Methods
We captured female wild turkeys using rocket nets baited with corn from January – March 2014–2019. We aged each individual using the presence of barring on the 9th and 10thprimaries (Pelham and Dickson 1992). We fitted all females with a uniquely identifiable aluminum rivet tarsal band and backpack style GPS/VHF transmitter (Biotrack Limited, Wareham, Dorset, UK; Guthrie et al. 2011). We programmed GPS units to collect data at 1-hour intervals (Cohen et al. 2018) between 05:00 to 20:00 daily with one location at night (23:59:58) to identify roosts until the battery died or the unit was recovered. We immediately released individuals at the capture location following processing. Capture, handling, and marking procedures were approved by the Louisiana State University Agricultural Center Animal Care and Use Committee (Permits A2015-07 and A2018-13). We monitored live-dead status daily during the reproductive season using handheld Yagi antennas and Biotracker receivers (Biotrack Ltd., Wareham, Dorset, UK). We downloaded GPS locations once per week via a VHF/UHF handheld command unit receiver (Biotrack Ltd., Wareham, Dorset, UK).
When winter flocks disband, social groups of wild turkeys alter space use and focus efforts on reproduction (Badyaev et al. 1996b , Thogmartin 2001). We assumed that all females within a social group had access to the same mates, and presumably the same dominant males. Therefore, we defined a social group as a group of females captured together during January to March as GPS data indicated that turkeys did not disperse from wintering flocks before reproduction started, contrary to suggestions in Badyaev et al. (1996a ). While we defined females captured together as a social group, we acknowledge that we may not have captured all of the females in the same breeding group. We assumed, based on estimates of daily movements by females (Conley et al. 2016, Bakner et al. 2019), that individuals captured within 2 km of each other were members of the same social group as these individuals regularly interacted as detailed herein. To further ensure we accurately defined social groups, we used a dynamic Brownian Bridge movement model (dBBMM) to create 99% utilization distributions (UDs) for each individual (Byrne et al. 2014) for the 21 days before the first female in each group laid the first egg at an eventual nest site. We chose a 21-day period because we were interested in overlap in space use during the time immediately preceding initiation of the first nest in the social group, under the assumption that this individual was the dominant female in the group (Watts and Stokes 1971). We calculated all UDs in program R version 3.2.5 (R Core Development Team 2020) using package move (Kranstauber and Smolla 2013). We used a window and margin size equal to 21 and 9 respectively, and a location error of 10 m (Byrne et al. 2014). Individuals that share space may constitute a single social unit (Brown 1975), therefore we calculated the percentage of utilization distributions that overlapped at least one other UD within a defined social group during the 21-day period to quantify shared space use (Kernohan et al. 2001). We assumed that any females who did not maintain an overlapping range with at least one other female within a social group or individuals within subgroups were of lower rank, and as such should subsequently nest later (Ringgenberg et al. 2015). We defined smaller groups that contained 2-3 individuals with ranges that overlapped each other, but did not overlap with the main social group, as a subgroup (Figure 2).
We determined locations of each nesting attempt for each female when an individual’s locations became concentrated around a single point for several days (Guthrie et al. 2011, Conley et al. 2015, Yeldell et al. 2017, Wood et al. 2019). We defined the first date of nest incubation as the first day we recorded the nightly roost location at the nest site, indicating the female continued incubation during the night (Bakner et al. 2019). To determine the first date of egg laying (hereafter nest initiation), we evaluated GPS locations to determine when a female initially visited the nest site as female wild turkeys do not visit their nest site until they lay their first egg (Conley et al. 2016, Collier et al. 2019). We monitored each nesting attempt following Bakner et al. (2019) and after nest termination, located nest sites using VHF telemetry and GPS data to confirm the nest location and determine nest fate. We considered a nest to have been depredated or abandoned if the female left the nest ≤25 days into incubation, or if only intact eggs, no eggs, or egg fragments were found at the nest bowl. We considered a nest successful if ≥1 live poult hatched, and was confirmed visually during subsequent brood surveys following methods outlined in Chamberlain et al. (2020).
We scaled the initiation date of the first nest attempt to each social group, where the date of the first nest initiation was noted as day 1. We delineated subsequent nest attempts based on the number of days after the first nest was initiated. We subtracted the initiation day of the second nest from the initiation day of the first nest, and then subtracted the initiation day of the third nest from the initiation day of the second nest, and so on for each first nest attempt within each social group. We then calculated mean number of days between each nest initiation attempt within each social group. We speculated that social groups with more individuals would have more days between subsequent nest attempts compared to smaller groups. Presumably, larger groups would contain more females competing to copulate with dominant males (Orbach et al. 2015), whereas smaller groups would have less competition and thus be able to copulate in a shorter temporal window, resulting in a narrower time window during which nests were initiated by females in that group (Dewsbury 1982, Foster 1983, Avery 1984, Trail 1985, Gratson et al. 1991, Möller 1992).
Females that attempt reproduction earlier within a season are expected to have greater annual reproductive success compared to later breeding individuals (Lack 1968, Perrins 1970) and previous research has noted that in lekking birds, dominant females breed first (Robel and Ballard 1974, Foster 1983). Dominant females can presumably select nest sites that could confer fitness advantages through improved nest success (Sӕther 1990, Martin 1995a , Martin 1995b ), compared to subordinate females that nest later and may be forced to nest in suboptimal parts of their ranges or travel farther distances to find suitable nest sites. To test the prediction that dominant females would travel shorter distances within their ranges prior to onset of nest initiation, we used the distance between a female’s nest location and the centroid of the UD range of the 21-day period before the first nest of each social group was initiated as our metric. We measured the distance between the centroid of each female’s 99% UD range to each of her nest attempts in ArcGIS 10.6 (Environmental Systems Research Institute, Inc., Redlands, California, USA; Figure 1). To locate the centroids of each 99% UD, we calculated the x and ycentroid of each UD in the attribute table. We then created a line between each nest attempt and the centroid and calculated the distance between each nest attempt within the 99% UD. To test for differences in mean distance traveled for females with successful versus failed nests, we used an independent 2-group t -test with an α=0.05 in R (R Core Team 2020). Likewise, we used a binomial generalized linear model (GLM) in R (R Core Team 2020) to estimate nest success as a function of first nest initiation date. We then used a Poisson GLM to estimate the rate (in days) at which females left their social groups and initiated their first and second nesting attempts as a function of group size and year. Finally, we used linear regression to evaluate the effect of social group size on the number of days between nesting attempts within social groups.