The role of human hunters and natural predators in shaping the selection of behavioural types in male wild turkeys

Abstract

The expression of behaviour can vary both among (i.e. behavioural types (BTs)) and within individuals (i.e. plasticity), and investigating causes and consequences of variation has garnered significant attention. Conversely studies quantifying harvest-induced selection (HIS) on behavioural traits have received significantly less attention, and work investigating HIS and natural selection simultaneously is rare. We studied sources of variation in three movement traits that represented risk-taking and one trait that represented exploration in male eastern wild turkeys (Meleagris gallopavo silvestris). We used data from 109 males in two hunted populations located in Georgia and South Carolina, USA. We assessed how both hunters and natural predators simultaneously influenced the selection of male turkey BTs. We found significant among-individual variation in all movement traits and adjustments in risk-taking and exploration relative to whether hunting was occurring. We observed that predators selected against similar BTs across both populations, whereas hunters selected for different BTs across populations. We also demonstrated that significant HIS acts on risk-taking behaviours in both populations, which could render wild turkeys more difficult to harvest if these traits are indeed heritable.

Headnote

Keywords:

animal personality, harvest-induced selection, among-individual variation, natural selection, hunting, behavioural type

The expression of behaviour can vary both among (i.e. behavioural types (BTs)) and within individuals (i.e. plasticity), and investigating causes and consequences of variation has garnered significant attention. Conversely studies quantifying harvest-induced selection (HIS) on behavioural traits have received significantly less attention, and work investigating HIS and natural selection simultaneously is rare. We studied sources of variation in three movement traits that represented risk-taking and one trait that represented exploration in male eastern wild turkeys (Meleagris gallopavo silvestris). We used data from 109 males in two hunted populations located in Georgia and South Carolina, USA. We assessed how both hunters and natural predators simultaneously influenced the selection of male turkey BTs. We found significant among-individual variation in all movement traits and adjustments in risk-taking and exploration relative to whether hunting was occurring. We observed that predators selected against similar BTs across both populations, whereas hunters selected for different BTs across populations. We also demonstrated that significant HIS acts on risk-taking behaviours in both populations, which could render wild turkeys more difficult to harvest if these traits are indeed heritable.

1. Introduction

Phenotypic traits expressed in wild populations experience dynamic selective pressures that lead to shifts in trait optima over time [1,2]. Modifications in phenotypic traits are shaped by the environment and agents of natural selection, such as predators [3,4]. Predators play a significant role in driving phenotypic changes in prey species by non-randomly selecting for certain phenotypes that are easier to exploit [5-7]. Similarly, evidence is accumulating that humans are another major evolutionary force acting on phenotypes that often outpaces selection from natural drivers [8-11]. Human-induced changes in phenotypes are usually the most conspicuous when humans act as the primary predator by harvesting wild populations [12-15]. Harvesting can exert significant selective pressures by decreasing survival rates [16-18], altering population dynamics [19,20] and selectively targeting specific phenotypes in a process known as harvest-induced selection (HIS) [12,21-26].

An aspect of HIS that has received less attention in the literature involves the non-random selection of behavioural traits [12]. Non-random selection of behaviours stemming from harvest can be attributed to among-individual differences in behaviour (i.e. behavioural types (BTs)), wherein certain behavioural characteristics of individuals influence the likelihood of being harvested [12,21-23,26]. For instance, research on the movement behaviour of elk (Cervus elaphus) has demonstrated that individuals who express riskier behaviours, such as a preference for open areas, and those with greater exploratory tendencies are more likely to be harvested [21]. In terrestrial systems, HIS on behavioural traits has received limited attention collectively [12,21,22,26,27], but studies conducted underscore the harvest of individuals exhibiting a fast phenotype, characterized by greater risk-taking, activity and exploration [12,21-23,26]. Furthermore, most research on HIS has been conducted within the context of fisheries [15,25,28]. This body of work, including a recent review by Leclerc et al. [12], provided a detailed framework for exploring hypotheses and predictions about the influence of HIS on behavioural traits. For example, it is hypothesized that in hunted populations, traits that increase detectability to hunters, such as occurring closer to open areas, displaying boldness or greater activity levels, could make animals more susceptible to harvest [12]. Likewise, several studies within the fisheries literature have concurrently explored the impact of HIS and natural selection on behavioural traits, revealing divergent preferences for behavioural phenotypes between predators and humans [28,29]. Despite the insights revealed by examining HIS and natural selection simultaneously [27,28], there remains a notable absence of research on the combined effects of HIS and natural selection on the behaviour of terrestrial game species.

Male wild turkeys (Meleagris gallopavo silvestris) provide a unique opportunity to simultaneously assess HIS and natural selection, stemming from the fact that male wild turkeys are primarily harvested during the reproductive period, which is when natural predation for males is also heightened [16,17,30-32]. Wild turkeys are a non-migratory upland game bird that use a polygamous-promiscuous mating system in which females select the most attractive males for copulation [30]. To attract and secure mating opportunities, males engage in courtship behaviours such as vocalizations and elaborate displays [30-35], which are inherently risky behaviours that can attract predators. Males that engage in courtship behaviours will often use open landcover for displaying and vocalizing [3638] as open areas facilitate increased visual perception and reduced sound attenuation, increasing the opportunity for attracting potential receptive females at greater distances [39-41].

Previous studies have found that male wild turkeys alter movements during the hunting season and increase movements prior to and during peak incubation to search for receptive females [33,4244]. This period of increased competition and mate searching coincides with increased mortality risk from hunters and predators, with hunting accounting for most male mortalities [16,17,45]. Earlier work on the movement behaviour of wild turkey hunters demonstrated that hunters do not move far from hunting access points, and primarily use secondary roads which are comprised of extensive edges that contain early successional plant communities used by wild turkeys [46,47]. Moreover, edge landcover functions as travel corridors and foraging areas for significant wild turkey predators like bobcats (Lpnx rufus) and coyotes (Canis latrans), with the latter exhibiting increased interactions with male wild turkeys [42,48]. As such, behaviours involving the use of edge landcover, open landcover, closer proximity to hunting access points and greater exploratory tendencies pose significant mortality risks, as these actions increase the likelihood of encountering both predators and hunters [42,4648]. Nevertheless, despite the mortality risks linked to these behaviours, our understanding of how risk-taking and exploratory behaviours affect the survival of male wild turkeys remains limited.

Wild turkey abundance, productivity and harvest across broad areas of the species range have been steadily declining [49,50], whereas predators of wild turkeys have exhibited increases in abundance [51]. Despite numerous studies exploring survival rates, harvest rates and movement characteristics of male wild turkeys [16,17,42-44], there is still a limited understanding of the effects of HIS and natural selection on behavioural phenotypes. Moreover, although recent research has explored how hunting pressure affects the movement behaviour of male wild turkeys [42], there is a notable absence of studies comparing differences in movement behaviour across multiple locations subjected to hunting pressure. Our objectives were to investigate (i) whether wild turkeys are being harvested based on their BT (i.e. average phenotypic expression); (ii) whether wild turkeys with specific BTs were being non-randomly removed by predators; (iii) whether hunters and predators selected similar or divergent BTs; (iv) if hunters and predators were selecting for similar BTs across two hunted populations; and (v) whether behavioural responses to hunting pressure were similar across both hunted populations. We hypothesized that hunters and predators would both non-randomly select certain behavioural phenotypes that were easier to exploit. Similarly, we hypothesized that both predators and hunters would select similar BTs across both populations. We predicted that in both populations, hunters and predators would select BTs characterized by greater levels of risk-taking and exploration. Our logic for this prediction was that greater expressions of risk-taking and exploration would lead to increased interactions with both hunters and predators, ultimately negatively impacting male turkey survival. Finally, we hypothesized that male wild turkeys would adjust their behaviour in response to hunting pressure, and we predicted that adjustments of movement behaviour would be in a similar direction for each behavioural trait across both populations.

2. Methods

2.1. Study area

During 2017-2023, we conducted research on two wildlife management areas (WMAs) and surrounding private lands located in the Piedmont region of Georgia, USA. Cedar Creek WMA (CCWMA) was approximately 16 187 ha, B.F. Grant WMA (BFGWMA) was approximately 4613 ha and both were managed in cooperation with the Georgia Department of Natural Resources-Wildlife Resources Division (GADNR). The CCWMA was owned by the United States Forest Service, whereas BFGWMA was owned by the Warnell School of Forestry and Natural Resources at the University of Georgia. For analysis, we considered males in BFGWMA and CCWMA as belonging to a single population (hereafter referred to as Georgia) since the two WMAs were separated by 4.55 km and individuals from these sites clustered genetically into a single population [52]. During 2014-2018, we also conducted work on the Webb WMA Complex (hereafter referred to as South Carolina), which was approximately 10 483 ha and consisted of three contiguous WMAs (Hamiliton Ridge WMA, Palachacola WMA and Webb WMA) located near the Savannah River in Garnett, South Carolina, USA, and managed by the South Carolina Department of Natural Resources (SCDNR). In Georgia, a total of 1316 ha of open areas were available, featuring large agricultural fields and pastures (>10 ha), alongside smaller openings (<2 ha) that featured planted plots designed to attract wildlife. In contrast, in South Carolina only 180 ha of open areas were present, and openings primarily consisted of small food plots (<2 ha) established to attract wildlife. For further information on the study sites and hunting seasons at each site, see electronic supplementary material, SI.

2.2. Data collection

We captured wild turkeys from January to March annually using rocket nets baited with cracked corn, and upon capture each individual was marked with an aluminium leg band, and fitted with a remotely downloadable backpack-style GPS-VHF-UHF transmitter (Biotrack Ltd, Wareham, Dorest, UK and e-obs GmbH, Gruenwald, Germany [53]). We aged wild turkeys based on the presence of barring on the 9th and 10th primary feathers, and characterized individuals as either adults or juveniles [30]. Capture, handling and marking procedures were approved by the Institutional Animal Care and Use Committee at the University of Georgia (A2019 01-025 R2) and Louisiana State University (A2014-13).

Due to battery constraints of the GPS transmitters, data were only collected for approximately 1 year on each individual; thus we used data collected from 1 March until 31 July for each year of the study. We used data during this time period because it spanned the breeding season and hunting season in their entirety, which are when male wild turkeys face the greatest levels of mortality risk [16]. All GPS transmitters were programmed to record one location nightly (23:58:58), birds captured in Georgia recorded locations hourly from 05.00 to 20.00 hours, whereas in South Carolina locations were recorded every 30 minutes in accordance with project objectives [54]. We located each individual and downloaded GPS data using handheld antennas and receivers once a month and monitored survival 3-5 times a week. At mortality sites, we used field signs to identify predator guilds responsible for killing wild turkeys, such as avian predators like great horned owls (Bubo virginianus) and red-tailed hawks (Buteo jamaicensis) and mammalian predators like bobcats (Lynx rufus) and coyotes (Canis latrans). However, we combined all predator-related deaths due to insufficient sample sizes for analysing survival rates specific to predator guilds. We also sent wild turkeys that died with no visible trauma to the Southeastern Cooperative Wildlife Disease Study (SCWDS) at the University of Georgia, where they performed necropsies and X-rays. We assigned hunting mortalities to those recovered and reported directly by hunters or when we determined the cause of death resulted from crippling loss under the following three circumstances: a transmittered bird that was with another harvested transmittered bird and was found dead within 48 hours of the harvest and the carcass was indicative of crippling; a transmittered bird was known to be shot at by a hunter and visually wounded based on a conversation with the hunter, which likely resulted in mortality; and a transmitter bird was recovered dead during the wild turkey hunting season with no apparent signs of predation, and during necropsy we found evidence of gunshot pellets throughout the bird's carcass. We considered instances where the transmitter harness had been cut and discarded to be illegal kills and we included such mortalities as harvest. If a wild turkey was found dead or determined a hunter harvest, we truncated data to include only GPS points when the individual was alive. We cleaned raw GPS data and excluded any GPS locations that did not have adequate satellite coverage and had dilution of precision (DOP) values >7 [55].

2.3. Behavioural data

Juvenile male wild turkeys were rarely harvested in our study populations and contemporary evidence suggests that adults are primarily targeted by hunters [16]. Therefore, we used GPS data from adult male wild turkeys and focused on three metrics that we believed represented risk-taking (i.e. distance to edge landcover, distance to open landcover and distance to hunter access) and one metric that represented exploratory tendencies (average hourly speed). To calculate our risk-taking metrics, we first generated individual raster layers for forested (pine, hardwood, shrub and mix pine/hardwoods) and open landcover types using 30 m x 30 m raster layers from the United States Department of Agriculture (USDA), National Agricultural Statistical Service for each year of the study [56]. The USDA cropland data layers update annually and can account for changes in forest cover from active timber harvest, which occurred on both study sites. Next, we obtained road data from GADNR and SCDNR for roads inside the WMAs and used USGS Tiger/Line data (Topologically Integrated Geographic Encoding and Referencing) for roads outside the WMAs that traversed private lands. We characterized secondary roads as either logging roads and/or unpaved gravel roads that did not allow vehicular access, whereas primary roads were paved/gravelled and vehicle access was not limited. We created an edge landcover feature by identifying all areas of the landscape where the forested raster layers were adjacent to open raster layers for either road type. We then used ArcGIS 10.5.1 (Environmental System Research Institute, Inc., Redlands, GA, USA) to calculate a distance-based raster for edge and open landcover on both study sites. We completed the remainder of the analyses using R v. 4.3.1 [57]. We used package raster [58] and sf [59] to calculate the distance from GPS points to nearest open and edge landcover types, which allowed us to obtain continuous behavioural measurements rather than using classification or categorical approaches [60].

We recognize that at our Georgia study site, certain areas were unreachable to hunters from primary roads due to private property blocking access to public land, whereas in South Carolina, no private properties restricted access to public land. However, at both study sites, seasonal road closures occur to reduce vehicle travel rendering portions of both study sites more difficult to access. Therefore, we also computed the closest distance to hunter access points (see above methods) where the general public could park a vehicle and access public hunting areas, assuming that areas farther from hunter access points would experience less hunting activity [42,46,47]. For each distance to metric (i.e. distance to open landcover, edge landcover, hunter access points), we calculated the average daily distance for each male turkey and used these metrics for further analysis. Finally, to compute our metric for exploratory behaviour (i.e. average hourly speed) we calculated the total distance travelled within a

given day by each male turkey using package adehabitatLT [61] and sp [62]. We then divided that value by 15 (i.e. hours within a day that the GPS transmitter collected data) for each individual to calculate their average hourly speed.

2.4. Statistical analysis

We conducted our analysis in three steps, first by constructing four separate univariate models, each incorporating one of the four behavioural traits, using the brms package [63]. We used the function scaleO to standardize each response variable prior to analysis to assist with model convergence. Subsequently, we categorized GPS data into three separate time frames: the pre-hunt stage (from March to the beginning of the hunting season), the hunt stage (during the hunting season) and the post-hunt stage (from the day after the hunting season ended until 31 July). We then fitted random intercepts for individual wild turkeys, and fixed effects of hunting stage (pre-hunt, hunt, post-hunt), study population (Georgia, South Carolina) and an interaction between hunting stage and study population. We felt justified in including a fixed effect for hunting stage since prior research has demonstrated that male wild turkeys adjust their movement behaviour before, during and after the hunting season [42]. Likewise, we included an interaction between hunting stage and study population to statistically control and test differences in behavioural traits between study populations. Next, to assess the existence of among-individual variation (i.e. BTs) in each behavioural trait, we calculated repeatability, which measures the amount of phenotypic variance attributed to individual variation [64,65]. We considered any behavioural trait with a repeatability value of r > 0.3 as showing moderate repeatability [64]. To calculate repeatability, we extracted the posterior distribution for the among-individual variance and residual variance, and we used the following formula [64]:

(ProQuest: ... denotes formula omitted.) (2.1)

where Vindg was the among-individual variance and Veg was the residual within-individual variance.

For all univariate models, we performed 15 300 iterations of each model with a burn in of 300, a thinning interval of 15, and we used two Markov chains per model. We used the default prior in the brms package [63], and assessed convergence using R-hat values and visually inspected trace and density plots. We considered R-hat values < 1.1 to show adequate model convergence. For all models, we report the median (ß) of the posterior distribution, 95% credible intervals (Cri) and probability of direction (PD) (i.e. proportion of the posterior distribution that is in the direction of the estimated effect) to evaluate the support for each effect [66,67]. We describe effects that did not have CrI overlapping zero as showing strong support for an effect, and we considered CrI centred around zero (with an equal distribution on both sides of zero) with an effect estimate close to zero and a PD <89% as strong support for no effect. We considered CrI that overlapped zero with a PD >89% as showing moderate support for an effect [66,67].

2.5. Survival analysis

We first attempted to fit an additional multivariate model that estimated the among-individual correlation between behaviour (trait with repeated measurements) and survival (single measure of survival per individual), but the model would not converge. Therefore, we used a generalized linear model (GLM) with a binary response variable for survival as a function of best linear unbiased predictors (BLUPs) for each behaviour extracted from univariate models [68-71]. This approach allowed us to incorporate the uncertainty associated with using single values of BLUPs by iterating through the entire distribution of BLUP values for each individual [68], and resulted in 2000 unique GLMs for each behavioural trait that we used to assess the effects of behaviours on survival. We used the 2000 estimated effects to calculate a total effect size for all 2000 iterations [69-71]. We report the median effect size for all 2000 iterations and 95% CrI for the relationship between each behavioural trait and survival and used the same interpretation for significance of effects as noted above.

To ease interpretation, we refitted risk-taking models (average daily distance to hunter access, average daily distance to open landcover, average daily distance to edge landcover) and multiplied each trait by -1 prior to analysis. This allowed us to interpret a positive log odds ratio as more risky individuals having greater survival, and shyer individuals having lower survival, whereas we interpreted a negative log odds ratio as more risky individuals having lower survival and shyer individuals having greater survival. Next, we subset the data to include individuals that either survived or were harvested in the year they were monitored (Georgia-и = 36 survived, n = 13 harvested; South Carolina-и = 29 survived, n = 14 harvested). We then fitted survival as a binary response as a function of BLUP values for each behavioural trait separately using the approach stated above, which allowed us to directly estimate if greater levels of risk-taking and exploratory behaviour were unintentionally selected for by hunters. Likewise, we used the same approach to assess if BTs were selected for by predators by only including data on individuals that survived or were predated (Georgia-n = 36 survived, n = 10 predated; South Carolina-и = 29 survived, n = 7 predated).

3. Results

3.1. Univariate models assessing behavioural types

From Georgia, we used data on 59 adult males, resulting in 5185 daily values with an average of 88 repeated measures per individual for each behavioural trait (minimum = 7, maximum = 153). For South Carolina, we used data on 50 adult males, resulting in 4557 daily values with an average of 91 repeated measures per individual for each behavioural trait (minimum = 13, maximum = 155). Collectively, there were 3284 behavioural observations during the pre-hunt stage, ТПТ observations during the hunting stage and 3686 in the post-hunt stage (see electronic supplementary material, tables SI and S2, for more detailed summary statistics on observations per site and per individual).

We found for each behavioural trait moderate to high repeatability, which confirmed the existence of consistent among-individual differences in risk-taking and exploration (table 1). Furthermore, we found that on average, individuals from South Carolina displayed riskier behaviour and occurred closer to edge landcover and hunter access points (distance to edge-ß = -0.88, 95% Crl = -1.14, -0.64, PD = 100%; distance to hunter access-ß = -0.56, 95% Crl = -0.87, -0.25, PD = 100%), whereas in Georgia we found a greater investment in only one risk-taking measure, distance to open landcover ф = -0.42, 95% Crl = -0.68, -0.14, PD = 99.95%). We also detected adjustments in behaviour in response to hunting stage that varied in strength and direction between the two study populations (table 1; figure 1). For instance, in South Carolina individuals had increased exploratory tendencies during the hunting season compared to the pre-hunting stage ф = 0.50, 95% Crl = 0.44, 0.57, PD = 100%), whereas in Georgia we failed to detect a difference in exploration between the pre-hunt and hunting stage since ROPE (region of practical equivalence) values were >98% which demonstrated no effect ф = -0.05, 95% Cri = -0.10, -0.00, PD = 97.25%, ROPE = 98.58%). However, we found that exploration did decrease after the hunting stage in both study populations (South Carolina-ß = -0.49, 95% Cri = -0.55, -0.43, PD = 100%; Georgia-ß = -0.43, 95% Cri = -0.48, -0.38, PD = 100%). In addition, for both study populations we found a similar response to distance to open landcover during the hunting stage, whereas individuals from both study populations increased risk-taking as the hunting stages progressed and occurred closer to open landcover (table 1; figure 1). We also found that in both study populations, individuals increased risk-taking and occurred closer to edge landcover after the hunting stage (table 1; figure 1). However, we did not detect any differences in distances to edge landcover between the pre-hunt and hunt stage in South Carolina (ß = 0.02, 95% Cri = -0.03, 0.08, PD = 81.39%), but in Georgia we did detect an increase in risk-taking during this same period ф = 0.43, 95% Cri = 0.38, 0.48, PD = 100%). Finally, in South Carolina, we failed to detect differences in distance to hunter access across all hunting stages (table 1; figure 1). In Georgia, we did detect a decrease in risk-taking with individuals occurring farther from hunter access after the hunting stage compared to the pre-hunt stage ф = -0.11, 95% CrI = -0.15, -0.08, PD = 100%, ROPE = 22.11%).

3.2. Survival analysis

For each behavioural trait, we analysed the influence of HIS on BTs. In Georgia, we found that hunters harvested individuals exhibiting more risky BTs, such as occurring closer to hunter access points (log odds ratio = -0.40, 95% CrI = -0.48, -0.33, PD = 100%) and closer to edge landcover (log odds ratio = -1.40, 95% CrI = -1.88, -1.00, PD = 100%). We did not find support that hunters harvested individuals that occurred closer to open landcover types (log odds ratio = -0.10, 95% CrI = -0.09, 0.29, PD = 85.10%), but hunters did harvest individuals that were fast explorers (log odds ratio = -0.16, 95% CrI = -0.28, -0.03, PD = 99.35%). In contrast, in South Carolina, we generally found the opposite pattern, as hunters harvested individuals exhibiting less risky BTs such as occurring farther away from hunter access points (log odds ratio = 1.10, 95% CrI = 0.86, 1.37, PD = 100%), and being slow explorers (log odds ratio = 0.58, 95% CrI = 0.15, 0.98, PD = 99.40%). However, like individuals in Georgia (figure 2; table 2), we did find that hunters were more likely to harvest individuals that occurred closer to edge landcover types (log odds ratio = -1.17, 95% CrI = -1.65,-0.79, PD = 100%).

We found that for both populations, natural selection acted in a similar direction for all four behavioural traits but varied in strength among traits (table 2; figure 2). Individuals that expressed riskier BTs, such as occurring closer to open landcover (Georgia-log odds ratio = -1.75, 95% CrI = -2.50, -1.09, PD = 100%; South Carolina-log odds ratio = -0.40, 95% CrI = -0.58, -0.25, PD = 100%) and edge landcover (Georgia-log odds ratio = -1.79, 95% CrI = -2.70, -1.08, PD = 100%; South Carolina- log odds ratio = -0.63, 95% CrI = -1.02, -0.35, PD = 100%), were more likely to suffer mortality from predators. Predators selected for slow explorers (Georgia-log odds ratio = 0.67, 95% CrI = 0.47, 0.89, PD = 100%; South Carolina-log odds ratio = 2.23, 95% CrI = 1.78, 2.74, PD = 100%) and individuals that occurred farther from hunter access points (table 2; figure 2).

4. Discussion

We demonstrated significant repeatable variation across all three risk-taking behaviours, as well as our single measure of exploratory behaviour, which is similar to other studies using individual movement data to partition phenotypic variance into among- and within-individual components [22,72,73]. In addition, we observed site-specific differences in behavioural trait repeatability (see electronic supplementary material, table S3) and found that all four traits adjusted in response to hunting stages, with the strength and direction of these adjustments varying across populations. Similar findings of behavioural adjustments in response to hunting pressure have been documented in numerous studies of other game species [74-76], including wild turkeys [42,43,77].

We found that in both study populations, male wild turkeys exhibited increased risk-taking behaviours by moving closer to open landcover as the hunting stages advanced, an observation reported previously [42]. Prior research on resource selection of female wild turkeys has demonstrated that females selected open landcover throughout the duration of the reproductive period [78]. Selection of open landcover by females likely influences male wild turkeys in both of our study populations, because selecting open landcover would enhance their chances to encounter reproductively active and receptive females, which is consistent with findings from other species that use a mating system similar to wild turkeys [79,80]. Our findings also indicate that male wild turkeys modify their proximity to edge landcover based on the hunting stages, as we observed that males from both populations exhibited increased risk-taking from the pre-hunt to the post-hunt stage. We offer three plausible, non-mutually exclusive, explanations for these behaviours. Males might be using edge landcover as corridors to more efficiently locate reproductively active females. Alternatively, considering that females often choose nesting sites near roads and edge landcover [81,82], males could be attracted to these areas to monitor the nesting status of females. Finally, previous resource selection studies noted that edge landcover was important for female wild turkeys by offering quality foraging and escape cover [82], a relationship that likely extends to males as well.

Our findings lend support to a hypothesis outlined in Leclerc et al. [12], wherein individuals occurring closer to open landcover are more likely to be detected and harvested by hunters. However, we only found support for this hypothesis in South Carolina, whereas in Georgia, we found no evidence of HIS influencing risk-taking. Given that openings are important to wild turkeys [36-38], and that wild turkeys persistently seek out open landcover even when it is scarce [83], our observations in South Carolina suggest a unique scenario. Males in South Carolina increased risk-taking during the hunting season by occurring closer to open land cover, but open areas were limited to smaller food plots (<2 ha) meant to attract wildlife. We speculate that this behaviour led males to become more predictable and conspicuous to wild turkey hunters, which negatively impacted male survival. Conversely, in Georgia, the landscape contained large agricultural fields and pastures (>10 ha) in conjunction with smaller openings that were designated as food plots. As such, it was not surprising that we failed to observe HIS for males in Georgia, since a diverse set of openings were available for male wild turkeys to use, which likely made their use of openings less predictable and conspicuous to hunters.

Our research demonstrates that edge landcover influences the propensity of male wild turkeys being harvested. In both populations, individuals exhibiting riskier behaviour (i.e. occurring closer to edge landcover) had lower survival rates. Lower survival rates for individuals occurring closer to edge landcover was not unexpected, considering that previous research on hunter movements using GPS data indicated that hunters primarily used secondary roads to locate wild turkeys [46,47], which are surrounded by substantial edge landcover. Nevertheless, our study shows significant HIS affecting risk-taking behaviours (Georgia-proximity to edge landcover, distance to hunter access; South Carolina-distance to open landcover, distance to edge landcover) in both study populations. Given that contemporary research on movements of wild turkey hunters has shown similar patterns of space use and hunting tactics across various sites and landscapes, such as using open and edge landcover types to locate wild turkeys [46,47,84], hunters may need to adjust their hunting strategies if wild turkeys alter their behaviour due to the increased mortality risks associated with these landcover types. Additionally, if risk-taking behaviour is found to be heritable and results in evolutionary change (i.e. changes in mean population values of these traits), hunters may need to further adapt [85,86]. We encourage future research to explore the heritability of risk-taking behaviours, especially considering that previous studies on roe deer (Capreolus capreolus) have demonstrated relatively high heritability in GPS-based movement behaviours using comparable metrics, such as average speed, distance to roads and use of open landcover [87].

Our research contributes to an increasing body of evidence suggesting that harvest non-randomly targets specific behavioural traits [21-23,25,26]. For male wild turkeys in Georgia, we found that hunters selected to harvest wild turkeys that were fast explorers and took more risks, especially those closer to hunter access points. Conversely, in South Carolina, we observed a different trend: individuals who were faster explorers and riskier (i.e. closer proximity to hunter access points) experienced higher survival rates. Prior research on terrestrial game species has exclusively examined HIS within a single study site [21-23,26]. Conversely, our findings highlight the complexity of HIS, demonstrating that it can differ in magnitude and direction across different study populations. To our knowledge, our study is the first to reveal divergent selection pressures on behavioural traits imposed by hunters across two different populations subjected to similar methods of harvest. Consistent with previous research, our study also confirms that predators selectively target specific BTs [5,7]. Notably, we observed that predators exploited the same BTs across both populations. For example, we found that being closer to edge and open landcover adversely affected survival across both populations. We surmise that male wild turkeys using edge and open landcover types might equate to higher encounter rates with key predators [42], leading to a decrease in survival. Additionally, our data indicate that faster explorers generally had better survival rates than their slower counterparts. Thus, individuals with faster, more superficial exploratory behaviours might not only have more effective anti-predator escape tactics but could also acquire greater spatial memory of potential refuges where predation pressure is lower, consistent with work on a hunted population of brown bears (Ursus arctos) where more exploratory individuals had greater survival rates compared to less mobile ones [26].

Similar to suggestions outlined in Leclerc et al. [12], we recommend that future investigations into HIS focus on selection regimes at the individual site level, as overlooking this aspect might lead to incorrect management actions intended to mitigate selection exerted on certain phenotypes through harvesting. We acknowledge that such monitoring programmes require continuous funding and a considerable amount of time. Nonetheless, quantifying selection pressures from both hunting and natural selection can be advantageous for devising harvesting regimes [13]. Several previous studies have highlighted that shifting harvest regimes to mimic natural selection could contribute to enhancing the behavioural diversity within a population [12,28,29,88,89]. However, empirical studies that evaluate trait distributions following the manipulation of harvesting regimes to mimic natural selection are currently lacking, although simulation-based studies have shown that such harvest regimes can amplify directional changes in trait frequencies [90]. Therefore, until research demonstrates that directional selection is not amplified under harvest, we recommend adopting conservative harvesting practices that avoid bias towards specific behavioural phenotypes. Preserving behavioural diversity is essential for enhancing a population's resilience to environmental changes and emerging pathogens. Simultaneously, harvest regimes that are unbiased towards specific behavioural phenotypes could improve hunter satisfaction by preventing the elimination of phenotypes that are easier to harvest, thus maintaining a continuum of individuals that vary in their degree of harvestability within the population [12]. Moreover, additional studies should aim to measure the ecological consequences of HIS, specifically regarding its effects on social hierarchy and sexual selection. These aspects have been shown to be influenced by HIS in other contexts and warrant detailed investigation [91,92]. Additionally, we recognize that differences in available landcover types across study sites may also contribute to the observed patterns. Therefore, we recommend that future studies investigate whether variations in landcover availability influence common movement metrics. Finally, we recommend more studies in terrestrial game species that properly partition phenotypic variance into its amongand within-individual components. An individual-based approach is crucial for comprehending the complex dynamics of HIS and natural selection, ultimately aiding in the refinement of theoretical frameworks.

Ethics. Capture, handling and marking procedures were approved by the Institutional Animal care and Use Committee at the University of Georgia (A2019 01-025-R2) and Louisiana State University (A2014-13).

Data accessibility. Data and relevant code for this research work are stored in Dryad [93] and GitHub [94] and have been archived within the Zenodo repository [95].

Electronic supplementary material is available online [96].

Declaration of Al use. We have not used Ai-assisted technologies in creating this article.

Authors' contributions. N.A.G.: conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing-original draft; P.H.W.: conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing-review and editing; B.A.C.: conceptualization, data curation, funding acquisition, project administration, supervision, writing-review and editing; MJ.C.: conceptualization, data curation, funding acquisition, project administration, supervision, writing-review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration. We declare we have no competing interests.

Funding. This work was funded and supported by South Carolina Department of Natural Resources, United States Department of Agriculture Forest Service, Louisiana State University School of Renewable Natural Resources, Georgia Department of Natural Resources-Wildlife Resources Division, National Wild Turkey Federation and the Warnell School of Forestry and Natural Resources at the University of Georgia.

Acknowledgements. The authors are grateful to the biologists, technicians and graduate students at Webb WMA Complex, Cedar Creek WMA and B.F. Grant WMA for their assistance with data collection and logistics.

Sidebar

Cite this article: Gulotta NA, Wightman PH, Collier BA, Chamberlain MJ. 2024 The role of human hunters and natural predators in shaping the selection of behavioural types in male wild turkeys. R. Soc Open Sei. 11:240788. https://doi.org/10.1098/rsos.240788

Received: 13 May 2024

Accepted: 1 October 2024

Subject Category:

Organismal and evolutionary biology

Subject Areas:

behaviour, ecology, evolution

Author for correspondence:

Nick A. Gulotta

e-mail: [email protected]

Electronic supplementary material is available online at https://doi.org/10.6084/ m9.figshare.c.7524685.

Footnote

1Present address: NOAA Great Lakes Environmental Research Laboratory, Ann Arbor, Ml 48101, USA.

References

References

1. Endler JA. 1986 Natural selection in the wild. Princeton, NJ: Princeton University Press.

2. Grant PR, Grant BR. 2002 Unpredictable evolution in a 30-year study of Darwin's finches. Science 296,707-711. (doi:10.1126/science.1070315)

3. Losos JB, Schoener TW, Langerhans RB, Spiller DA. 2006 Rapid temporal reversal in predator-driven natural selection. Science 314,1111. (doi: 10.1126/science.1133584)

4. Dingemanse NJ, Both C, Drent PJ, Tinbergen JM. 2004 Fitness consequences of avian personalities in a fluctuating environment. Proc. R. Soc. Lond. В 271,847-852. (doi:10.1098/rspb.2004.2680)

5. Réale D, Festa-Bianchet M. 2003 Predator-induced natural selection on temperament in bighorn ewes. Anim. Behav. 65, 463-470. (doi:10. 1006/anbe.2003.2100)

6. Moodie GEE. 1972 Predation, natural selection and adaptation in an unusual threespine stickleback. Heredity 28,155-167. (doi:10.1038/hdy. 1972.21)

7. Lapiedra 0, Schoener TW, Leal M, Losos JB, Kolbe JJ. 2018 Predator-driven natural selection on risk-taking behavior in anole lizards. Science 360, 1017-1020. (doi: 10.1126/science.aap9289)

8. Darimont CT, Carlson SM, Kinnison MT, Paquet PC, Reimchen ТЕ, Wilmers CC. 2009 Human predators outpace other agents of trait change in the wild. Proc. Natl Acad. Sci. USA 106,952-954. (doi:10.1073/pnas.0809235106)

9. Allendorf FW, Hard JJ. 2009 Human-induced evolution caused by unnatural selection through harvest of wild animals. Proc. Natl Acad. Sci. USA 106,9987-9994. (doi:10.1073/pnas.0901069106)

10. Allendorf FW, England PR, Luikart G, Ritchie PA, Ryman N. 2008 Genetic effects of harvest on wild animal populations. Trends Ecol. Evol. 23, 327-337. (doi:10.1073/pnas.0901069106)

11. Ciuti S, Northrup JM, Muhly TB, Simi S, Musiani M, Pitt JA, Boyce MS. 2012 Effects of humans on behaviour of wildlife exceed those of natural predators in a landscape of fear. PLoS ONEI, e50611. (doi:10.1371 /journal.pone.0050611 )

12. Leclerc M, Zedrosser A, Pelletier F. 2017 Harvesting asa potential selective pressure on behavioural traits J. Appl. Ecol. 54,1941-1945. (doi:10. 1111/1365-2664.12893)

13. Milner JM, Nilsen EB, Andreassen HP. 2007 Demographic side effects of selective hunting in ungulates and carnivores. Conserv. Biol. 21,36-47. (doi:10.1111/j.1523-1739.2006.00591.x)

14. Fügére V, Hendry AP. 2018 Human influences on the strength of phenotypic selection. Proc. Natl Acad. Sei. USA 115, 10070-10075. (doi:10. 1073/pnas.1806013115)

15. Kuparinen A, Festa-Bianchet M. 2017 Harvest-induced evolution: insights from aquatic and terrestrial systems. Phil. Trans. R. Soc. В 372, 20160036. (doi:10.1098/rstb.2016.0036)

16. Wightman PH etai. 2024 Survival and cause-specific mortality of male wild turkeys across the southeastern United States J. Wildl. Manage. 88, e2253l.(doi:10.1002/jwmg.22531)

17. Vangilder LD, Kurzejeski EW. 1995 Population ecology of the eastern wild turkey in northern Missouri. Wildlife Monogr. 130,3-50. (dokhttps:// www.jstor.org/stable/3830761)

18. Sandercock BK, Nilsen EB, Brøseth H, Pedersen HC. 2011 Is hunting mortality additive or compensatory to natural mortality? Effects of experimental harvest on the survival and cause-specific mortality of willow ptarmigan J. Anim. Ecol. 80,244-258. (doi:10.1111/1.1365-2656. 2010.01769.x)

19. Jonzen N, Ranta E, Lundberg P, Kaitala V, Lindén H. 2003 Harvesting-induced population fluctuations? Wildlife Biol. 9,59-65. (doi:10.2981 /wlb. 2003.008)

20. Bunnefeld N, Reumán DC, Baines D, Milner-Gulland EJ. 2011 Impact of unintentional selective harvesting on the population dynamics of red grouse. J. Anim. Ecol. 80,1258-1268. (d o i : 10.1111 /j. 1365-2656.2011.01862.x)

21. Ciuti S, Muhly ТВ, Paton DG, McDevitt AD, Musiani M, Boyce MS. 2012 Human selection of elk behavioural traits in a landscape of fear. Proc. R. Soc. В 279,4407-4416. (doi : 10.1098/rsp b.2012.1483)

22. Ofstad EG, Markussen SS, Saether BE, Solberg EJ, Heim M, Haanes H, Røed KH, Herfindal 1.2020 Opposing fitness consequences of habitat use in a harvested moose population. J. Anim. Ecol. 89,1701-1710. (d o i : 10.1111/1365-2656.13221 )

23. Madden JR, Whiteside MA. 2014 Selection on behavioural traits during 'unselective' harvesting means that shy pheasants better survive a hunting season. Anim. Behav. 87,129-135. (doi:10.1016/j.anbehav.2O13.10.021)

24. Diaz Pauli B, Sih A. 2017 Behavioural responses to human-induced change: why fishing should not be ignored. Evol. Appl. 10,231-240. (doi:10. 1111/eva.12456)

25. Biro PA, Post JR. 2008 Rapid depletion of genotypes with fast growth and bold personality traits from harvested fish populations. Proc. Natl Acad. Sei. USA 105,2919-2922. (doi:10.1073/pnas.0708159105)

26. Leclerc M, Zedrosser A, Swenson JE, Pelletier F. 2019 Hunters select for behavioral traits in a large carnivore. Sei. Rep. 9,12371. (doi:10.1038/ s41598-019-48853-3)

27. Grzegorczyk E, Caizergues A, Fraud C, Francesiaz C, Le Rest K, Guillemain M. 2024 Demographic and evolutionary consequences of hunting of wild birds. Biol. Rev. 99,1298-1313. (doi:10.1111/brv.13069)

28. Monk CT, Bekkevold D, Klefoth T, Pagel T, Palmer M, Arlinghaus R. 2021 The battle between harvest and natural selection creates small and shy fish. Proc. Natl Acad. Sei. USA 118, e2009451118. (doi:10.1073/pnas.2009451118)

29. Edeline E, Carlson SM, Stige LC, Winfield IJ, Fletcher JM, James JB, Haugen TO, Vøllestad LA, Stenseth NC. 2007 Trait changes in a harvested population are driven by a dynamic tug-of-war between natural and harvest selection. Proc. Natl Acad. Sei. USA 104,15799-15804. (doi:10. 1073/pnas.0705908104)

30. Dickson JG. 1992 The wild turkey: biology and management. Mechanicsburg, PA: Stackpole Books.

31. Healy W, Powell S. 1999 Wild turkey harvest management: biology, strategies, and techniques. Biological Technical Publication. Washington, DC: US Fish and Wildlife Service.

32. Williams LE. 1981 The book of the wild turkey. Washington, DC: Winchester Press.

33. Miller DA, Hurst GA, Leopold BD. 1997 Chronology of wild turkey nesting, gobbling, and hunting in Mississippi. J. Wildl. Manage. 61,840. (doi: 10.2307/3802192)

34. Krakauer AH. 2008 Sexual selection and the genetic mating system of wild turkeys. Condor 110,1-12. (doi:10.1525/cond.2008.110.1.1)

35. Buchholz R. 1995 Female choice, parasite load and male ornamentation in wild turkeys. Anim. Behav. 50, 929-943. (doi:10.1016/00033472(95)80095-6)

36. Barwick LH, Speake DW. 1973 Seasonal movements and activities of wild turkey gobblers in Alabama. In Proc. National Wild Turkey Symp., pp. 125-134.

37. Godwin K, Hurst G, Leopold B, Kelley R. 1992 Habitat use of wild turkey gobblers on Tallahala Wildlife Management Area, Mississippi. In Proc. Southeastern Association of Fish and Wildlife Agencies, pp. 249-259.

38. Holdstock D et al. 2005 Male Rio Grande wild turkey habitat characteristics in the Texas Panhandle and southwestern Kansas. In Proc. National Wild Turkey Symp., pp. 217-229.

39. Ey E, Fischer J. 2009 The 'acoustic adaptation hypothesis'-a review of the evidence from birds, anurans and mammals. Bioacoustics 19,21-48. (d o i : 10.1080/09524622.2009.9753613)

40. Ey E, Rahn C, Hammerschmidt К, Fischer J. 2009 Wild female olive baboons adapt their grunt vocalizations to environmental conditions. Ethology 115,493-503. (doi:10.1111/j.1439-0310.2009.01638.x)

41. Koda H, Shimooka Y, Sugiura H. 2008 Effects of caller activity and habitat visibility on contact call rate of wild Japanese macaques {Macaca fuscata). Am. J. Primato!. 70,1055-1063. (doi:10.1002/ajp.20597)

42. Wightman PH, Martin JA, Kohl MT, Collier BA, Chamberlain MJ. 2023 Effects of human and nonhuman predation risk on antipredator movement behaviors of an upland game bird. Ecosphere 14, e4581. (doi: 10.1002/ecs2.4581 )

43. Chamberlain MJ, Wightman PH, Cohen BS, Collier BA. 2018 Gobbling activity of eastern wild turkeys relative to male movements and female nesting phenology in South Carolina. Wildl. Soc. Bull. 42,632-642. (doi:10.1002/wsb.932)

44. Collier BA, Wightman P, Chamberlain MJ, Cantrell J, Ruth C. 2017 Hunting activity and male wild turkey movements in South Carolina. J. Southeast. Assoc. Fish Wildl. Ag. 4,85-93.

45. Chamberlain MJ, Grisham BA, Norris JL, Stafford NJ III, Kimmel FG, Olinde MW. 2012 Effects of variable spring harvest regimes on annual survival and recovery rates of male wild turkeys in southeast Louisiana J. Wildl. Manage. 76,907-910. (doi:10.1002/jwmg.341)

46. Gross JT, Cohen BS, Prebyl TJ, Chamberlain MJ. 2015 Movements of wild turkey hunters during spring in Louisiana. J. Southeast. Assoc. Fish. Wildl. Ag. 2,130.

47. Gerrits AP, Wightman PH, Cantrell JR, Ruth C, Chamberlain MJ, Collier BA. 2020 Movement ecology of spring wild turkey hunters on public lands in South Carolina, USA. Wildl. Soc. Bull. 44,260-270. (doi:10.1002/wsb.1094)

48. McNitt DC, Alonso RS, Cherry MJ, Fies ML, Kelly MJ. 2020 Influence of forest disturbance on bobcat resource selection in the central Appalachians. For. Ecol. Manage. 465,118066. (doi:10.1016/j.foreco.2020.118066)

49. Byrne M, Chamberlain M, Collier B. 2015 Potential density dependence in wild turkey productivity in the southeastern United States. In Proc. National Wild Turkey Symp., pp. 329-351.

50. Chamberlain MJ, Hatfield M, Collier BA. 2022 Status and distribution of wild turkeys in the United States in 2019. Wildl. Soc. Bull. 46, e1287. (doi:10.1002/wsb.1287)

51. Prugh LR, Stoner CJ, Epps CW, Bean WT, Ripple WJ, Laliberte AS, Brashares JS. 2009 The rise of the mesopredator. Bioscience 59,779-791. (doi: 10.1525/bio.2009.59.9.9)

52. Watkins SA. 2022 Fine-scale genetic structure of eastern wild turkeys (Meleagris gallopavo silvestris) and influences of kinship on female sociality. MSc thesis, University of Georgia, Athens, GA, USA.

53. Guthrie JD et al. 2011 Evaluation of a global positioning system backpack transmitter for wild turkey research. J. Wildl. Manage. 75,539-547. (doi:10.1002/jwmg.!37)

54. Cohen BS, Prebyl TJ, Collier BA, Chamberlain MJ. 2018 Home range estimator method and GPS sampling schedule affect habitat selection inferences for wild turkeys. Wildl. Soc. Bull. 42,150-159. (doi:10.1002/wsb.845)

55. Gupte PR, Beardsworth CE, Spiegel 0, Lourie E, Toledo S, Nathan R, Bijleveld Al. 2022 A guide to pre-processing high-throughput animal tracking data. J. Anim. Ecol. 91,287-307. (doi:10.1111 /1365-2656.13610)

56. Boryan C, Yang Z, Mueller R, Craig M. 2011 Monitoring US agriculture: the US Department of Agriculture, National Agricultural Statistics Service, cropland data layer program. Geocartolnt. 26,341-358. (doi:10.1080/10106049.2011.562309)

57. R Core Team. 2024 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.

58. Hijmans RJ etai. 2Q2A Package raster. R package. See https://rspatial.org/raster/pkg/RasterPackage.pdf.

59. Pebesma EJ. 2018 Simple features for R: standardized support for spatial vector data. R J. 10,439. (doi:10.32614/RJ-2018-009)

60. Conner LM, Smith MD, Burger LW. 2003 A comparison of distance-based and classification-based analyses of habitat use. Ecology 84,526-531. (doi:10.1890/0012-9658(2003)084[0526:AC0DBA]2.0.C0;2)

61. Calenge C. 2006 The package a 'dehabitat' for the R software: a tool for the analysis of space and habitat use by animals. Ecol. Modell. 197,516519. (doi:10.1016/j.ecolmodel.2006.03.017)

62. Pebesma E, Bivand RS. 2005 S classes and methods for spatial data: the sp package. R News 5,9-13.

63. Bürkner PC. 2017 brms: an R package for Bayesian multilevel models using stan J. Stat. Softw. 80,1-28. (doi:10.18637/jss.v080.i01 )

64. Bell AM, Hankison SJ, Laskowski KL. 2009 The repeatability of behaviour: a meta-analysis. Anim. Behav. 77,771-783. (doi:10.1016/j.anbehav. 2008.12.022)

65. Dingemanse NJ, Dochtermann NA. 2013 Quantifying individual variation in behaviour: mixed-effect modelling approaches. J. Anim. Ecol. 82, 39-54. (doi:10.1111/1365-2656.12013)

66. Makowski D, Ben-Shachar MS, Chen SHA, Lüdecke D. 2019 Indices of effect existence and significance in the Bayesian framework. Front. Psychol. 10,2767. (doi:10.3389/fpsyg.2019.02767)

67. Makowski D, Ben-Shachar MS, Lüdecke D. 2019 bayestestR: describing effects and their uncertainty, existence and significance within the Bayesian framework J. Open Source Softw. 4,1541. (doi:10.21105/joss.01541)

68. Hadfield JD, Wilson AJ, Garant D, Sheldon BC, Kruuk LEB. 2010 The misuse of BLUP in ecology and evolution. Am. Nat. 175,116-125. (doi:10. 1086/648604)

69. Mathot KJ, Arteaga-Torres JD, Wijmenga JJ. 2022 Individual risk-taking behaviour in black-capped chickadees {Poecile atricapillus) does not predict annual survival. R. Soc. Open Sei. 9,220299. (doi:10.1098/rsos.220299)

70. LaRocque M, Arteaga-Torres JD, Haave-Audet E, Sridharan S, Wijmenga JJ, Mathot KJ. 2023 An investigation of personality-related recapture bias in black-capped chickadees, Poecile atricapillus. Anim. Behav. 196,103-112. (doi:10.1016/j.anbehav.2022.12.007)

71. Haave-Audet E, Martin JGA, Wijmenga JJ, Mathot KJ. 2024 Information gathering is associated with increased survival: a field experiment in black-capped chickadees {Poecile atricapillus). Am. Nat. 203,109-123. (doi:10.1086/727509)

72. Hertel AG, Leclerc M, Warren D, Pelletier F, Zedrosser A, Mueller T. 2019 Don't poke the bear: using tracking data to quantify behavioural syndromes in elusive wildlife. Anim. Behav. 147,91-104. (doi:10.1016/j.anbehav.2018.11.008)

73. Hertel AG, Royauté R, Zedrosser A, Mueller T. 2021 Biologging reveals individual variation in behavioural predictability in the wild. J. Anim. Ecol. 90,723-737. (doi:10.1111/1365-2656.13406)

74. Ordiz A, Støen OG, Sæbø S, Kindberg J, Delibes M, Swenson JE. 2012 Do bears know they are being hunted? Biol. Conserv. 152,21-28. (doi:10. 1016/j.biocon.2012.04.006)

75. Stillfried M, Belånt JL, Svoboda NJ, Beyer DE, Kramer-Schadt S. 2015 When top predators become prey: black bears alter movement behaviour in response to hunting pressure. Behav. Processes 120,30-39. (doi:10.1016/j.beproc.2015.08.003)

76. Little AR, Webb SL, Domarais S, Gee KL, Riffell SK, Gaskamp JA. 2016 Hunting intensity alters movement behaviour of white-tailed deer. Basic Appl. Ecol. 17,360-369. (doi:10.1016/j.baae.2O15.12.003)

77. Wakefield CT, Wightman PH, Martin JA, Bond ВТ, Lowrey DK, Cohen BS, Collier BA, Chamberlain MJ. 2020 Hunting and nesting phenology influence gobbling of wild turkeys. J. Wildl. Manage. 84,448-457. (doi:10.1002/jwmg.21804)

78. Wood JD, Cohen BS, Prebyl TJ, Conner LM, Collier BA, Chamberlain MJ. 2018 Time-since-fire and stand serai stage affect habitat selection of eastern wild turkeys in a managed longleaf pine ecosystem. For. Ecol. Manage. 411,203-212. (doi:10.1016/j.foreco.2018.01.033)

79. Kotrschal A, Taborsky B. 2010 Resource defence or exploded lek? A question of perspective. Ethology 116,1189-1198. (doi:10.1111/j.14390310.2010.01831.x)

80. Morales MB et al. 2014 Density dependence and habitat quality modulate the intensity of display territory defence in an exploded lekking species. Behav. Ecol. Sociobiol.66,1493-1504. (doi:10.1007/s00265-014-1758-z)

81. Veideli NA, Cohen BS, Little AR, Collier BA, Chamberlain MJ. 2017 Nest site selection and nest survival of eastern wild turkeys in a pyric landscape. J. Wildl. Manage. 81,1073-1083. (doi:10.1002/jwmg.21267)

82. Wood JD, Cohen BS, Conner LM, Collier BA, Chamberlain MJ. 2019 Nest and brood site selection of eastern wild turkeys. J. Wildl. Manage. 83, 192-204. (doi:10.1002/jwmg.21562)

83. Miller D, Leopold B, Hurst G. 1997 Seasonal home ranges of wild turkeys in central Mississippi. In Proc. Annual Conf, of the Southeastern Association of Fish and Wildlife Agencies, pp. 414-425.

84. Thomas JW, Gill JD, Pack JC, Healy WM, Sanderson HR. 1976 Influence of forestland characteristics on spatial distribution of hunters. J. Wildl. Manage. 40,500. (doi:10.2307/3799954)

85. Law R. 2000 Fishing, selection, and phenotypic evolution. ICES J. Mar. Sei. 57,659-668. (doi:10.1006/jmsc.2000.0731)

86. Dochtermann NA, Schwab T, Sih A. 2015 The contribution of additive genetic variation to personality variation: heritability of personality. Proc. R. Soc. В 282,20142201. (doi:10.1098/rspb.2014.2201 )

87. Gervais L, Hewison AJM, Morellet N, Bernard M, Merlet J, Cargnelutti B, Chaval Y, Pujol B, Quéméré E. 2020 Pedigree-free quantitative genetic approach provides evidence for heritability of movement tactics in wild roe deer. J. Evol. Biol. 33,595-607. (doi:10.1111/jeb.13594)

88. Carlson SM, Edeline E, Asbjørn Vøllestad L, Haugen TO, Winfield IJ, Fletcher JM, Ben James J, Stenseth NC. 2007 Four decades of opposing natural and human-induced artificial selection acting on Windermere pike {Esox lucius). Ecol. Lett. 10, 512-521. (doi:10.1111/j.1461-0248. 2007.01046.x)

89. Swain DP. 2011 Life-history evolution and elevated natural mortality in a population of Atlantic cod {Gadus morhua). Evol. Appl. 4,18-29. (doi: 10.1111/j.1752-4571.2010.00128.x)

90. Bischof R, Mysterud A, Swenson JE. 2008 Should hunting mortality mimic the patterns of natural mortality? Biol. Lett. 4, 307-310. (doi:10. 1098/rsbl.2008.0027)

91. Palkovacs EP, Kinnison MT, Correa C, Dalton CM, Hendry AP. 2012 Fates beyond traits: ecological consequences of human-induced trait change. Evol. Appl. 5,183-191. (doi:10.1111/j.1752-4571.2011.00212.x)

92. Palkovacs EP, Moritsch MM, Contolini GM, Pelletier F. 2018 Ecology of harvest-driven trait changes and implications for ecosystem management. Front. Ecol. Environ. 16,20-28. (doi:10.1002/fee.1743)

93. Gulotta NA, Wightman PH, Collier BA, Chamberlain MJ. 2024 Data for: The role of human hunters and natural predators in shaping the selection of behavioral types in male wild turkeys. Dryad. (doi:10.5061/dryad.5mkkwh7dj)

94. Gulotta N. 2024 Popular repositories. Github. https://github.com/nickgulotta77

95. Gulotta N, Wightman P, Collier B, Chamberlain M. 2024 The role of human hunters and natural predators in shaping the selection of behavioral types in male wild turkeys. Zenodo. (doi:10.5281 /zenodo. 13285430)

96. Gulotta NA, Wightman PH, Collier BA, Chamberlain M. 2024. Supplementary material from: The role of human hunters and natural predators in shaping the selection of behavioral types in male wild turkeys. FigShare (doi:10.6084/m9.figshare.c.7524685)