Abstract

Body size is an important influence on the life history of males of polygynous mammals because it is usually highly correlated with fitness and is under intense selection. In this paper, we investigated the effect of high-risk foraging behavior (crop raiding) and genetic heterozygosity on male body size in a well-studied population of African elephants. Crop raiding, the foraging on cultivated food crops by wildlife is one of the main causes of wildlife human conflict and is a major conservation issue for many polygynous mammals that live in proximity to agriculture or human habitation. Body size was estimated using hind foot size, a measure strongly correlated with stature and mass. Crop raiding predicted male size in adulthood, with raiders being larger than nonraiders. However, elephants that became raiders were neither larger nor smaller for age when young. Enhanced growth rates and size among raiders suggest that taking risks pays off for males. Lastly, genetic heterozygosity had no effect on size for age in male elephants, most likely because low-heterozygosity males were rare. Risky foraging behavior can evolve as a result of strong sexual selection for large size and condition-dependent mating success in males. We discuss the implications of these results for managing human–wildlife conflict.

INTRODUCTION

Body size is an important lifehistory–related trait in polygynous mammals, often influencing survival early in life (Loison et al. 1999), fighting ability, and dominance at maturity and consequently reproductive success (McElligott et al. 2001; Zedrosser et al. 2007). In polygynous mammals, males generally experience more intense selection for large body size than females, as a result of intrasexual reproductive competition as well as female preferences for larger males during mate choice (Poole 1989; Bowyer et al. 2007; Charlton et al. 2007). As a result of this more intense intrasexual competition among males than among females, there is a larger variance in reproductive success among males than females (Struhsaker and Pope 1991; Alberts et al. 2003; Setchell et al. 2005).

In most polygynous species, males have delayed reproductive maturation compared with females. This delay creates an opportunity for males to invest in growth in order to enhance their reproductive competitiveness (Whitehead 1994; Post et al. 1999; Isaac 2006; Field et al. 2007). For example, male elephants are physiologically capable of siring offspring at 15–17 years of age (Owen-Smith 1988) but typically do not sire their first offspring until they are 26–30 years of age (Poole 1989; Hollister-Smith et al. 2007). In spite of strong selection for large size, male size remains variable even within populations. Identifying factors that generate this variation is critical for understanding male reproductive success and life-history strategies.

In this paper, we investigate the influence of 2 factors on male size for age in African elephants. One of these factors, crop raiding (foraging on cultivated food crops by wildlife) is a major conservation issue for many polygynous mammals that live in proximity to agriculture or human habitation. Crop raiding occurs in all elephant populations that border agricultural areas and is a major source of human–elephant conflict (Williams et al. 2001). Crop raiding is sex biased and is undertaken more by males than females (Sukumar and Gadgil 1988; Chiyo and Cochrane 2005). However, not all male elephants raid crops even when their home ranges abut agricultural areas (Sukumar 1995; Williams et al. 2001). Crop raiding by independent males is initiated after they have dispersed from their maternal families at approximately 14 years of age (Lee and Moss 1999), and we therefore predicted raiding to affect adult growth and hence size after dispersal. The second factor we examined is multilocus heterozygosity at microsatellite loci. Multilocus heterozygosity has been demonstrated to influence infant survival and to correlate with body size or growth in some species (Coltman et al. 1998; Charpentier et al. 2006). We therefore expected the effects of heterozygosity on male body size to occur throughout life.

Crop raiding as a high-risk foraging behavior

Several studies on sexually dimorphic mammals have demonstrated that males seek more abundant, high-quality forage at the risk of predation, whereas females may sacrifice forage abundance to minimize predation risk when there is a positive correlation between food abundance and predation risk (Bleich et al. 1997; Apollonio et al. 2005; MacFarlane and Coulson 2007; Hay et al. 2008). The high proportion of males undertaking high-risk foraging with regard to crop or livestock raiding has been reported in other species, such as chimpanzees (Wilson et al. 2007) and lynxes (Odden et al. 2002), respectively. Although male elephants engage in a variety of high-risk foraging tactics more than females, the propensity of individuals to raid crops is variable (Sukumar 1995). Crop raiding is a high-risk foraging strategy for elephants because crop raiders are often killed or injured by farmers and sometimes by wildlife authorities when they are detected raiding (Haigh et al. 1979; Cheeran et al. 2004; Mpanduji et al. 2004; Obanda et al. 2008). In the Amboseli elephant population in Southern Kenya, 10% of raiders were seen with spear injuries during this study (Chiyo PI, unpublished data). Elephants appear to recognize these risks. For example, although elephants naturally forage both during the day and night, foraging on crops invariably occurs at night (Graham et al. 2009) and particularly during moonless nights (Barnes et al. 2007), probably to minimize risks of detection by farmers. On the other hand, crop raiding offers high nutritional returns compared with foraging on wild plants (Sukumar 1990; Rode et al. 2006), and raiders can obtain 38% of their daily forage intake in 10% of their foraging time while raiding crops compared with foraging on wild plants (Chiyo and Cochrane 2005).

Trading-off safety from predators for forage abundance has been speculated to provide a nutritional pay off that could accelerate growth or increase body size and consequently reproductive success (Corti and Schackleton 2002; Mooring et al. 2003; Hay et al. 2008). In a species with indeterminate growth and high energetic costs of sustained growth, such as elephants, sustaining growth rates may be especially critical to reproductive output. However, no study has demonstrated a causal relationship between body size and such high-risk foraging behavior. Here, we test for the effect of high-risk foraging behavior on body size by specifically examining whether independent male crop raiders are larger for age compared with nonraiders and whether males that raid crops are initially smaller or larger for age.

Microsatellite heterozygosity effects on life-history traits

Multilocus heterozygosity at microsatellite markers has been shown to be linked to measures of fitness, such as growth (Liu et al. 2006), infant survival (Da Silva et al. 2006; Cohas et al. 2009), and reproductive success (Slate et al. 2000; Zedrosser et al. 2007). These multilocus heterozygosity and fitness correlations have been shown to result either from locus-specific effects for loci linked to functional genes that influence fitness and are under balancing selection or from inbreeding depression associated with genome wide loss in allelic variants that influence fitness (Hansson et al. 2004; Gage et al. 2006). Some studies have found positive correlations between microsatellite heterozygosity and growth rates or size for age in vertebrates (Hildner et al. 2003), whereas other studies have found no correlations between heterozygosity and size or growth rate (Curik et al. 2003; Overall et al. 2005; Zedrosser et al. 2006). A limited number of studies have investigated the effects of multilocus heterozygosity of neutral loci on size for age or growth in natural free-ranging large mammals, specifically males (Curik et al. 2003; Charpentier et al. 2006; Zedrosser et al. 2007), but no similar studies have been done on elephants. Here, we test the hypothesis that elephants with a high microsatellite multilocus genetic heterozygosity are larger for age.

MATERIALS AND METHODS

Study population and age estimation

This study focused on the Amboseli elephant population in southern Kenya, currently consisting of ∼1400 elephants. This population has been studied continuously since 1972 by the Amboseli Elephant Research Project (Moss 2001; Croze and Moss 2011). All elephants born to the Amboseli population are individually known and recognizable from natural tears, notches, holes, and vein patterns on pinnae, tusk characteristics, natural body marks, and shape. This population is free ranging and uses an area nearly 8000 km2 of Maasai ranches (Croze and Moss 2011) surrounding Amboseli National Park and is connected to elephant populations from Kimana, Tsavo, and Chyulu in the east and those of Kilimanjaro and Longido controlled hunting area in the south and southwest (Douglas-Hamilton et al. 2005; Croze and Moss 2011). All known Amboseli elephants have ages assigned to them; elephants born since 1975 have their ages estimated to within 2 weeks, those born between 1972 and 1974 have ages estimated to ±3 months, and elephants born between 1969 and 1971 have ages estimated to within 1 year. Elephants born before 1969 have ages estimated to within 2–5 years depending on the time difference between when a mother was last seen without a calf and when she was first seen with a new born calf. All age estimations are validated from long-term observations of growth and body shape as well as tooth ages when dead (Moss 2001).

Estimation of elephant size from footprint measurements

Footprint measurement is a reliable and well-established noninvasive method for determining elephant size (Western et al. 1983). Elephant hind footprint length is highly correlated with height at the shoulder, accounting for up to 93% of the variance in male elephant height in African elephants (Lee and Moss 1995) and 94% in Asian elephants (Kanchanapangka et al. 2007). Footprint length is also highly correlated with body mass in both species of elephants. In Asian elephants, forefoot circumference was found to account for 86% of variance in body weight (Kanchanapangka et al. 2007). We estimated the length of the hind foot from measurements of footprint impressions left on the soil by known individuals. Specifically, we measured the linear distance perpendicular to the short foot axis from the outer rear edge of the footprint to the internal arch of the toe excluding the toenail imprint. For any given sighting, we took several opportunistic measurements of footprints whenever the soil substrate allowed delineation of footprints and whenever we were able to observe footprints for target individual elephants. We measured 650 footprint sizes from 302 unique individuals between 1976 and 2007. These include 120 measurements collected from 36 unique individuals that we observed to raid crops in 2005–2007. Only 46 unique males measured in 1976/1991 were measured again in 2005/2007.

Genetic analysis and sample collection

We used 2 sets of genetic data in this paper. First, we used genotypes of known individuals from recent genetic studies on this population (See Archie et al. 2007, 2008) for whom we had footprint measurements (n = 119 males). Second, we collected fecal samples from 50 known males that were not previously genotyped for a genetic analysis. From these fecal samples, we extracted DNA using a QIAamp DNA Stool Mini Kit (Qiagen, Germantown, MD) following a modified protocol (Archie et al. 2003). All individuals were genotyped at a minimum of 8 loci and up to 11 loci from previous studies. These included 1 dinucleotide locus (LAFMS02; Nyakaana and Arctander 1998) and 10 tetranuclotide loci (LaT05, LaT07, LaT08, LaT13, LaT16, LaT17, LaT18, LaT24, LaT25, and LaT26; Archie et al. 2003). We used the polymerase chain reaction (PCR) protocols detailed in Archie et al. (2003) to amplify DNA from the loci of interest. PCR products were separated using Applied Biosystems 3730XL DNA Analyzer and analyzed using Genemapper v.3.7 (Applied Biosystems, Beverly, MA). Microsatellites alleles were scored using GeneMarker v.1.6. (SoftGenetics State College, PA). For each sample, we ran PCR and genotyping twice if the initial PCR product was scored as a heterozygote and 3–4 times if it was a homozygote, in order to minimize error associated with spurious alleles and allelic drop out (Archie et al. 2006). We tested all loci for Hardy–Weinberg equilibrium and for the presence of null alleles (non amplifying alleles) using CERVUS software (Marshall et al. 1998; Kalinowski et al. 2007; See Table 1). We calculated a weighted multilocus homozygosity index for 119 male elephants where we had both footprint measurements and genetic data. We used the method of Aparicio et al. (2006) to calculate homozygosity. We then derived multilocus heterozygosity by subtracting the homozygosity value for each individual from one.

Table 1

An analysis of allele frequency in the Amoseli elephant population showing that all loci did not differ significantly (as shown in the column of probability values) from the expected frequency under the Hardy-Weinberg equilibrium

Locus Number of alleles per locus Number of individuals typed Observed frequency Expected frequency P value 
LAF2 575 0.722 0.743 0.574 
LAT5 14 586 0.865 0.857 0.691 
LAT7 20 580 0.916 0.917 0.261 
LAT8 15 586 0.851 0.841 0.311 
LAT13 586 0.706 0.717 0.627 
LAT16 10 582 0.831 0.842 0.272 
LAT17 13 575 0.857 0.821 0.134 
LAT18 12 573 0.819 0.827 0.337 
LAT24 11 584 0.818 0.827 0.128 
LAT25 556 0.817 0.808 0.208 
LAT26 13 530 0.868 0.874 0.882 
Locus Number of alleles per locus Number of individuals typed Observed frequency Expected frequency P value 
LAF2 575 0.722 0.743 0.574 
LAT5 14 586 0.865 0.857 0.691 
LAT7 20 580 0.916 0.917 0.261 
LAT8 15 586 0.851 0.841 0.311 
LAT13 586 0.706 0.717 0.627 
LAT16 10 582 0.831 0.842 0.272 
LAT17 13 575 0.857 0.821 0.134 
LAT18 12 573 0.819 0.827 0.337 
LAT24 11 584 0.818 0.827 0.128 
LAT25 556 0.817 0.808 0.208 
LAT26 13 530 0.868 0.874 0.882 

Identification of crop raiders

Identification of crop raiders was a multistep process (Chiyo et al. 2011). In the field, we either identified crop raiders by tracking elephants for several hours following a farm raid or, when we were not able to track raiders, by collecting their dung from raided farms. For elephants that we tracked and located, we could visually identify known members of the Amboseli-born elephant population. In cases, where we were not immediately able to ascertain their identities, we took photos and later matched these photos with a database of all Amboseli-born males. Dung collected from raided farms was preserved in 95% ethanol in the field and later brought to the lab at Duke University. In the lab, we extracted and genotyped DNA from the dung of these unknown crop raiders that we collected from raided farms. We genotyped on average 6 loci (LAFMS02, LaT05, LaT08, LaT13, LaT16, and LaT24) for samples collected from farms. We then compared the genotypes of crop-raiding elephants collected over a period of 2 years with genotypes of 586 known male and female elephants.

In order to match genotypes, we had to determine the minimum number of loci required to discriminate between genetic samples collected from different individuals. We did this by calculating the probability of identity (PI), that is, the probability that a pair of animals will match at a specified number of loci. Previous studies have identified a PI threshold of 0.0001 as sufficient for discriminating between genotypes of different individuals (Waits et al. 2001; Creel et al. 2003), we therefore sought to identify the number of loci that would provide a similar threshold for our study population. We calculated PI from allele frequency using the formula provided by Waits et al. (2001). This formula is based on a theoretical expectation of Hardy–Weinberg equilibrium. From calculated PI values, genotyping 4 loci (i.e., PI = 0.00004) was sufficient for individual identification. We therefore treated 2 genotype samples as coming from the same individual if 4 or more loci matched. We also allowed for a mismatch at a maximum of 1 additional locus for matched pairs to account for possible genotyping error. Using this criterion, we matched similar genotypes using the CERVUS software (Marshall et al. 1998; Kalinowski et al. 2007).

Using the fecal genotyping and elephant observations, we confirmed the identities of 52 Amboseli elephants that were crop raiders. In combination with population simulations (Chiyo et al. 2011), we estimated that a total of approximately 84 elephants raid crops in Amboseli, suggesting that there are likely a few raiders among the elephants we classified as nonraiders (false negatives). We do not expect this to significantly affect our result because the proportion of unidentified raiders that might occur in the sample of “nonraiders” was small relative to true nonraiders (∼15%), and any effect of these false negatives will be to weaken our power to detect the effect of raiding on size for age.

Statistical analyses

Our goal was to test the hypothesis that raiders are large for size as a result of raiding crops. However, our footprint size data were unbalanced; we had genetic data for some but not all individuals, and we had repeated measurements from a small fraction of individuals taken at irregular intervals. Our ability to estimate growth rate directly by fitting slopes of age for size for individual animals was limited because even among animals with repeated measures we had an average of only 2 measurements per individual. Consequently, we used a linear mixed-effects model framework to determine the effects of raiding behavior and multilocus heterozygosity on elephant size for age. Elsewhere, residuals of foot length for age from sex-specific von Bertalanffy growth curve (Lee and Moss 1995) have been used, but the mixed effects model was used here because it is an effective formal way to deal with repeated measurements and an unbalanced data set. We carried out these analyses on log-transformed male age ln(Xi + 1) and footprint ln(Yi) data because the relationship between footprint size and age was nearly asymptotic. We added 1 to each male age (Xi) because some values for male age were close to zero at the time of footprint measurement, making it impossible to normalize data through a log transformation. Multilocus heterozygosity was also transformed using arcsine transformation.

We ran 2 separate analyses based on male life stages because we expected crop raiding to affect male size only during a later life stage (or 16 years and older) and not during an early life stage (or 10 years and younger). The first analysis included males 16 years and older because males at this age are independent and spend most of the time away from their mothers with other males where they may initiate raiding. Sixteen years of age is also the youngest age an elephant was observed to raid in Amboseli and follows the mean age at independence from the natal family in Amboseli (average age is 14 years, and range is from 8–19 years; Lee and Moss 1999). In a second analysis, we included only males aged 0–10 years old because males in this age category are unlikely to raid because they are still with their natal family and are thus spending most of their time with their mothers; we found no evidence for raiding by females in Amboseli (Chiyo et al. 2011).

For each male life stage (0–10 years and 16 years and above), we first fitted a full mixed-effects model that included footprint size as a response variable and male age, male raiding status, multilocus heterozygosity, interaction between age and raiding status, and the interaction between age and heterozygosity as fixed effects. For each model, we used a random intercept and individual elephant ID as a nesting factor because we had repeated measurements for some elephants. We then dropped nonsignificant fixed effects from the full model in a stepwise manner starting with the least significant until we were left with a model with only significant fixed effects (Pinheiro and Bates 2004). In all cases in which we used 3 or more fixed effect factors, age was the only significant fixed effect for each elephant life stage analysis. As a result, we ran 2 models for each life stage analysis; a model including raiding status and age, and a model including heterozygosity and age. We fitted these 2 models separately in order to avoid Type II error because of low statistical power, which could result from the inclusion of many parameters in the model or from a reduction in sample size after the elimination of unbalanced covariate data.

All statistical analyses were carried out in S+, 8.1 (TIBCO Software Inc, Palo Alto, CA).

RESULTS

Crop-raiding effects on size

Crop raiding had a positive effect on foot size for age for males 16 years and older but not for males 0–10 years (Table 2), such that males observed to raid crops were larger for age than those not observed to raid (Figure 1). However, males that became raiders were not larger for age when young (0–10 years old) compared with same-aged males that were not subsequently observed to raid (Figure 1, Table 2). In other words, raiders became large for age only after they began raiding, suggesting that raiding affected male body size rather than male body size determining raiding status.

Table 2

Coefficients for the effect of crop-raiding status and genetic heterozygosity on foot size for age in male elephants presented separately for males 0–10 years old, before individuals that became raiders initiated raiding and 16 and more years of age after they initiate raiding

Covariate Model: fixed effects Coefficient (SE) df P value 
Crop-raiding effects: 
    On males 0–10 years Intercept 2.856 (0.016) 29 <0.001 
    Age 0.291 (0.010) 29 <0.001 
    Raiding status −0.012 (0.022) 101 0.604 
    On males 16+ years Intercept 3.141 (0.025) 236 <0.001 
    Age 0.219 (0.008) 236 <0.001 
    Raiding status 0.028 (0.008) 195 <0.001 
Heterozygosity effects: 
    On males 0–10 years Intercept 2.910 (0.052) 11 <0.001 
    Age 0.274 (0.019) 11 <0.001 
    Heterozygosity −0.015 (0.041) 26 0.721 
    On males 16+ years Intercept 3.145 (0.033) 154 <0.001 
    Age 0.218 (0.010) 154 <0.001 
    Heterozygosity 0.007 (0.014) 97 0.646 
Covariate Model: fixed effects Coefficient (SE) df P value 
Crop-raiding effects: 
    On males 0–10 years Intercept 2.856 (0.016) 29 <0.001 
    Age 0.291 (0.010) 29 <0.001 
    Raiding status −0.012 (0.022) 101 0.604 
    On males 16+ years Intercept 3.141 (0.025) 236 <0.001 
    Age 0.219 (0.008) 236 <0.001 
    Raiding status 0.028 (0.008) 195 <0.001 
Heterozygosity effects: 
    On males 0–10 years Intercept 2.910 (0.052) 11 <0.001 
    Age 0.274 (0.019) 11 <0.001 
    Heterozygosity −0.015 (0.041) 26 0.721 
    On males 16+ years Intercept 3.145 (0.033) 154 <0.001 
    Age 0.218 (0.010) 154 <0.001 
    Heterozygosity 0.007 (0.014) 97 0.646 

The coefficients indicated for age and intercept represent the baseline coefficient for non crop-raiding males, and the coefficient for raiding status represents an additional effect of raiding on size. Elephant ID is incorporated into the model as a random intercept. Male age (+1) in years and foot size in centimeters were transformed into the natural logarithm, and we used arcsine transformation for heterozygoszity. We used a random intercept for all the models displayed. SE, standard error; df, degrees of freedom.

Figure 1

Effect of crop raiding on size for age in males 16+ years of age showing that both raiders (solid line) and nonraiders (dashed line) have the same slope (see Age and Intercept in Table 2) but different intercepts (see raiding status in Table 2). Size is presented as a log-transformed measure of footprint size (see main text). Data for raiders are shown as solid diamonds, whereas data for nonraiders are shown as open circles.

Figure 1

Effect of crop raiding on size for age in males 16+ years of age showing that both raiders (solid line) and nonraiders (dashed line) have the same slope (see Age and Intercept in Table 2) but different intercepts (see raiding status in Table 2). Size is presented as a log-transformed measure of footprint size (see main text). Data for raiders are shown as solid diamonds, whereas data for nonraiders are shown as open circles.

Microsatellite heterozygosity effects on size

Multilocus heterozygosity had no significant effect on size for age for males 10 years or younger or males 16 years and older (Table 2). We did not explore locus-specific effects because of a lack of a general effect or trend in the predicted direction as well as the lack of a priori expectations of locus-specific heterozygosity on size for age. The distribution of multilocus heterozygosity in the Amboseli population was skewed toward higher values (Figure 2). The mean (± standard deviation) heterozygosity for 119 males from this population was (0.8 ± 0.147).

Figure 2

Distribution of mean multilocus heterozygosity using microsatellites in male Amboseli elephants, showing that multilocus heterozygosity was generally high in this population. Heterozygosity was calculated as 1-(HL, homozygosity by loci) estimated using the method of Aparicio, et al. (2006).

Figure 2

Distribution of mean multilocus heterozygosity using microsatellites in male Amboseli elephants, showing that multilocus heterozygosity was generally high in this population. Heterozygosity was calculated as 1-(HL, homozygosity by loci) estimated using the method of Aparicio, et al. (2006).

DISCUSSION

Our results indicate that foraging behavior such as raiding crops with a high risk of injury or mortality can lead to gains in body size. Although we cannot unambiguously identify the causal link between body size and crop raiding (and we did not explicitly model interactions because of the unbalance measurements we had for raiders), the fact that our data indicated that raiders were not largefor age before they began raiding suggests that crop raiding caused males to be large for age. This observation supports our hypothesis that high-risk foraging has an energetic pay off and suggests that the relative dearth of females among crop raiders relates to the fact that males, the sex experiencing high variance in reproductive success (Poole 1989), are more likely to take on these mortality risks in order to obtain the energy and growth payoff. An enhanced size for age as a result of crop raiding is likely to confer reproductive benefits to male elephants for 2 reasons.

First, their age at the onset of musth, the physiological state of heightened sexual and aggressive behavior, will be younger for larger males resulting in a longer breeding lifespan (Lee et al. 2011). Male elephants experience delayed siring of offspring (Hollister-Smith et al. 2007) that extends many years past their maturation in reproductive physiology (Owen-Smith 1988). The delay in siring by young reproductively mature males due to intense competition from older males provides an opportunity for an extended period of investment in growth and size in young males.

Second, annual reproductive performance is positively correlated with musth duration (Poole 1989; Hollister-Smith et al. 2007), and the duration of musth is dependent on condition and body size as well as age of individual males. Males with access to reliable, easily digested, and high energy human crops experience longer musth episodes while those with limited energy are less likely to experience musth (Poole 1989; Sukumar 2003). Larger males may also have increased reproductive success because they are preferred by females as mates (Moss 1983; Poole 1989; Hollister-Smith et al. 2007).

Microsatellite heterozygosity has been shown to be positively correlated with fitness traits in many populations, but in this study, mean individual heterozygosity of microsatellites loci was not correlated with size. Correlations between heterozygosity of neutral markers and fitness traits, such as size and growth rate are weak or absent in outbred populations and stronger in inbred populations (Rowe and Beebee 2001; Hildner et al. 2003; Overall et al. 2005). The lack of a relationship between multilocus heterozygosity and size for age in the Amboseli elephant population may simply reflect the high heterozygosity and extensive outbreeding in this population. This outbreeding is a consequence of inbreeding avoidance (Archie et al. 2007) and extensive gene flow between Amboseli and adjacent elephant populations.

Our findings on the sex-specific nature of high-risk behavior and the consequences of such behavior to gains in body size may generalize to other sexually dimorphic species with high reproductive variance because high-risk foraging is widespread among males of many sexually dimorphic species. Examples include fallow deer, (Apollonio et al. 2005) roe deer (Mysterud et al. 1999), moose (Miquelle et al. 1992), elk (Winnie and Creel 2007), Dall's sheep (Corti and Schackleton 2002), bighorn sheep (Berger 1991; Mooring et al. 2003), mountain sheep (Bleich et al. 1997), and African buffalo (Hay et al. 2008). The prevalence of additional physiological traits in species exhibiting this behavior such as condition-dependent mating strategies (musth and rut) and prolonged growth in males suggests that these traits are related to the evolution of high-risk foraging behavior in males of sexually dimorphic species.

Our findings also relate to the fact that high-risk foraging behavior such as crop raiding is a major cause of wildlife–human conflict. Wildlife–human conflict is a major conservation issue for many polygynous mammals like primates and carnivores that live in proximity to agriculture or livestock (Bunnefeld et al. 2006; Wilson et al. 2007; Hazzah et al. 2009). Because these findings suggest that crop raiding or livestock killing by wildlife is a male life-history tactic commonly observed in polygynous mammals, conflict will likely be a recurring problem in most wildlife populations found in proximity to humans. Managing such conflict will require measures that separate agriculture or livestock from conservation areas through use of wildlife barriers (Hoare 1995). It will also require land use practices that are compatible with wildlife conservation in proximity to protected areas through zoning (Linnell et al. 2005; Woodroffe et al. 2005). Alternatively, coexistence between wildlife and humans can also be encouraged by use of measures that increase public tolerance for the damage caused by wildlife (Woodroffe et al. 2005). On the other hand, approaches that directly target males such as control shooting are likely to have limited effect in reducing conflict (e.g., Hoare 2001) and will negatively influence the evolution of body size in these populations. Because high-risk foragers grow large, culling of males for purposes of reducing conflict is likely to eliminate individuals that are larger and reproductively competitive. Evolutionary consequences of selective killing have been demonstrated elsewhere. For example, hunting pressure has been demonstrated to decrease body size and horn size of male bighorn sheep (Coltman et al. 2003) and to increase the incidence of tusklessness in African and Asian elephants (Jachmann et al. 1995).

FUNDING

US Fish and Wildlife Service (grant number AFE-0314/6-G085) and the National Science Foundation (grant number IBN0091612) to S.C.A. Further support was provided by the Amboseli Trust for Elephants.

We thank the Office of the President of the Government of Kenya for permission to conduct this research through Ministry of Education, department of Science and Technology permit number 13/001/35C 225. We also thank Kenya Wildlife Services and Amboseli National Park staff for their hospitality and collaboration during the study. Finally, we thank Norah Njiraini, Katito Sayialel, and Soila Sayialel for their invaluable knowledge on elephant identification and for their support in the field.

References

Alberts
SC
Watts
HE
Altmann
J
Queuing and queue-jumping: long-term patterns of reproductive skew in male savannah baboons, Papio cynocephalus
Anim Behav
 , 
2003
, vol. 
65
 (pg. 
821
-
840
)
Aparicio
JM
Ortego
J
Cordero
PJ
What should we weigh to estimate heterozygosity, alleles or loci?
Mol Ecol
 , 
2006
, vol. 
15
 (pg. 
4659
-
4665
)
Apollonio
M
Ciuti
S
Luccarini
S
Long-term influence of human presence on spatial sexual segregation in fallow deer (Dama dama)
J Mammal
 , 
2005
, vol. 
86
 (pg. 
937
-
946
)
Archie
EA
Hollister-Smith
JA
Poole
JH
Lee
PC
Moss
CJ
Maldonado
JE
Fleischer
RC
Alberts
SC
Behavioural inbreeding avoidance in wild African elephants
Mol Ecol
 , 
2007
, vol. 
16
 (pg. 
4138
-
4148
)
Archie
EA
Maldonado
JE
Hollister-Smith
JA
Poole
JH
Moss
CJ
Fleischer
RC
Alberts
SC
Fine-scale population genetic structure in a fission-fusion society
Mol Ecol
 , 
2008
, vol. 
17
 (pg. 
2666
-
2679
)
Archie
EA
Moss
CJ
Alberts
SC
Characterization of tetranucleotide microsatellite loci in the African savannah elephant (Loxodonta africana africana)
Mol Ecol Notes
 , 
2003
, vol. 
3
 (pg. 
244
-
246
)
Archie
EA
Moss
CJ
Alberts
SC
The ties that bind: genetic relatedness predicts the fission and fusion of social groups in wild African elephants
Proc R Soc B Biol Sci
 , 
2006
, vol. 
273
 (pg. 
513
-
522
)
Barnes
RFW
Dubiure
UF
Danquah
E
Boafo
Y
Nandjui
A
Hema
EM
Manford
M
Crop-raiding elephants and the moon
Afr J Ecol
 , 
2007
, vol. 
45
 (pg. 
112
-
115
)
Berger
J
Pregnancy incentives, predation constraints and habitat shifts: experimental and field evidence for wild bighorn sheep
Anim Behav
 , 
1991
, vol. 
41
 (pg. 
61
-
77
)
Bleich
VC
Terry
BR
Wehausen
JD
Sexual segregation in mountain sheep: resources or predation?
Wildl Monogr
 , 
1997
, vol. 
134
 (pg. 
3
-
50
)
Bowyer
RT
Bleich
VC
Manteca
X
Whiting
JC
Stewart
KM
Sociality, mate choice, and timing of mating in American bison (Bison bison): effects of large males
Ethology
 , 
2007
, vol. 
113
 (pg. 
1048
-
1060
)
Bunnefeld
N
Linnell
JDC
Odden
J
van Duijn
MAJ
Andersen
R
Risk taking by Eurasian lynx (Lynx lynx) in a human dominated landscape: effects of sex and reproductive status
J Zool
 , 
2006
, vol. 
270
 (pg. 
31
-
39
)
Charlton
BD
Reby
D
McComb
K
Female red deer prefer the roars of larger males
Biology Letters
 , 
2007
, vol. 
3
 (pg. 
382
-
385
)
Charpentier
M
Setchell
JM
Prugnolle
F
Wickings
EJ
Peignot
P
Balloux
F
Hossaert-Mickey
M
Life history correlates of inbreeding depression in mandrills (Mandrillus sphinx)
Mol Ecol
 , 
2006
, vol. 
15
 (pg. 
21
-
28
)
Cheeran
JV
Zachariah
A
Subash
CK
Easwaran
EK
Jayewardene
J
Interpretation of wildlife diseases as manifestation of human elephant conflict in Wayanad Wildlife Sanctuary-Kerala, India
Endangered elephants, past present and future
 , 
2004
Colombo (Sri Lanka)
Biodiversity & Elephant Conservation Trust
(pg. 
187
-
188
)
Chiyo
PI
Cochrane
EP
Population structure and behaviour of crop-raiding elephants in Kibale National Park, Uganda
Afr J Ecol
 , 
2005
, vol. 
43
 (pg. 
233
-
241
)
Chiyo
PI
Moss
CJ
Archie
EA
Hollister-Smith
JA
Alberts
SC
Using molecular and observational techiniques to estimate the number and raiding patterns of crop-raiding elephants
J Appl Ecol
 , 
2011
 
(in press)
Cohas
A
Bonenfant
C
Kempenaers
B
Allaine
D
Age-specific effect of heterozygosity on survival in alpine marmots, Marmota marmota
Mol Ecol
 , 
2009
, vol. 
18
 (pg. 
1491
-
1503
)
Coltman
DW
Bowen
WD
Wright
JM
Birth weight and neonatal survival of harbour seal pups are positively correlated with genetic variation measured by microsatellites
Proc R Soc Lond Ser B Biol Sci
 , 
1998
, vol. 
265
 (pg. 
803
-
809
)
Coltman
DW
O'Donoghue
P
Jorgenson
JT
Hogg
JT
Strobeck
C
Festa-Bianchet
M
Undesirable evolutionary consequences of trophy hunting
Nature
 , 
2003
, vol. 
426
 (pg. 
655
-
658
)
Corti
P
Schackleton
DM
Relationship between predation-risk factors and sexual segregation in Dall's sheep (Ovis dalli dalli)
Can J Zool
 , 
2002
, vol. 
80
 (pg. 
2108
-
2117
)
Creel
S
Spong
G
Sands
JL
Rotella
J
Zeigle
J
Joe
L
Murphy
KM
Smith
D
Population size estimation in Yellowstone wolves with error-prone noninvasive microsatellite genotypes
Mol Ecol
 , 
2003
, vol. 
12
 (pg. 
2003
-
2009
)
Croze
H
Moss
C
Moss
C
Croze
H
Lee
PC
Patterns of occupancy in time and space
The Amboseli Elephants: a long-term perspective on a long-lived mammal
 , 
2011
Chicago (IL)
University of Chicago Press
(pg. 
89
-
105
)
Curik
I
Zechner
P
Solkner
J
Achmann
R
Bodo
I
Dovc
P
Kavar
T
Marti
E
Brem
G
Inbreeding, microsatellite heterozygosity, and morphological traits in Lipizzan horses
J Hered
 , 
2003
, vol. 
94
 (pg. 
125
-
132
)
Da Silva
A
Luikart
G
Yoccoz
NG
Cohas
A
Allaine
D
Genetic diversity-fitness correlation revealed by microsatellite analyses in European alpine marmots (Marmota marmota)
Conserv Genet
 , 
2006
, vol. 
7
 (pg. 
371
-
382
)
Douglas-Hamilton
I
Krink
T
Vollrath
F
Movements and corridors of African elephants in relation to protected areas
Naturwissenschaften
 , 
2005
, vol. 
92
 (pg. 
158
-
163
)
Field
IC
Bradshaw
CJA
Burton
HR
Hindell
MA
Differential resource allocation strategies in juvenile elephant seals in the highly seasonal southern ocean
Mar Ecol Prog Ser
 , 
2007
, vol. 
331
 (pg. 
281
-
290
)
Gage
MJG
Surridge
AK
Tomkins
JL
Green
E
Wiskin
L
Bell
DJ
Hewitt
GM
Reduced heterozygosity depresses sperm quality in wild rabbits, Oryctolagus cuniculus
Curr Biol
 , 
2006
, vol. 
16
 (pg. 
612
-
617
)
Graham
MD
Douglas-Hamilton
I
Adams
WM
Lee
PC
The movement of African elephants in a human-dominated land-use mosaic
Anim Conserv
 , 
2009
, vol. 
12
 (pg. 
445
-
455
)
Haigh
JC
Parker
ISC
Parkinson
DA
Archer
AL
An elephant extermination
Environ Conserv
 , 
1979
, vol. 
6
 (pg. 
305
-
310
)
Hansson
B
Westerdahl
H
Hasselquist
D
Ãkesson
M
Bensch
S
Hey
J
Does linkage disequilibrium generate heterozygosity-fitness correlations in great reed warblers?
Evolution
 , 
2004
, vol. 
58
 (pg. 
870
-
879
)
Hay
CT
Cross
PC
Funston
PJ
Trade-offs of predation and foraging explain sexual segregation in African buffalo
J Anim Ecol
 , 
2008
, vol. 
77
 (pg. 
850
-
858
)
Hazzah
L
Borgerhoff Mulder
M
Frank
L
Lions and warriors: social factors underlying declining African lion populations and the effect of incentive-based management in Kenya
Biol Conserv
 , 
2009
, vol. 
142
 (pg. 
2428
-
2437
)
Hildner
KK
Soulé
ME
Min
M-S
Foran
DR
The relationship between genetic variability and growth rate among populations of the pocket gopher, Thomomys bottae
Conserv Genet
 , 
2003
, vol. 
4
 (pg. 
233
-
240
)
Hoare
R
Management implications of new research on problem elephants
Pachyderm
 , 
2001
, vol. 
30
 (pg. 
44
-
48
)
Hoare
RE
Options for the control of elephants in conflict with people
Pachyderm
 , 
1995
, vol. 
19
 (pg. 
54
-
63
)
Hollister-Smith
JA
Poole
JH
Archie
EA
Vance
EA
Georgiadis
NJ
Moss
CJ
Alberts
SC
Age, musth and paternity success in wild male African elephants, Loxodonta africana
Anim Behav
 , 
2007
, vol. 
74
 (pg. 
287
-
296
)
Isaac
JL
Sexual dimorphism in a marsupial: seasonal and lifetime differences in sex-specific mass
Aust J Zool
 , 
2006
, vol. 
54
 (pg. 
45
-
50
)
Jachmann
H
Berry
PSM
Imae
H
Tusklessness in African elephants: a future trend
Afr J Ecol
 , 
1995
, vol. 
33
 (pg. 
230
-
235
)
Kalinowski
ST
Taper
ML
Marshall
TC
Revising how the computer program CERVUS accommodates genotyping error increases success in paternity assignment
Mol Ecol
 , 
2007
, vol. 
16
 (pg. 
1099
-
1106
)
Kanchanapangka
S
Supawong
S
Koedlab
K
Kaewpannarai
J
Khawnual
P
Tummaruk
P
Sajjarengpong
K
Body weight formulation in Asian elephant
Thai J Vet Med
 , 
2007
, vol. 
37
 (pg. 
49
-
58
)
Lee
PC
Moss
CJ
Statural growth in known-age African elephants (Loxodonta africana)
J Zool
 , 
1995
, vol. 
236
 (pg. 
29
-
41
)
Lee
PC
Moss
CJ
Box
HO
Gibson
KR
The social context for learning and behavioral development among wild African elephants
Mammalian social learning: comparative and ecological perspectives
 , 
1999
London
Cambridge University Press
(pg. 
102
-
125
)
Lee
PC
Poole
JH
Njiraini
N
Moss
CJ
Moss
CJ
Croze
HJ
Lee
PC
Male social dynamics: Independence and beyond
The Amboseli elephants: a long-term perspective on a long-lived mammal
 , 
2011
Chicago (IL)
University of Chicago Press
(pg. 
260
-
271
)
Linnell
JDC
Nilsen
EB
Lande
US
Herfindal
I
Odden
J
Skogen
K
Andersen
R
Breitenmoser
U
Woodroffe
R
Thirgood
S
Rabinowitz
A
Zonning as a means of mitigating conflict with large carnivores: principles and reality
People and wildlife: conflict or coexistence?
 , 
2005
Cambridge (UK)
Cambridge University Press
(pg. 
162
-
175
)
Liu
GQ
Jiang
XP
Wang
JY
Wang
ZY
Correlations between heterozygosity at microsatellite loci, mean d(2) and body weight in a Chinese native chicken
Asian Australas J Anim Sci
 , 
2006
, vol. 
19
 (pg. 
1671
-
1677
)
Loison
A
Langvatn
R
Solberg
EJ
Body mass and winter mortality in red deer calves: disentangling sex and climate effects
Ecography
 , 
1999
, vol. 
22
 (pg. 
20
-
30
)
MacFarlane
AM
Coulson
G
Sexual segregation in western grey kangaroos: testing alternative evolutionary hypotheses
J Zool
 , 
2007
, vol. 
273
 (pg. 
220
-
228
)
Marshall
TC
Slate
J
Kruuk
LEB
Pemberton
JM
Statistical confidence for likelihood-based paternity inference in natural populations
Mol Ecol
 , 
1998
, vol. 
7
 (pg. 
639
-
655
)
McElligott
AG
Gammell
MP
Harty
HC
Paini
DR
Murphy
DT
Walsh
JT
Hayden
TJ
Sexual size dimorphism in fallow deer (Dama dama): do larger, heavier males gain greater mating success?
Behav Ecol Sociobiol
 , 
2001
, vol. 
49
 (pg. 
266
-
272
)
Miquelle
DG
Peek
JM
Vanballenberghe
V
Sexual segregation in Alaskan moose
Wildl Monogr
 , 
1992
, vol. 
122
 (pg. 
1
-
57
)
Mooring
MS
Fitzpatrick
TA
Benjamin
JE
Fraser
IC
Nishihira
TT
Reisig
DD
Rominger
EM
Sexual segregation in desert bighorn sheep (Ovis canadensis mexicana)
Behaviour
 , 
2003
, vol. 
140
 (pg. 
183
-
207
)
Moss
CJ
Oestrous behaviour and female choice in the African elephant
Behaviour
 , 
1983
, vol. 
86
 (pg. 
167
-
195
)
Moss
CJ
The demography of an African elephant (Loxodonta africana) population in Amboseli, Kenya
J Zool
 , 
2001
, vol. 
255
 (pg. 
145
-
156
)
Mpanduji
DG
Hahn
R
Siege
L
Baldus
RD
Hildebrandt
TB
Goeritz
F
East
ML
Hofer
H
Jayewardene
J
Conflicts between humans, elephants and other wildlife in Songea rural district, southern Tanzania
Endangered elephants, past present and future
 , 
2004
Colombo (Sri Lanka)
Biodiversity and Elephant Conservation Trust
(pg. 
82
-
85
)
Mysterud
A
Lian
L-B
Hjermann
Scale-dependent trade-offs in foraging by European roe deer (Capreolus capreolus) during winter
Can J Zool
 , 
1999
, vol. 
77
 (pg. 
1486
-
1493
)
Nyakaana
S
Arctander
P
Isolation and characterization of microsatellite loci in the African elephant, Loxodonta africana
Mol Ecol
 , 
1998
, vol. 
7
 (pg. 
1436
-
1437
)
Obanda
V
Ndeereh
D
Mijele
D
Lekolool
I
Chege
S
Gakuya
F
Omondi
P
Injuries of free ranging African elephants (Loxodonta africana africana) in various ranges of Kenya
Pachyderm
 , 
2008
, vol. 
44
 (pg. 
54
-
58
)
Odden
J
Linnell
JDC
Moa
PF
Herfindal
I
Kvam
T
Andersen
R
Lynx depredation on domestic sheep in Norway
J Wildl Manag
 , 
2002
, vol. 
66
 (pg. 
98
-
105
)
Overall
ADJ
Byrne
KA
Pilkington
JG
Pemberton
JM
Heterozygosity, inbreeding and neonatal traits in soay sheep on St. Kilda
Mol Ecol
 , 
2005
, vol. 
14
 (pg. 
3383
-
3393
)
Owen-Smith
NR
Megaherbivores: the influence of very large body size on ecology
 , 
1988
Cambridge
Cambridge University Press
Pinheiro
JC
Bates
DM
Mixed-effects models in S and S-Plus
 , 
2004
New York
Springer Science and Business Media, LLC
Poole
JH
Mate guarding, reproductive success and female choice in African elephants
Anim Behav
 , 
1989
, vol. 
37
 (pg. 
842
-
849
)
Post
E
Langvatn
R
Forchhammer
MC
Stenseth
NC
Environmental variation shapes sexual dimorphism in red deer
Proc Natl Acad Sci U S A
 , 
1999
, vol. 
96
 (pg. 
4467
-
4471
)
Rode
KD
Chiyo
PI
Chapman
CA
McDowell
LR
Nutritional ecology of elephants in Kibale National Park, Uganda, and its relationship with crop-raiding behaviour
J Trop Ecol
 , 
2006
, vol. 
22
 (pg. 
441
-
449
)
Rowe
G
Beebee
TJC
Fitness and microsatellite diversity estimates were not correlated in two outbred anuran populations
Heredity
 , 
2001
, vol. 
87
 (pg. 
558
-
565
)
Setchell
JM
Charpentier
M
Wickings
EJ
Sexual selection and reproductive careers in mandrills (Mandrillus sphinx)
Behav Ecol Sociobiol
 , 
2005
, vol. 
58
 (pg. 
474
-
485
)
Slate
J
Kruuk
LEB
Marshall
TC
Pemberton
JM
Clutton-Brock
TH
Inbreeding depression influences lifetime breeding success in a wild population of red deer (Cervus elaphus)
Proc R Soc Lond Ser B Biol Sci
 , 
2000
, vol. 
267
 (pg. 
1657
-
1662
)
Struhsaker
TT
Pope
TR
Mating system and reproductive success—a comparison of two African forest monkeys (Colobus badius and Cercopithecus ascanius)
Behaviour
 , 
1991
, vol. 
117
 (pg. 
182
-
205
)
Sukumar
R
Ecology of the Asian elephant in southern India. 2. Feeding habits and crop raiding patterns
J Trop Ecol
 , 
1990
, vol. 
6
 (pg. 
33
-
53
)
Sukumar
R
Elephant raiders and rogues
Nat Hist
 , 
1995
, vol. 
104
 (pg. 
52
-
60
)
Sukumar
R
The living elephants: evolutionary ecology, behavior and conservation
 , 
2003
Oxford
Oxford University Press
Sukumar
R
Gadgil
M
Male-female differences in foraging on crops by Asian elephants
Anim Behav
 , 
1988
, vol. 
36
 (pg. 
1233
-
1235
)
Waits
LP
Luikart
G
Taberlet
P
Estimating the probability of identity among genotypes in natural populations: cautions and guidelines
Mol Ecol
 , 
2001
, vol. 
10
 (pg. 
249
-
256
)
Western
D
Moss
CJ
Georgiadis
N
Age estimation and population age structure of elephants from footprint dimensions
J Wildl Manag
 , 
1983
, vol. 
47
 (pg. 
1192
-
1197
)
Whitehead
H
Delayed competitive breeding in roving males
J Theor Biol
 , 
1994
, vol. 
166
 (pg. 
127
-
133
)
Williams
AC
Johnsingh
AJT
Krausman
PR
Elephant-human conflicts in Rajaji National Park, north-western India
Wildl Soc Bull
 , 
2001
, vol. 
29
 (pg. 
1097
-
1104
)
Wilson
ML
Hauser
MD
Wrangham
RW
Chimpanzees (Pan troglodytes) modify grouping and vocal behaviour in response to location-specific risk
Behaviour
 , 
2007
, vol. 
144
 (pg. 
1621
-
1653
)
Winnie
JJ
Creel
S
Sex-specific behavioural responses of elk to spatial and temporal variation in the threat of wolf predation
Anim Behav
 , 
2007
, vol. 
73
 (pg. 
215
-
225
)
Woodroffe
R
Thirgood
S
Rabinowitz
A
Woodroffe
R
Thirgood
S
Rabinowitz
A
The future of coexistence: resolving human-wildlife conflicts in a changing world
People and wildlife: conflict or coexistence?
 , 
2005
Cambridge (UK)
Cambridge University Press
(pg. 
388
-
405
)
Zedrosser
A
Bellemain
E
Taberlet
P
Swenson
JE
Genetic estimates of annual reproductive success in male brown bears: the effects of body size, age, internal relatedness and population density
J Anim Ecol
 , 
2007
, vol. 
76
 (pg. 
368
-
375
)
Zedrosser
A
Dahle
B
Swenson
JE
Population density and food conditions determine adult female body size in brown bears
J Mammal
 , 
2006
, vol. 
87
 (pg. 
510
-
518
)