Influence of Vegetation Characteristics at and Near Nests on Female Prairie Vole (Microtus ochrogaster) Survival and Reproductive Success.
Nest location can affect survival and reproductive success in animals (Belles-Isles and Pieman, 1986; Li and Martin, 1991; Cudworth and Koprowski, 2011) which suggests, within a particular environment, some habitat features associated with nest sites will differ significantly from those in randomly selected sites. Differences in habitat characteristics in nest versus non-nest sites have been confirmed for a few mammalian species (e.g., eastern chipmunk, Tamias striatus, Mahan and Yahner, 1996; hazel dormouse, Muscardinus avellanarius, Berg and Berg, 1998; garden dormouse, Eliomys quertinus, Bertolino and Montezemolo, 2007). Previous studies suggested risk of predation (Berg and Berg, 1998; Ross et al., 2010), food quality and quantity (Forsyth and Smith, 1973; Hackett and Pagels, 2003), and microclimate surrounding the nest site (Carey et al., 1997) affected nest site location in rodents.
As vegetation associated with a nest site can affect an animal's behavior, individuals might sometimes have to choose, for example, between protection from predation and access to high quality food plants located close to the nest (wood mice, Apodemus sylvaticus, Rosalino et al., 2011; hazel dormouse, Juskaitis et al, 2013). There are only a few studies of the relationship between vegetation at nests and survival and reproductive success in mammals (yellow mongoose, Cynicitis penicillate, Blaum et al, 2007; Iberian hare, Lepus granatensis, Sanchez-Garcia et al., 2012; African wild dog, Lycaon pictus, Davies et al, 2016). Documenting the relationships between specific habitat characteristics associated with nest locations and survival and reproduction is essential for understanding the evolutionary pressures that can influence nest-site selection.
Herein, we determined if vegetation near prairie vole (Microtus ochrogaster) nests differed from that of randomly sampled areas in the same habitat and if vegetation at nest locations and in habitat containing nests predicted female survival, offspring production, and pup survival. Prairie voles are socially monogamous rodents in which male-female pairs cooperatively defend a territory. Within their territory they construct surface or underground nest(s) where they live and care for young (Getz et al., 1993). As herbivorous rodents prairie voles tend to prefer forbs, which often have a patchy distribution (Cole and Batzli, 1979; Sueatfeild et al, 2011; Dejaco and Batzli, 2013). Voles suffer substantial mortality from mammalian and avian predators (Baker and Brooks, 1982; Lin and Batzli, 1995). Therefore, the type of vegetation at or near a vole nest might influence survival and reproductive success by providing food, cover, or both (Solomon et al, 2005).
In a previous study, female voles in seminatural enclosures did not nest at sites with preferred food resources or the tallest vegetation (Solomon et al, 2005). This might have occurred because voles avoided nesting at sites where there had been a pre- existing underground nest (Solomon et al 2005). Proximity to food resources also did not affect nest location in a natural population (Getz et al, 1992). The lack of a relationship between vegetation and nest-sites might occur because some females are forced to occupy lower quality territories due to intraspecific competition. Females in lower quality territories likely have decreased survival and reproductive success but this has not been examined.
If selection of nest locations is influenced by vegetation, we would expect vegetation associated with nest sites would differ from vegetation at random sites within the same habitat. Specifically, we predicted nests would be associated with greater food resources (dicots) and other vegetation that decreases the risk of predation (e.g., taller, woodier, thornier). We also tested the possible influence of vegetation near nests on female survival and reproductive success. In particular we predicted that female survival, total number of pups produced, and pup survival would be greater at nest locations associated with greater food resources (dicots) and features of vegetation that might deter predation (taller, woodier, and thornier plants and less bare ground; Bimey et al, 1976; Baker and Brooks, 1982). Because females that are pregnant or have weaned pups might not be as constrained in the time spent away from the nest foraging as females nursing pups, female survival and reproductive success could be related to food resources available at a larger scale than the area immediately surrounding the nest. Therefore, female survival and reproductive success would be expected to be associated with food resources (dicots or preferred food) at this larger scale.
We conducted this study at the Miami University Ecology Research Center located near Oxford, Ohio (39[degrees]31'41"N, 84[degrees]43'21"W) during summer and autumn of 2016 utilizing two sets of enclosures (n = 8 in each set). The sixteen 0.1 ha (~32 m X 32 m) enclosures were constructed to prevent voles from escaping (Cochran and Solomon, 2000). Each set of enclosures was encircled by an electric wire to exclude mammalian predators. Vegetation within each enclosure was primarily perennial grasses and forbs that provided food and cover for voles (see Solomon et al, 2009). Honey locust (Gleditsia triacanthos) saplings and invasive species including Russian olive (Elaeagnus angastifolia), common teasel (Dipsacus fullonum), and Japanese honeysuckle (Lonicera japonica) also were present in the enclosures. We mowed and maintained the vegetation in a 1 m strip bordering the inside walls of enclosures at a height of 5-10 cm to prevent voles from digging near the walls. Before starting our study, we live-trapped each enclosure to remove any meadow voles, Microtus pennsylvanicus. A few meadow voles that we missed during this period were captured and removed during the study.
Founders released into each enclosure were descendants of voles collected from a wild population at the University of Kansas's Nelson Environmental Study Area (15 km northeast of Lawrence, Kansas; 39[degrees]03'03"N, 95[degrees]11'32"W) and from captive-bred animals descended from voles captured in Illinois. Voles were housed at Miami University's animal care facility. We used [F.sub.2] generation individuals from wild-caught Kansas (KS) voles and seventh or greater generation individuals from the captively bred Illinois (IL) animals. We released four IL males and four KI (IL sire-KS dam) males into eight enclosures and eight KI males into an additional eight enclosures because the objective of a simultaneously occurring study was to examine potential differences in male social and reproductive behavior between the KI and IL males (Lambert, 2018). In every enclosure all founding females (n = 8) were the offspring of Illinois parents because IL females are more socially monogamous than KS females (Roberts et al, 1998; Cushing and Kramer 2005).
EXPERIMENTAL DESIGN AND FIELD METHODS
In June 2016 we released the founding prairie voles (eight males; eight females) into each of 16 enclosures. This initial density (160 voles/ha) is within the range of densities reported for natural populations of prairie voles in Illinois (Getz et al, 2001) and is considered to be a moderate to high density (McGuire and Getz, 1998). Each vole was fitted with a radio-collar (model PD-2C; Holohil Systems Ltd., Carp, Ontario, Canada) weighing about 3 g (<10% of body mass for a typical adult) before release. Before release we also gave all founders a unique toe-clip for identification (no more than one toe per foot) and collected a small sliver of the pinnae from one ear. Tissue samples were stored at -20 C for subsequent genetic analysis of parentage.
One week after release, we began radio-telemetry to locate female nests. Females were tracked 1-10 times a day between 1100 h and 1700 h for up to 5 d a week (Monday-Friday) until we located their nests. The variation in the number of times a female was tracked each day was because females often were moving when tracked. After locating a nest site, we examined each nest daily Monday through Friday between 1100 and 1800 h for the duration of the study for the radio signal from all founding males and females within every enclosure. If a female's radio signal was not consistently detected at a specific nest, we attempted to locate her new nest using the same procedures used previously.
As prairie voles have a ~21 d gestation and are weaned at ~21 d of age (Hasler, 1975; Nadeau, 1985), the first pups should have started to appear above ground by ~ week six of the study. To capture pups born in an enclosure, we began live-trapping within each enclosure starting the 6th wk using Ugglan multiple capture traps (Grahnab, Hillerstorp, Sweden). Traps were covered with an aluminum shield and also with vegetation to protect animals from extreme weather. We baited traps with cracked com, a low-quality food (Desy and Batzli, 1989) used in numerous previous field studies (Keane et al., 2007; Solomon et al., 2009).
We alternated between trapping at nests and in a grid pattern. For nest-trapping we placed three traps in surface runways within 1 m of surface nests or entrance(s) to an underground nest. Nest-trapping was conducted during weeks 6-7, 9-10, and 12-13 of the study. During nest-trapping weeks there were 10 trap-examinations per week. For grid trapping traps were arranged in a 5 X 5 m grid (1 trap near each grid stake; 25 traps per enclosure). Grid trapping was conducted during weeks 8, 11, 14, and 15 with five trap-examinations per week. We released all animals that had been live-trapped at their point of capture.
Because vegetation changed throughout the study and females could potentially change nest sites, we divided both our vegetation sampling and our analyses into three periods corresponding to our trapping protocol: period one comprised weeks 6-8, period two weeks 9-11, and period three weeks 12-14. Therefore, each period was two consecutive weeks of nest trapping followed by a week of grid trapping. We considered a female to be a resident at a specific nest if she was trapped/located with radio-telemetry at least two times during a period and [greater than or equal to] 75% of her detections were at the same nest within that period.
When pups were first captured, they were given a unique toe-clip (no more than one toe per foot) for identification and tissue samples were stored at -20 C for parentage analysis. We also weighed pups every time that they were captured. Animals < 21 g in mass were classified as juveniles, animals 21-29 g were classified as subadults, and animals > 29 g were classified as adults and considered sexually mature (Gaines et al., 1979; Getz et al., 1993). After pups reached 30 g, we removed them from the enclosure to prevent them from reproducing and to keep adult density relatively constant. Between the time pups attained 30 g and the end of the actual experiment, we removed approximately 3.1 offspring/day. All remaining offspring were captured during the 18 d trap-out period after the end of the 15 wk experiment at the rate of 15.5 offspring/day. We were fairly confident that we caught all surviving offspring.
At the end of the study, we live-trapped all surviving voles, removed them from their enclosures, and euthanized them with C[O.sub.2]. All the methods involving live animals were in accordance with the guidelines provided by the American Society of Mammalogists (Sikes et al, 2016) for the use of wild mammals in research and were approved by Miami University's Institutional Animal Care and Use Committee.
For vegetation sampling we divided each enclosure into 36 sections of approximately 25 [m.sup.2]. During the beginning of weeks 6, 9, and 14 of the study, we measured the same vegetation at all nests and at randomly selected non-nest locations within 10 different randomly selected sections in each enclosure. By week six females should have established nests, and the first pups born in the enclosures could be trappable. We resampled vegetation during each period because some vegetation characteristics (e.g., height) change throughout the season.
We sampled vegetation by placing a 0.5 m x 0.5 m (0.25 [m.sup.2]) quadrat over the center of a nest site or at a randomly selected site. For each sample we measured the height of the tallest plant and estimated the percentage of the quadrat that was bare ground, dicots, thorny plant cover, woody plant cover, and preferred food plants similar to the methods used by Solomon et al (2005). Based on previous studies (Cole and Batzli, 1979; Lin and Batzli, 2001), we categorized brome grass (Bromus sp.), wild carrot (Daucus carotus), dandelion (Taraxacum officinale), ragweed (Ambrosia sp.), goldenrod (Solidago sp.), and clover (Trifolium sp.) as preferred food plants.
We collected vegetation data at 160 random sites (10 per enclosure) during each of the three vegetation sampling periods: 97 nest sites during period one, 65 nest sites during period two, and 70 nest sites during period three. Of the 97 nest locations sampled in period one, 24 (25%) were used by females during all three periods and 22 (23%) were used by females only during periods one and two. There were 19 (of 65 or 29%) new nests sampled during period two that were not present in period one and 14 (of 19 or 74%) of these also were used by females in period three. Thirty-two (of 70 or 46%) new nests were sampled during period three that were not present in period one or two. Based on our criteria for determining residency at a nest site, no female was considered a resident at more than one nest during the study.
To assess parentage of voles born in enclosures, we extracted genomic DNA from tissue samples using DNeasy extraction kits (Qiagen, Valencia, CA, U.S.A.) and genotyped individuals at six microsatellite loci previously shown to be polymorphic in prairie voles using Polymerase Chain Reaction (PCR; for details see Keane et al., 2007; Solomon et al., 2009). PCR products were diluted and combined with an internal size standard (LIZ GS500, Applied Biosystems, Foster City, CA), and fragments were detected using an ABI 3730 DNA sequencer (Applied Biosystems, Foster City, CA, U.S.A.). Base-pair lengths of the fluorescently labeled DNA fragments were analyzed using Peakscanner (Version 1.0) fragment analysis software (Applied Biosystems). Microsatellite allele lengths were compiled into discreet size classes using FlexiBin (Amos et al., 2006).
We assigned parentage using the computer program Cervus 3.0 (Kalinowski et al., 2007) using the parent-pair analysis option, which uses a simulation to calculate a likelihood ratio score for each candidate parent to identify the male and female most likely to be the biological parents of a particular pup. We treated each enclosure as a discrete population and conducted a separate simulation and parentage analysis for each enclosure to assess the level of statistical confidence in parentage assignments. All simulations were performed for 10,000 cycles with a genotyping error rate of 0.02. This error rate was based on empirical estimates of two potential sources of error: mutation and mis-scoring of alleles (Solomon et al., 2004). All founders within an enclosure that were alive within 10 d of the estimated conception date of pups were used as candidate parents. Because prairie voles are considered sexually mature at approximately 40 d of age (Solomon, 1991), it is possible that pups born during period one could produce young of their own during period three. However, because pups were removed from enclosures upon reaching 30 g, when they were first classified as adults (Gaines et al, 1979; Getz et al, 1993), therefore they were never considered to be candidate parents. Prairie vole pups typically weigh 2-3 g at birth and gain approximately 0.6-1.0 g daily until weaning at 21 d (Solomon pers. comm.), therefore we estimated the birth date of a pup based on its body mass at first capture. Pup conception dates were estimated by subtracting 21 d from the estimated birth date, as that is the average length of gestation for prairie voles (see Keane et al, 2014 for details). We accepted a parentage assignment only when the confidence level among a male-femalejuvenile trio was >95%, and only these pups were used in the analyses of female reproductive success. We used the genetic parentage data to determine the total number of pups produced by a female within each of the three periods with pups assigned to a specific period based on their birth date.
VEGETATION AT NEST AND RANDOM LOCATIONS
We used forward stepwise regression to determine which combination of vegetation variables best distinguished nest sites from random sites. We used generalized linear mixed models with a binomial logit response and enclosure as a random effect in each model. We first compared models that included each individual vegetation trait using Akaike information criterion (AIC). The model including the vegetation trait with the lowest AIC was then used for the next 'step' where we added the remaining vegetation variables one at a time to this 'best' model, and these models were again compared by AIC. We continued this process until no additional vegetation characteristic decreased the AIC (Meier et al., 2010; Fisher et al, 2013). A model was considered a good model if the AAIC between the model in question and null model was >2.
NEST SITE VEGETATION AND FEMALE SURVIVAL AND REPRODUCTIVE SUCCESS
Because all founding females were equipped with radio-collars, we could determine when a female died to within a week. We used the number of weeks alive (maximum = 15) as our response variable to determine if vegetation characteristics at nest sites affected adult female survival. If females were residents of a nest during more than one period, we calculated means for the nest vegetation data across these periods. For our analysis we used Cox (proportional hazards) forward stepwise regression mixed-effects models, which allows for right censoring of data to account for pups still alive at 40 or 20 d of age, depending on the period. Enclosure was included as a random effect and the various measures of nest vegetation were classified as potential fixed effects for our model selection process. We added new variables until no additional variable decreased the AIC (Meier et al., 2010; Fisher et al, 2013).
We determined the number of pups born to a female while at a particular nest within each period and the estimated the number of days those pups survived to determine if nest vegetation influenced female reproductive success. Only females assigned to a nest within a period were included in these analyses even if they had 0 offspring. Although this might underestimate the number of pups produced by females because mortality can occur before pups are captured, Shuster et al (2013) argues it is critical to include all individuals when estimating fitness. To estimate pup survival, we used the estimated birth date and either the last date a pup was trapped or the date it was removed from the enclosure to calculate the number of days each pup survived. For our analyses we truncated the maximum number of days a pup could survive to 40 d for pups born during the first and second periods, because prairie voles are considered sexually mature at approximately 40 d of age (Solomon, 1991). Although we removed pups from enclosures when they reached a body mass of 30 g, none was removed before 40 d because no pups reached 30 g in this time. For period three we truncated the maximum number of days of survival to 20 d given this was the maximum number of days that pups born during period three could have survived from birth until the time when we began to remove voles from enclosures at the end of the study. To determine the influence of nest vegetation on pup survival, we used the same model selection procedure with Cox survival mixed effects models.
To determine if nest vegetation influenced the number of pups per female, we used forward stepwise regression to determine the best fit mixed effects model (Meier et al., 2010; Fisher et al, 2013). We used a zero-inflated Poisson distribution (Zeileis et al., 2008) and first determined the best fit "base" model of random effects for the model using AIC model selection with potential random effects of period, nest, female ID, and enclosure. We then used forward selection with all of the potential vegetation characteristics until the AIC increased with any additional variable.
ENCLOSURE VEGETATION AND FEMALE SURVIVAL .AND REPRODUCTIVE SUCCESS
As vegetation further from nests might influence survival and reproductive success, we determined if vegetation within enclosures predicted female survival, reproductive success, and pup survival. For these analyses we used the mean value for each vegetation characteristic from 10 random sites sampled within each enclosure. We included all pups born within an enclosure regardless of whether we could assign parentage or a nest to a pup. By eliminating the constraint of matching a pup to a female and nest, we were able to increase our sample size while still examining the potential influence of vegetation on female survival, reproductive success, and pup survival.
To examine the effect of vegetation within an enclosure on the number of pups produced per enclosure, we used forward stepwise regression with negative binomial mixed-effects models to determine the best model with period, enclosure, and number of females alive in the enclosure during the period as potential random effects and all vegetation measures as potential fixed effects. We then used AIC methods to determine the best model. A similar process was repeated, but using stepwise regression via Cox mixed effects models with female and pup survival, to determine which enclosure-wide vegetation traits might influence survival. For female survival vegetation within an enclosure was averaged across all three sampling periods, whereas the pup survival analysis was analyzed with period one, two, or three as a random effect.
Variation for vegetation ranged between 0.3 and 3.5 SE with random sites generally being less variable than nest sites (Table 1). Because cover from thorny and woody plants were highly correlated at both random and nest sites (r = 0.58, P < 0.00001), these two variables were never included in the same model during our stepwise regression procedures. For period one taller and thornier vegetation best distinguished a nest site from a random site (Table 2). A smaller percentage of preferred food plants and a greater percentage of bare ground best distinguished nests from random sites during period two, and the model that best predicted nest sites for period three included a greater percentage of bare ground and thorny vegetation (Table 2).
We were able to assign both parents at 95% confidence to 437 pups (88% of 496 total pups). For 128 (29%) of the pups assigned parentage, we were able to assign them to a particular nest (meaning their mother was assigned residency within the period of their birth; n = 72 nests). The remaining 309 pups were not included in the analyses of nest vegetation and female survival, reproductive success, and pup survival because their mothers were not assigned residency during the period in which they were born.
For the analyses including the vegetation at nest sites, the best model for female survival included vegetation height; shorter vegetation at nest sites predicted greater female survival (Table 3). The best model for pup survival included only the percentage of woody plants at nest sites. A greater percentage of woody vegetation was positively related to pup survival (Table 3). However, our model selection examining the number of pups born at a nest yielded no model that was better than the null model (Table 3) indicating that none of the vegetation characteristics we measured at nest sites was an important predictor of the number of pups produced.
For the enclosure level analyses, the best model for female survival included woody and preferred food plants, with a smaller percentage of woody and a greater percentage of preferred food plants associated with increased female survival (Table 3). There were significant negative correlations between woody or thorny plants and preferred food plants within enclosures (woody/food, r =-0.40, P < 0.001; thorny/food, r =-0.50, P < 0.001). Pup survival was best predicted by a model that included the percentage of dicot cover within an enclosure (Table 3). A greater percentage of dicots predicted increased pup survival. The best model for the number of pups produced per enclosure included only the percentage of thorny plants. A greater percentage of thorny cover predicted more pups produced within an enclosure (Table 3).
Vegetation associated with prairie vole nests differed from that at randomly selected sites although the specific vegetation traits that predicted nest sites varied across sampling periods. Given the differences in vegetation between nests and random sites, it was not surprising that some vegetation associated with a nest site as well as within the enclosure where the nest was located was related to female survival, reproductive success, and pup survival.
In two of three sampling periods, vegetation at nests was taller and/or consisted of a greater percentage of thorny plants compared to random sites. Taller plants could reduce the risk of predation at nests because they might provide more cover whereas thornier plants might impede predator movement. Previous studies reported vegetation influences predation risk of rodents (Townsend's ground squirrels, Urocitellus' toumsendii, Schooley et al., 1996; house mice, Mus musculus, Arthur et al, 2004). Several studies examining the habitat characteristics of den sites of mammals also have suggested females might locate dens in areas where the vegetation reduces predation risk (wolverine, Gulo gulo, Magoun and Copeland, 1998; American marten, Maites americana, Ruggiero et al., 1998; Iberian lynx, Lynx pardinus, Fernandez and Palomares, 2000; yellow mongoose, Cynictis penicillata, Blaum et al, 2007; Pallas cat, Otocolobus manul, Ross et al, 2010).
Contrary to our expectations, nest sites were not associated with high quality food resources, either dicots or preferred food plants, compared to random sites. It is likely that several different selective pressures (e.g., predation risk, access to quality food resources, thermoregulation) are simultaneously influencing nest locations (e.g., Rosalino et al, 2011; Juskaitis et al., 2013). For prairie voles living in our enclosures, vegetation near the nest that reduced predation risk might have been more important than access to high quality food resources as preferred food plants comprised an average of 91% of the total plant cover at all sites (random and nest sites) in all three sampling periods.
During periods two and three, the percentage of bare ground was significantly greater near nests than at random sites. This result seems counterintuitive because a greater percentage of bare ground results in less vegetation to provide protection from predators or in fewer food resources. One possible explanation for the increase in bare ground at nests as compared to random sites is that the taller and more thorny vegetation, that was more prevalent at nest sites during period one, reduced the ability of plants to grow beneath them. Nodding thistle (Carduus nutans L.), which is a relative of Canada thistle (Cirsium arvense) and the perennial Sowthistle (Sonchus arvensis) growing in our enclosures, reduces the rate of seed germination and growth in nearby species (Wardle et al, 1991). It is also possible that the continued use of these nests had a negative impact on vegetation growth near nests through digging by prairie voles or trampling vegetation from repeated live-trapping and radio-tracking at nests by members of the field crew. However, we detected no increase in the percentage of bare ground over time among the 24 nests used by females during all three periods ([chi square] = 0.86, df = 1, P = 0.35).
Solomon and colleagues measured the vegetation height, dicot biomass, monocot biomass, and preferred food biomass at sites with and without prairie vole nests in the same enclosures that we used in our study and found none of these variables predicted the location of nest sites (Solomon et al., 2005). In contrast we found vegetation was taller at nest sites during period one, and there were fewer preferred food plants at nest sites during period two. The discrepancies between studies might be partly the result of annual variability in weather and plant community composition as well as other differences. In the Solomon et al. (2005) study, vegetation was measured in late May, whereas in our study we measured vegetation at three different times starting in mid-June. Therefore, vegetation might have been taller by the time we first sampled it in this study compared to the measurements recorded by Solomon et al. (2005). Additionally, in half the enclosures (n = 3) used in the Solomon et al. (2005) study, the initial density of voles was 50% greater than in our enclosures. Increased intraspecific competition due to the greater population density might have resulted in more voles nesting in sites with less preferred vegetation.
A female's fitness might be influenced by vegetation at two spatial scales: (1) the vegetation near the nest and (2) the vegetation in the habitat patch (e.g., enclosure) where the nest is located (Bowman and Harris, 1980; Martin and Roper, 1988). To our knowledge only a few studies have demonstrated an effect of vegetation associated with mammalian nests/dens on survival and reproductive success (yellow mongoose, Blaum et al., 2007; Iberian hares, Sanchez-Garcia et al, 2012; African wild dogs, Davies et al, 2016). Vegetation at prairie vole nest sites did not affect the number of pups produced per nest, but females in enclosures with more thorny vegetation produced more pups. Pup survival was greater at nests with more woody vegetation and in enclosures with more dicots. These results suggest nests located near more woody vegetation and with more thorny vegetation and dicots in the habitat patch where they are located would yield the greatest number of pups reaching sexual maturity. The thorny and woody vegetation might be providing protection from predators, whereas the dicots would include high-quality food resources.
In general selection of a nest site by prairie voles could result from conflicting selection pressures. Female survival was greatest in enclosures with a greater percentage of preferred food plants but smaller percentage of woody vegetation, whereas juvenile survival was enhanced by a greater percentage of woody vegetation in enclosures. In our enclosures woody vegetation is negatively correlated with preferred food resources indicating more woody vegetation in an area leads to fewer available preferred food resources. Perhaps predation risk is a greater selective pressure for juvenile prairie voles, whereas the energetic demands of reproduction means that access to food resources is more important for breeding females. Such a trade-off seems conceivable because mortality of nestling prairie voles is estimated to be three to six times higher than that of reproductive adults (Getz et al, 1979). Additionally, the costs of reproduction in mammals are high during pregnancy and even higher during lactation (Bronson, 1989). The apparent trade-off in selection between predation risk and need for high quality food resources is seen in other mammalian species. For example the fitness consequences of den location appear to be a trade-off between the reduced risk of predation on pups by locating dens under large Acacia (Acacia sp.) shrubs and the increased prey availability in home range patches with fewer shrubs in yellow mongooses (Blaum et al, 2007).
Our study suggests the selection of nest sites by prairie voles might be an adaptive response to the fitness consequences resulting from vegetation near nests and on a larger spatial scale in the habitat patch where nests are located. Although, in this study, we only examined the association between nest sites and vegetation characteristics, it is likely other habitat features influence nest location. Solomon et al (2005) found voles avoided sites with pre- existing nests. In addition other factors such as soil moisture or hardness might be important for placement of prairie vole nests, which are usually underground. These factors were important for woodland voles, Microtus pinetorum (Fisher and Anthony, 1980; Rhodes and Richmond, 1985). Subsequent studies are needed to examine other habitat characteristics associated with prairie vole nests to better understand the evolutionary pressures that can influence nest-site selection within this species.
Acknowledgments.--We thank Jeremy Fruth at Miami University's Ecological Research Center and Dean Kettle al the University of Kansas s Nelson Environmental Study Area for providing logistical assistance at the field sites. Andor Kiss at Miami University's Center for Bioinformatics and Functional Genomics provided logistical support with the microsatellite analyses. We thank James Lichter, Imani Smith, Shay Spelman, Ryan Ihrig, Morgan Lindsey, Tyler Neu, Teresa Slonaker and numerous other undergraduate students for assistance with field and laboratory research. Financial support for this study was provided by awards from the National Institute of Health (1R15HD075222-01A1) to BK and NGS, Sigma Xi and American Society of Mammalogists to CTL, and an Undergraduate Summer Scholarship awarded to MBO by Miami University.
MALORY B. OWEN and CONNOR T. LAMBERT Department of Biology and the Center for Animal Behavior, Miami University, Oxford, Ohio 45056
BRIAN KEANE Department of Biology and the Outer for Animal Behavior, Miami University, Hamilton, Ohio 45011
NANCY G. SOLOMON Department of Biology and the Center for Animal Behavior, Miami University, Oxford, Ohio 45056
AMOS, YV., J. I. HOFFMAN, A. FRODSHAM, E. ZHANG, S. BKST, AND A. V. S. HILL. 2006. Automated binning of microsatellite alleles: problems and solutions. Mot. Ecol. Notes, 7:10-14.
ARTHUR, A. D., R. P. PECH, AND C. R. DICKMAN. 2004. Habitat structure mediates the non-lethal effects of predation on enclosed populations of house mice. J. Anim. Ecol., 73:867-877.
BAKER, J. A. AND R. J. BROOKS. 1982. Impact of raptor predation on a declining vole population. J. Mammal., 63:297-300.
BELLES-ISLES, J. C. AND J. PICMAN. 1986. Nesting losses and nest preferences in house wrens. Condor, 88:483486.
BERG, L. AND A. BERG. 1998. Nest site selection by the dormouse Muscardinus avellanarius in two different landscapes. Ann. Zool. Fennici, 35:115-122.
BERTOLINO, S. AND N. CORDERO DI MONTEZEMOLO. 2007. Garden dormouse (Eliomys quncinus) nest site selection in an alpine forest. EthoL Ecol. Extol., 19:51-60.
BIRNEY, E. C., W. E. GRANT, AND D. D. BAIRD. 1976. Importance of vegetative cover to cycles of Microtus populations. Ecology, 57:1043-1051.
BLAUM, N., E. ROSSMANITH, G. FLEISSNER, AND F. JELTSCH. 2007. The conflicting importance of shrubby landscape structures for the reproductive success of the yellow mongoose (Cynictis penicillata). J. Mammal.. 88:194-200.
BOWMAN, G. B. AND E. D. HARRIS. 1980. Effect of spatial heterogeneity on ground-nest depredation. J. Wildl. Manage., 44:806-813.
BRONSON, F.H. 1989. Mammalian reproductive biology. The University of Chicago Press, Chicago, Illinois. 336 p.
CAREY, A. B, T. M. WILSON, C. C. MAGI IRE, AND B. E. BISWELI., 1997. Dens of northern flying squirrels in the Pacific northwest. J. Wildl. Manage., 61:684-699.
COCHRAN, G. R. AND N. G. SOLOMON. 2000. Effects of food supplementation on the social organization of prairie voles (Microtus ochrogaster). J. Mammal. 81:746-757.
COLE, R. AND G. O. BATZLI. 1979. Nutrition and population dynamics of the prairie vole, Microtus ochrogaster, in central Illinois. J. Anim. Ecol., 48:455-470.
CUDWORTH, N. L. AND J. L. KOPROVVSKI. 2011. Importance of scale in nest site selection in Arizona gray squirrels. J. Wildl. Manage., 75:1668-1674.
GUSHING, B. S. AND K. M. KRAMER. 2005. Microtines: a model system for studying the evolution and regulation of social monogamy. Ada Theriol. Sinica. 25:182-199.
DAMES, A. B., D. G. MARNEWECK, D. J. DRUCE, AND G. P. ASNER. 2016. Den site selection, pack composition and reproductive success in endangered African wild dogs. Behav. Ecol., 27:1869-1879.
DEJACO, C.E. AND G. O. BATZLI. 2013. Palatability of plants to small mammals in nonnative grasslands of east-central Illinois. J. Mammal., 94:427-435.
DESY, E. A. AND G. O. BATZLI. 1989. Effects of food availability and predation on prairie vole demography: a field experiment. Ecology, 70:411-421.
FERNANDEZ, N. AND F. PALOMARES. 2000. The selection of breeding dens by the endangered Iberian lynx (Lynx pardinus): implications for its conservation, Biol. Conserv., 94:51-61.
FISHER, A. R. AND R. G. ANTHONY. 1980. The effect of soil texture on distribution of pine voles in Pennsylvania orchards. Am. Midi. Nat., 104:39-46.
FISHER, J. T., B. ANHOLT, S. BRADBURY, M. WHEATLEY, AND J. P. VOLPE. 2013. Spatial segregation of sympatric marten and fishers: the influence of landscapes and species-scapes. Urography, 36:240-248.
FORSYTH, D.J. AND D. A. SMITH. 1973. Temporal variability in home ranges of eastern chipmunks (Tamias striatus) in a southeastern Ontario woodlot. Am. Midi. Nat., 90:107-117.
GAINES, M. S., A.M. VIVAS, AND C. L. BARER. 1979. An experimental analysis of dispersal in fluctuating vole populations: demographic parameters. Ecology, 60:814-828.
GETZ, I.. L., L. VERNER, F. R. COLE, J. E. HOFMANN, AND D. E. AVALOS. 1979. Comparisons of population demography of Microtus ochrogaster and M. pennsylvanicus. Acta Theriol., 24:319-349.
--, D. F. GUDERMUTH, AND S. M. BENSON. 1992. Pattern of nest occupancy of the prairie vole Microtus ochrogaster in different habitats. Am. Midi. Nat., 128:197-202.
--. B. MCGUIRE, T. PIZZUTO. J. E. HOFMANN, AND B. FRASE. 1993. Social organization of the prairie vole (Microtus ochrogaster). J. Mammal., 74:44-58.
--, J. E. HOFMANN, B. MCGUIRE, AND T. W. DOLAN. 2001. Twenty-five years of population fluctuations of Microtus ochrogaster and M. pennsylvanicus in three habitats in east-central Illinois. J. Mammal., 82:22-34.
HACKETT, H. M. AND J. F. PAGELS. 2003. Nest site characteristics of the endangered northern flying squireel (Glaucomys sabrinus coloratus) in southwest Virginia. Am. Midi. Nat., 150:321-331.
HASLER, J. 1975. A review of reproduction and sexual maturation in the microtine rodents. Biologist, 57:52-86.
JUSKAITIS, R., L. BALCAIUSKAS, AND V. SLOZINYTE. 2013. Nest site selection by the hazel dormouse Muscardinus avellanarius: is safety more important than food? Xool. Stud., 52:53-61.
KALINOWSKI, S. T., M. L. TAPER, AND T. C. MARSHALL. 2007. Revising how the computer program CERVUS accommodates genotyping error increases success in paternity assignment. Mol. Ecol., 16:1099-1106.
KEANE, B., I.. BRYANT, U. GOYAL, S. WILLIAMS, S. I.. KORTERING, K. E. LUCIA, A. R. RICHMOND, AND N. (5. SOLOMON. 2007. Effect of body condition al weaning on survival and reproduction in prairie voles. Can. J. Zool., 85:718-727.
--, S. PARSONS, B. J. SMICKER, AND N. G. SOLOMON. 2014. Length polymorphism al the avpr1a locus is correlated with male reproductive behavior in a natural population of prairie voles (Microtus ochrogaster). Beliav. Ecol. Sociobiol.. 68:1951-1964.
LI, P. AND T. E. MARTIN. 1991. Nest-site selection and nesting success of cavity-nesting birds in high elevation forest drainages. Auk, 108:405-418.
LIN. Y. K. AND G. O. BATZLI. 1995. Predation on voles: an experimental approach. J. Mammal 76:1003-1012.
--AND--. 2001. The influence of habitat quality on dispersal, demography, and population dynamics of voles. Ecol. Monograph., 71:245-275.
MAGOUN, A. J. AND J. P. COPELAND. 1998. Characteristics of wolverine reproductive den sites. J. Wildl. Manage., 62:1313-1320.
MAHAN, C. G. AND R. H. YVHNER. 1996. Effects of forest fragmentation on burrow-site selection by the Eastern Chipmunk (Tamias striatus). Am. Midi. Nat., 136:352-357.
MARTIN, T. E. AND J.J. ROPER. 1988. Nest Predation and Nest-Site Selection of a Western Population of the hermit thrush. Condor, 90:51-57
MEIER, E. S., F. KIFNAST, P. B. PEARMAN. J. C. SVENNING, W. THUILLER, M. B. ARAUJO, A. GUISAN, AND N. E. ZIMMERMANN. 2010. Biotic and abiotic variables show little redundancy in explaining tree species distributions. Erography, 33:1038-1048.
MCGUIRE, B. AND L. L. GETZ. 1998. The nature and frequency of social interactions among free-living prairie voles (Microtus ochrogaster). Behav. Ecol. SociobioL, 43:271-279.
NADEAU, J. 1985. Ontogeny, p. 254-285. In: R. H. Tamarin (ed.). The biology of new world Microtus. Special Publication No. 8, The American Society of Mammalogists.
RHODES, D. H. AND M. E. RICHMOND. 1985. Influence of soil texture, moisture and temperature on nest-site selection and burrowing by the pine vole, Microtus pinetorum. Am. Midi. Nat., 113:102-108.
ROBERTS, R. L, J. R. WILLIAMS, A. K. WANG, AND C. S. CARTER. 1998. Cooperative breeding and monogamy in prairie voles: influences of the sire and geographic variation. Anim. Behav., 55:1131-1140.
ROSALINO, L. M., D. FKRRIERA, I. LEITAO, AND M. SANTOS-REIS. 2011. Selection of nest sites by wood mice Apodemus sylvaticus in a Mediterranean agro-forest landscape. Ecol. Res., 26:445-452.
ROSS, S., R. KAMNITZER, B. MUNKHTSOG, AND S. HARRIS. 2010. Den-site selection is critical for Pallas's cats (Otocolobus manul). Can. J. ZooL, 88:905-913.
RUGGIERO, L. F., E. PEARSON, AND S. E. HENRY. 1998. Characteristics of American marten den sites in Wyoming. J. Wildl. Manage., 62:663-673.
SANCHEZ-GARCIA, C., M. E. ALONSO, D. J. BARTOLOME, J. A. PEREZ, R. T. LARSEN, AND V. R. GAUDIOSO. 2012. Survival, home range patterns, probable causes of mortality, and den site selection of the Iberian hare flepus, Leporidae, Mammalia) on arable farmland in north-west Spain. Ital. J. ZooL, 79:590-597.
SCHOOLEY R. L., P. B. SHARPE, AND B VAN HORNE. 1996. Can shrub cover increase predation risk for a desert rodent? Can. J. Zool., 74:157-164.
SHUSTER, S. M., W. R. BRIGGS, AND J. A. DENNIS. 2013. How multiple mating by females affects sexual selection. Phil. Trans. Roy. Soc. B, 368:20120046.
SIKES, R. S., J. A. BRYAN, D. BYMAN, B. J. DANIELSON, J. EGGLESTON, M. R. GANNON, W. E. GANNON, D. W. HALE, B. R. JESMER, D. K. ODELL, L. E OLSON, R. D. STEVENS, T. A. THOMPSON, R. M. TIMM, S. A. TREWHITT, AND J. R. WILLOUGHBY. 2016. 2016 Guidelines of the American Society of Mammalogists for the use of wild mammals in research and education. J. Mammal., 97:663-688.
SOLOMON, N. G. 1991. Age of pairing affects reproduction in prairie voles. Lab. Anim.. 25:232-235.
--, B. KEANE, E. R. KNOCH, AND P.J. HOGAN. 2004. Multiple paternity in socially monogamous prairie voles (Microtus ochrogaster). Can. J. Zool., 82:1667-1671.
--, A. M. CHRISTIANSEN, Y. KIRK LIN, AND E. D. HAVES. 2005. Factors affecting nest location of prairie voles (Microtus ochrogaster). J. Mammal., 86:555-560.
--, A. R. RICHMOND, P. A. HARDING, A. FRIES, S. JACQUEMIN, R. E. SCHAEFER, K. E. LUCIA, AND B. KEANE. 2009. Polymorphism at the avpr1a locus in male prairie voles correlated with genetic but not social monogamy in field populations. Mot. Ecol., 18:4680-4695.
STREVTFEILD, C. A., K. E. MABRY, B. KEANE, T. O. CRIST, AND N. G. SOLOMON. 2011. Intraspecific variability in the social and genetic mating systems of prairie voles, Microtus ochrogaster. Anim. Behav., 82:1387-1398.
WARDLE, D. A., M. AHMED, AND K. S. NICHOLSON. 1991. Allelopathic influence of nodding thistle (Carduus nutans L.) seeds on germination and radicle growth of pasture plants. N. Z. J. Agric. Res., 34:185-191.
ZEILEIS, A., C. KLEIBER, AND S. JACKMAN. 2008. Regression models for count data in R. J. Stat. Softw., 27:1-25.
Submitted 15 June 2018
Accepted 20 December 2018
TABLE 1.--Mean ([+ or -] se) height (cm) of the tallest plant, and mean ([+ or -] se) percentage of dicots, bare ground, thorny plant cover, woody plant cover and preferred food plants at random sites and prairie vole nest sites in outdoor enclosures at Miami University's Ecology Research Center during summer and autumn 2016. Data are shown for residency period I, period 2, and period 3 within the 15 wk study season Height (cm) Dicots Bare ground Period 1 Random 72.6 [+ or -] 1.9 28.0 [+ or -] 2.1 4.3 [+ or -] 1.5 Nest 84.0 [+ or -] 2.6 24.9 [+ or -] 2.7 6.6 [+ or -] 1.1 Period 2 Random 80.9 [+ or -] 2.2 21.8 [+ or -] 1.9 3.4 [+ or -] 0.7 Nest 86.7 [+ or -] 3.4 26.7 [+ or -] 3.3 7.8 [+ or -] 1.1 Period 3 Random 88.8 [+ or -] 2.3 21.3 [+ or -] 1.8 1.4 [+ or -] 0.3 Nest 92.0 [+ or -] 3.5 26.9 [+ or -] 3.4 5.1 [+ or -] 1.0 Thorny cover Woody Preferred food Period 1 Random 3.1 [+ or -] 0.7 1.4 [+ or -] 0.4 87.0 [+ or -] 1.7 Nest 8.4 [+ or -] 1.8 1.5 [+ or -] 0.9 87.0 [+ or -] 1.9 Period 2 Random 2.5 [+ or -] 0.7 2.0 [+ or -] 0.8 94.2 [+ or -] 0.9 Nest 5.0 [+ or -] 1.5 5.7 [+ or -] 1.9 86.5 [+ or -] 2.4 Period 3 Random 1.9 [+ or -] 0.4 1.9 [+ or -] 0.5 95.2 [+ or -] 0.8 Nest 8.3 [+ or -] 2.4 6.3 [+ or -] 1.9 88.1 [+ or -] 2.2 TABLE 2.--The top model from AIC stepwise regression of vegetation characteristics that best distinguish prairie vole nest sites from random sites in outdoor enclosures at Miami University's Ecology Research Center during summer-autumn 2016. A (+) indicates a positive association with nest sites, and a (-) indicates a negative association. The [DELTA]AIC represents the change from a null model with no vegetative variables to the selected model Period Model parameter [DELTA]AIC coefficient SE Period 1 Height (+) -16.03 0.02 [+ or -] 0.01 Thorny Cover (+) 0.03 [+ or -] 0.01 Period 2 Preferred Food -12.47 -0.03 [+ or -] 0.01 (-) Bare Ground (+) 0.04 [+ or -] 0.02 Period 3 Bare Ground (+) -30.96 0.13 [+ or -] 0.03 Thorny Cover (+) 0.05 [+ or -] 0.02 SE = Standard Error TABLE 3.--The top Cox mixed effect models and stepwise regression models of the vegetative characteristics associated with prairie vole nests and with random sites within enclosures that best predict pup production and female and pup survival at Miami University's Ecology Research Center during summer-autumn 2016. A (+) indicates a positive association with the response variable, and a (-) indicates a negative association. The [DELTA]AIC represents the change from a null model with no vegetative variables to the selected model Response Model parameters [DELTA]AIC coefficient Pups per Nest Null 0 -- Pup Survival Woody (+) -3.39 0.03 Female Survival Height (-) 0.5 -0.01 Pups per Thorny Cover (+) 2.31 0.07 Enclosure Pups Survival Dicot (+) 0.17 -0.01 by Enclosure Female Survival Woody (-) 3.35 -0.12 by Enclosure Preferred Food (+) 0.04 Response Model parameters SE Pups per Nest Null -- Pup Survival Woody (+) [+ or -] 0.01 Female Survival Height (-) [+ or -] 0.01 Pups per Thorny Cover (+) [+ or -] 0.03 Enclosure Pups Survival Dicot (+) [+ or -] 0.01 by Enclosure Female Survival Woody (-) [+ or -] 0.08 by Enclosure Preferred Food (+) [+ or -] 0.03 SE = Standard Error
|Printer friendly Cite/link Email Feedback|
|Author:||Owen, Malory B.; Lambert, Connor T.; Keane, Brian; Solomon, Nancy G.|
|Publication:||The American Midland Naturalist|
|Date:||Apr 1, 2019|
|Previous Article:||Grassland Bird and Butterfly Responses to Sericea Lespedeza Control via Late-Season Grazing Pressure.|
|Next Article:||Presence of both Active and Inactive Colonies of Prairie Dogs Contributes to Higher Vegetation Heterogeneity at the Landscape Scale.|