Southeastern Naturalist 267 S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 22001155 SOUTHEASTERN NATURALIST 1V4o(2l.) :1246,7 N–2o8. 02 Wildlife Visitation on a Multi-unit Educational Livestock Facility in Northwestern Georgia Susanna E. Kitts-Morgan1, Reneé E. Carleton2,*, Stuart L. Barrow3, Katharine A. Hilburn4, and Amanda K. Kyle5 Abstract - Wildlife visitation of livestock facilities results in economic losses through feed consumption and a potential for disease transmission through fecal contamination of feeds and associated facilities. In order to assess wildlife visitation among livestock-management teaching units on a college campus, we monitored feeding and feed-storage areas via direct observations, live-trapping, and motion-detecting cameras. We also examined visitation patterns and fecal contamination and consumption of grain-based feed and hay. Nine species of wildlife visited the livestock units during the course of the study. Birds and Odocoileus virginianus (White-tailed Deer) were the most frequent visitors in less-enclosed facilities, and rodents, Didelphis virginiana (Virginia Opossum), and Procyon lotor (Raccoon) were mostly documented in more-enclosed facilities. Birds visited daily throughout the year, but documented visitations by Raccoons, Virginia Opossums, and White-tailed Deer occurred only during summer months. Marmota monax (Groundhog) were present each month except for January, February, and March. Of 827 feed samples examined, 16.8% were contaminated by wildlife feces, primarily from birds. Grain-based feed was consumed or removed more frequently than hay, and loss declined during the winter and increased in spring and summer. Introduction Peridomestic wildlife species are attracted to livestock-feeding and feed-storage facilities because of the readily available and easily accessible food resources these areas offer (Daniels et al. 2003). Wildlife visitation and associated consumption and fecal contamination of livestock feed not only results in economic losses to producers (Johnson and Timm 1987, Pimental et al. 2000), but has the potential to increase disease transmission risks between wildlife and both humans and livestock animals (Carlson et al. 2011, Corn et al. 2005, Daniels et al. 2003, Kirk et al. 2002, Tolhurst et al. 2009, Ward et al. 2006). Sturnus vulgaris L. (European Starling) are frequent visitors to feedlots and other animal-feeding operations and have been shown to carry pathogens such as Salmonella enterica (ex Kauffmann & Edwards) Le Minor & Popoff. This situation is of great concern because their droppings have been shown to contaminate cattle feed and water and have the potential to cause disease in humans and cattle (Carlson et al. 2011, Kirk et al. 2002). Likewise, birds, deer, and other mammalian 1Department of Animal Science, Berry College, Mount Berry, GA 30149. 2Department of Biology, Berry College, Mount Berry, GA 30149. 3Kolomoki Mounds Historic State Park, Blakely, GA 39823. 4College of Veterinary Medicine, Auburn University, Auburn, AL 36849. 5UGA/Clinical Pharmacy Program, Georgia Regents University, Augusta, GA 30912. *Corresponding author - rcarleton@berry.edu. Manuscript Editor: Andrew Edelman Southeastern Naturalist S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 268 pests frequenting feedlots and other livestock areas may promote transmission of diseases, such as paratuberculosis (also known as Johne’s disease; Corn et al. 2005, Davidson et al. 2004), bovine tuberculosis (Böhm et al. 2009, Schmitt et al. 2002, Ward et al. 2006), and rabies (Chipman et al. 2013, Dyer et al. 2013), to domestic animals and humans. Certain wildlife species are commonly reported on livestock operations such as dairies and feedlots: Procyon lotor L. (Raccoon; Corn et al. 2005, Ikeda et al. 2004), Odocoileus virginianus Zimmerman (White-tailed Deer; Berentsen et al. 2013, Johnson and Timm 1987, Philips et al. 2012, VerCauteren et al. 2008), small rodents (Johnson and Timm 1987), European Starlings (Johnson and Timm 1987, Linz et al. 2007), and other grain-feeding birds (Palmer 1976). An understanding of factors promoting wildlife visitation is useful when considering means of reducing wildlife–livestock and wildlife–human interactions or minimizing feed losses due to wildlife consumption. Our objectives were to 1) document wildlife presence within livestock-management teaching units on a college campus and 2) investigate their patterns of visitation, consumption, and associated fecal contamination of livestock feed. We expected grain-feeding birds and small rodents to be regular visitors within all of the feeding and feed-storage areas. We also expected these species would be the most common source of deposited feces. Because of the proximity of the livestock units to wildlife habitats, we also expected visitation, to a slightly lesser degree, by medium-sized wild mammals, such as Raccoons. For all wildlife species, we expected a higher rate of visitation during winter months, when natural food sources would be reduced, than in summer months. Field-Site Description We conducted this study (June 2011–April 2012) within 3 livestock-management teaching units at Berry College (34º17'43.82"N, 85º11'12.6"W) in northwestern Georgia. The college property is expansive and consists of approximately 10,500 ha, including a wildlife management area and wildlife refuge located adjacent to the livestock units (Fig. 1). Mixed Quercus (oak)/Carya (hickory) and Pinus (pine) spp. forests dominate the land tract and border livestock- grazing pastures and concentrations of campus buildings. All of the units had been actively housing livestock for more than 10 years at the time the study was conducted. Each unit, located on separate areas of the campus, featured different livestock (horses, dairy cattle, and beef cattle), supportive feeding and feed-storage areas, and feedstuffs of differing composition. The units also differed in the design of associated buildings (i.e., fully enclosed by 4 complete walls and a roof, partially enclosed having at least 2 walls and a roof, or open-air with less than 2 solid walls but under roof), type of feed used, and method of livestock feeding. The equine unit consisted of 2 partially enclosed and connected barns containing multiple stalls, 2 adjoining partially enclosed 10 m × 4 m hay sheds, and surrounding paddocks. Grain-based equine feed was stored in fully enclosed bunks. Horses were fed grain from individual buckets and hay from racks in their stalls. A working dairy, dairy-cattle feeding and holding areas, feed and hay storage, and Southeastern Naturalist 269 S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 adjacent pastures comprised the dairy unit. The feeding and holding barn for milking cows was of an open-air design. Dairy cows were fed Total Mixed Ration (TMR; Purina Mills, Gainesville, GA) and hay exclusively. The TMR, which was constantly available and replenished frequently, was deposited in a 20-m long, linear row in the feeding area. Hay was stored in the barn and fed to dairy cows in the holding area. In addition to beef cattle grazing pastures, a fully enclosed 30 m × 17 m feed- and hay-storage barn and open-air cattle pens made up the beef cattle unit. Corn gluten pellets were the only grain-based ration fed to the beef cattle. The pellets were distributed into troughs placed within beef-cattle pastures and stored within an open bunk inside the barn. Hay was distributed in beef-cattle pastures during winter months to supplement available pasture grasses. Amount and timing of human activity also varied by unit, ranging from near-constant daylight-hour activity in the equine barn to peaks of activity for milking and feeding activity at the dairy or a limited twice-per-day visit at the beef-unit storage building. Methods Wildlife documentation We employed various methods to document wildlife presence in order to maximize the likelihood of detecting species of differing sizes and habits. These methods included remote-imaging, live-trapping, direct visual observation, and fecal evidence. For each instance of visitation, we noted the species, the livestock unit in which it was observed, the design of facility, and feed type used at that unit in addition to the date of visitation. We used motion-detecting game cameras (Moultrie DGS-200, Global Point Products, Farmington, NY) as our most-constant means Figure 1. Map of study site in northwest Georgia (Berry College) showing location of livestock units and adjacent wildlife refuge and wildlife management areas (WMA). Southeastern Naturalist S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 270 of monitoring to remotely record medium- to larger-sized animals that we expected would visit the units when human activity was minimal. Six cameras were deployed from mid-June 2011 through April 2012 for a total of 302 days (mean = 27.5 days per month). We placed 2 cameras at each unit; 1 within a livestock-feeding area and 1 within a feed-storage area. Based on trials we conducted beforehand, the cameras were aimed at feeding areas or stored feed in a position to take maximal advantage of the cameras’ 42° field of view. Cameras were active for 24 hr/day throughout the study, except for short periods when we retrieved them to replenish their batteries and download images. We programmed the cameras to take still photographs at 30- sec intervals during activation. For 31 nights between 11 July and 10 August 2011, when the college’s student population was minimal, we set live traps at each unit, in addition to the cameras, to capture and document elusive small- to medium-sized wildlife, such as rodents, Raccoons, and Virginia Opossums. We placed a Havahart 82.28-cm, single-door trap (Woodstream Corporation, Lititz, PA) baited with sugar-coated apple pieces and fermented cantaloupe in each of two areas at the dairy and equine units for a combined total of 124 trap nights. Four 7.6 cm × 8.9 cm × 22.8 cm folding traps (H.B. Sherman, Tallahassee, FL), baited with rolled oats to target small rodents, were placed at least 4 m apart but adjacent to feeding or within feed-storage areas at each unit for a combined total of 372 trap nights. Traps were open approximately dusk to dawn. Since we could not be certain of how many unique individuals of a particular species were recorded by camera each day, we compared mammal visitations among units in terms of numbers of detection/no detection. Captured rodents, Raccoons, and Virginia Opossums were humanely killed following American Society of Mammologists guidelines (Sikes et al. 2011) because they were considered nuisance animals, i.e., those causing damage to the facilities or capable of spreading disease to livestock or personnel. Relocation was not considered an option per United States Department of Agriculture recommendations (USDA APHIS 2011). Tissues from these animals were collected for a separate, regional pathogenprevalence study currently in progress. We also documented wildlife presence by recovering and identifying feces deposited in samples of feed (methodology described below). Our protocol for manipulation of live animals was approved by the Institutional Animal Care and Use Committee of Berry College and carried out under a State of Georgia Scientific Collecting permit. Evaluation of feed contamination and consumption We used 50 cm × 40.5 cm × 2.5 cm fast-food restaurant-style trays to hold feed or hay samples to evaluate fecal contamination and feed removal by wildlife. Before placing feed samples into the trays, we thoroughly examined fresh quantities of feed or hay for feces or other contaminants and removed any if found. After weighing each tray to the nearest 0.01 g, we completely filled them with the sample and then reweighed the tray. At the equine unit, we randomly placed 4 trays adjacent to or on stored Cynodon sp. (Bermuda grass) hay bales. In the beef unit, corn gluten pellets were stored in an enclosed building but on the Southeastern Naturalist 271 S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 floor of an open bunk consisting of a 6.5 m × 6 m concrete pad bounded by low concrete block walls on 3 sides. We placed 1 tray containing pellets adjacent to and 1 tray on the pile of stored pellets, and 2 trays filled with hay were placed on and adjacent to Bermuda grass square bales. At the dairy barn, we placed 2 trays containing TMR at either end of the feeding area and 2 trays containing Bermuda grass hay on square bales stored at one end of the barn. With the exception of the TMR samples, we left trays exposed to wildlife for approximately 1 week before collecting them. Due to its water content and propensity to develop mold, we recovered the TMR samples each day after 12 hrs (17:00–05:00) and replaced them with fresh samples at the beginning of another 12-hr period. After retrieving the samples from their respective locations, we reweighed them and performed a thorough visual examination for fecal material. Feces were removed and later visually identified (by S.L. Barrow or A.K. Kyle) based on appearance as having been produced by a bird (small size of irregular shape and possibly containing white urates), small rodent (small-sized oblong pellets), or other mammal (larger in size). We determined the amount of feed presumably removed or consumed by calculating the difference between weight before placement and after recovery. For a small number of samples, there was an increase in weight, probably due to incidental addition of feed caused by animal movement, unit worker interaction, or urination into the tray by Felis catus L. (Domestic Cat), and we did not include those in our analysis of feed removal. To compensate for temperature-associated water evaporation from TMR samples, we used dry-matter weight to determine the weight lost from each sample. The differing nature of each unit in terms of building design, type of feedstuffs used, manner of feed storage or distribution, human activity, and number of sample trays we could place without disrupting feeding activities presented us with a number of confounding factors including unequal sample sizes and differences in the number of days each feed type was available. More samples of TMR were exposed to wildlife than other feed samples because of its propensity to mold if not consumed within 24 hours. This reduced the time during which the samples could have been visited compared to the other feed types. While not ideal, we felt that collecting a greater number of samples of TMR per week would sufficiently compensate for the difference. Analyses JMP® 8.0.2 statistical software was used for all analyses (α = 0.05). We compared wildlife visitation among units for live-trapping and camera-detection methods by means of likelihood ratio chi-square tests with unit (the independent variable) and visitation (0 = no visitation recorded and 1 = visitation recorded) assigned categorical roles. In JMP®, the negative log-likelihood of the response probabilities is used to compute the likelihood ratio (Sall et al. 2005). We used Fisher’s exact test to perform post-hoc pair-wise comparisons of number of visitations between units (Sall et al. 2005). To determine changes in visitation patterns for each mammal species, we also used likelihood ratio chi-square tests with month as the independent Southeastern Naturalist S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 272 variable and whether or not visitations were recorded for each day of the month as the response variable (Sall et al. 2005). We standardized fecal-contamination and consumption data by dividing the number of contaminated samples or weight of feed removed from samples by days of sample availability. Distributions of both the fecal-contamination data and feed-removal data were skewed as the majority of samples either did not contain feces or did not have feed removed from them. We used Mann-Whitney U tests for nonparametric distributions to make relevant comparisons among explanatory variables (Sall et al. 2005). Numbers of standardized fecal-contaminated samples were compared among feed types, with livestock unit as an additional explanatory variable, among the months of the study. Standardized weights of feed removed were also compared among the explanatory variables feed type, livestock units, and months of the study. We also calculated percentages of samples containing feces by month and for the study period and percentage of feed lost by weight by month and for the study period after combining all samples of each type from the units. Results Wildlife observations We documented 9 species of wildlife by direct observation, motion-detecting camera, and live trapping (Table 1). Images of one or more individual mammals were recorded among the cameras on 64 days (21.2% of recording days). The number of visitations documented by camera varied by unit (likelihood ratio chisquared test: G = 95.13, df = 2, G2 < 0.0001; Table 1) with more wildlife visits at the beef unit compared to both the dairy and equine units (Fisher’s exact test: df = 1, for the probability of camera documentation at the beef unit greater than the dairy unit P < 0.0001, and probability of camera documentation at the beef unit greater than the equine unit P < 0.0001); more visits were documented by camera at the dairy unit than the equine unit (Fisher’s exact test: df = 1, P = 0.018). Eleven Table 1. Visitations by mammalian wildlife species to feeding and/or feed-storage areas within 3 livestock-management teaching units on a college campus in Georgia. Visitations (presence or absence per day) were documented by visual observation (V; June 2011–April 2012), remote camera (C; June 2011–April 2012), or live-trapping (T; July–August 2011). Zero values indicated no species were detected by any of the methods used, and + indicates individuals were detected but not counted. Livestock unit Wildlife species Equine Dairy Beef Procyon lotor (Raccoon) 4 (T), 1 (C) 1 (T) 0 Didelphis virginiana (Virginia Opossum) 1 (T) 0 1 (C) Marmota monax (Groundhog) 0 0 58 (C) Mus musculus (House Mouse) 0 1 (T) 4 (T) Odocoileus virginianus (White-tailed Deer) 0 8 (C) 0 Passer domesticus (House Sparrow) + (V) + (V) + (V) Sturnus vulgarus (European Starling) 0 + (V) + (V) Columba livia (Rock Dove) 0 + (V) 0 Hiurndo rustica (Barn Swallow) + (V) 0 0 Southeastern Naturalist 273 S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 mammals were trapped during 10 of 23 days (43.5% of trapping days) during July and August 2011 (mean =11.5 days per month). No animals were trapped on the same day that animal images were recorded on camera. There was no significant difference in the number of days that mammals were trapped among the livestock units (likelihood ratio chi-squared test: G = 1.61, df = 2, G2 = 0.44; Table 1). Although we trapped Mus musculus L. (House Mouse) or other small rodents, we also noted their presence by fecal evidence in the feed samples we examined. Other wild mammals visiting the units included White-tailed Deer (Fig. 2A), Marmota monax L. (Groundhog; Fig. 2B), Raccoons, and Virginia Opossums. After House Mice, Groundhogs were the most frequent visitors, followed by White-tailed Deer, Raccoons, and Virginia Opossums (Table 1). Most of the medium-sized mammals tended to be secretive and thus we only captured or documented them by photography at night and in barns or buildings that were mostly enclosed and had limited human activity. Mammals visited the units primarily after sunset, but birds were seen most often during daylight hours. Birds were present at all units throughout the study. We observed them frequently in the open-air barns of the dairy, but also in the equine unit and inside the enclosed feed-storage barn. We visually observed Passer domesticus L. (House Sparrow) and European Starlings feeding alongside dairy cows inside the dairy barn each time we collected feed samples. House Sparrows and Hirundo rustica L. (Barn Swallow) also nested in the equine and dairy barns in proximity to stored hay, and Columba livia Gmelin (Rock Dove) roosted in the beef feed- and hay-storage barn. Documented visitation by camera of mammals occurred during summer and fall months but not during mid-to-late winter months. The number of documented visits per month by Groundhogs (likelihood ratio chi-squared test: G = 113.67, df = 10, G2 < 0.0001), but not White-tailed Deer (likelihood ratio chi-squared test: G = 15.10, df = 10, G2 = 0.1284), differed significantly (Fig. 3); Raccoons and Virginia Opossums were only documented by camera once, and because trapping was limited, comparisons to demonstrate seasonal patterns were not possible. Figure 2. Wildlife remotely documented consuming feed within livestock units located on a college campus in Georgia: (A) White-tailed deer photographed within the dairy unit, September 2011 at 03:37. (B) Groundhogs photographed on corn gluten pellets in a beef feed-storage building, June 2011 at 19:49. Southeastern Naturalist S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 274 Feed contamination We examined a combined total of 827 feed and hay samples for wildlife fecal contamination (Table 2). Most (83.2%) of the samples were not contaminated with feces. In total (combined across all locations and feed types), we recovered bird feces more frequently (10.04% of contaminated feed) than feces from small rodents (1.8%) and other unidentified non-rodent mammals (0.5%). Insects, including beetles, small flies, ants, and spiders, were also recovered from some (4.5%) samples. Figure 3. Number of visits per month by Groundhogs and White-tailed Deer at livestock units on a college campus in Georgia, June 2011–April 2012, as documented by remote camera. Table 2. Percentages of total livestock feed samples contaminated by wildlife feces or insects grouped by taxon, livestock unit, and feed type, and mean number of samples contaminated per day with standard error (SE). Feed samples consisted of total mixed ration (TMR), hay, or corn gluten pellets. Trays containing samples were positioned on or adjacent to feed stored or distributed within livestock units of a college campus in Georgia, June 2011–April 2012. Dairy unit Equine unit Beef unit Wildlife contaminator TMR Hay (hay only) Pellets Hay None 80.3% 78.6% 94.5% 75.8% 89.5% Bird 14.6% 8.6% 1.4% 3.2% 5.3% Small rodent 0.4% 0.3% 2.7% 8.1% 2.6% Other mammal 0.0% 0.0% 0.7% 16.0% 2.6% Insect 0.47% 10.0% 0.7% 11.3% 0.0% Total samples (n =827) 472 70 147 62 76 Mean samples contaminated/day 0.144 0.025 0.008 0.032 0.08 (SE) ( 0.016) (0.008) (0.003) (0.009) (0.006) Southeastern Naturalist 275 S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 Birds were the most-frequent contaminators by percentage of contaminated samples of both TMR (97.0%) and stored hay (41.1%). However, more contaminated samples of corn gluten pellets from the beef feed-storage barn contained feces from mammals (80.0%) than birds (20.0%). Few samples of stored hay from the equine barn contained feces of any kind, and no samples from the dairy barn contained feces from non-rodent mammals. There were significant differences in the types of feed contaminated (Mann-Whitney U test: U = 7.84, df = 2, P = 0.0198), with more samples of TMR contaminated than pellets or hay (Table 2); livestock unit as a factor did not contribute to the model (Fig. 4A). More samples of TMR were contaminated in November and December than any other month (Mann-Whitney U test: U = 56.75, df = 10, P < 0.001). Contamination of hay, the only feed type used in all units, varied by month (Mann-Whitney U test: U = 21.22, df = 10, P < 0.019; Fig. 4A), but did not vary among units (Mann-Whitney U test: U =5.276, df = 2, P = 0.071). Contamination of corn gluten pellets, which were used only in the beef unit, also varied by month (Mann-Whitney U test: U = 19.03, df = 10, P = 0.039; Fig. 4A). Feed consumption/loss We compared weights of 721 feed samples before and after exposure to potential consumption by wildlife (Table 3). Consumption or loss of grain-based feed was significantly greater than hay (Mann-Whitney U test: U = 436.53, df = 2, P less than 0.001); livestock unit was not an appropriate factor because only hay samples were examined from the equine unit. Approximately 22 kg (35.3% of the total sample weight) of corn gluten pellets were removed or consumed from trays located in the beef feed-storage barn and a total of 15.68 kg (21.6%) of TMR (dry matter) was lost or consumed from dairy-barn samples. Consumption or loss of hay did vary among the units (Mann-Whitney U test: U =11.28, df = 2, P = 0.0036), with more loss occurring at the beef (24.0% of total sample weight) and dairy (19.5%) units than the equine unit (10.6%). The amount of hay lost from sample trays by consumption or removal was less than 1 kg per unit over the course of the study, but when combined, represented a loss of 15.1% of the total sample weight. There was a loss of feedstuffs of all types each month during the study (Fig. 4B). Seasonally, loss of TMR was greatest from June through September of 2011 compared to other months (Mann-Whitney U test: U = 128.94, df = 10, P < 0.0001). Table 3. Summary statistics of feed consumed or removed by wildlife (g/day) by feed type at livestock units on a college campus in Georgia (June 2011–April 2012). Livestock Unit Dairy Equine Beef Mean Mean Mean Feed type n (g/day) SE n (g/day) SE n (g/day) SE Hay 37 1.49 0.29 113 0.76 0.14 70 2.36 0.46 Corn gluten pellets - - - - - - 53 74.12 8.98 Total mixed ration 448 93.03 4.46 - - - - - - Southeastern Naturalist S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 276 Figure 4. (A) Percentages of livestock-feed samples containing wildlife feces grouped by month and feed type. (B) Percentages of feed removed from sample trays by wildlife visiting livestock units grouped by month and feed type. Samples were exposed to potential wildlife visitation for a period of 12 hours (total mixed ration (TMR) or 7 days (hay and corn gluten pellets) during the study period June 2011–April 2012 within mixed livestock units on a college campus in Georgia. Southeastern Naturalist 277 S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 Similarly, more corn gluten pellets were lost from June through August 2011, but also in November 2011 and March and April 2012, than in other months (Mann- Whitney U test: U = 32.16, df = 10, P = 0.0004). Monthly variation in loss of hay was significant at the beef unit (Mann-Whitney U test U = 30.82, df = 10, P = 0.0006), with losses greater in June, July, and August 2011, but not at the dairy (Mann-Whitney U test: U = 10.51, df = 10, P = 0.39) or the equine unit (Mann- Whitney U test: U = 17.03, df = 10, P = 0.073). Discussion A variety of wildlife species visited the feeding and feed-storage areas of the livestock units, and many were the same as those reported in other studies (Anderson et al. 2007, Berentsen et al. 2013, Carlson et al. 2011, Corn et al. 2005, Daniels et al. 2003, Kirk et al. 2002, Philips et al. 2012, VerCauteren et al. 2008). The patterns of wildlife visitation and fecal contamination of feeds observed in this study and others are most likely reflective of their ecological phenology. For example, studies conducted in the United Kingdom found that farm-building visitations by Meles meles L. (Eurasian Badger) were more likely to occur during the spring and summer and mainly during periods of low rainfall when their earthworm prey were less available (Garnett et al. 2002, Tolhurst et al. 2009). The Groundhogs we observed were the dominant mammal within the beef feed- and hay-storage areas. While these animals were active mostly at night, they were occasionally seen during the day depending on time of year and weather conditions. We rarely observed them during the winter months, which coincided with the minimal loss of corn gluten pellets we documented from December 2011 through February 2012. Groundhogs typically enter hibernation as early as November and may not emerge until March (Davis 1967, Grizzell 1955). Their tendency to hibernate and only occasionally emerge from their burrows during these months (Davis 1967) is a reasonable explanation for their absence. During the month of April, we recorded no Groundhog presence, and it was possible some of them had emerged from hibernation to breed in March and left the immediate area to find mates farther away (Davis 1967). White-tailed Deer have been previously documented visiting cattle-feeding areas (Berentsen et al. 2013, Philips et al. 2012). White-tailed Deer visitations at these operations have been observed to increase during January and peak in June coinciding with the fawning season and lactation (Berentsen et al. 2013). We documented visits by 1 or more White-tailed does within the dairy barn in June, July, and also a single visit in September, but not in other months. As the visits occurred during fawning season, we hypothesize the increased energy demands attributed to lactation prompted the visitations. We found no reports of them entering a partially enclosed structure to feed, as was the case on our site, which is atypical of their normal behavior. Similar to other findings (Daniels et al. 2003), we found birds and small rodents to be frequent fecal contaminators of grain-based feeds. Because mammals other than rodents tend to defecate away from food sources, we found very little evidence of fecal contamination by them. Visitation by Raccoons was not unexpected because they can often be found in the vicinity of artificial food sources Southeastern Naturalist S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 278 (Prange et al. 2004). Although we trapped Raccoons in the equine-housing areas and found a latrine in an adjacent barn, we were surprised none were documented in the hay-storage shed given its proximity to the trapping location. It is possible that the lack of secluded areas in the hay-storage area, abundance of hiding places in adjacent barns, and the simple lack of appeal of grass hay compared to other foodstuffs discouraged them to venture there. The necessity of not interfering with normal feeding operations restricted the number of feed sample trays we could employ. We were limited to using only 2 sample trays per sampling area within each unit and were restricted in where we could place the trays to avoid consumption by livestock or interference by and disturbance of unit workers. Both of these factors may have reduced the likelihood of wildlife contact with the samples. Our finding of fecal contamination in 16.8% of samples may seem minor but, given the total amount of feed or hay to which birds and rodents had access, we conclude that much of the stored feed at the units becomes contaminated by 1 or more wildlife species, and hence, a potential for disease transmission exists. In their study examining fecal contamination of stored livestock feeds, Daniels et al. (2003) found in Scotland that a single cow could potentially ingest well over 1000 rodent and/or bird feces during the winter and, based on infectious-disease modeling, ingestion of contaminated feed would increase risk of infection by diseases known to be carried by wildlife and prevalent in that country. Cows housed in the dairy unit of this study consume approximately 13 kg of TMR daily (I. Peeler, Berry College, Mount Berry, GA, pers. comm.). Based on our estimate that 14.4% of TMR samples were contaminated with bird or rodent feces, 1.83 kg of their daily ration could be contaminated. As we had no disease-prevalence data for the wildlife species visiting the units, calculation of infection risk was not possible. The amount of feed consumed or removed by wildlife, which we reported as percentage lost, is likely much greater than the loss our sample trays represented because the amount of feed placed in the trays, the basis of the analyses, was only a minimal proportion of the total feed exposed to wildlife. Groundhogs, by virtue of their size and the regularity of their visits, appeared to consume more feed, namely corn gluten pellets, than any of the other wildlife species we documented. The easy accessibility of the pellets, along with the relative seclusion of the storage barn and ability of the animals to feed undisturbed, probably contributed to the amount of sample lost. Groundhogs frequently inhabit burrows under farm buildings, as was the case at the beef feed-storage barn, and so, like other species of rodents, would focus much of their feeding in these areas. The results of our study suggested that visitation by a variety of wildlife species occurs commonly within the livestock units and results in consumption and fecal contamination of feedstuffs. This situation is cause for concern, mainly because of a potential for disease transmission from wildlife to livestock and the unit workers through fecal contamination of feedstuffs and unit areas. The design of the facilities and manner in which livestock were fed and feedstuffs stored appeared to influence visitation patterns, but other factors, such as human activity and weather, may Southeastern Naturalist 279 S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 have also contributed. A mark-and-release study would increase our understanding of visitation patterns and help estimate disease prevalence and transmission risks. While it is not possible to completely prevent wildlife from entering feeding or storage areas, we recommend instituting protection by storing feed in sealed containers or secure, non-accessible enclosures, covering potential entry ways, or using deterrent devices, such as electrical fencing, to reduce consumption loss and risk of contamination. Literature Cited Anderson, J.L., J.K. Meece, J.J. Koziczkowski, D.L. Clark, R.P. Radcliff, C.A. Nolden, M.D. Samuel, and J.L. Ellingson. 2007. Mycobacterium avium subsp. paratuberculosis in scavenging mammals in Wisconsin. Journal of Wildlife Diseases 43:302–308. Berentsen, A.R., R.S. Miller, R. Misiewicz, J.L. Malmberg, and M.R. Dunbar. 2013. Characteristics of White-tailed Deer visits to cattle farms: Implications for disease transmission at the wildlife-livestock interface. European Journal of Wildlife Research 60:161–170. Böhm, M., M.R. Hutchings, and P.C.L. White. 2009. Contact networks in a wildlife–livestock host community: Identifying high-risk individuals in the transmission of bovine TB among badgers and cattle. PLoS ONE 4:e5016. Carlson, J.C., A.B. Franklin, D.R. Hyatt, S.E. Pettit, and G.M. Linz. 2011. The role of Starlings in the spread of Salmonella within concentrated animal-feeding operations. Journal of Applied Ecology 48:479–486. Chipman, R.B., T.W. Cozzens, S.A. Shwiff, R. Biswas, J. Plumley, J. O'Quinn, T.P. Algeo, C.E. Rupprecht, and D. Slate. 2013. Costs of Raccoon rabies incidents in cattle herds in Hampshire County, West Virginia, and Guernsey County, Ohio. Journal of the American Veterinary Medical Association 243:1561–1567. Corn, J.L., E.J. Manning, S. Sreevatsan, and J.R. Fische. 2005. Isolation of Mycobacterium avium subsp. paratuberculosis from free-ranging birds and mammals on livestock premises. Applied and Environmental Microbiology 71:6963–6967. Daniels, M.J., M.R. Hutchings, and A. Greig. 2003. The risk of disease transmission to livestock posed by contamination of farm-stored feed by wildlife excreta. Epidemiology and Infection 130:561–568. Davidson,W.R., E.J.B. Manning, and V.F. Nettles. 2004. Culture and serologic survey for Mycobacterium avium subsp. paratuberculosis infection among southeastern Whitetailed Deer (Odocoileus virginianus). Journal of Wildlife Diseases 40:301–306. Davis, D.E. 1967. The role of environmental factors in hibernation of Woodchucks (Marmota monax). Ecology 48:83–689. Dyer, J.L., R. Wallace, L. Orciari, D. Hightower, P. Yager, and J.D. Blanton. 2013. Rabies surveillance in the United States during 2012. Journal of the American Veterinary Medical Association 243:805–815. Garnett, B.T., R.J. Delahay, and T.J. Roper. 2002. Use of cattle-farm resources by Badgers (Meles meles) and risk of bovine tuberculosis (Mycobacterium bovis) transmission to cattle. Proceedings of the Royal Society of London, Series B 269:1487–1491. Grizzell, Jr., R.A. 1955. A study of the Southern Woodchuck, Marmota monax monax. American Midland Naturalist 53:257–293. Ikeda, T., M. Asano, Y. Matoba, and G. Abe. 2004. Present status of invasive alien Raccoon and its impact in Japan. Global Environmental Research 8:25–131. Southeastern Naturalist S.E. Kitts-Morgan, R.E. Carleton, S.L. Barrow, K.A. Hilburn, and A.K. Kyle 2015 Vol. 14, No. 2 280 Johnson, R.J., and R.M. Timm. 1987. Wildlife damage to agriculture in Nebraska: A preliminary cost assessment. Pp. 57–65, In N.R. Holler (Ed.). Proceedings of the Third Eastern Wildlife Damage Control Conference. University of Nebraska-Lincoln, Lincoln, NE. 357 pp. Kirk, J.H., C.A. Holmberg, and J.S. Jeffrey. 2002. Prevalence of Salmonella spp. in selected birds captured on California dairies. Journal of the American Veterinary Medical Association 220:359–362. Linz, G.M., H.J. Homan, S.M. Gaulker, L.B. Penry, and W.J. Bleier. 2007. European Starlings: A review of an invasive species with far-reaching impacts. Pp. 378–386, In G.W. Whitmer, W.C. Pitt, and K.A. Fagerstone (Eds.). Managing Vertebrate Invasive Species: Proceedings of an International Conference. US Department of Agriculture-Animal and Plant Health Service-Wildlife Services, Washington, DC. 481 pp. Palmer, T.K. 1976. Pest-bird damage control in cattle feedlots: The integrated systems approach. Pp. 17–21, In C.C. Siebe (Ed.). Proceedings of the 7th Vertebrate Pest Conference. University of California, Davis, CA. 317 pp. Philips, G.E., M.J. Lavelle, J.W. Fischer, J.J. White, S.J. Wells, and K.C. VerCauteren. 2012. A novel bipolar electric fence for excluding White-tailed Deer from stored livestock feed. Journal of Animal Science 90:4090–4097. Pimentel, D., L. Lach, R. Zuniga, and D. Morrison. 2000. Environmental and economic costs of nonindigenous species in the United States. BioScience 50:53–65. Prange, S., S.D. Gehrt, and E.P. Wiggers. 2004. Anthropogenic resources on Raccoon (Procyon lotor) movements and spatial distribution. Journal of Mammalogy 85:483–490. Sall, J., L. Creighton, and A. Lehman. 2005. JMP Start Statistics, 3rd Edition. Brooks/Cole, Monterey, CA. 560 pp. Schmitt, S.M., D.J. O’Brien, C.S. Brunning-Fahn, and S.D. Fitzgerald. 2002. Bovine tuberculosis in Michigan wildlife and livestock. Michigan bovine tuberculosis bibliography and database. Available online at http://digitalcommons.unl.edu/michbovinetb/114. Accessed 7 July 2014. Sikes, R.S., and W.L. Gannon. 2011. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 92:235–253. Tolhurst, B.A., R.J. Delahay, N.J. Walker, A.I. Ward, and T.J. Roper. 2009. Badger (Meles meles) behavior in farm buildings: Opportunities for transmission of Mycobacterium bovis? Applied Animal Behaviour Science 117:103–113. United States Department of Agriculture, Animal and Plant Health Inspection Service, Wildlife Services Division (USDA APHIS). 2011. Relocating wildlife requires caution. Available online at http://www.aphis.gov/publications/fs_relocating_wildlife_2011.pdf. Accessed 2 May 2011. VerCauteren, K.C., M.J. Lavelle, and G.E. Phillips. 2008. Livestock-protection dogs for deterring deer from cattle and feed. The Journal of Wildlife Management 72:1443–1448. Ward, A.L., B.A. Tolhurst, R.J. Delahay. 2006. Farm husbandry and the risks of disease transmission between wild and domestic mammals: A brief review focusing on bovine tuberculosis in badgers and cattle. Animal Science 82:767–773.