Breeding behavior of female white‐tailed deer relative to conception: Evidence for female mate choice

Abstract Female white‐tailed deer (Odocoileus virginianus) are thought to choose between two behavioral strategies to maximize the quality of potential mates: sit and wait, characterized by concentrating activity within a restricted area, and excursive behavior, characterized by increased activity and excursions outside the home range. As movement patterns may influence conception, our goal was to examine the patterns of female white‐tailed deer movements to evaluate which breeding strategy was employed. We equipped 36 female white‐tailed deer with GPS collars from August 2013 to December 2015. We found that movement rate and probability of activity were greatest near the peak of the breeding season, and we observed increases in both metrics during the 40 days prior to estimated conception. Peak size of home range and core area occurred in the days surrounding conception. We found that 11 deer performed an excursion, ranging from 43 days before until 36 days after conception, with the peak probability of being outside of an individual home range occurring 1 day prior to conception. Our results suggest that female white‐tailed deer may attempt to maximize the quality of their mates by advertising availability for breeding through excursive behaviors just prior to entering estrus.


| INTRODUCTION
An animal's fitness, or the extent to which it contributes genetic material to subsequent generations, is determined by reproductive success and the success of progeny (Freeman & Herron, 2004). Due to inequalities in parental investment found in polygynous species (Maynard Smith, 1977;Raven & Johnson, 2001;Trivers, 1972), advantageous methods to increase fitness may vary between the sexes.
Male reproductive output is limited only by breeding success, and as sperm is relatively inexpensive to produce, males can increase fitness by minimizing selectivity and breeding as many females as possible (Raven & Johnson, 2001). Females face greater investment due to the demands of gestation and lactation (Clutton-Brock, 1989;Speakman, 2008) and possible trade-offs between current and future reproduction (Del Guidice, Gangestad, & Kaplan, 2015;Stearns, 1992). Therefore, a female can improve fitness by maximizing mate quality, generally achieved through a high degree of selectivity (Raven & Johnson, 2001). By breeding with a high-quality male, females provide offspring with advantages in areas such as future reproductive potential and immunocompetence (Ditchkoff, Lochmiller, Masters, Hoofer, & Van Den Bussche, 2001;von Schantz, Grahn, & Goransson, 1994).
However, as females near the end of receptivity, fitness benefits accrued from breeding with a high-quality male may be lessened relative to benefits of conceiving during first estrus (predator swamping and increased growth of offspring before their first winter; Zwank & Zeno, 1986;Whittaker & Lindzey, 1999;Gray, Ditchkoff, Causey, & Cook, 2002), leading her to breed with a lower-quality male rather than risk not reproducing in the first estrous cycle.
Due to the benefits of breeding with a high-quality male, behavior of females approaching receptivity is likely a function of herd demography. In white-tailed deer (Odocoileus virginianus), previous studies have documented two behavioral strategies employed by females during the breeding season. The first strategy, known as "sit and wait," has been observed near the peak of the breeding season and is characterized by a decrease in movement rate and range size, and an increase in core area use (Beier & McCullough, 1990;Holzenbein & Schwede, 1989;Ivey & Causey, 1981). The sit-and-wait strategy is advantageous for the female as it provides reduced energy expenditure and limits mortality risk by remaining in a familiar area (Labisky & Fritzen, 1998).
Unfortunately, there still exists a major gap in our understanding of deer breeding behavior. Each study that has previously documented the movement patterns of free-ranging female white-tailed deer during the breeding season has performed so only at the broad scale of population breeding stages (i.e., prerut, rut, postrut; Sawyer, 1981;Holzenbein & Schwede, 1989;Beier & McCullough, 1990;Labisky & Fritzen, 1998;D'Angelo et al., 2004;Kolodzinski et al., 2010). While the aforementioned works have identified general behavioral patterns of females during the breeding season, the behavioral and spatial tendencies of females relative to their specific conception date remain unknown. Therefore, evaluating the movement of female white-tailed deer relative to conception is a clear next step in understanding not only how female large mammals utilize space during a key life-history event, but also in exploring the possibility of female mate choice. The goal of this project was to examine the movement and space use patterns of female white-tailed deer during periods relevant to breeding at the population and individual levels. Our objectives were to: (1) characterize activity patterns of free-ranging female white-tailed deer relative to the populations' peak of conception, (2) characterize activity patterns relative to individual conception, and (3) determine whether the observed behaviors are breeding related.  (Collier, Ditchkoff, Raglin, & Smith, 2007). Hardwood drains were found throughout the property with mixed pine-hardwood areas comprised of loblolly (Pinus taeda), slash (Pinus elliottii), and pond (Pinus serotina) pine, along with oak (Quercus spp.), sweetgum (Liquidambar styraciflua), and red maple (Acer rubrum).

| Study area
Bottomland drains included oak, sweetgum, black gum (Nyssa sylvatica), and yellow poplar (Liriodendron tulipifera). The majority of forest stands were actively managed for wildlife and timber production and were burned on a 2-to 3-year rotation to maintain an open understory (Collier et al., 2007;Lauerman, 2007). Food plots on the study area ranged in size from 0.6 to 8.1 ha and comprised a total of 119 ha. While a majority of plots were planted annually with a cool

| Capture
During May 2013-August 2015, a total of 36 female white-tailed deer (≥1 year old) were immobilized via a 2 cc transmitter dart (Pneudart Inc., Williamsport, Pennsylvania, USA) containing a Xylazine (Lloyd Laboratories, Shenandoah, Iowa; 100 mg/ml given at a rate of 2.2 mg/kg) and Telazol (Fort Dodge Animal Health, Fort Dodge, Iowa; 100 mg/ml given at a rate of 4.5 mg/kg) mixture. Deer were fit with an ATS G2110D GPS Collar (Advanced Telemetry Systems, Isanti, Minnesota) tightened to within approximately two finger widths of the neck, allowing the collar to stay in the proper upright position, thereby improving data accuracy (D'Eon & Delparte, 2005). After processing was completed, a 3-ml intramuscular injection of Tolazoline (Lloyd Laboratories, Shenandoah, Iowa; 100 mg/ml given at a rate of 6.6 mg/kg) was administered to act as a reversal to the Xylazine. All protocols involving animals were approved by the Auburn University Animal Care and Use Committee (PRN# 2013-2205).

| Data collection and manipulation
Collars were programmed to take GPS fixes at 30-min intervals from where harvest location and time, age, weight, and pregnancy status were recorded. All deer were aged via tooth wear and replacement (Severinghaus, 1949). Fetuses were removed from pregnant females, and date of conception was back calculated using a fetal aging scale (Hamilton, Tobin, & Moore, 1985).
All spatial data were classified using two separate approaches.
First, following the classification system of previous studies (Beier & McCullough, 1990;D'Angelo et al., 2004;Holzenbein & Schwede, 1989;Kolodzinski et al., 2010;Labisky & Fritzen, 1998;Sawyer, 1981), we assigned each fix as being either prerut, rut, or postrut. Breeding periods on Brosnan Forest ranged from 16 August to 18 September, 19 September to 28 October, and 29 October to 1 December, respectively. These dates were determined such that the rut period would account for 80% of all conceptions previously documented on the study site (McCoy et al., 2013). Second, for all deer that conceived during our study, we created a covariate for days since conception for each fix ranging from 45 days prior to conception to 45 days postconception.
We used 45 days as it would encompass movements up to and after conception and is approximately 2 weeks greater than the 28-day estrous cycle of female white-tailed deer (Plotka, Seal, Schmoller, Karns, & Keenlyne, 1977). Examining activity at both population breeding stages and relative to date of conception allowed us to compare movements of individuals to population-level trends from this study and previously reported studies and to determine whether behaviors observed in this population were site specific or potentially generally representative of white-tailed deer.
We evaluated six spatial metrics (movement rate, home range size, core area size, core area:home range ratio, probability of being outside the seasonal home range, and the probability of activity) that we hypothesized could be impacted by breeding-related activity.
Movement rate was calculated by finding the average Euclidean distance between consecutive half hour fixes throughout a single day (Labisky & Fritzen, 1998;Root, Fritzell, & Giessman, 1988;Webb, Gee, Strickland, Demarais, & DeYoung, 2010). Following Webb et al. (2010), only days with at least half of the possible consecutive fixes were utilized. We calculated home range and core use areas for each deer using the Brownian bridge method (Horne, Garton, Krone, & Lewis, 2007) for our study period (seasonal home range), as well as for the prerut, rut, and postrut periods. The 95% and 50% isopleths were used for the home range and core area estimates, respectively.
Additionally, for females that conceived, home ranges and core area sizes were calculated for 13, 7-day periods from 45 days before until 45 days after conception. The area of each home range and core area was calculated using ArcMap 10.2 (ArcMap version 10.2, ESRI Inc., Redlands CA, ESRI, 2013). Core area:home range ratio was calculated by dividing core area size by home range size and allowed for characterization of the intensity of an individual's use of space, where a ratio close to 1 indicated an individual was using its entire home range equally and a ratio close to 0 suggested that an individual was using small portions of its home range much more intensively (Monsarrat et al., 2013;Rotem, Berger, King, Bar, & Saltz, 2011). To find the probability of being outside the seasonal home range (SHR), each fix was classified as within or not within the SHR. The probability of a deer being outside the SHR provides information about an individual's propensity to travel outside of its home range and engage in exploratory movements (Kolodzinski et al., 2010). Probability of activity was calculated daily for each 30-min interval within three separate periods: the 24-hr day, diurnal hours, and nocturnal hours. Diurnal hours ranged from one half hour before sunrise until one half hour after sunset, with nocturnal being the remainder of the 24-hr day.
While white-tailed deer are crepuscular by nature, their reproductive behaviors do not necessarily follow the same pattern. Therefore, crepuscular periods were not analyzed separately. Probability of activity was treated as a binary variable, and an individual was considered to be active in a 30-min interval if the distance between consecutive locations was greater than the predetermined threshold distance of 51.78 m according to Sullivan, Ditchkoff, Collier, Ruth, and Raglin (2016).
Finally, movement paths of all individuals found to have conceived were inspected for excursions outside of the seasonal home range.
Similar to previous studies (Karns, Lancia, DePerno, & Conner, 2011;Kolodzinski et al., 2010), an excursion was defined as (1) any series of fixes outside of the seasonal home range which extended >0.5 km, or (2) when half of an animal's daily fixes were outside of the seasonal home range. When an excursion was identified, the date and days from conception were recorded.

| Data analysis
Movement rate, home range and core area size, and core area:home range ratio were analyzed relative to population breeding stages using linear regression (metric~rut stage), while likelihoods of activity and being outside the SHR were analyzed via logistic regression. Linear regression was used to evaluate variation in home range size, core area size, and core area:home range ratio relative to date of conception, while movement rate was analyzed via generalized additive modeling (metric~days to conception) with the model specified as a Gaussian distribution and a cubic regression spline smoother (Wood, 2006).
Probability of activity and probability of being outside the SHR were also analyzed via generalized additive modeling but with distributions specified as binomial. We estimated the probability of an excursion occurring each day relative to conception via a general linear model with a second-degree polynomial and used a generalized linear model to determine whether age class influenced the probability of an excursion. All analyses were conducted in R (v3.1.3; R Core Team, 2015).

| RESULTS
During the study, the 36 collars had an average fix success rate, after data censoring, of 87.61% (n = 163,508), while the stationary collars had an average fix success rate of 99.87% (n = 792). Thirty-two collared deer with an average age of 3 years (SE = 0.2) were harvested by hunters following the study period, and conception was documented in 21 of these animals. Fetuses ranged from 40 to 114 days postconception with calculated conception dates from 11 September to 2 November. We also found that 80.95% of conceptions occurred during the rut period.
There was a general trend of increasing activity from the prerut to the rut when examining movement rate, probability of activity, probability of being outside the SHR, home range size, core area size, and core area:home range (Table 1). Movement rate, probability of activity, and probability of being outside the SHR all increased from prerut to rut, and while movement rate and probability of activity declined from rut to postrut, probability of being outside the SHR continued to rise during this period. The space use metrics of core area and home range size both increased from prerut to rut, while core area:home range did not. While home range size continued to increase from rut to postrut, core area size and core area:home range values declined. When examined relative to date of conception, there was a tendency for activity metrics (movement rate and probability of activity) to steadily increase during the 40 days prior to conception, level off or slightly decrease in the days around conception, and then increase once again until 15-20 days following conception (Figures 1 and 2). Home range (17.284 ha, SE = 0.826) and core area (2.966 ha, SE = 0.165) were both greatest during the weeks surrounding conception, but the core area:home range surrounding conception was not found to be different from any other period (Figure 3).
We identified 23 excursions from the 21 deer that conceived with 2, 14, and 7 excursions performed during the prerut, rut, and postrut periods, respectively. Excursions ranged from 43 days before to T A B L E 1 Mean measures of movement rate, probability of activity during any given 30-min interval, probability of being outside of the seasonal home range, home range size, core area size, and core area: home range ratio for female white-tailed deer at Brosnan Forest, SC during prerut, rut, and postrut periods of 2013-2015 36 days postconception (Table 2)

| DISCUSSION
Our results suggest that adjustments in female behavior and space use appear to be closely linked to conception. When examining the spatial metrics relative to breeding stages, an increase in movement rate and probability of activity from prerut to rut suggests that females were actively advertising their presence during the breeding season (D'Angelo et al., 2004;Labisky & Fritzen, 1998;Sawyer, 1981).
The conclusion that females are increasing movement to serve as a method of advertisement is supported by an increase in both metrics from approximately 40 days before until just prior to conception.
Increases in both home range and core area from prerut to rut, paired with no change in the core area:home range ratio over the same time period, suggest that deer were expanding movements over a larger area while using this area uniformly (Monsarrat et al., 2013;Rotem et al., 2011). Our results indicate that peak home range and core area sizes were observed near conception, suggesting that the increase was driven by breeding behaviors, and the peak during the rut period occurs due to most deer breeding within that window of time. Such space use is directly opposed to the behavior expected in a sit-andwait strategy where deer would concentrate movement into a portion of the core area (Beier & McCullough, 1990;Holzenbein & Schwede, 1989;Ivey & Causey, 1981).
Perhaps the trends that most clearly demonstrate the connection between behavioral changes and breeding are those seen through the probability of being outside the seasonal home range and the daily probability of performing an excursion. The probability of being outside the seasonal home range increased approaching conception and declined postconception, suggesting that deer were attempting to advertise their presence to as many potential mates as possible (Kolodzinski et al., 2010). Additionally, while the peak probability of an excursion was after conception, the cluster of excursions within 3 days of conception suggests that excursive behaviors and increases in activity before conception were related to reproductive activities. However, the occurrence of several excursions both before and after conception indicates that breeding is likely not the only driver of excursions, even when occurring during the breeding season. An additional potential explanation for the observed behaviors is harassment by rutting males or isolation during the tending bond (Kilpatrick & Lima, 1999;Kolodzinski et al., 2010 ies would require proximity loggers to answer such questions. The potential also exists that these behaviors could be a response to hunting pressure as opposed to reproductive activities (Kilpatrick & Lima, 1999;Kolodzinski et al., 2010). Hunting pressure on our study site was very consistent both spatially (a lack of large refuge area) and temporally (hunting occurred equally throughout the week and season) due to the properties role as an outdoor recreational facility for large conference groups (Sullivan, 2016). Therefore, we deem it unlikely that these temporally specific behaviors were a response to the long and sustained hunting seasons that took place on this property during our research.
Increases in movement rate and probability of activity postconception were unexpected, as it is not advantageous to maintain mateseeking behavior once breeding has occurred (Labisky & Fritzen, 1998).
However, it is possible that a female is unaware whether she has conceived until approximately 20 days postconception and therefore behaves as though she will be entering a second estrus. In ruminants, pregnancy may be recognized following the release of interferon tau, which signals the stoppage of luteolysis, and subsequent increase in progesterone levels (Bazer, Ott, & Spencer, 1998;Demmers, Kaluz, Deakin, Jabbour, & Flint, 1999;Spencer, Burghardt, Johnson, & Bazer, 2004). It is likely these changes in progesterone level, referred to as the pregnancy hormone (Spencer et al., 2004) (Holzenbein & Schwede, 1989;Labisky & Fritzen, 1998).
However, behavioral patterns observed during the three breeding stages closely resembled those reported by Labisky and Fritzen (1998) for a low-density deer population (five deer/km 2 ). As females should have experienced no difficulty finding a mate on our study site, we suggest that the conclusion of Labisky and Fritzen (1998) that excursions serve only as a last resort to find a mate may not be generally applicable. Similar to Kolodzinski et al. (2010), we conclude that excursive behavior in female white-tailed deer is an adaptive behavior to maximize the quality of the pool of potential mates and is an expression of female mate choice.
While it appears that female white-tailed deer are performing excursive behavior as a form of mate choice, it is not likely that these excursions are an attempt to locate a particular mate. Stopher et al.
(2011) rejected the notion of females using excursions to reach preferred males in red deer as the excursions appeared random.
Furthermore, as males are highly mobile during the rut, the likelihood of an excursion allowing a female to locate a specific male is low (DeYoung & Miller, 2011;Labisky & Fritzen, 1998). Instead, a straight line search, which has been observed for excursions, would maximize the likelihood of encountering any given male, while an intensive search would be required to locate a particular male (Zollner & Lima, 1999). The impracticality of a female utilizing an excursion to locate a specific male suggests that excursive behaviors may be an attempt to incite competition by alerting more males of the female's presence, as suggested by Cox and Le Boeuf (1977).
Our results provide insight into the movement ecology and behavioral strategies employed by reproductively active female white-tailed deer, and our examination of spatial metrics relative to an individual's date of conception provides a more detailed understanding of breeding behaviors. While the results of the breeding stage analysis in this study supported trends seen relative to date of conception, they failed to identify unique behaviors which provide a greater degree of insight into possible implications of animal activities. Therefore, we recommend that future behavioral studies focus on events at the individual scale as opposed to population-level analyses.