OUP user menu

Repeatability of antler characteristics in mature white-tailed deer in South Texas: consequences of environmental effects

Aaron M. Foley, Randy W. DeYoung, Steven D. Lukefahr, John S. Lewis, David G. Hewitt, Mickey W. Hellickson, Don A. Draeger, Charles A. DeYoung
DOI: http://dx.doi.org/10.1644/11-MAMM-A-183.2 1149-1157 First published online: 14 September 2012


Antler traits are both genetically determined and environmentally influenced. However, the degree to which environmental factors affect antler expression has rarely been quantified. We captured 30 to 150 male white-tailed deer (Odocoileus virginianus) annually at 7 South Texas sites during 1985 to 2009 to determine repeatability of antler traits from a semiarid environment with variable rainfall. Repeatability is defined as the intraclass correlation between repeated measures of the same trait over time. Repeatability was moderate to high (0.42–0.82) for all antler traits. Overall, number of antler points had the lowest repeatability, whereas inside spread of main beams and length of main beams had the highest repeatability. Repeatability of total antler score and number of antler points from sites with variable rainfall was 16% and 24% lower than sites with consistent rainfall, respectively. Sites with variable rainfall had 13–18% higher repeatability when enhanced nutrition was available. Studies of cervids reveal a tendency for lower repeatability of antler traits as the environmental conditions become more variable. The association between repeatability and variable environmental conditions illustrates the magnitude of environmental effects and supports the role of antlers as an honest advertisement of individual condition or quality. Our results help to understand potential of microevolution in antlers and have implications for sexual selection and harvest management.

Key words
  • antlers
  • honest advertisement
  • nutrition
  • Odocoileus virginianus
  • rainfall
  • repeatability
  • sexual selection
  • white-tailed deer

The nature and extent of individual variation in quantitative traits has long been a topic of interest in the ecology and management of wildlife (Nussey et al. 2007). Traits may be expressed more than once during an animal's life, through temporal or spatial repetition of growth. In animals, temporal repetition of traits is more common, and studies of domestic animals focused on production traits such as litter size, lactation performance, or wool length (Falconer and Mackay 1996). Temporal repetition of quantitative traits varies among and within individuals. Therefore, understanding the magnitude and cause of individual variation in trait expression in wildlife is important for assessing the potential for genetic selection and how heritable traits respond to environmental variation (Hayes and Jenkins 1997).

Antlers are a sexually selected trait unique to most cervids, and are cast and regrown each year. Antlers appear to serve as honest advertisements of individual condition or quality (Ditchkoff et al. 2001) and fit the definition of a handicap trait (Zahavi 1975). Antler size and overall conformation are partially genetically determined (Goss 1983; Lukefahr and Jacobson 1998), but phenotypic expression is also influenced by environment. For instance, antler development may be affected by maternal effects (Monteith et al. 2009) or birth date, where late-born individuals may be phenotypically stunted (Gray et al. 2002). Population density (Couturier et al. 2010) and nutrition or habitat quality (Bowyer et al. 2002; Strickland and Demarais 2000) also influence antler development through temporary and permanent effects on expression. Temporary effects on antler expression may be caused by resource limitation (Clutton-Brock et al. 1982; Geist 1986; Simard et al. 2008) such that when nutritional availability increases, improved antler growth occurs (Ashley et al. 1998; Leberg and Smith 1993). Permanent environmental effects may occur for example when nutritional conditions during year of birth influence general development of young males (Kirkpatrick and Lande 1989; Mech et al. 1991), which may have long-lasting effects on antler traits throughout life (Monteith et al. 2009). Finally, populations of ungulates may modify their life history to adjust body mass and reproduction in response to population density and resource limitation (e.g., Gaillard et al. 1998; Simard et al. 2008). As a result, it is difficult to separate the effects of environmental and genetic components of variation on the expression of antler traits.

The effects of environment on antler growth have been assessed mainly in northern populations of cervids (Mysterud et al. 2005; Schmidt et al. 2001). Variable precipitation in arid environments influences population dynamics of ungulates (Marshal et al. 2002; Owen-Smith 1990) but the influence of climate on antler growth in semiarid environments has not been quantified. Southern Texas and northern Mexico experience dramatic year-to-year variation in rainfall (coefficient of variation >30%). The 60 years preceding the year 2006 produced 23 wet summers and 37 dry summers; total rainfall ranged from 40 to 120 cm (Norwine and John 2007). This semiarid environment provides opportunities to quantify how environmental variability affects individual variation in a sexually selected quantitative trait. The source and magnitude of variation in antler expression, as an individual characteristic, under different environmental conditions has implications for honest advertisement of condition or quality. The handicap theory states that sexually selected traits should be costly to produce and maintain (Zahavi 1975). The cost associated with the trait curtails cheating, and thus functions as an honest advertisement of condition or quality. Consistency in expression of antler traits in areas with variable environmental conditions may reveal the degree to which antlers conform to the expectations for honest advertisement. If antler traits reflects male quality or affect male–male competition, variation in antler traits within and among individuals has implications for sexual selection (Andersson and Simmons 2006). If breeding success is associated with antler size, directional selection may occur; however, if environmental influences on antler expression is great then masked genetic potential may result in static evolution of antler traits (e.g., Kruuk et al. 2002).

Male–male confrontations are a frequent occurrence during the breeding season and competitors often utilize antlers for leverage (Goss 1995). Resource allocation toward developing “standard” antlers (i.e., the typical appearance of antlers) during nutritionally difficult years may result in lower antler strength, rendering antlers as a less efficient tool for leverage. The potential for trade-offs in antler trait investment is plausible; an alternative strategy to an overall reduction in antler size might be to alter antler conformation, particularly if there is an antler characteristic used for male–male competition. It is unknown whether resources are allocated toward specific antler traits (e.g., Mysterud et al. 2005). Assessing the consistency of certain antler traits may reveal whether males allocate resources toward certain antler traits.

Finally, antlers are both targets and tools in the management of cervids (Demarais and Strickland 2011; Miller and Marchinton 1995). Understanding the magnitude of environmental effects on antler traits has practical applications for the consideration of antler characteristics in harvest decisions (Demarais and Strickland 2011; Miller and Marchinton 1995; Mysterud and Bischof 2010).

We assessed the role of environmental effects on antler growth using a quantitative genetic approach applied to a spatially and temporally extensive data set of free-ranging white-tailed deer (Odocoileus virginianus). We estimated repeatability, the correlation between repeated measurements on a quantitative trait separated by space or time. Repeatability is defined as ratio of among-individual variance to total variance of a measured trait, which also sets the upper limit of heritability (Falconer and Mackay 1996). We had three main objectives: derive repeatability estimates for antler traits in wild populations; determine how variable rainfall and enhanced nutrition affect repeatability of antler traits; and evaluate whether trade-offs in antler conformation may be a valid hypothesis. The results of the study should improve our understanding of how environmental variation affects expression of a phenotypic trait, with implications for honest advertisement, sexual selection, and the design of harvest management strategies.


Study sites.—We had 7 study sites in South Texas; 1 in Kleberg County, 4 in Webb County, and 2 in Dimmit County (27°27′N, 97°33′W to 28°12′N, 100°11′W; Fig. 1). The Kleberg site was located 16 km east of Kingsville, Texas and was part of the Southern Subhumid Gulf Coastal Prairie ecoregion (Environmental Protection Agency [EPA] 2010). Typical woody plants included huisache (Acacia farnesiana) and mesquite (Prosopis glandulosa). Sites in Webb County (Webbl to Webb4) were located 43 km east, 29 km east, 15 km northeast, and 24 km northeast of Laredo, Texas, respectively. The sites were part of Texas-Tamaulipan Thornscrub ecoregion (EPA 2010); typical woody vegetation included mesquite, brasil (Condalia hookeri), and lime prickly ash (Zanthoxylum fagara). Dimmit County sites 1 and 2 were 48 km southwest and 23 km northwest of Carrizo Springs, respectively. Both sites were part of the Texas-Tamaulipan Thornscrub ecoregion (EPA 2010).

Fig. 1

Locations of 7 study sites in South Texas and average rainfall isohyets for the region.

Deer capture and antler measurements.—We captured deer using the helicopter net-gun (Barrett et al. 1982; Webb et al. 2008) or drive-net techniques (DeYoung 1988). We captured 30 to 150 male white-tailed deer annually from each site for 4 to 10 years, depending on the site. Helicopter pilots were instructed to capture the first antlered deer encountered. After net-gunning and physically restraining deer, we aged deer according to tooth replacement and wear (Severinghaus 1949) and measured antlers (see below). We either inserted unique numbered microchips (Avid Microchip ID Systems, Mandeville, Louisiana) subcutaneously in the front leg and at the base of one ear or attached unique numbered and colored livestock ear tags to both ears. We released deer near capture locations. Hunting occurred on some of the sites during the study period and served as additional means for resampling previously captured individuals. Deer captures were approved by Texas A&M University–Kingsville Institutional Animal Care and Use Committee (Animal Use Protocol numbers 3-98-09, 99-5-2, 2003-5-14, 2009-05-6A) and were consistent with guidelines approved by the, American Society of Mammalogists (Sikes et al. 2011).

We measured length and mass of antlers using the Boone and Crockett system (B&C—Nesbitt and Wright 1981), widely used in North America as an index of antler size. Antler size using the B&C system is the sum of 4 circumferences of main beams, lengths of main beams, lengths of antler points (if ≥2.54 cm), and the spread between the main beams. Measurements were taken using a metal tape to the nearest 0.31 cm, as per B&C rules. Total number of antler points was the count of points ≥2.54 cm in length. The sole exception to our antler measurement protocol was site Dimmit1, where only 1 basal circumference of both main beams was recorded to minimize processing time during drive net capture (see below for analysis).

Repeatability.—Repeatability is the sum of additive and nonadditive genetic variation and permanent environmental variation divided by total variation (Falconer and Mackay 1996). The equation is as follows: Embedded Image where VA is additive genetic variation, VEg is general environmental variance (environmental variance contributing to the between-individual component, attributed to permanent or nonlocalized environmental effects), and VP is the total phenotypic variation. Additive genetic variation, the basis of heritability, is the sum of the independent effects of individual genes’ influence on an expressed trait of an individual (Falconer and Mackay 1996). Permanent environmental effects may include nonadditive genetic effects (e.g., epistatis), maternal condition, range conditions (e.g., birth year), or physical injuries that have permanent positive or negative consequences on the individual (e.g., Gaillard et al. 1993). Residual variation may include temporary environmental effects, such as drought effects on antler traits in 1 year. Instances of temporary environmental effects also may include health status and position in social hierarchy (Lukefahr and Jacobson 1998). Repeatability defines the upper limit to the broad-sense heritability, the extent that phenotypes are determined by genotypes (VG/VP, where VG is the genotypic variance), and the narrow-sense heritability, expressed as VA/VP (Falconer and Mackay 1996). In wild or free-ranging populations, repeatability will often be easier to estimate than heritability due to the lack of pedigreed individuals; repeatability also allows insights into the effects of environment on the traits.

We estimated repeatability of antler traits using an animal model (Henderson 1953). Animal identification number, trait measurement, and capture year were imported into the computer program LSMLMW (least-squares maximum-likelihood mean weighted—Harvey 1987), and we used capture year as our fixed effect and animal as our random effect. Repeatability is estimated as the intraclass coefficient r, derived from the ratio of the between-individual variation to the total phenotypic variation.

Variation of a measured trait is separated into 2 categories: repeatability and residual variation, which always sum to 1. Repeatability values range from 0 to 1; values of 0 indicate that the average of repeated antler measurements from all individual males is identical and variation is entirely within individuals. Values of 1 indicate that the same antler measurements are obtained every time an individual is captured and all variation is among individuals (Hayes and Jenkins 1997).

Males typically develop small antlers as yearlings and continue to increase in antler size until about 5 years of age (after attaining physical maturity), then may not change dramatically until senescence at >7 years old (Lewis 2010). Antler size thereafter usually decreases. Due to the dramatic change in antler size from 1 to 3 years old as a function of growth and physical maturation, we used measurements only from prime-age males (ages 3 to 6 years old) on 6 of 7 sites. Data collected from Dimmit1 were previously reported in DeYoung (1998) and included males >6 years old. Inaccuracies of the tooth wear and replacement method (Gee et al. 2002; Lewis 2010) complicate the assignment of specific year classes due to variation among individuals within age classes. However, white-tailed deer in South Texas can be assigned to age classes 2.5 years, 3.5 to 5.5 years, and >6.5 years of age with acceptable accuracy (72%, 73%, and 68%, respectively; Lewis 2010).

Most individuals were captured before the breeding season and broken antlers were rare (Webb et al. 2008). Antler points or main beams broken before or during capture were removed from specific analyses. For instance, an individual with a broken main beam was removed from analyses incorporating main-beam measurements but was considered for analyses of individual traits not affected by the broken portions, such as inside spread or number of antler points. Broken antler points ≥2.54 cm were included in analyses of total number of antler points, but not in analyses requiring length of antler points. Last, to minimize data removal due to broken antlers and asymmetry between a set of antlers from an individual, we treated left and right antlers separately (measurements from left and right antlers were not combined). The spread between main beams is not a measurement of antler size or mass. Therefore, we subtracted the inside spread measurement from the B&C score and termed the modified score as “total antler length.” Because site Dimmit1, an unfed site, did not record all measurements, we compared the correlation among different combinations of antler trait measurements from the Kleberg population, another unfed site. Sum of antler point length in Dimmitl had the highest correlation with total antler length (r = 0.84) in Kleberg; thus, we used sum of antler point length in lieu of total antler length for Dimmit1. Repeatability values for traits with multiple measurements (basal circumference and main beam length) were averaged for each site.

Rainfall and enhanced nutrition.—We used a quasi-experimental approach (Morrison et al. 2008) to assess the effects of environmental variation on repeatability of antler traits. Our treatments were not randomly assigned to study sites; thus, our experiment was not designed with equal numbers of controls and treatments. Furthermore, sites were exposed to uncontrolled environmental conditions for the duration of capture. Antler growth in male white-tailed deer begins during spring and ceases in early autumn (Sauer 1984). Consequently, rainfall during March to May is important because of its impact on forage production and quality during the antler growing period. It is possible that nutrition availability before antler development can influence antler expression; however, we also use March to May rainfall to illustrate differences among sites. Total rainfall in March to May during years of capture on each site averaged 15.2 cm (variance [var] = 39.9), 9.6 cm (var = 11.8), 9.6 cm (var = 11.8), 9.6 cm (var = 11.8), 9.2 cm (var = 14.9), 14.6 cm (var = 51.2), and 18.1 cm (var = 60.3) for sites Kleberg, Webbl, Webb2, Webb3, Webb4, Dimmitl, and Dimmit2, respectively (NOAA Satellite and Information Service 2010; Departmènt of Atmospheric Sciences, Texas A&M University 2010). Webb sites were considered to be independent sites because they were high-fenced to minimize deer immigration and emigration for deer management purposes. We first grouped sites on the basis of relative rainfall variance during the 4 to 10 years of data collection; sites Kleberg, Dimmitl, and Dimmit2 were categorized as variable rainfall (var ≥ 39.9 cm) and sites Webbl, Webb2, Webb3, and Webb4 were categorized as consistent rainfall (var ≤ 14.9 cm).

Many private landowners in South Texas have implemented enhanced nutrition as part of deer management programs (Jacobson et al. 2011; McBryde 1995). Sites in this study had different deer management objectives and some chose to provide supplemental feed. Intensity of feeding varied among sites, where some sites had a greater density of feed stations or provided feed year-round versus during a portion of the year. We categorized sites into quasi-treatment groups on the basis of feed intensity (unfed, moderate, and intensive). Sites Kleberg and Dimmit1 did not provide enhanced nutrition and served as unfed controls. Sites Webb1, Webb2, and Dimmit2 had intensive feed programs, defined as constant feeding year-round. Sites Webb3 and Webb4 had moderate-intensity feed programs, defined as not feeding year-round. Enhanced nutrition programs consisted of commercial pelleted feed rations provided ad libitum via various types of feeder troughs. Pellets contained no less than 16% crude protein, no more than 12% fiber, no less than 2% fat, and contained minerals (calcium, phosphorus, salt, etc.) and vitamins (A and E).

We used 95% confidence intervals to evaluate differences in mean repeatability of same antler traits among feed (n = 3) and rainfall (n = 2) quasi-treatments.


We captured 233 to 856 unique individuals per site; 98 to 235 individuals per site were recaptured ≥1 times. Number of records removed from analysis because of broken antlers was low, ranging from 0 to 6 occurrences per site for each trait. Number of antler points had lowest average repeatability ( = 0.55, SD = 0.09) compared with repeatability of other traits, regardless of site (Table 1). Inside spread ( = 0.69, SD = 0.07) and beam length ( = 0.66, SD = 0.05) had the highest average repeatability. Total antler length (0.59 to 0.82), antler points (0.42 to 0.64), and spread (0.58 to 0.80) displayed the greatest variation in repeatability among sites. Main beam length and basal circumference were less variable, ranging from 0.60 to 0.74 and 0.54 to 0.70, respectively (Table 1).

View this table:
Table 1

Repeatability estimates for antler traits (SE) estimated from free-ranging, prime-aged male white-tailed deer on 7 South Texas sites during 1985 to 2009. Sites are categorized by rainfall during March–May and presence of supplemental nutrition. Data from a previous study on captive white-tailed deer are used for comparison.

SiteYearsN total antler lengthFeedaRainbTotal antler lengthAntler pointsBeam lengthAntler spreadBasal circumference
Kleberg1999–2009321278.3 (50.0)NoneVar0.59 (0.07)0.42 (0.09)0.68 (0.08)0.58 (0.07)0.54 (0.07)
Dimmit22007–2009186249.3 (45.4)IntVar0.68 (0.07)0.44(0.11)0.64 (0.08)0.66 (0.07)0.66 (0.07)
Webbl1998–2007633255.5 (48.7)IntCon0.64 (0.05)0.61 (0.05)0.62 (0.05)0.66 (0.05)0.66 (0.05)
Webb21998–2008737259.5 (51.6)IntCon0.68 (0.04)0.59 (0.05)0.70 (0.04)0.67 (0.04)0.70 (0.04)
Webb31998–2007313236.2 (46.7)ModCon0.81 (0.04)0.51 (0.10)0.61 (0.08)0.72 (0.06)0.65 (0.07)
Webb41998–2005307260.8 (50.1)ModCon0.82 (0.03)0.64 (0.06)0.74 (0.05)0.80 (0.04)0.69 (0.05)
Captivec1977–1993469N/A---0.48 (0.14)0.58 (0.15)0.60 (0.12)0.57 (0.14)
Additive genetic (h2, or narrow-sense heritability)0.39 (0.14)c0.14 (0.15)c0.03 (0.11)c0.29 (0.14)c
Permanent environmental0.09 (0.13)0.44 (0.15)0.57 (0.12)0.28 (0.13)
  • a Supplemental nutrition; Int = feed provided year-round, Mod = feed during part of the year.

  • b Rainfall; Var = variable rainfall (variance ≥ 39.9 cm); Con = consistent rainfall (variance ≤ 14.9 cm).

  • c From Lukefahr and Jacobson 1998. Repeatability is sum of additive and permanent environmental effects. N/A = not applicable.

Each antler trait had lower average repeatability in variable rainfall sites than consistent rainfall sites. Repeatability of total antler length was 16% lower in variable rainfall sites ( = 0.62, SD = 0.05, 95% confidence interval [CI] = 0.57–0.68) than consistent rainfall sites ( = 0.74, SD = 0.09, 95% CI = 0.65–0.83).

Repeatability of number of antler points appeared to be affected by rainfall regardless of enhanced nutrition (Table 1). Repeatability of number of antler points was 24% lower in sites with variable rainfall ( = 0.45, SD = 0.03, 95% CI = 0.41–0.48) versus consistent rainfall ( = 0.59, SD = 0.06, 95% CI = 0.53–0.64).

Antler traits also displayed lower average repeatability in unfed sites versus fed sites. Unfed sites had numerically lower total antler score repeatability ( = 0.60, SD = 0.01, 95% CI = 0.59–0.60) than sites with high feed intensity ( = 0.66, SD = 0.02, 95% CI = 0.64–0.69) and with moderate feed intensity ( = 0.82, SD = 0.01, 95% CI = 0.81–0.82, Table 1). The availability of enhanced nutrition in variable rainfall sites appeared to moderate the environmental effects for some antler traits. Repeatability estimates for total antler length and basal circumference in variable rainfall sites were 13% and 18% higher when feed was available, respectively. Sites with moderate feed intensity and consistent rainfall had 24% higher total antler score repeatability estimates ( = 0.82) than sites with high feed intensity and consistent rainfall ( = 0.66, SD = 0.04, 95% CI = 0.62–0.69, Table 1).


In semiarid regions, variable rainfall affects forage quality and quantity. Forage quality is important for a concentrate selector such as the white-tailed deer, because the digestive system does not afford the opportunity to offset low-quality diets by increasing forage intake (Barboza et al. 2009). In our study, sites exposed to variable rainfall had lower repeatability than sites with consistent rainfall. However, enhanced nutrition appeared to moderate the effects of rainfall variation. Sites with enhanced nutrition had higher average repeatability estimates ( = 0.66) than sites without supplementation ( = 0.58), irrespective of rainfall. Of 3 sites exposed to variable rainfall, the site with enhanced nutrition had 15% higher total antler length repeatability.

Unexpectedly, the highest observed repeatability values occurred where nutrition was provided at moderate intensity, rather than year-round. A decline in repeatability might occur if antler size continued to increase later in life as a function of enhanced nutrition. However, we detected no statistical difference in average antler size among sites (Table 1) or in ages of peak antler size between the 2 nutrition treatments (data not shown). Behavioral interactions or competition among deer at feed sites might plausibly result in biased access to feed among individuals (Bartoskewitz et al. 2003; Donohue 2010). However, standard deviation of total antler score did not appear to covary with feed intensity (Table 1). It is possible that our study sites have some inherent differences in soil or habitat quality that we were unable to assess due to the quasi-experimental approach; the nature of the study precluded random assignment of treatments and controls. Nonetheless, it appears that enhanced nutrition reduces some of the environmental effects on antler expression.

Repeatability estimates for antler traits are similar for populations of cervids maintained under like conditions. For instance, captive and supplemented populations consistently have statistically higher repeatability for antler traits than unsupplemented or free-ranging populations (Table 2). Captive conditions provide food, shelter, protection from predators, and veterinary treatment, all of which may reduce environmental variation due to temporary effects in antler expression and effectively increase repeatability. Thus, environmental factors appear to exert similar effects on antler expression in diverse populations of cervids.

View this table:
Table 2

Repeatability estimates (SE) of antler mass (g) and total antler length for populations of cervids maintained under different environmental conditions.

LocationHabitat descriptionRepeatabilityAuthors
ScotlandaFree-ranging, unfed, temperate0.57bKruuk et al. 2002
KlebergFree-ranging, unfed, semiarid0.59cThis study
Dimmit1Free-ranging, unfed, semiarid0.60cThis study
New ZealandaCaptive, pasture, some feed0.64bvan den Berg and Garrick 1997
Webb1Free-ranging, fed, semiarid0.64cThis study
Dimmit2Free-ranging, fed, semiarid0.67cThis study
Webb2Free-ranging, fed, semiarid0.68cThis study
Czech RepublicaCaptive, pens, fed0.75bBartos et al. 2007
MississippiCaptive, pens, fed0.76bLukefahr and Jacobson 1998
Webb3Free ranging, some feed, semiarid0.81cThis study
Webb4Free ranging, some feed, semiarid0.82cThis study
  • a Red deer.

  • b Mass (g).

  • c Total antler length.

Repeatability was lower in areas with variable environmental conditions. The influence of environmental factors on the predictability of antler expression is consistent with theoretical expectations for sexually selected traits. The handicap principle states that if antlers function as an honest advertisement of individual condition or quality, the trait must be costly to produce or maintain (Zahavi 1975). Individuals in good condition can afford to devote additional resources toward antler expression, but cannot maintain the investment in times of poor nutrition. In this manner, cheating strategies are curtailed in part by risking overallocation of scarce resources to a deciduous trait in poor years. Therefore, the lower predictability in antler traits in variable environmental conditions supports antlers as an indicator of honest advertisement.

Number of antler points had the lowest repeatability compared with other antler traits. Antler points are a discrete trait defined by presence or absence, whereas the other antler characters were continuous. Therefore, the lower trait repeatability for antler points may be influenced by differences in trait measurement. However, repeatability of antler points was noticeably lower than other antler traits in our study and also in other studies (Bartos et al. 2007; Lukefahr and Jacobson 1998). Visual appearance is a factor in judging rivals before broadside threats (Clutton-Brock 1982; Lincoln 1972), yet the visual appearance of individual white-tailed deer with similar antler height, spread, and beam length may be nearly identical within a range of antler points (i.e., 8 versus 10 antler points). Branched antlers are used as leverage during pushing and shoving when battling with other males (Goss 1995), which suggests that antler strength, not necessarily the number of antler points, would be advantageous. If this reasoning is correct, males might sacrifice number of antler points when nutrition is limited to maintain overall visual appearance and breaking strength. Thus, males may invest resources toward primary antler development (spread and beam length) to ensure sufficient strength rather than antler points (e.g., Mysterud et al. 2005). Furthermore, the use of antlers in fighting limits cheating in that antlers may need to maintain a minimum threshold of strength or risk breakage, hampering the ability to compete with conspecifics during the breeding season. Limitations on cheating are supported by empirical observations from wild populations. For instance, within-year breaking strength, density, and proportion of spongy bone in antlers did not differ among white-tailed deer in a free-ranging population (McDonald et al. 2005). Further research should test the hypothesis that males allocate available resources toward main antler characteristics for the purpose of male–male competition.

From an applied perspective, antler trait repeatability has implications for hunter selection and harvest management. Because repeatability is a correlation between prior measurements, lower repeatability values imply less predictability in future performances. Many state agencies have established harvest criteria on the basis of antler traits, usually aimed at protecting young males from harvest (Demarais and Strickland 2011). Elsewhere, antler traits are considered in harvest decisions, where sportsmen may preferentially harvest the largest-antlered males, or where antler traits are used in culling decisions (Jacobson et al. 2011; Mysterud and Bischoff 2010). Commonly used criteria are number of points and antler spread due to ease of visual confirmation by hunters. On average, regardless of site, number of antler points had the lowest repeatability values ( = 0.53) in our study and in previous studies (Bartos et al. 2007; Lukefahr and Jacobson 1998). Spread ( = 0.69) and main beam length ( = 0.66) had the highest average repeatability values; thus, these may be more useful for establishing harvest criteria in mature male white-tailed deer.

The potential for hunter selection to affect trait evolution has raised concerns (Coltman et al. 2003; Darimont et al. 2009; Harris et al. 2002; Mysterud and Bischof 2010). Traits with nonzero heritability values have sufficient additive genetic variance to respond to selection (Hayes and Jenkins 1997). Heritability of antler traits varies widely among age classes of males and among specific antler traits (Lukefahr and Jacobson 1998). Our results suggest that traits with moderate to high narrow-sense heritability (h2) appeared to be more sensitive to environmental conditions than traits with low heritability (h2 = < 0.20). Number of points (h2 = 0.39), basal circumference (h2 = 0.29), and total antler score (h2 = 0.43—Lukefahr and Jacobson 1998) in variable rainfall sites had lower repeatability (31%, 15%, and 21% lower, respectively) than in sites with consistent rainfall. The aforementioned traits were 24%, 17%, and 19% lower in unfed sites than in fed sites. Antler spread (h2 = 0.03) was 8% and 5% lower in variable rainfall sites and unfed sites than consistent rainfall sites and fed sites, respectively. The negative association between heritability and repeatability of certain antler traits appears counterintuitive. It is possible that traits with low heritability, such as spread of main beams, are influenced more than permanent or other environmental effects early in life and are less sensitive to annual variation for that reason. Because the environmental covariance hypothesis (Kruuk et al. 2002) states that sexual selection is based on traits associated with environmental components (i.e., body condition, nutritional state) rather than genetic components, the potential for trait evolution via selection, either by harvest or by sexual selection, appears lower in variable environments.

Environmental effects on antler expression appear to be pervasive in the aggregate, but the effects of specific Stressors are difficult to quantify. For instance, 60% to 70% of variation in antler yield of farm-reared cervids was due to nonadditive and environmental effects, even under controlled conditions (Wang et al. 1999). Similarly, we were able to detect variation in antler expression associated with environmental variation, but could not quantify the specific causal factors involved. Although white-tailed deer are continuously distributed in South Texas, we observed relatively large differences in repeatability among populations in proximity (e.g., Webb County). In field conditions, variation in repeatability is probably due to environmental factors and not genetic differences; otherwise, repeatability would be similar in sites that are in proximity. Additional research is clearly needed to understand the nature of the environmental influence on antler expression, especially the role of maternal and cohort effects. Our study provides the basis to construct research hypotheses and guide further experimentation.


King Ranch Inc., Texas Parks and Wildlife Department, IBC Bank, Caesar Kleberg Wildlife Research Institute, and Texas A&M University–Kingsville provided financial support. A. R. Sanchez, Jr., C. Rush, C. Y. Benavides, II, K. Shepard, and J. Finley also contributed financial support. We thank numerous student volunteers from Texas A&M University–Kingsville for assistance with deer capture and S. Stedman for insightful discussions about environmental effects on antler repeatability. This is publication 12–120 of Caesar Kleberg Wildlife Research Institute.


  • Associate Editor was Harald Beck.

Literature Cited

View Abstract