OUP user menu

Inconsistent association of male body mass with breeding success in captive white-tailed deer

Phillip D. Jones, Bronson K. Strickland, Stephen Demarais, Randy W. DeYoung
DOI: http://dx.doi.org/10.1644/10-MAMM-A-294.1 527-533 First published online: 9 June 2011


Breeding success among male white-tailed deer (Odocoileus virginianus) has been correlated with numerous physical and physiological variables. We investigated the effect of relative body mass on breeding success in captive male deer from 21 breeding trials. Contrary to our prediction, larger males on average had identical levels of breeding success (59% of breeding opportunities) in the 1st and 2nd halves of the breeding season. The variability of breeding success increased with greater relative mass; thus deer with substantial mass disadvantages had only limited success, and males weighing ≥92% of the pen average had success ranging from 0% to 100% of potential breeding opportunities. Although constant proximity limited males primarily to a strategy of direct confrontation, some subordinate males evidently mated opportunistically when >1 female was receptive. We conclude that ≥1 uncontrolled variable, possibly related to aggressiveness or life-history strategy, remained influential and limited the ability of some males to compete successfully in spite of greater relative body mass.

Key words
  • animal personality
  • body mass
  • breeding success
  • Odocoileus virginianus
  • quantile regression
  • white-tailed deer

Breeding success of male cervids has been associated with age (Ditchkoff et al. 2001b; Mysterud et al. 2003), antler development (Hirth 1977; Kruuk et al. 2002; Severinghaus and Cheatum 1956), body mass (Clutton-Brock et al. 1982; McElligott et al. 2001), nutritional status (Ditchkoff et al. 2001a), and physical condition (Mulvey and Aho 1993). Age, antler characteristics, and body mass are correlated, with the largest antlers grown by prime-aged males in good condition (Anderson and Medin 1969; Bowyer 1986; McCorquodale et al. 1989; Stewart et al. 2000). These correlations among physical variables render it difficult to distinguish the influence of each variable on breeding success.

The mating system of white-tailed deer (Odocoileus virginianus; hereafter, deer) is polygynous, where males form a temporary tending bond with individual estrous females (Hirth 1977). Previously, a small number of mature males were assumed to dominate younger males and sire most offspring (Hirth 1977; Marchinton and Hirth 1984; McCullough 1979; Miller and Ozoga 1997). However, recent assessments of paternity patterns in free-ranging deer populations have revealed that males ≤2.5 years old can be responsible for up to one-third of reproduction (DeYoung et al. 2009; Sorin 2004). Alternative strategies to contest competition have been suggested to explain much of the breeding success of younger males (DeYoung et al. 2009; Sorin 2004).

By controlling the range of alternative breeding strategies and facilitating determination of breeding success, captive-breeding experiments can elucidate factors responsible for breeding success and minimize the limitations of observation-based field studies. For example, DeYoung et al. (2002, 2006) initially used captive-breeding studies to examine the effect of age on male breeding success substantiated in later field studies (DeYoung et al. 2009). Recent studies of ungulate mating indicate that the association of age with breeding success is mostly an artifact of the correlation of age with body mass, antler size, and other physical and behavioral characteristics (Coltman et al. 2002; Johnson et al. 2007; McElligott et al. 2001; Pelletier and Festa-Bianchet 2006; Preston et al. 2001). Differences among individuals in experience, aggression, or other behavioral attributes that might influence reproductive success have long been suspected (Clutton-Brock et al. 1982; Kruuk et al. 2002), but focused research in this area is a relatively recent development (Dingemanse et al. 2009; Réale et al. 2009). By controlling experimentally for physical correlates of age, we can quantify the effects of physical characters on breeding success and reveal the magnitude of unmeasured or intangible variables, such as behavioral traits.

DeYoung et al. (2006) reported that breeding success appeared to be related to body mass; however, no statistical comparisons were possible due to low sample size. Since that report we have conducted 15 additional breeding trials that we combined with the previous data set to allow statistical evaluation of the effect of body mass on breeding success. Because we maintained all deer on the same pelleted ration, examined deer for health status, and removed antlers prior to introducing males into breeding pens, the potential influences of nutrition, health, and antler characteristics were minimized. We predicted that males with greater body mass relative to their competitors would obtain greater breeding success. In addition, if larger males are more actively involved in breeding, they may suffer reduced breeding success as the rut progresses due to declining condition or mass loss (Mysterud et al. 2004; Ozoga and Verme 1985). We therefore also predicted that larger males would have greater success in the 1st half of the breeding season than in the 2nd half.

Materials and Methods

Deer handling and sample collection.—We conducted this study at the Mississippi State University Rusty Dawkins Memorial Deer Unit in Starkville, Mississippi. Details of procedures for deer handling and paternity assignment used in the 1999–2001 trials are found in DeYoung et al. (2006). Following are procedures for handling and paternity assignment during the 2006–2009 breeding trials. We used both wild-caught and captive-bred females; although some males were wild-caught as fawns, all males were captive-raised. We housed 10–24 deer in breeding pens of 1.0–1.3 ha. All pens had natural ground cover and contained water and 2 feeders supplied with ad libitum 20% protein deer pellets (Purina AntlerMax Professional High Energy Breeder 59UB; Purina, St. Louis, Missouri). We seeded additional forages within the pens, including Durana Clover (Trifolium repens; Pennington Seed Co., Madison, Georgia) and Max-Q Fescue (Festuca arundinacea shreb; Pennington Seed Co.); volunteer grasses and forbs also were available.

Males and females were housed separately outside of the breeding season, with females housed at Mississippi State University and males at satellite facilities. Breeding males were moved to breeding pens in early October, coinciding with antler hardening. Males were selected without regard to previous housing; thus some males were exposed to another breeding male before antler removal. For each pen we selected 2–4 males that would avoid potential father–daughter pairings. Breeding activity typically began in mid-November to early December; thus males had 6–8 weeks to develop dominance relationships after antler removal. Breeding activity ended by late February, and males were removed from the breeding pens by 15 March.

We sedated selected breeders with telazol (4.4 mg/kg) and xylazine (2.2 mg/kg) in combination (Kreeger 1996) using a Pneu-Dart projection system (Pneu-Dart, Inc., Williamsport, Pennsylvania). We weighed each deer to the nearest 0.45 kg, removed antlers 2.5 cm above the pedicel, and inspected each male for injury or illness that might influence interactions with other males. We administered appropriate amounts of the antibiotic Nuflor (Schuring-Plough Animal Health Corp., Summit, New Jersey), the endectocide Noromectin (Norbrook Laboratories, Ltd., Down, Northern Ireland, United Kingdom), the clostridial vaccine Vision 7 with SPUR (Ivesco LLC, Iowa Falls, Iowa), and the leptospirosis vaccine Leptoferm-5 (Pfizer, Inc., New York, New York). We then moved the -deer into the selected breeding pen and reversed anesthesia with intravenously injected yohimbine (0.125 mg/kg—Kreeger 1996) or tolazoline (4.0 mg/kg—Miller et al. 2004).

For DNA analysis we obtained a blood sample from all adult deer by venipuncture and stored samples at 4°C in vacuum tubes containing ethylenediaminetetraacetic acid (Vacutainer; Becton-Dickson and Company, Franklin Lakes, New Jersey). We manually restrained offspring within 1– 3 days postpartum and obtained DNA samples through ear notches that we preserved in 70%ethanol. DeYoung et al. (2003) evaluated a panel of 19 microsatellite markers for white-tailed deer from 13 geographically distinct populations. Although 2–9 loci in each population deviated from Hardy– Weinberg equilibrium, removal of those loci did not reduce overall exclusion probabilities in any population to<99 %; only 1 locus exhibited significant linkage disequilibrium (DeYoung et al. 2003). Seventeen markers from this panel were selected subsequently for parentage assignments in the 1999–2001 trials (DeYoung et al. 2006).

For assigning parentage in the 2006–2009 trials we used 14 microsatellite markers, 11 of which were selected from the original panel. Deer from the 2006–2009 breeding trials represented 3 geographically distinct regions, and we bred deer only within populations. Therefore, we calculated summary statistics for each locus by population in Cervus 3.0 (Kalinowski et al. 2007), including mean allelic diversity, observed and expected heterozygosity, polymorphic information content, and a nonexclusion probability for candidate parent pairs (Table 1). We determined that either 1 (2 populations) or 2 (1 population) loci deviated from Hardy– Weinberg equilibrium; however, overall exclusion probabilities for candidate parent pairs were >99.9%. The likelihood of correct parentage assignment probably was enhanced by the limited number of potential sires and dams in each pen and a DNA profile obtained for all potential parents.

View this table:
Table 1

Locus-specific summary statistics for 14 microsatellite DNA loci amplified for 661 individual white-tailed deer from 3 populations representing different soil resource areas of Mississippi.

LocusAllelic richness (SD)Expected (SD)Observed (SD)PICa(SD)NE-PPb(5D)HWEc
BL254.3 (0.6)0.63 (0.06)0.61 (0.06)0.58 (0.07)0.43 (0.06)0
BM643814.0 (1.7)0.88 (0.01)0.77 (0.12)0.86(0.02)0.10 (0.02)1
BM650611.3 (0.6)0.87 (0.02)0.76 (0.10)0.85 (0.03)0.11 (0.03)1
Cervidl13.7 (1.2)0.87 (0.02)0.86 (0.03)0.85 (0.02)0.10 (0.02)0
D10.7 (0.6)0.83 (0.03)0.78 (0.03)0.80(0.03)0.16 (0.04)0
INRA0116.7 (0.6)0.53 (0.11)0.50 (0.10)0.48 (0.10)0.54 (0.10)0
5.7 (0.6)0.30 (0.18)0.27 (0.14)0.28 (0.16)0.72(0.15)0
L10.3 (1.2)0.76 (0.04)0.74 (0.08)0.73 (0.04)0.23 (0.04)0
17.7(1.5)0.90 (0.01)0.89 (0.02)0.89 (0.01)0.06 (0.01)0
O7.3 (0.6)0.61 (0.06)0.59 (0.03)0.55 (0.07)0.47 (0.09)0
OAR14.0(1.7)0.87 (0.01)0.86 (0.04)0.85 (0.01)0.11 (0.01)0
P14.7 (0.6)0.88 (0.01)0.81 (0.04)0.86 (0.01)0.09 (0.01)2
Q19.7 (0.6)0.88 (0.04)0.87 (0.05)0.87 (0.05)0.09(0.05)0
S24.0 (3.5)0.90 (0.01)0.89(0.02)0.89(0.01)0.06 (0.01)0
  • a Polymorphic information content.

  • b Nonexclusion probability of a candidate parent pair.

  • c Number of populations in which locus deviates from Hardy—Weinberg equilibrium.

Of the 27 males studied from 2006 to 2009, 14 were captured from the wild as fawns and presumably were unrelated. Of the 13 pen-conceived males 2 half-siblings and 2 full siblings were assigned subsequently to the same breeding pen. However, because they shared no common alleles at either 6 (half-siblings) or 4 (siblings) loci, we had no difficulty distinguishing between them as potential sires.

Parentage was assigned using DNA samples by DNA Solutions (Oklahoma City, Oklahoma), using a proprietary, nonstatistical, custom-structured query language database, DASM. In the pairwise allele comparison the parentage assignments were made when we were able to exclude all but 1 sire and 1 dam based upon a shared allele from each parent at all loci tested (B. G. Cassidy, DNA Solutions, pers. comm.). All procedures followed guidelines established by the American Society of Mammalogists (Gannon et al. 2007) and were approved by the Mississippi State University Institutional Animal Care and Use Committee (protocols 98-033, 04-068, and 07-036).

Statistical analysis.—We used data from 17 trials with complete fawn birth date records to evaluate whether body mass affected the order of matings. To reconstruct the timing of breeding efforts we assumed a uniform gestation of 200 days and backdated from fawn birth date to estimate conception date (DeYoung et al. 2002). We used the results to estimate when females were most likely in estrus and thus when males would have mating opportunities. We used a 1-tailed paired t-test to determine if breeding success of the heaviest male in each pen was greater in the 1st half of the breeding season relative to the 2nd half, with breeding season determined individually for each pen. We examined a frequency histogram of differences between paired observations and determined that it met the assumption of normality. Because dominance is used to establish access to receptive females, we considered the breeding season half over when half the females were bred. If the pen contained an odd number of females, we assigned the middle female to the 1st or 2nd half of the breeding season based on whether she was bred before or after the midpoint of all breeding. If the female's litter was sired by 2 males, we assigned 0.5 successful breeding attempts to each male, regardless of the number of fawns produced. DeYoung et al. (2006) reported the incidence of multiple paternity in compound litters for the 1999–2001 trials; in this paper we report multiple paternity results from the 2006–2009 trials.

We used the complete data set of 21 breeding trials to analyze effects of body mass on breeding success, including 48 breeding season observations from 42 different males. Six males were used twice: 4 were paired with different males in subsequent years, and 2 were paired together twice, at 1 year and again at 3 years of age. We estimated the effects of body mass on male breeding success using quantile regression (Koenker and Bassett 1978). Quantile regression is an alternative to ordinary least-squares regression for use with data sets that exhibit heterogeneity of variance, a common situation in ecology due to unmeasured or unknown limiting factors (Cade and Noon 2003). Because unequal variance indicates there is more than 1 rate of change describing the relationship between dependent and independent variables, the application of a single mean response fitted across the entire range of the independent variable(s) (as in ordinary least-squares regression) could mischaracterize the relationship in parts of the response distribution. Furthermore, although heteroscedasticity can invalidate ordinary least-squares results, quantile regression does not require adherence to a particular error distribution (Cade and Noon 2003). Quantile regression calculates separate slopes for the response variable within investigator-defined quantiles τ (proportions ranging between 0 and 1) of the response distribution, more thoroughly describing the relationship between independent and response variables. For each specified τ the algorithm finds the function for whichτ of the data points fall at or below the value predicted. Because we had n = 48 samples, the highest quantile we could estimate conservatively was τ = 0.75 (Scharf et al. 1998). We were particularly interested in this highest quantile because it describes a function nearest the upper edge of the response distribution, which is presumably least influenced by limiting factors other than the variable of interest (relative body mass). Because we also were interested in quantifying the influence of these other potential limiting factors, we estimated quantiles from τ = 0.25-0.75 at 0.05 intervals. The quantile τ = 0.50 describes the rate of change of the median of the response variable and is similar to ordinary least-squares regression, and the lower quantiles extend information about the variance of the data set.

Within each breeding pen we calculated a breeding index (dependent variable) for each male by dividing the proportion of observed : expected breeding opportunities into the sum of proportions for that pen to standardize reproductive success among trials with varying numbers of males. The potential breeding index therefore ranged from 0 to 1. As in the previous test, we assigned 0.5 successful breeding attempts to each male in the case of shared paternity. We calculated relative body mass of each male by dividing individual body mass by the average male body mass within each pen; thus values >1.0 indicate males of above-average mass. We regressed breeding index against relative body mass (independent variable) using the QUANTREG procedure in SAS 9.2 (SAS Institute, Cary, North Carolina). We report regression coefficients and 90% confidence intervals (C/s) for each quantile.


Breeding trials produced 358 fawns from 206 females, and we successfully assigned parentage to all fawns. Eighteen of 124 compound litters (14.5%) in the 2006–2009 trials had >1 sire. In the 3 pens from this period with >2 males 2 of the 12 compound litters were jointly sired by smaller males only.

Average breeding success of the largest male from each breeding trial was identical between the 1st (X = 58.9% ± 1.8% SE) and 2nd (X̄58.9%± 1.8%SE) halves of the breeding season (t16 = –0.004, P = 0.498). Although 4 cases occurred where breeding success of the larger male was reduced by 33–75% between season halves, this was counterbalanced by 2 males that were relatively unsuccessful in the early season (0% and 25%) but substantially improved their breeding success in the late season (64% and 92%, respectively).

Relative mass ranged from 0.60 to 1.27, and breeding index from 0.0 to 1.0. The response of breeding index to change in relative mass did not differ from 0 for τ = 0.25, but relative mass was a significant predictor of breeding index beginning at τ = 0.30 (Fig. 1). Slope of the response generally increased to τ = 0.60 and formed a plateau to τ = 0.70. The slope of the line described by τ = 0.75 was 1.25 times greater than at τ = 0.50 and 1.68 times greater than at τ = 0.25, indicating that the maximum function was more responsive to changes in relative mass than were the median or minimum functions. The increasing slopes of the quantile lines confirmed that greater relative mass was associated with greater variability in breeding index (Fig. 2). The relationship between relative body mass and breeding index was such that low relative mass was associated with low breeding success. However, greater relative mass was associated with an increasingly broad range of success, and maximum dispersion of breeding index values occurred when relative mass reached approximately 0.92. Thus, the effect of relative mass seemed to depend on an interaction with some other limiting factor(s). The calculated response of breeding success for τ = 0.75 differed by 0.76 over the range of relative mass. Thus, at the τ expected to yield the clearest indication of the effect of body mass, response to increasing relative mass covered most of the possible range of breeding success.

Fig. 1

Parameter estimates βP(τ ) (upper) and β1(τ ) (lower), with 90% C1s, of a breeding index (range 0–1) versus relative mass (body mass as a percentage of the pen mean) of 48 captive male white-tailed deer from 21 breeding enclosures in Mississippi, 1999–2008. The regression coefficient (β1(τ)) at a given quantile indicates the effect of a unit change in relative mass on the breeding index.

Fig. 2

Breeding index (range 0–1) versus relative mass (body mass as a percentage of the pen mean) of 48 captive male white-tailed deer from 21 breeding enclosures in Mississippi, 1999–2008. Quantiles shown exhibit the range of slopes attributed to quantile functions tested and demonstrate the increasing variability in breeding success associated with greater relative mass.


Our results confirm that relative body mass was a significant factor determining breeding success of penned male white-tailed deer limited primarily to the strategy of direct contest competition. Low relative body mass clearly placed substantial limitations on breeding success. However, relative body mass alone did not consistently ensure greater breeding success, but it was associated with greater variability of success. This increasing variance of breeding success likely was due to interaction(s) of body mass with some unmeasured, uncontrolled variable(s). Our results are not conclusive in the identification of these potential variables, but these variables probably include some behavioral, personality, or other similar traits. The lack of significant relationship between relative mass and breeding index at the lowest quantile demonstrates the strength of these unmeasured variables. We expect the response in our highest quantile (τ = 0.75) to be most representative of the effect of body mass unconstrained by other limiting factors (Cade et al. 1999). Our sample size did not allow estimation at higher quantiles that might have yielded greater slopes, and additional breeding trials would help test this supposition. However, because any measurement of breeding success will have a maximum value, higher quantiles might have their slopes adjusted downward to accommodate the flat-topped nature of the response distribution Therefore, although the slope associated with τ = 0.75 might underestimate the true effect of increasing relative body mass, use of greater quantiles might not improve our understanding of the relationship between body mass and breeding success. We therefore encourage exploration of potential limiting factors to enable multivariate modeling that will obviate the need for modeling with quantile regression.

The apparent breeding success of subordinate males in wild populations might be explained by the use of an array of alternative mating tactics. Subordinate males of nearly equal mass with a dominant male might wait for the dominant male to lose sufficient body condition to enable a successful challenge (endurance rivalry), thus gaining access to a greater percentage of estrous females at some point in the season. As an example, in a pen with two 3-year-old males of similar body mass (71.4 and 69.1 kg), the smaller male was initially dominant, siring the first 8 litters over a 45-day period. We estimated that litters 9–12 were conceived during days 46–49, during which the smaller male sired litter 9 and the males shared paternity of litter 10. The larger male successfully displaced the smaller male during this period and sired the remaining 4 litters alone. These 2 males exhibited evidence of a fight on day 50, evidently won by the larger male. Within the artificially close proximity of the breeding pen the situation is perhaps more akin to that of harem breeders such as elk (Cervus elaphus), with a dominant male attempting to monopolize breeding opportunities of a group of females while defending his status against an ever-present competitor. Thus, endurance rivalry would perhaps best explain the dominance shifts in these breeding trials. Using a portion of the animals included here, DeYoung et al. (2006) reported that heavier males held initial dominance regardless of age in all 6 pens studied; however, 2 were displaced during the breeding season by rivals they outweighed by 2.1–10.0%, whereas dominant males that were not displaced outweighed rivals by 2.7–34.1%. Mass loss during the rut most likely favored smaller males and increased their relative mass over time (Mysterud et al. 2004). Thus, greater initial mass advantages seemingly favored dominance retention and limited substantially smaller subordinate males to tactics that avoided direct confrontation.

An alternative to direct confrontation is for males to pursue females that higher-ranking males are unaware of or unable to tend due to time constraints. In the wild this could be similar to scramble competition, based on the ability of males to locate receptive females due to greater investment in searching or prior experience of female locations. If a receptive female can be located before other males are aware of her, a subordinate male could sire her litter. However, it is unlikely that any such advantage existed within the relatively small confines of our breeding pens. The displacement hypothesis (Sorin 2004) suggests that multiple paternities can be attributable to situations wherein a younger, smaller male initially tends a receptive female and is displaced by a larger, older male while the female is still receptive. If we assume that length of gestation was identical among females, we noted that many instances of successful breeding by subordinate males, both cases of single or multiple paternity, occurred when ≥2 females were in estrus. Because tending males typically do not leave a female until she is no longer receptive (Hirth 1977; Marchinton and Hirth 1984), subordinate males could take advantage of such opportunities at least to begin tending unpaired receptive females. Another possible factor is that of sperm depletion in larger males during periods of near-constant female receptivity, which would increase the likelihood of smaller males with presumably greater sperm reserves siring offspring in females that copulated with both males (Lambiase et al. 1972; Preston et al. 2001). Because we could not predict when females would become receptive, we could not control for use of these breeding tactics.

Although age has long been credited as a primary driver of breeding success of male white-tailed deer, recent studies of captive and wild deer reveal a more complex situation (DeYoung et al. 2002, 2009). The confounding of age with body mass and other potentially influential factors likely explains much of the apparent effect of age (Coltman et al. 2002; Johnson et al. 2007; McElligott et al. 2001; Pelletier and Festa-Bianchet 2006; Preston et al. 2001). However, although age itself might not be an important consideration, age potentially could affect or act as a surrogate for other variables, such as experience, that are more difficult to quantify.

Because we minimized or excluded potential physical influences other than body mass, much of the variability in breeding success probably was due to physiological or behavioral factors. Expression of rutting behaviors and aggressive displays in deer have been correlated with testosterone (Bubenik et al. 1977; Ketterson and Nolan 1992; Miller et al. 1987); and greater testosterone levels might have allowed smaller deer to mitigate their mass disadvantage and compete successfully with larger males. Additionally, glucocorticoid levels have been related to dominance hierarchies in many species, but with inconsistent patterns (Creel 2001). It is unclear whether higher glucocorticoid levels in deer are related consistently to dominance (Taillon and Cô té 2008) or to what degree they might influence breeding success.

Individual behavior also can play a substantial role in determining breeding success (Smith and Blumstein 2008). Life-history strategies differ markedly among individuals of the same population, rooted in genetic differences and environmental variability (Dingemanse et al. 2009; Schuett et al. 2010). For example, recent work in bighorn sheep (Ovis canadensis) indicated that lifetime reproductive strategies among rams correlated with measures of boldness and docility (Réale et al. 2009). In our study 2 males were paired together as yearlings and again at age 3. Although the mass advantage of the larger male was virtually identical between years (15– 16%), the smaller male sired only 29% of the offspring as a 1-year-old but sired 97% as a 3-year-old, indicating some potential for differing life-history strategies between the 2 males. The effects of animal personality and behavioral plasticity on breeding strategy offer promising fields of study to explain differences in breeding success not captured by more easily measured variables.

Although we controlled for antler size by removing antlers before introducing males into breeding pens, wild deer are under no such constraints, and the degree of influence that antler characteristics have on breeding success in deer is unknown. Antler size could represent an inefficient method for males to predict the outcome of confrontations, particularly as the breeding season progresses and males lose body mass and condition (Clutton-Brock and Albon 1979). Antler size reportedly had no effect on harem-holding status, fighting success, or male-male assessment in tule elk (Cervus elaphus nannodesJohnson et al. 2007). Thus, although antler size generally correlates with body mass, it could be a secondary consideration in determining breeding success. However, antlers might influence breeding success by providing opportunity for males to establish dominance relationships prior to breeding season, particularly through prerut sparring bouts (Barrette and Vandal 1990; Mattiangeli et al. 1999; McElligott et al. 1998). By limiting opportunities for males to assess an opponent's body mass and strength through sparring, antler removal might have increased the influence of other factors in establishing dominance, perhaps explaining some of the variability in breeding success. Controlled experiments using males of variable mass and antler size could help quantify the level of influence antlers have on determining breeding success in deer.


This study was funded by the Mississippi Department of Wildlife, Fisheries, and Parks through a Federal Aid in Wildlife Restoration grant and the National Council for Air and Stream Improvement, Inc. We thank A. Blaylock, M. Dye, E. Flinn, and J. Flinn for sample collection and deer husbandry. This manuscript is contribution WF322 of the Mississippi State University Forest and Wildlife Research Center.


  • Associate Editor was Christine R. Maher.

Literature Cited

View Abstract