. However, body mass of males was significantly affected by season (F = 3.08, d.f. = 3, 33, P = 0.041), being lower in summer than in autumn and winter, and by individual (F = 5.19, d.f. = 33, 33, P < 0.001). Body mass of females was not significantly affected by age (F= 3.96, d.f = 1, 27, P = 0.057) or season (F = 2.20, d.f = 3, 27, P = 0.111), but was significantly affected by individual (F = 5.26, d.f = 26, 27, P < 0.001). Consequently, body mass of males, but not females, was standardized to mass in autumn for all subsequent analyses.
|Male||Absolute mass (kg)||6.23 ± 0.2||6.08 ± 0.2||6.48 ± 0.3||6.70 ± 0.2|
|Waller–Duncan post hoc test||ab||b||a||a|
|X̄ (± SE) ratio||—||0.98 ± 0.02||—||—|
|Female||Absolute mass (kg)||5.64 ± 0.7||5.31 ± 0.6||5.63 ± 0.7||5.54 ± 0.7|
|Male||Absolute mass (kg)||6.23 ± 0.2||6.08 ± 0.2||6.48 ± 0.3||6.70 ± 0.2|
|Waller–Duncan post hoc test||ab||b||a||a|
|X̄ (± SE) ratio||—||0.98 ± 0.02||—||—|
|Female||Absolute mass (kg)||5.64 ± 0.7||5.31 ± 0.6||5.63 ± 0.7||5.54 ± 0.7|
Influence of body mass on male territory size and life-history tactics.—Data on territory size were available for 13 males. Male territory size was significantly related to standardized body mass (t = 2.51, d.f. = 9, P = 0.033; Fig. 2) and mean annual territory size (t = 2.64, d.f. = 9, P = 0.027), but not age (t = 0.07, d.f = 10, P = 0.94; multiple linear regression: R2= 0.61, F = 4.76, d.f = 3, 9, P = 0.030). There was a significant negative relationship between male territory size and boundary pressure (t = −5.18, d.f. = 4, P = 0.007) and annual male territory size (t = 5.75, d.f = 4, P = 0.005; R2 = 0.92, F = 21.53, d.f. = 2, 4, P = 0.007), that is, smaller males had greater boundary pressure from surrounding males, and so had smaller territories. Similarly, both the mean asymmetry in body mass (t = −6.36, d.f = 4, P = 0.003) and annual male territory size (t = 6.33, d.f. = 4, P = 0.003) explained male territory size (R2 = 0.94, F = 31.91, d.f = 2, 4, P = 0.002).
Winter-range sizes (n = 9) were positively related to standardized body mass (t = 3.15, d.f. = 5, P = 0.020) and mean annual territory size (t = 2.85, d.f. = 5, P = 0.036), but not age (t = 0.90, d.f. = 5, P = 0.402; multiple linear regression: R2 = 0.76, F = 5.41, d.f. = 3, 5, P = 0.050; Fig. 3), indicating that heavier males explored greater areas while making extraterritorial movements. Males contacted a median of 2.0 (interquartile range [IQR] = 1.0–2.5) territories but heavier males did not appear to contact more territories (rp = 0.69, P = 0.197), although sample sizes were small. Most movements were into the neighboring territory (7 of 9 movements, distance X̄ = 0.88 ± 0.22 territories), but heavier males moved farther away from their territory boundaries (n = 9, rp = 0.86, P = 0.003) and moved greater distances per night during reproductive movements (range 2.9–11.2 km, rp = 0.67, P = 0.047). Heavier males did not move any faster, although the speed of movement was approaching significance (rp = 0.66, P = 0.055).
In total, 117 putative offspring were sampled across 25 group-years; 72 (61%) were captured as cubs, 29 (25%) as subadults, and 16 (14%) as adults. Paternity was assigned for 38 of these offspring; both paternity and maternity were assigned for 34 (49%), and paternity only was assigned for 4 (6%). Of 15 males with adequate knowledge of presence– absence of breeding, we recorded 14 litters sired by 9 males (median n sired = 1.0, IQR = 1.0–3.0) and 6 nonbreeders. The number of litters produced by individual males was significantly correlated with standardized mass in autumn (t = 2.24, P = 0.047; Fig. 4), but not age (t = 1.47, P = 0.169) or the interaction between age and mass (t = −1.38, P = 0.195; R2 = 0.67, F = 7.73, d.f. = 3, 11, P = 0.005). Six litters were from extrapair matings involving 4 individual males. The body mass of males that sired cubs in groups other than the one in which they were resident (6.9 ± 0.2 kg; n = 4) was close to being significantly greater than that of males that reproduced only in their own group or that were cuckolded (6.0 ± 0.3 kg; n = 11, U = 8.5, exact P = 0.060) but sample sizes were small and therefore this result should be treated with caution. For 3 males we knew the body mass of the cuckolded male; in all cases the cuckolded male was smaller.
Influence of body mass on female territory size and life-history tactics.—Male territories were on average 14.3% ± 29.4% larger than the territories of the females with which they associated (matched-pairs t-test: t = 1.88, P = 0.049, n = 9). There was a high degree of congruence between territories of the dominant pair (n = 9, 86.5% ± 6.6%), which was significantly different from the overlap between neighboring dyads (F = 265.4, d.f. = 3,40, P < 0.001). Dominant pairs overlapped significantly more than male–male neighbors (Tukey: P < 0.001, 6.0% ± 5.7%), female–female neighbors (Tukey: P < 0.001, 10.9% ± 8.1%), and male–female neighbors (Tukey: P < 0.001, 10.7% ± 6.9%). Overlap between different categories of neighbors was not significantly different (male–male versus female–female, Tukey: P = 0.48; male–male versus male–female, Tukey: P = 0.57; female– female versus male–female, Tukey: P = 1.00). Unlike male territory size, female territory size (n = 13) was not significantly related to body mass (t = 1.48, d.f. = 9, P = 0.17) and, like male territory size, female territory size was not related to age (t = 1.03, d.f. = 9, P = 0.33) and was significantly related to mean annual territory size (t = 4.18, d.f. = 9, P = 0.004; multiple linear regression: R2 = 0.71, F = 7.26, d.f. = 3, 9, P = 0.009). Where we had data for male– female pairs, female territory size was significantly related to the size of the range of the male with which they were associated (linear regression: R2 = 0.88, F = 53.28, d.f. = 1,7, P < 0.001, n = 9; both male and female territory size log transformed; Fig. 5).
For 15 pregnant females, mean number of fetuses was 4.2 ± 1.6. There was no relationship between number of fetuses and age (t = −0.48, P = 0.64), body mass (t = −1.48, P = 0.16), or their interaction (t = 0.62, P = 0.55; R2 = 0.09, F = 1.52, d.f. = 2, 12, P = 0.26). Of 117 putative offspring sampled, maternity was assigned for a total of 65 individuals; both paternity and maternity were assigned for 34 (49%), and maternity only was assigned for 31 (45%). Body mass (t = −1.71, P = 0.162) and the interaction between both body mass and age (t = 1.64, P = 0.176) did not influence the presence or absence of breeding. In contrast, age significantly affected the presence or absence of breeding (t = 10.70, P < 0.001), with fewer yearling females breeding (55%, n = 11) than older females (100%, n = 11).
Neither maternal mass (t = 3.32, P = 0.186), maternal age (t = 3.26, P = 0.189), nor their interaction (t = −3.31, P = 0.187) influenced minimum litter size (observed litter size X̄ = 3.6 ± 0.3, litter size determined genetically X̄ = 3.0 ± 0.2). Similarly, maternal mass (t = 0.72, P = 0.483), maternal age (t = −0.85, P = 0.407), and litter size (t = –0.56, P = 0.581) were not found to influence mass of cubs significantly; only sex of cubs (t = 3.54, P = 0.002) was significant. Body mass may not be expected to influence female reproductive success if there is no link to nutritional status; neither age (t = 1.19, P = 0.24), body mass (t = 1.81, P = 0.07), or their interaction (t = –1.15, P = 0.251; n = 126 cadavers, R2 = 0.04, F = 1.45, d.f. = 3, 122, P = 0.232) significantly affected kidney fat scores. Nor was there any relationship between kidney fat scores and the number of fetuses (r = −0.29, P = 0.266). Thus, body mass was not a reliable predictor of nutritional condition.
Body mass and life-history tactics of males.—Greater body mass appeared to confer significant advantages to male foxes in the population under study. Contrary to our predictions, body mass of males was positively related to territory size and boundary pressure on neighboring foxes. Consistent with our predictions, body mass of males was positively related to the effort put into achieving copulations (area explored, distance moved, and, partially, to speed of movement) and to reproductive success (number of litters sired) within and outside the resident social group. Size-mediated reproductive tactics can often be found within the same population (Bachman and Widemo 1999; Lee 2005). The fact that body mass predicted copulation efforts of males is not a common finding, but movement parameters are not generally considered; therefore, we suggest that these parameters be explored more widely. In this population, one-half of the social groups were monogamous, although reproductive strategies ranged from behavioral and genetic monogamy to behavioral and genetic polygyny (Iossa 2005). Larger or heavier males in many species are more likely to seek extraterritorial matings (Bachman and Widemo 1999; Cavallini 1998; Lee 2005; Sandell 1986, 1989) and extrapair copulations appear to be common in many socially monogamous mammals (e.g., Goossens et al. 1998; Reichard 1995; Schülke et al. 2004). In a diverse range of polygynous species, body mass confers an advantage in male–male competitive interactions and larger males achieve greater reproductive success (e.g., marsupials [Clinchy et al. 2004], perissodactyls [Gamier et al. 2001], pinnipeds [Hoelzel et al. 1999], carnivores [Kovach and Powell 2003], artiodactyls [McElligott et al. 2001], and primates [Wickings et al. 1993]). In foxes, increased levels of bite wounding and mortality during the mating period (Harris and Smith 1987; White and Harris 1994) suggest that males compete directly for access to females. Examination of our data indicated that heavier males gained a competitive advantage in these interactions and fathered more litters. This is the 1st study to report this relationship in a socially monogamous species, but given the widespread evidence of extrapair copulations in many monogamous species, it is likely that body mass may determine life-history tactics in other species and may explain why some male-biased sexual dimorphism occurs even in monogamous species.
Asymmetries in competitive pressure (e.g., Petrie 1984) may allow larger individuals to defend larger territories that provide them with access to a greater amount, or more consistent supply, of resources. Such increased resource availability may, in turn, be expected to have concomitant benefits such as increased litter size or offspring survival (Tannerfeldt and Angerbjörn 1998). In our study, heavy male red foxes held larger territories but, because female territories were congruent with male territories, this did not increase access to females. The amount of resources available on each territory was substantially greater than the amount required to raise 1 litter of cubs (0.6–1.3% of total resources available—Ansell 2004). Instead, the benefits of larger territories were related to lower intruder pressure. Boundary pressure was lower for males holding larger territories, suggesting that there were fewer intrusions from other males and, thus, a smaller chance of getting cuckolded. Unfortunately, we knew the identity of the cuckolded males only in 3 cases so we cannot test this hypothesis. One potential advantage of larger territories is that they may allow portions to be split off by nondispersing offspring (Lindström 1988), thereby allowing these individuals to forego the costs of dispersal (Soulsbury et al. 2008a). Territory splitting has been recorded in several mammal species (Boutin et al. 1993; Meier et al. 1995), including in this fox population pre–sarcoptic mange (Baker et al. 1998) and in 4 cases postmange. However, the costs and benefits of maintaining nonminimal territories warrant future examination.
Body mass and life-history tactics of females.—According to our predictions, territory size of females was not determined by body mass but by the size of the territory of the male with whom they consorted. This suggests that male–male interactions were most important in determining spatial patterns within the population. Red foxes are aggressive to intruding individuals from neighboring groups (White and Harris 1994), although it is not clear whether patterns of aggression are directed solely or predominantly to same-sex individuals, as recorded in other carnivores (Boydston et al. 2001; McComb et al. 1994; Sillero-Zubiri and Macdonald 1998). The fact that male–male competition was evident in patterns of boundary pressure, and that males had slightly larger territories than females, suggests that males may primarily determine territory size in this species.
Body mass has been linked to reproductive output in females, because nutritional condition is thought to determine the likelihood of breeding or the number of offspring produced, or both (e.g., Noyes et al. 2002). Yet larger females may only have an advantage over smaller ones when or where resources are scarce (Festa-Bianchet et al. 1998). Nonbreeding female coyotes (Canis latrans) are lighter than breeding females (Sacks 2005) and female red foxes with larger than average litters are heavier than those with smaller than average litters (Cavallini and Santini 1996). In contrast, we did not find any effect of body mass on minimum litter size or number of fetuses but, like Cavallini and Santini (1996), we found no relationship between nutritional status and number of fetuses. The reasons for the similarities and differences in the 2 studies are unclear.
It is thought that reproductive performance in red foxes is linked to resource abundance (Englund 1970; Lindström 1988), with the greatest effects being on 1st-year females (Englund 1970). For example, in most populations, a higher proportion of lst-year females do not breed relative to other age-classes (Allen 1984; Harris 1979; Mcllroy et al. 2001; Zapata et al. 1998), and this pattern also was evident in this study. However, food is not a limiting resource in the Bristol population (Ansell 2004), potentially explaining why body mass was not correlated with reproductive parameters. Furthermore, body mass of adults is determined primarily by environmental conditions during growth (Soulsbury et al. 2008b), and therefore is not a reliable proxy for body condition per se (Cavallini 1996b). Indeed, we found no relationship between body mass and kidney fat, suggesting that this is the case. It is possible that lst-year females that did not reproduce were in poorer body condition, even though their body mass was not significantly different from females that bred; this pattern may be even more complex, because a significantly greater proportion of dispersing 1st-year females did not reproduce compared to philopatric females (Soulsbury et al. 2008a). Thus, a number of factors linked to dispersal may cause this disparity, including a short-term reduction in accessing resources or stress (Johnson 1986), both of which may lead to abortion (Hartley et al. 1994).
In conclusion, body mass of males was positively related to a number of life-history traits and other parameters likely to affect individual reproductive potential, such as territory size, boundary pressure exerted on neighboring territories, and the number of litters sired both within and outside their resident group. Despite this, male-biased sexual size dimorphism was low, males being 15.6% heavier than females on average. Although variables other than sexual selection can affect size dimorphism (e.g., Clutton-Brock and Harvey 1978), the implication from our study is that body mass may mediate male reproductive strategies in other socially monogamous species. In contrast, life-history traits of females were not significantly related to body mass, suggesting that other factors are likely to affect female reproductive success; these should be the focus of future studies.
We thank the Natural Environment Research Council (CDS), Rotary Foundation of Rotary International, Newby Trust Ltd., and the Sir Richard Stapley Educational Trust (GI), the International Fund for Animal Welfare (PJB), and The Dulverton Trust (SH) for financial support, and Keith Edwards and Jane Coghill of the University of Bristol Transcriptomics Unit for help in analyzing the genetic data.