Habitat Interference by Axis Deer on White-Tailed Deer - BioOne
Journal of Wildlife Management 74(4):698–706; 2010; DOI: 10.2193/2009-135
Management and Conservation Article
Habitat Interference by Axis Deer on White-Tailed Deer CLINTON J. FAAS, Department of Biology, Texas State University, San Marcos, TX 78666, USA FLOYD W. WECKERLY,1 Department of Biology, Texas State University, San Marcos, TX 78666, USA
ABSTRACT Many studies of interactions between exotic and native ungulates have not had temporal and spatial controls nor have they considered the types of competitive interactions that would allow coexistence. For exotic axis deer (Axis axis) and native white-tailed deer (Odocoileus virginianus) to coexist one species should be superior at interference competition and the other species should be superior at exploitative competition. We generated and tested predictions, based on body size and diet breadth, about habitat selection by white-tailed deer in the presence and absence of axis deer, dominance relationships, and time at sites provisioned with high quality forage. We conducted our study in treatment (axis and white-tailed deer) and control (white-tailed deer only) areas when both species were present and after axis deer were removed. We conducted vehicle surveys to determine habitat use of both species. At provisioned feeding sites we recorded aggressive behaviors and amount of time species spent at feeding sites alone and together. In the treatment area white-tailed deer selection for wooded habitat increased 2.1 times after axis deer were removed, whereas habitat selection by white-tailed deer was constant in the control area over the same time. At feeding sites axis deer were dominant to white-tailed deer; both species spent a significantly greater amount of time alone than at feeders together, but amount of time that individuals of each species spent at feeders did not differ. Axis deer were superior at interference competition, but white-tailed deer were not superior at exploitative competition; thus, species coexistence is unlikely. Whether white-tailed deer are negatively impacted by axis deer at spatial scales larger than our experiment probably depends on abundance of axis deer at larger spatial scales. Experiments of species interactions with temporal and spatial controls that consider types of competitive interactions increase a manager’s understanding of when and how native ungulates may be negatively impacted by exotic ungulates.
KEY WORDS Axis axis, coexistence, exotics, habitat, interspecific competition, Odocoileus virginianus, temporal and spatial controls, Texas.
Understanding how distribution of ungulates is affected by habitat selection is necessary for examining resource partitioning and competition (Stewart et al. 2002). However, competitive interaction studies among cervids, especially between exotics and native species, have rarely been conducted (Bartos et al. 2002). Stewart et al. (2002) reported that studies of competition between large herbivores are complicated because addition or removal experiments are not feasible in most cases, largely due to wideranging distributions of large herbivores and difficulty in conducting manipulative experiments that have spatial and temporal controls (White and Garrot 1990, Gabor and Hellgren 2000). Since the time of their original introduction into highfenced enclosures, many exotic species have escaped or been released and free-ranging populations continue to increase (Butts 1979, Mungall and Sheffield 1994, Demarais et al. 1998). With increasing numbers of exotic species there is an increased need to determine negative effects on native white-tailed deer (Odocoileus virginianus). Because exotic species in Texas, USA, are regulated as livestock and not game animals, there has been little done to control numbers of animals outside of managed herds within game ranches (Mungall and Sheffield 1994). Axis deer (Axis axis) are the most common exotic species in Texas (Demarais et al. 1998). With the largest concentration of axis deer being in those areas of the state with the highest densities of whitetailed deer, there is potential for axis deer to negatively affect white-tailed deer (McGhee and Baccus 2006).
1
E-mail: [email protected]
698
White-tailed deer and axis deer overlap in habitat use and food selection. Axis deer tend to occupy open habitat more often than white-tailed deer but axis deer use habitat with dense cover (Putman 1988, Mungall and Sheffield 1994, Demarais et al. 1998, Geist 1998). Axis deer consume more grass than white-tailed deer yet axis deer also consume dietary items such as browse, mast, and forbs that comprise the bulk of white-tailed deer diets in central Texas (Ables 1977; Henke et al. 1988; Demarais et al. 1998, 2000; Fulbright and Ortega 2006). Overlap in food selection between axis deer and white-tailed deer is expected when considering rumen-reticular morphology (Hofmann 1985, Clauss et al. 2009). White-tailed deer have greater rumen papillation than axis deer, probably reflecting digestive processes that allow axis deer to digest more grasses than white-tailed deer. If predation is reduced or both species are exposed to the same risks from predators then interactions between the species are likely to be mediated by competition. Further, dietary overlap and competition for limited resources has often been attributed as the mechanism for competition between white-tailed deer and exotic species (Demarais et al. 1998). Interspecific competition between exotic and native species can be indirect (exploitative) or direct (interference; Case and Gilpin 1974, Carothers and Jaksic 1984, Amarasekare 2002). When one species is superior in both exploitative and interference, competition exclusion or displacement of the other species is likely to occur (Amarasekare 2002). For axis deer and white-tailed deer to coexist in an area, one of 2 scenarios must exist: 1) axis deer are superior in one form of competition and white-tailed deer are superior in the other form of competition; or 2) the 2 species The Journal of Wildlife Management N 74(4)
use different resources within the same habitat (Carothers and Jaksic 1984, Ziv et al. 1993). The latter possibility seems unlikely for white-tailed deer and axis deer. We tested whether presence of axis deer affects habitat selection of white-tailed deer by comparing habitat selection of white-tailed deer in an area where axis deer were present and then after axis deer were removed (temporal control) and in another area where only white-tailed deer were present (spatial control). We wanted to assess whether axis deer and white-tailed deer could coexist via each species being superior in a different form of competition. Axis deer are 1.2–1.4 times larger than white-tailed deer and should be superior in interference competition (Ables 1977). If axis deer are superior at interference competition then we would expect 3 conditions to be met: 1) after removal of axis deer, there should be a shift in habitat selection of white-tailed deer because in the presence of axis deer white-tailed deer were displaced from their preferred habitat (Morse 1974, Carothers and Jaksic 1984, Amarasekare 2002, Stewart et al. 2002); 2) axis deer should be socially dominant to whitetailed deer and displace them (Carothers and Jaksic 1984, Amarasekare 2002); 3) at stations with a highly desired food source, each species alone (i.e., only axis deer or only whitetailed deer) should occur more frequently than both species together. If both species occur together more frequently than each species alone, it is unlikely that a displacement would have taken place. If each species was present more frequently at the food source alone, it would suggest that those species were partitioning the resource over time and one may be avoiding the other (Case and Gilpin 1974, Carothers and Jaksic 1984, Bartos et al. 2002). Conversely, white-tailed deer may excel at exploitative competition. White-tailed deer are native and have a narrower diet breadth than do axis deer. White-tailed deer may have greater awareness of distribution of desired forages and may consume them before axis deer. If white-tailed deer are superior at exploitative competition 3 conditions should be met: 1) after removal of axis deer, there should be no change in habitat selection of white-tailed deer (Case and Gilpin 1974, Carothers and Jaksic 1984, Bartos et al. 2002); 2) axis deer should be socially dominant to white-tailed deer; 3) at stations with a highly desired food source, white-tailed deer should occur more frequently than axis deer (Case and Gilpin 1974, Ziv et al. 1993).
STUDY AREA We conducted the study on the Flying A Ranch, Bandera County, Texas. The ranch was approximately 3,763 ha of noncontiguous land dispersed over 68 km2. Temperatures ranged from a low of 2u C in January to a high of 35u C in July, with an average annual precipitation of 73.6 cm (National Oceanic and Atmospheric Administration [NOAA] 2007). Vegetation varied due to soil type and depth and past management practices. Primary habitat types were live oak (Quercus virginiana)–ashe juniper (Juniperus ashei) woodlands, post oak (Q. stellata)–elm (Ulmus spp.) woodlands with mixed hardwoods, and open grasslands. Grasslands were dominated by little bluestem (SchizachyrFaas and Weckerly N Habitat Interference
ium scoparium), king ranch bluestem (Bothriochloa ischaemum), threeawns (Aristida spp.), gramas (Bouteloua spp.), and Texas wintergrass (Nassella leucotricha). Soils types included clay, clay loam, and silty clay. Predator control for coyotes (Canis latrans) was continually conducted on the ranch. Along with shooting predators when they were seen during the day, ranch employees used snares set at travel corridors on fences to remove as many predators as possible.
METHODS Design We surveyed 2 areas, a treatment and control, for 6 months for each of 2 seasons. In the treatment area, axis and whitetailed deer were present in survey season 1, but axis deer were removed prior to survey season 2, which allowed us to assess changes in white-tailed deer habitat use due to axis deer. Conducting surveys in 2 study areas over 2 survey periods provided both spatial and temporal controls, which reduced possibilities that environmental differences between survey seasons or possible peculiarities of the treatment area influenced conclusions (White and Garrot 1990). We assessed competition across survey periods but not within survey periods. As such we could not estimate the influence of seasonal changes in forage availability or mating behavior on the 2 forms of competition. Yet, if competition caused shifts in habitat selection that affected reproduction and survival we would have detected it when we pooled data across survey periods. The approximately 130-ha treatment area was enclosed in 2001 by a 2.4-m-high fence that was a complete barrier to deer movement (Fig. 1). Approximately 10 axis deer were trapped in the area when we constructed the fence and we added 15 more shortly thereafter. White-tailed deer found within the area were either part of the initial herd trapped when we erected the fence or offspring of that herd. We introduced no white-tailed deer to the area after enclosure. During the 6 years from fence construction to conclusion of our study, both species were hunted through a commercial hunting operation. The control area was 360 ha of a 2,023-ha pasture located 6 km south of the treatment area (Fig. 1). We enclosed the larger pasture in a 2.4-m-high fence (also a complete barrier to deer movement) erected immediately after the treatment area fence in 2001. The area we selected as the control was not fully enclosed by game-proof fence; however, it did have fencing on the north and east sides. We selected this area because of its similarities to the treatment area in habitat type and topography. White-tailed deer found within the area were either part of the initial herd trapped when we erected the fence or offspring of that herd. Because axis deer were not present in the area, it served as an adequate representation of white-tailed deer in their native habitat. As with the treatment area, white-tailed deer in the control area were hunted through a commercial hunting operation. Spatial Data We collected spatial data over 2 survey seasons: July– December 2006 and 2007. In each survey season we 699
Figure 1. Map of treatment and control areas with survey routes for of our study of habitat interference by axis and white-tailed deer in Bandera County, Texas, USA, 2006–2007. Shaded area is wooded habitat.
conducted vehicle surveys on both the treatment and control areas to obtain spatial locations of white-tailed deer. Prior to data collection, we established survey routes in both the treatment and control areas for measuring habitat use of both species. We selected these routes to maximize the amount of area surveyed while minimizing the risk of double-counting animals (Fig. 1). Once we established survey routes, we designated 150-m-diameter (1.77-ha) circular detection plots: 21 in the treatment area and 33 in the control area. We set centers of these plots approximately 200 m apart along the survey route and identified them with florescent surveyor’s tape. We used detections of animals in these plots to calculate detection probabilities (below). We drove each survey route in the morning and evening, one day per week. Morning surveys began approximately 30 minutes before sunrise and evening surveys began 90 minutes before sunset so we could complete surveys before dark. Each week, we rotated the order and direction in which we drove the routes so that we did not observe a given area more than once at the same time of day. We drove surveys at 17 km/hour (Sanders 1963). When we detected an animal, we recorded the Universal Transverse Mercator (UTM) coordinates of our location as well as a distance and compass bearing to the animal. We obtained UTM coordinates from a Global Positioning System unit and measured distance with a range finder (NikonTM ProStaff Laser 440H; Nikon, Inc., Melville, NY). We used this location information to determine whether the animal was within any of our detection plots. We also recorded time of day, number of animals, species, sex, and age class (juv [ 1 yr old], ad). M
Because it was unrealistic to assume we counted all animals during surveys, we estimated probability of detecting an animal when it was present to obtain unbiased estimates (MacKenzie 2005). We used occupancy models to estimate detection probabilities in Program PRESENCE (MacKenzie 2005). One critical assumption of occupancy modeling was constant residence status, that is, animals were either present or not present in individual detection plots throughout an entire survey season (MacKenzie et al. 2002). Because detection plots were smaller than the daily movements of axis and white-tailed deer, the assumption of constant residence status was violated. To meet this assumption, we pooled data into 1-month intervals to allow time for animal movement in and out of detection plots (Kendall 1999, Longoria and Weckerly 2007). We used pooled data to calculate detection probability. Distance between detection plots was probably insufficient to ensure that the assumption of independence was met for occupancy estimators (MacKenzie 2005). Violation of independence results in underestimates of variances but the estimates themselves are probably unbiased (Draper and Smith 1998). Because we used estimates of detection, we did not correct variances. We evaluated white-tailed deer data using 5 models. Occupancy was constant in each model. We modeled detection probabilities (pˆ) as influenced by diel period (morning or evening), habitat (wooded [ 50% closed canopy], open), season (1 or 2), area (treatment, control), or no covariate. We evaluated axis deer data using 3 models where again, occupancy was constant. We modeled detection probability as influenced by diel period, habitat, season, or no covariate. We selected models based on Akaike’s
700
The Journal of Wildlife Management N 74(4)
L
M
Information Criterion corrected for small sample size (AICc), Akaike weights, and numbers of parameters estimated (Burnham and Anderson 2002). We used detection probabilities to correct data to obtain unbiased estimates of animals present (Lancia et al. 2000). To calculate this correction, we used the formula C Nˆ ~ pˆ where Nˆ was the population estimate or uncorrected count and C was the mean of raw counts in a season (Lancia et al. 2000). If number of animals in the corrected count for any given day exceeded the presumed population size, then we adjusted the number of animals in the raw counts for that day to not exceed the presumed population size. We estimated population size by dividing mean number of animals seen throughout the survey season by detection probability (Lancia et al. 2000). We used these corrected data to determine habitat use. We obtained orthophotos (geo-referenced aerial photographs) from Texas Natural Resources Information System (2007) and uploaded for analysis using ArcGISH 9.1. We classified habitat as wooded or open. Using the orthophotos, we digitized the wooded habitat in each study area. To obtain an accurate percentage of wooded habitat, we performed ground-truthing based on a thorough knowledge of the study area. We clipped the digitized wooded habitat with a 150-m buffer on each side of the survey route to obtain the percent of wooded habitat on that route. We chose 150-m buffer size because it represented the furthest distance that we could detect animals during surveys. We considered the area within the buffer to be the actual portion of the study areas that we surveyed. We then entered corrected data into ArcGIS as individual detections. We calculated number of detections in both wooded and open habitats in each season to determine percentage (and 95% CI) of white-tailed deer using wooded habitat. After calculating the 95% confidence interval for this percentage, we compared percentage of white-tailed deer using wooded habitat to percentage of wooded habitat available to assess whether white-tailed deer were using wooded habitat in proportion to its availability in each survey season. We concluded that white-tailed deer used wooded habitat in proportion to its availability if the confidence interval of percentage used overlapped percentage available (Sokal and Rohlf 1995). We also calculated 95% confidence intervals for the difference between percentage of wooded habitat used in survey seasons 1 and 2, which allowed us to determine whether there were changes in habitat use by white-tailed deer relative to presence of axis deer. We determined there was no change in habitat use by whitetailed deer throughout the study if the confidence interval of the difference overlapped zero (Sokal and Rohlf 1995). We expressed 95% confidence intervals as the estimate 650% of the confidence interval. We conducted this process for both the treatment and control areas to assess differences in habitat use patterns in the 2 areas over the 2 survey seasons. Faas and Weckerly N Habitat Interference
To express change in habitat selection between survey seasons we calculated W2 =A2 W1 =A1 where W is wooded habitat use, A is percent wooded habitat in the area, and 1 and 2 are survey seasons 1 and 2, respectively. Behavioral Data Because axis deer were only present during the first survey season, we only collected behavioral data during that time. We recorded observations twice per day, one day per week from July to December 2006. We recorded observations from 2 deer blinds installed about 0.7 km apart in the enclosed area. Each week we switched blinds. To make observations when deer were most active, morning observations took place for 2 hours starting 0.5 hours before sunrise and evening observations took place for 2 hours starting 1.5 hours before sunset (Halls 1984). Each blind had a corn feeder located within 100 m in grasslands near (,20 m) wooded habitat. Whole shelled corn is a desired food source because it is rapidly fermented by both deer species (Wheaton and Brown 1983, Van Soest 1994). We set feeders to feed each day when we were making observations, and feeders dispersed approximately 2.3 kg of corn in a 14m-diameter area. By baiting deer to the area, there was greater potential to observe interspecific interactions for a highly desired food for an extended period of time (Thomas et al. 1965, Ables 1977, Hirth 1977, McGhee and Baccus 2006). When we observed an animal, we recorded time of day, number of animals, species, sex, and age class. We made observations using 10 3 42 binoculars. When both axis and white-tailed deer were present at a feeder at the same time, we recorded any aggressive behavior resulting in a displacement of one species by another. Displacement was categorized as complete and incomplete (Weckerly 1999). We categorized behavior resulting in one animal leaving the area as complete displacement and behavior resulting in one animal moving away from another but not leaving the feed area as incomplete displacement (Weckerly 1999). We used only those times when both axis deer and white-tailed deer were present at the feeder for the displacement analysis, because we were not concerned with intraspecific interactions. We calculated rate of displacement as the number of displacements per animal per minute. For example, if 5 white-tailed deer were at a feeder with axis deer for 10 minutes and 2 displacements took place, rate of displacement would be 0.04 displacements/animal/minute. Due to nonnormal distribution of the data, we used a randomization test to determine whether axis deer displaced white-tailed deer to a greater extent than white-tailed deer were displaced by axis deer (Manly 1997). Because we predicted that most displacements were by axis deer, we conducted a one-tailed randomization test for paired data (Manly 1997). We conducted these randomization tests for both complete and incomplete displacements and when both displacements were pooled. We conducted 10,000 permuta701
tions. The test statistic (ts) for the randomization test was the sum of the differences between the rate of displacement of white-tailed deer and axis deer for each sampling occasion. We also evaluated length of time axis deer and whitetailed deer were at feeders alone and together. We recorded time when 1 animal from either species was present. We did not include number of animals present. We divided data into morning and evening observations and calculated mean amount of time for each species alone and together. We calculated a likelihood ratio test to determine whether variances differed between species composition (axis deer only, white-tailed deer only, and axis and white-tailed deer together) and across diel period (morning, evening; Sokal and Rohlf 1995). If variances were significantly different, we conducted a linear contrast in a restricted maximum likelihood analysis of variance (ANOVA) that accounted for heteroscedasticity to determine whether the amount of time species spent alone at feeders differed from the amount of time both species were present together (Sokal and Rohlf 1995, Pinheiro and Bates 2000). We used the restricted maximum likelihood ANOVA because least-squares ANOVA is unreliable when data are heteroscedastic (Sokal and Rohlf 1995, Pinheiro and Bates 2000). To determine whether the 2 species spent different amounts of time at feeders we determined deer-minutes by multiplying number of minutes a species was present at the feeder by number of animals of that species. We again used a likelihood ratio test to determine whether variances differed among combinations of species and diel period. If variances differed we conducted a restricted maximum likelihood ANOVA that accounted for heteroscedasticity to determine whether there were differences in number of deer-minutes each species spent at the feeders in morning and evening. Likelihood ratio tests often lack statistical power, so to increase statistical power of the test we divided the P-value by 2 (Hedeker and Gibbons 2006). Removal of axis deer began immediately after conclusion of the first data collection period. The first technique we used for removal was a catch pen, a 2.4-m-high fencing formed into a funnel and lane leading up to the facility for capturing animals. We used a field planted in oats, wheat, and triticale (Triticale hexaploide) to bait deer to the fenced area. We also scattered shelled corn on the ground one week prior to capture to lure deer to the area. After 2 captures in the catch pen (26 Jan and 30 Jan 2007) we closed the bait area. One week prior to capture, we set up and baited a drop-net with corn (Rongstad and McCabe 1984). We dropped the net on a group of axis deer and removed it on 27 February 2007. On 2 March, a helicopter capture team flew the area, captured axis deer with net guns, and removed as many axis deer as possible (Webb et al. 2008). Remaining deer were harvested with rifles either by sitting in hunting blinds over bait or out of a vehicle at night using artificial light. Because axis deer were not regulated by Texas Parks and Wildlife Department, we could use any legal means to remove axis deer (Texas Parks and Wildlife Department 2007). L
702
RESULTS Habitat Use We conducted 50 surveys (morning and evening) in both the treatment and control areas during both survey seasons. We tallied 261 (per survey; ¯x 5 5.22, s 5 6.29) axis deer and 108 ¯ 5 2.16, s 5 3.03) white-tailed deer in the treatment area (x during the first survey season. In the second survey season we ¯ 5 0.88, s 5 1.49) white-tailed deer in the observed 44 (x treatment area. In the control area, we counted 53 (x ¯ 5 1.06, ¯ 5 2.58, s 5 2.86) white-tailed deer in s 5 1.61) and 129 (x the first and second survey seasons, respectively. In the treatment area, 26 axis deer were located within 67% of detection plots. Also in the treatment area, during the first survey season, 39% of detection plots were occupied by 9 and 12 white-tailed deer in the mornings and evenings, respectively. During the second survey season, 60% of detection plots in the treatment area were occupied by 5 and 15 white-tailed deer in the mornings and evenings, respectively. In the control area, we counted 4 and 13 whitetailed deer in mornings and evenings, respectively, in 62% of detection plots during the first survey season. In the second survey season, 48% of detection plots were occupied by 15 and 17 white-tailed deer in mornings and evenings, respectively. Between survey seasons, we removed 50 axis deer from the treatment area, 25 with the catch pen, 3 with the drop-net, and 9 by the helicopter crew. We spent approximately 70 man-hours in blinds and spotlighting at night, resulting in harvest of 10 axis deer. During the first 2 weeks of surveys, we observed 4 axis deer females on the survey route. After this initial observation 3 axis deer females were harvested within the first 2 months of surveys in the second survey season. We are confident that only one axis deer remained in the treatment area. Using the model with no covariate we estimated detection probability for axis deer to be 0.123 (Table 1). Based on activities to remove axis deer, minimum population size of axis deer was 51, close to the population size estimated from the ratio of the mean of survey counts and detection probability (5.17/0.123 5 42). The agreement in the 2 estimates of population size indicates that detection probabilities were not strongly biased. The model with no covariate was also selected for whitetailed deer (Table 1). From this model, we obtained a detection probability estimate of 0.069. From the detection probability and mean of survey counts, estimated population sizes were 21 white-tailed deer in the treatment area and 26 in the control area. The treatment area was 53.85% wooded habitat. Whitetailed deer used wooded habitat 34.6 6 4.2% during the first survey season and 79.7 6 4.5% during the second survey season, less and more than its availability, respectively (Fig. 2). The difference between percentage habitat use in the 2 survey seasons was 244.8 6 4.5%. Axis deer used wooded habitat 30.4 6 2.6% of the time in the first survey season, which was less than its availability. The control area consisted of 34.31% wooded habitat. White-tailed deer used wooded habitat 23.1 6 3.6% and The Journal of Wildlife Management N 74(4)
Table 1. Model selection analyses to examine whether detection probabilities of white-tailed deer and axis deer in Bandera County, Texas, USA, 2006– 2007, were influenced by diel period (diel), open or wooded habitat (hab), treatment or control area (site), survey season (yr), or not influenced by any of these variables (.). y 5 occupancy and p 5 detection probability. Model selection was based on Akaike Information Criterion coefficients (AICc), Akaike weights (wi), number of parameters (nPar), and 22 times the log likelihood (22LogLike). Modela White-tailed deer y(.),p(.) y(.),p(diel) y(.),p(hab) y(.),p(site) y(.),p(yr) Axis deer y(.),p(.) y(.),p(diel) y(.),p(hab) a
AICc
wi
nPar
22LogLike
631.30 631.69 632.80 633.13 635.15
0.33 0.30 0.17 0.15 0.05
2 3 3 3 3
627.06 625.21 626.32 626.65 628.67
171.15 173.72 173.74
0.51 0.20 0.20
2 3 3
166.48 166.31 166.33
For white-tailed and axis deer the y(.),p(.) model was selected.
25.1 6 3.1%, less than its availability, during survey seasons 1 and 2, respectively (Fig. 2). There was no difference in habitat use between the 2 survey seasons (22 6 4.8%). Because detection probabilities did not differ between diel periods, habitats, treatments, or survey seasons, estimates of wooded habitat use were similar for corrected or uncorrected counts. Hence, confidence intervals of the difference of percentages resulted in more conservative inferences, and a more robust analysis, when we used uncorrected counts, because sample sizes were much smaller. When we
calculated 95% confidence intervals of the difference of percentages from uncorrected counts our conclusions, however, were the same. Between survey seasons 1 and 2 there was an increase in use of wooded habitat by whitetailed deer (244 6 8%) in the treatment area but no change in use of wooded habitat in the control area (24 6 7%). In the treatment area selection by white-tailed deer for wooded habitat increased 2.2 (corrected counts) and 2.1 (uncorrected counts) times between survey seasons 1 and 2. In the control area selection by white-tailed deer for wooded habitat changed by (at most) 1.1 (corrected counts) and 1.2 (uncorrected counts) times between survey seasons 1 and 2. Displacements and Time Spent at Feeders We spent 58 hours observing feeders during 29 occasions. On 19 occasions we observed white-tailed deer or axis deer, but not both species (65.5%). Only during 8 occasions did we observe both white-tailed deer and axis deer at a feeder at the same time. During these sampling occasions, axis deer completely displaced white-tailed deer 17 times resulting in a mean rate of displacement of 0.140 (s 5 0.123). We did not observe complete displacement of axis deer by whitetailed deer. We observed incomplete displacements of white-tailed deer by axis deer 22 times (mean rate of displacement 5 0.086, s 5 0.046) and incomplete displacements of axis deer by white-tailed deer 2 times (mean rate of displacement 5 0.002, s 5 0.002). When we pooled both types of displacements, mean rate of axis deer displacing white-tailed deer was 0.226 (s 5 0.119) and mean rate of white-tailed deer displacing axis deer was 0.002 (s 5 0.002). Axis deer displaced white-tailed deer to a greater degree than white-tailed deer displaced axis deer for all displacement types (complete: ts 5 1.123, P , 0.001; incomplete: ts 5 0.666, P 5 0.005; combined: ts 5 1.789, P , 0.001). During all occasions when 1 animal of either species was present, we observed white-tailed deer at feeders alone ¯ 5 15.79, s 5 3.74), axis deer alone 300 minutes (x ¯ 5 11.21, s 5 3.85), and both species 213 minutes (x ¯ 5 5.26, s 5 1.85). Variance of each together 100 minutes (x species composition in each diel period differed (x25 5 29.89, P , 0.001) so we conducted the maximum likelihood ANOVA. There was no interaction between diel period and L
Figure 2. Proportion of wooded habitat used in survey seasons 1 and 2 by white-tailed deer compared to amount of wooded habitat available in treatment and control areas in Bandera County, Texas, USA, 2006–2007. Error bars represent 95% confidence intervals. Faas and Weckerly N Habitat Interference
703
species composition (t51 5 20.209, P 5 0.835). Amount of time axis and white-tailed deer each spent at feeders alone was longer than when both axis and white-tailed deer were present together (t51 5 21.89, P 5 0.064). When 1 animal of either species was present at a feeder, we recorded 1,542 deer-minutes for white-tailed deer (x ¯ 5 81.16, s 5 22.71) and 1,194 deer-minutes for axis deer (x ¯ 5 62.84, s 5 22.24). There were significant differences between variances in each diel period and species (x25 5 6.178, P 5 0.051) so we used the maximum likelihood ANOVA again. No interaction was present between diel period and species (t34 5 21.47, P 5 0.151), and number of deer-minutes whitetailed deer spent at feeders was not different from number of deer-minutes axis deer spent at feeders (t34 5 1.58, P 5 0.124). L
DISCUSSION Our study suggested that axis deer were superior at interference competition. First, we found that white-tailed deer shifted habitat use dramatically (36–52%) in the absence of axis deer, whereas white-tailed deer habitat use in the control area was constant over the same time span. Second, data from behavioral interactions clearly indicated that axis deer were socially dominant to white-tailed deer. We observed both male and female axis deer displacing male and female white-tailed deer. When we pooled both complete and incomplete displacements, 1 displacement of white-tailed deer by axis deer took place during every observation. In the 2 recorded observations of axis deer displacement by white-tailed deer, 2 male white-tailed deer made an aggressive advance toward 2 female axis deer resulting in those deer moving to a different location to feed. Under no circumstances was the rate of displacement of axis deer by white-tailed deer greater than that of the opposite. Third, at stations with a highly desired food source, each species occurred alone more frequently than both species together, which supports our hypothesis of interference competition of axis deer on white-tailed deer because the subordinate species should avoid areas used by the dominant species, or the subordinate species should change times it uses the resource to avoid interactions with the dominant species. Case and Gilpin (1974) and Carothers and Jaksic (1984) suggested that temporal partitioning between species may be one of the most recognizable symptoms of interspecific competition. Our study also suggests that the 2 species will not coexist because white-tailed deer were not superior at exploitative competition. Of the 3 conditions that should be met for white-tailed deer to be superior at exploitative competition, the 2 critical conditions were not met. There should have been no change in habitat use by white-tailed deer in the presence or absence of axis deer and white-tailed deer should have occurred more frequently at stations with a highly desired food source. If white-tailed deer had spent a greater amount of time at feeders than did axis deer, we would have expected white-tailed deer to be able to consume corn before or in a more efficient way than axis deer. Extending this finding to habitat use, white-tailed deer habitat use should L
704
be similar in the presence and absence of axis deer because foraging time in habitat is unaltered. To further refute the hypothesis that white-tailed deer are superior at exploitative competition, axis deer have an advantage over white-tailed deer in that axis deer are capable of using grasses, forbs, and browse whereas white-tailed deer mostly exploit only forbs and browse (Ables 1977, Putman 1988, Traweek and Welch 1992, Mungall and Sheffield 1994, Geist 1998). Food habits of axis deer vary according to the location in which the study is conducted. Whereas some have suggested grass is the primary food choice of axis deer, others have suggested that as much as half of axis deer diet may consist of forbs and browse (Ables 1977, Putman 1988, Mungall and Sheffield 1994, Geist 1998). The important conclusion is that axis deer are capable of using the same resources as white-tailed deer. Because white-tailed deer were not superior at exploitative competition, coexistence is unlikely to occur between whitetailed deer and axis deer (Amarasekare 2002, Stewart et al. 2002). Axis deer can interfere with resource selection of white-tailed deer and potentially exploit resources before white-tailed deer. Although we suggest that coexistence is unlikely, exploitative competition may not be as influential if densities of axis deer and white-tailed deer are kept low and habitat quality is high. Use of common resources manifests most when densities of animals are high and resources are limited (Stewart et al. 2002). When densities are low, there is greater potential for habitat partitioning, less interspecific competition, and fewer aggressive interactions, which should promote coexistence or niche partitioning to occur. When we designed this study, we thought densities of white-tailed deer and axis deer were approximately equal. However, we found that there were approximately twice as many axis deer as white-tailed deer in the treatment area, which possibly allowed axis deer to both interfere with and exploit resources before white-tailed deer (Amarasekare 2002, Stewart et al. 2002). If densities of white-tailed deer and axis deer were similar or if white-tailed deer were twice as abundant as axis deer in the treatment area, we suspect that axis deer would still be superior at interference competition but that white-tailed deer would be able to compete more strongly for available resources. Niche partitioning may take place to a greater degree, but over time, owing to interference from axis deer, white-tailed deer population sizes would decrease and axis deer would become more abundant. Our study is one of the first using both spatial and temporal controls to show habitat displacement of a native large mammal by an exotic large mammal (Berger 1985, Bartos et al. 2002, Stewart et al. 2002). Previous studies had only one type of control (usually spatial). Having both spatial and temporal controls was important because it allowed us to assess whether factors such as changes in the weather or differences in habitats in treatment and control areas affected results. For instance the first survey season was a below-average year for precipitation with only 61 cm recorded; however, the second survey season received .2.5 The Journal of Wildlife Management N 74(4)
times that amount (NOAA 2007). If the dramatic difference in habitat use by white-tailed deer in the treatment area could have been attributed to this difference in precipitation, a change in habitat use by white-tailed deer in the control area should have occurred, which did not happen. Although it was apparent that white-tailed deer shifted their habitat use because of axis deer, the underlying cause is unknown. The most likely conclusions were that axis deer displaced white-tailed deer from preferred habitat either by direct aggressive behavior resulting in interference competition or by increasing total deer density in wooded habitats to the point that white-tailed deer sought food in less preferred habitats. Our results suggest that interference competition took place; therefore, displacement by aggressive behaviors is most likely. Furthermore, given that axis deer are capable of consuming forbs and browse, their removal would decrease foraging pressure in wooded areas allowing white-tailed deer to use this habitat to a greater degree, assuming that white-tailed deer prefer and select for wooded habitat. However, this was not apparent in the control area where white-tailed deer used wooded habitat less than its availability. The difference in habitat selection of white-tailed deer between treatment and control areas may have been due to 2 possibilities. First, there could have been subtle differences in characteristics of habitat in the 2 areas (Fig. 1). For example, in the control area open habitat was characterized by more trees and shrubs scattered in the grassland and more broken terrain. More screening cover could have resulted in the deer spending greater amounts of foraging time away from the escape cover of wooded habitat. Second, densities of white-tailed deer probably differed by at least a factor of 2 between treatment (21 deer/130 ha 5 0.16) and control (25/ 360 5 0.07) areas. Most likely, habitat selection by deer is density dependent (Morris 1987). The dissimilar densities probably influenced habitat selection differently in the treatment and control areas.
MANAGEMENT IMPLICATIONS Due to their superiority at interference competition it is feasible to expect that axis deer densities will increase in central Texas. If axis deer densities actually increase throughout central Texas white-tailed deer will probably be displaced from higher quality to lesser quality habitats and abundance of white-tailed deer will decrease in this region because of their poor competitive abilities with axis deer. Whether axis deer densities will increase, however, is far from clear. How susceptible axis deer are to contemporary sources of predation in Texas has not been studied. Nor is it clear how spatial variation in habitat composition at different spatial scales and seasonal and annual variation in climatic conditions influences competitive interactions between the 2 species. The primary concern in the state of Texas is those areas with free-ranging axis deer. Currently, it is likely that there are fewer free-ranging axis deer than white-tailed deer throughout central Texas so negative effects are likely minimal. However, there are no population estimates or counts of axis deer throughout central Texas. Faas and Weckerly N Habitat Interference
Periodic monitoring of axis deer and white-tailed deer abundance in central Texas is warranted so that wildlife managers can determine whether white-tailed deer abundance is negatively associated with axis deer abundance.
ACKNOWLEDGMENTS We are grateful to C. Williams and the Flying A Ranch for access to the ranch and providing C. J. Faas with lodging and work during the field study. We also thank K. Lake, B. Trussell, and T. Reagan. Additional funding was provided from Dallas Safari Club, and C. J. Faas received 2 scholarships from the Houston Safari Club. R. Simpson, C. Green, and 2 anonymous reviewers commented on earlier drafts of the manuscript.
LITERATURE CITED Ables, E. 1977. The axis deer in Texas. Texas A&M Press, College Station, USA. Amarasekare, P. 2002. Interference competition and species coexistence. Proceedings of the Royal Society B: Biological Sciences 269:2541–2550. Bartos, L., D. Vankova, K. V. Miller, and J. Siler. 2002. Interspecific competition between white-tailed, fallow, red, and roe deer. Journal of Wildlife Management 66:522–527. Berger, J. 1985. Interspecific interactions and dominance among wild great basin ungulates. Journal of Mammalogy 66:571–573. Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Second edition. Springer Verlag, New York, New York, USA. Butts, G. L. 1979. The status of exotic big game in Texas. Rangelands 1:152–153. Carothers, J. H., and F. M. Jaksic. 1984. Time as a niche difference: the role of interference competition. Oikos 42:403–406. Case, T. J., and M. E. Gilpin. 1974. Interference competition and niche theory. Proceedings National Academy Sciences, USA 71:3073–3077. Clauss, M., R. R. Hofmann, J. Fickel, W. J. Streich, and J. Hummel. 2009. The intraruminal papillation gradient in wild ruminants of different feeding types: implications for rumen physiology. Journal of Morphology 270:929–942. Demarais, S., J. T. Baccus, and M. S. Traweek, Jr. 1998. Non-indigenous ungulates in Texas: long-term population trends and possible competitive mechanisms. Transactions North American Wildlife Natural Resource Conference 63:49–55. Demarais, S., K. V. Miller, and H. A. Jacobson. 2000. White-tailed deer. Pages 601–628 in S. Demarais and P. R. Krausman, editors. Ecology and management of large mammals in North America. Prentice Hall, Upper Saddle River, New Jersey, USA. Draper, N. R., and H. Smith. 1998. Applied regression analysis. Third edition. Wiley Interscience, New York, New York, USA. Fulbright, T. E., and J. A. Ortega-S, 2006. White-tailed deer habitat: ecology and management on rangelands. Texas A&M University Press, College Station, USA. Gabor, T. M., and E. C. Hellgren. 2000. Variation in peccary populations: landscape composition or competition by an invader? Ecology 81:2509– 2524. Geist, V. 1998. Deer of the world. Stackpole, Mechanicsburg, Pennsylvania, USA. Halls, L. K. 1984. White-tailed deer: ecology and management. Stackpole, Harrisburg, Pennsylvania, USA. Hedeker, D. R., and R. D. Gibbons. 2006. Longitudinal data analysis. Wiley Interscience, Hoboken, New Jersey, USA. Henke, S. E., S. Demarais, and J. A. Pfister. 1988. Digestive capacity and diets of white-tailed deer and exotic ruminants. Journal of Wildlife Management 52:595–598. Hirth, D. H. 1977. Social behavior of white-tailed deer in relation to habitat. Wildlife Monographs 53. Hofmann, R. R. 1985. Digestive physiology of deer—their morphological specialization and adaptation. Royal Society of New Zealand, Bulletin 22:393–407. 705
Kendall, W. L. 1999. Robustness of closed capture–recapture methods to violations of the closure assumption. Ecology 80:2517–2525. Lancia, R. A., C. S. Rosenberry, and M. C. Conner. 2000. Population parameters and their estimation. Pages 64–83 in S. Demarais and P. R. Krausman, editors. Ecology and management of large mammals in North America. Prentice Hall, Upper Saddle River, New Jersey, USA. Longoria, M. P., and F. W. Weckerly. 2007. Estimating detection probabilities of collared peccary from sign. Journal of Wildlife Management 71:652–655. MacKenzie, D. I. 2005. What are the issues with presence–absence data for wildlife managers? Journal of Wildlife Management 69:849–860. MacKenzie, D. I., J. D. Nichols, G. B. Lachman, S. Droege, J. A. Royle, and C. A. Langtimm. 2002. Estimating site occupancy rates when detection probabilities are less than one. Ecology 83:2248–2255. Manly, B. F. J. 1997. Randomization, bootstrap and Monte Carlo methods in biology. Chapman and Hall, London, United Kingdom. McGhee, J. D., and J. T. Baccus. 2006. Behavioral interactions between axis and fallow deer at high-value food patches. Southwestern Naturalist 51:358–367. Morris, D. W. 1987. Tests of density-dependent habitat selection in a patchy environment. Ecological Monographs 57:270–281. Morse, D. H. 1974. Niche breadth as a function of social dominance. American Naturalist 108:818–830. Mungall, E. C., and W. J. Sheffield. 1994. Exotics on the range: the Texas example. Texas A&M University, College Station, USA. National Oceanic and Atmospheric Administration [NOAA]. 2007. Climatological data annual summary, Texas. U.S. Department of Commerce, Washington, D.C., USA. Pinheiro, J. C., and D. M. Bates. 2000. Mixed-effects models in S and SPlus. Springer Verlag, New York, New York, USA. Putman, R. 1988. The natural history of deer. Cornell University Press, Ithaca, New York, USA. Rongstad, O. J., and R. A. McCabe. 1984. Capture techniques. Pages 655– 676 in L. K. Halls, editor. White-tailed deer ecology and management. Stackpole, Harrisburg, Pennsylvania, USA.
706
Sanders, C. L. 1963. Habitat preferences of white-tailed deer and several exotic ungulates in South Texas. Ecology 44:803–806. Sokal, R. R., and F. J. Rohlf. 1995. Biometry: the principles and practice of statistics in biological research. Third edition. W. H. Freeman, New York, New York, USA. Stewart, K. M., R. T. Bowyer, J. G. Kie, N. J. Cimon, and B. K. Johnson. 2002. Temporospatial distributions of elk, mule deer, and cattle: resource partitioning and competitive displacement. Journal of Mammalogy 83:229–244. Texas Natural Resources Information System. 2007. TNRIS data distribution system, Texas Natural Resources Information System, Austin, USA. ,http//www.tnris.state.tx.us.. Accessed 15 Feb 2007. Texas Parks and Wildlife Department. 2007. Hunting regulations. Texas Parks and Wildlife Department, Austin, USA. Thomas, J. W., R. M. Robinson, and R. G. Marburger. 1965. Social behavior in a white-tailed deer herd containing hypogonadal males. Journal of Mammalogy 46:314–327. Traweek, Jr., M. S., and R. Welch. 1992. Exotics in Texas. Texas Parks and Wildlife Department, Austin, USA. Van Soest, P. J. 1994. Nutritional ecology of the ruminant. Second edition. Comstock, Ithaca, New York, USA. Webb, S. L., J. S. Lewis, D. G. Hewitt, M. W. Hellickson, and F. C. Bryant. 2008. Assessing the helicopter and net gun as a capture technique for white-tailed deer. Journal of Wildlife Management 72:310–314. Weckerly, F. W. 1999. Social bonding and aggression in female Roosevelt elk. Canadian Journal of Zoology 77:1379–1384. Wheaton, C., and R. D. Brown. 1983. Feed intake and digestive efficiency of south Texas white-tailed deer. Journal of Wildlife Management 47:442–450. White, G. C., and R. A. Garrot. 1990. Analysis of wildlife radio-tracking data. Academic Press, San Diego, California, USA. Ziv, Y., Z. Abramsky, B. P. Kotler, and A. Subach. 1993. Interference competition and temporal and habitat partitioning in two gerbil species. Oikos 66:237–246. Associate Editor: Hudson.
The Journal of Wildlife Management N 74(4)
Management and Conservation Article
Habitat Interference by Axis Deer on White-Tailed Deer CLINTON J. FAAS, Department of Biology, Texas State University, San Marcos, TX 78666, USA FLOYD W. WECKERLY,1 Department of Biology, Texas State University, San Marcos, TX 78666, USA
ABSTRACT Many studies of interactions between exotic and native ungulates have not had temporal and spatial controls nor have they considered the types of competitive interactions that would allow coexistence. For exotic axis deer (Axis axis) and native white-tailed deer (Odocoileus virginianus) to coexist one species should be superior at interference competition and the other species should be superior at exploitative competition. We generated and tested predictions, based on body size and diet breadth, about habitat selection by white-tailed deer in the presence and absence of axis deer, dominance relationships, and time at sites provisioned with high quality forage. We conducted our study in treatment (axis and white-tailed deer) and control (white-tailed deer only) areas when both species were present and after axis deer were removed. We conducted vehicle surveys to determine habitat use of both species. At provisioned feeding sites we recorded aggressive behaviors and amount of time species spent at feeding sites alone and together. In the treatment area white-tailed deer selection for wooded habitat increased 2.1 times after axis deer were removed, whereas habitat selection by white-tailed deer was constant in the control area over the same time. At feeding sites axis deer were dominant to white-tailed deer; both species spent a significantly greater amount of time alone than at feeders together, but amount of time that individuals of each species spent at feeders did not differ. Axis deer were superior at interference competition, but white-tailed deer were not superior at exploitative competition; thus, species coexistence is unlikely. Whether white-tailed deer are negatively impacted by axis deer at spatial scales larger than our experiment probably depends on abundance of axis deer at larger spatial scales. Experiments of species interactions with temporal and spatial controls that consider types of competitive interactions increase a manager’s understanding of when and how native ungulates may be negatively impacted by exotic ungulates.
KEY WORDS Axis axis, coexistence, exotics, habitat, interspecific competition, Odocoileus virginianus, temporal and spatial controls, Texas.
Understanding how distribution of ungulates is affected by habitat selection is necessary for examining resource partitioning and competition (Stewart et al. 2002). However, competitive interaction studies among cervids, especially between exotics and native species, have rarely been conducted (Bartos et al. 2002). Stewart et al. (2002) reported that studies of competition between large herbivores are complicated because addition or removal experiments are not feasible in most cases, largely due to wideranging distributions of large herbivores and difficulty in conducting manipulative experiments that have spatial and temporal controls (White and Garrot 1990, Gabor and Hellgren 2000). Since the time of their original introduction into highfenced enclosures, many exotic species have escaped or been released and free-ranging populations continue to increase (Butts 1979, Mungall and Sheffield 1994, Demarais et al. 1998). With increasing numbers of exotic species there is an increased need to determine negative effects on native white-tailed deer (Odocoileus virginianus). Because exotic species in Texas, USA, are regulated as livestock and not game animals, there has been little done to control numbers of animals outside of managed herds within game ranches (Mungall and Sheffield 1994). Axis deer (Axis axis) are the most common exotic species in Texas (Demarais et al. 1998). With the largest concentration of axis deer being in those areas of the state with the highest densities of whitetailed deer, there is potential for axis deer to negatively affect white-tailed deer (McGhee and Baccus 2006).
1
E-mail: [email protected]
698
White-tailed deer and axis deer overlap in habitat use and food selection. Axis deer tend to occupy open habitat more often than white-tailed deer but axis deer use habitat with dense cover (Putman 1988, Mungall and Sheffield 1994, Demarais et al. 1998, Geist 1998). Axis deer consume more grass than white-tailed deer yet axis deer also consume dietary items such as browse, mast, and forbs that comprise the bulk of white-tailed deer diets in central Texas (Ables 1977; Henke et al. 1988; Demarais et al. 1998, 2000; Fulbright and Ortega 2006). Overlap in food selection between axis deer and white-tailed deer is expected when considering rumen-reticular morphology (Hofmann 1985, Clauss et al. 2009). White-tailed deer have greater rumen papillation than axis deer, probably reflecting digestive processes that allow axis deer to digest more grasses than white-tailed deer. If predation is reduced or both species are exposed to the same risks from predators then interactions between the species are likely to be mediated by competition. Further, dietary overlap and competition for limited resources has often been attributed as the mechanism for competition between white-tailed deer and exotic species (Demarais et al. 1998). Interspecific competition between exotic and native species can be indirect (exploitative) or direct (interference; Case and Gilpin 1974, Carothers and Jaksic 1984, Amarasekare 2002). When one species is superior in both exploitative and interference, competition exclusion or displacement of the other species is likely to occur (Amarasekare 2002). For axis deer and white-tailed deer to coexist in an area, one of 2 scenarios must exist: 1) axis deer are superior in one form of competition and white-tailed deer are superior in the other form of competition; or 2) the 2 species The Journal of Wildlife Management N 74(4)
use different resources within the same habitat (Carothers and Jaksic 1984, Ziv et al. 1993). The latter possibility seems unlikely for white-tailed deer and axis deer. We tested whether presence of axis deer affects habitat selection of white-tailed deer by comparing habitat selection of white-tailed deer in an area where axis deer were present and then after axis deer were removed (temporal control) and in another area where only white-tailed deer were present (spatial control). We wanted to assess whether axis deer and white-tailed deer could coexist via each species being superior in a different form of competition. Axis deer are 1.2–1.4 times larger than white-tailed deer and should be superior in interference competition (Ables 1977). If axis deer are superior at interference competition then we would expect 3 conditions to be met: 1) after removal of axis deer, there should be a shift in habitat selection of white-tailed deer because in the presence of axis deer white-tailed deer were displaced from their preferred habitat (Morse 1974, Carothers and Jaksic 1984, Amarasekare 2002, Stewart et al. 2002); 2) axis deer should be socially dominant to whitetailed deer and displace them (Carothers and Jaksic 1984, Amarasekare 2002); 3) at stations with a highly desired food source, each species alone (i.e., only axis deer or only whitetailed deer) should occur more frequently than both species together. If both species occur together more frequently than each species alone, it is unlikely that a displacement would have taken place. If each species was present more frequently at the food source alone, it would suggest that those species were partitioning the resource over time and one may be avoiding the other (Case and Gilpin 1974, Carothers and Jaksic 1984, Bartos et al. 2002). Conversely, white-tailed deer may excel at exploitative competition. White-tailed deer are native and have a narrower diet breadth than do axis deer. White-tailed deer may have greater awareness of distribution of desired forages and may consume them before axis deer. If white-tailed deer are superior at exploitative competition 3 conditions should be met: 1) after removal of axis deer, there should be no change in habitat selection of white-tailed deer (Case and Gilpin 1974, Carothers and Jaksic 1984, Bartos et al. 2002); 2) axis deer should be socially dominant to white-tailed deer; 3) at stations with a highly desired food source, white-tailed deer should occur more frequently than axis deer (Case and Gilpin 1974, Ziv et al. 1993).
STUDY AREA We conducted the study on the Flying A Ranch, Bandera County, Texas. The ranch was approximately 3,763 ha of noncontiguous land dispersed over 68 km2. Temperatures ranged from a low of 2u C in January to a high of 35u C in July, with an average annual precipitation of 73.6 cm (National Oceanic and Atmospheric Administration [NOAA] 2007). Vegetation varied due to soil type and depth and past management practices. Primary habitat types were live oak (Quercus virginiana)–ashe juniper (Juniperus ashei) woodlands, post oak (Q. stellata)–elm (Ulmus spp.) woodlands with mixed hardwoods, and open grasslands. Grasslands were dominated by little bluestem (SchizachyrFaas and Weckerly N Habitat Interference
ium scoparium), king ranch bluestem (Bothriochloa ischaemum), threeawns (Aristida spp.), gramas (Bouteloua spp.), and Texas wintergrass (Nassella leucotricha). Soils types included clay, clay loam, and silty clay. Predator control for coyotes (Canis latrans) was continually conducted on the ranch. Along with shooting predators when they were seen during the day, ranch employees used snares set at travel corridors on fences to remove as many predators as possible.
METHODS Design We surveyed 2 areas, a treatment and control, for 6 months for each of 2 seasons. In the treatment area, axis and whitetailed deer were present in survey season 1, but axis deer were removed prior to survey season 2, which allowed us to assess changes in white-tailed deer habitat use due to axis deer. Conducting surveys in 2 study areas over 2 survey periods provided both spatial and temporal controls, which reduced possibilities that environmental differences between survey seasons or possible peculiarities of the treatment area influenced conclusions (White and Garrot 1990). We assessed competition across survey periods but not within survey periods. As such we could not estimate the influence of seasonal changes in forage availability or mating behavior on the 2 forms of competition. Yet, if competition caused shifts in habitat selection that affected reproduction and survival we would have detected it when we pooled data across survey periods. The approximately 130-ha treatment area was enclosed in 2001 by a 2.4-m-high fence that was a complete barrier to deer movement (Fig. 1). Approximately 10 axis deer were trapped in the area when we constructed the fence and we added 15 more shortly thereafter. White-tailed deer found within the area were either part of the initial herd trapped when we erected the fence or offspring of that herd. We introduced no white-tailed deer to the area after enclosure. During the 6 years from fence construction to conclusion of our study, both species were hunted through a commercial hunting operation. The control area was 360 ha of a 2,023-ha pasture located 6 km south of the treatment area (Fig. 1). We enclosed the larger pasture in a 2.4-m-high fence (also a complete barrier to deer movement) erected immediately after the treatment area fence in 2001. The area we selected as the control was not fully enclosed by game-proof fence; however, it did have fencing on the north and east sides. We selected this area because of its similarities to the treatment area in habitat type and topography. White-tailed deer found within the area were either part of the initial herd trapped when we erected the fence or offspring of that herd. Because axis deer were not present in the area, it served as an adequate representation of white-tailed deer in their native habitat. As with the treatment area, white-tailed deer in the control area were hunted through a commercial hunting operation. Spatial Data We collected spatial data over 2 survey seasons: July– December 2006 and 2007. In each survey season we 699
Figure 1. Map of treatment and control areas with survey routes for of our study of habitat interference by axis and white-tailed deer in Bandera County, Texas, USA, 2006–2007. Shaded area is wooded habitat.
conducted vehicle surveys on both the treatment and control areas to obtain spatial locations of white-tailed deer. Prior to data collection, we established survey routes in both the treatment and control areas for measuring habitat use of both species. We selected these routes to maximize the amount of area surveyed while minimizing the risk of double-counting animals (Fig. 1). Once we established survey routes, we designated 150-m-diameter (1.77-ha) circular detection plots: 21 in the treatment area and 33 in the control area. We set centers of these plots approximately 200 m apart along the survey route and identified them with florescent surveyor’s tape. We used detections of animals in these plots to calculate detection probabilities (below). We drove each survey route in the morning and evening, one day per week. Morning surveys began approximately 30 minutes before sunrise and evening surveys began 90 minutes before sunset so we could complete surveys before dark. Each week, we rotated the order and direction in which we drove the routes so that we did not observe a given area more than once at the same time of day. We drove surveys at 17 km/hour (Sanders 1963). When we detected an animal, we recorded the Universal Transverse Mercator (UTM) coordinates of our location as well as a distance and compass bearing to the animal. We obtained UTM coordinates from a Global Positioning System unit and measured distance with a range finder (NikonTM ProStaff Laser 440H; Nikon, Inc., Melville, NY). We used this location information to determine whether the animal was within any of our detection plots. We also recorded time of day, number of animals, species, sex, and age class (juv [ 1 yr old], ad). M
Because it was unrealistic to assume we counted all animals during surveys, we estimated probability of detecting an animal when it was present to obtain unbiased estimates (MacKenzie 2005). We used occupancy models to estimate detection probabilities in Program PRESENCE (MacKenzie 2005). One critical assumption of occupancy modeling was constant residence status, that is, animals were either present or not present in individual detection plots throughout an entire survey season (MacKenzie et al. 2002). Because detection plots were smaller than the daily movements of axis and white-tailed deer, the assumption of constant residence status was violated. To meet this assumption, we pooled data into 1-month intervals to allow time for animal movement in and out of detection plots (Kendall 1999, Longoria and Weckerly 2007). We used pooled data to calculate detection probability. Distance between detection plots was probably insufficient to ensure that the assumption of independence was met for occupancy estimators (MacKenzie 2005). Violation of independence results in underestimates of variances but the estimates themselves are probably unbiased (Draper and Smith 1998). Because we used estimates of detection, we did not correct variances. We evaluated white-tailed deer data using 5 models. Occupancy was constant in each model. We modeled detection probabilities (pˆ) as influenced by diel period (morning or evening), habitat (wooded [ 50% closed canopy], open), season (1 or 2), area (treatment, control), or no covariate. We evaluated axis deer data using 3 models where again, occupancy was constant. We modeled detection probability as influenced by diel period, habitat, season, or no covariate. We selected models based on Akaike’s
700
The Journal of Wildlife Management N 74(4)
L
M
Information Criterion corrected for small sample size (AICc), Akaike weights, and numbers of parameters estimated (Burnham and Anderson 2002). We used detection probabilities to correct data to obtain unbiased estimates of animals present (Lancia et al. 2000). To calculate this correction, we used the formula C Nˆ ~ pˆ where Nˆ was the population estimate or uncorrected count and C was the mean of raw counts in a season (Lancia et al. 2000). If number of animals in the corrected count for any given day exceeded the presumed population size, then we adjusted the number of animals in the raw counts for that day to not exceed the presumed population size. We estimated population size by dividing mean number of animals seen throughout the survey season by detection probability (Lancia et al. 2000). We used these corrected data to determine habitat use. We obtained orthophotos (geo-referenced aerial photographs) from Texas Natural Resources Information System (2007) and uploaded for analysis using ArcGISH 9.1. We classified habitat as wooded or open. Using the orthophotos, we digitized the wooded habitat in each study area. To obtain an accurate percentage of wooded habitat, we performed ground-truthing based on a thorough knowledge of the study area. We clipped the digitized wooded habitat with a 150-m buffer on each side of the survey route to obtain the percent of wooded habitat on that route. We chose 150-m buffer size because it represented the furthest distance that we could detect animals during surveys. We considered the area within the buffer to be the actual portion of the study areas that we surveyed. We then entered corrected data into ArcGIS as individual detections. We calculated number of detections in both wooded and open habitats in each season to determine percentage (and 95% CI) of white-tailed deer using wooded habitat. After calculating the 95% confidence interval for this percentage, we compared percentage of white-tailed deer using wooded habitat to percentage of wooded habitat available to assess whether white-tailed deer were using wooded habitat in proportion to its availability in each survey season. We concluded that white-tailed deer used wooded habitat in proportion to its availability if the confidence interval of percentage used overlapped percentage available (Sokal and Rohlf 1995). We also calculated 95% confidence intervals for the difference between percentage of wooded habitat used in survey seasons 1 and 2, which allowed us to determine whether there were changes in habitat use by white-tailed deer relative to presence of axis deer. We determined there was no change in habitat use by whitetailed deer throughout the study if the confidence interval of the difference overlapped zero (Sokal and Rohlf 1995). We expressed 95% confidence intervals as the estimate 650% of the confidence interval. We conducted this process for both the treatment and control areas to assess differences in habitat use patterns in the 2 areas over the 2 survey seasons. Faas and Weckerly N Habitat Interference
To express change in habitat selection between survey seasons we calculated W2 =A2 W1 =A1 where W is wooded habitat use, A is percent wooded habitat in the area, and 1 and 2 are survey seasons 1 and 2, respectively. Behavioral Data Because axis deer were only present during the first survey season, we only collected behavioral data during that time. We recorded observations twice per day, one day per week from July to December 2006. We recorded observations from 2 deer blinds installed about 0.7 km apart in the enclosed area. Each week we switched blinds. To make observations when deer were most active, morning observations took place for 2 hours starting 0.5 hours before sunrise and evening observations took place for 2 hours starting 1.5 hours before sunset (Halls 1984). Each blind had a corn feeder located within 100 m in grasslands near (,20 m) wooded habitat. Whole shelled corn is a desired food source because it is rapidly fermented by both deer species (Wheaton and Brown 1983, Van Soest 1994). We set feeders to feed each day when we were making observations, and feeders dispersed approximately 2.3 kg of corn in a 14m-diameter area. By baiting deer to the area, there was greater potential to observe interspecific interactions for a highly desired food for an extended period of time (Thomas et al. 1965, Ables 1977, Hirth 1977, McGhee and Baccus 2006). When we observed an animal, we recorded time of day, number of animals, species, sex, and age class. We made observations using 10 3 42 binoculars. When both axis and white-tailed deer were present at a feeder at the same time, we recorded any aggressive behavior resulting in a displacement of one species by another. Displacement was categorized as complete and incomplete (Weckerly 1999). We categorized behavior resulting in one animal leaving the area as complete displacement and behavior resulting in one animal moving away from another but not leaving the feed area as incomplete displacement (Weckerly 1999). We used only those times when both axis deer and white-tailed deer were present at the feeder for the displacement analysis, because we were not concerned with intraspecific interactions. We calculated rate of displacement as the number of displacements per animal per minute. For example, if 5 white-tailed deer were at a feeder with axis deer for 10 minutes and 2 displacements took place, rate of displacement would be 0.04 displacements/animal/minute. Due to nonnormal distribution of the data, we used a randomization test to determine whether axis deer displaced white-tailed deer to a greater extent than white-tailed deer were displaced by axis deer (Manly 1997). Because we predicted that most displacements were by axis deer, we conducted a one-tailed randomization test for paired data (Manly 1997). We conducted these randomization tests for both complete and incomplete displacements and when both displacements were pooled. We conducted 10,000 permuta701
tions. The test statistic (ts) for the randomization test was the sum of the differences between the rate of displacement of white-tailed deer and axis deer for each sampling occasion. We also evaluated length of time axis deer and whitetailed deer were at feeders alone and together. We recorded time when 1 animal from either species was present. We did not include number of animals present. We divided data into morning and evening observations and calculated mean amount of time for each species alone and together. We calculated a likelihood ratio test to determine whether variances differed between species composition (axis deer only, white-tailed deer only, and axis and white-tailed deer together) and across diel period (morning, evening; Sokal and Rohlf 1995). If variances were significantly different, we conducted a linear contrast in a restricted maximum likelihood analysis of variance (ANOVA) that accounted for heteroscedasticity to determine whether the amount of time species spent alone at feeders differed from the amount of time both species were present together (Sokal and Rohlf 1995, Pinheiro and Bates 2000). We used the restricted maximum likelihood ANOVA because least-squares ANOVA is unreliable when data are heteroscedastic (Sokal and Rohlf 1995, Pinheiro and Bates 2000). To determine whether the 2 species spent different amounts of time at feeders we determined deer-minutes by multiplying number of minutes a species was present at the feeder by number of animals of that species. We again used a likelihood ratio test to determine whether variances differed among combinations of species and diel period. If variances differed we conducted a restricted maximum likelihood ANOVA that accounted for heteroscedasticity to determine whether there were differences in number of deer-minutes each species spent at the feeders in morning and evening. Likelihood ratio tests often lack statistical power, so to increase statistical power of the test we divided the P-value by 2 (Hedeker and Gibbons 2006). Removal of axis deer began immediately after conclusion of the first data collection period. The first technique we used for removal was a catch pen, a 2.4-m-high fencing formed into a funnel and lane leading up to the facility for capturing animals. We used a field planted in oats, wheat, and triticale (Triticale hexaploide) to bait deer to the fenced area. We also scattered shelled corn on the ground one week prior to capture to lure deer to the area. After 2 captures in the catch pen (26 Jan and 30 Jan 2007) we closed the bait area. One week prior to capture, we set up and baited a drop-net with corn (Rongstad and McCabe 1984). We dropped the net on a group of axis deer and removed it on 27 February 2007. On 2 March, a helicopter capture team flew the area, captured axis deer with net guns, and removed as many axis deer as possible (Webb et al. 2008). Remaining deer were harvested with rifles either by sitting in hunting blinds over bait or out of a vehicle at night using artificial light. Because axis deer were not regulated by Texas Parks and Wildlife Department, we could use any legal means to remove axis deer (Texas Parks and Wildlife Department 2007). L
702
RESULTS Habitat Use We conducted 50 surveys (morning and evening) in both the treatment and control areas during both survey seasons. We tallied 261 (per survey; ¯x 5 5.22, s 5 6.29) axis deer and 108 ¯ 5 2.16, s 5 3.03) white-tailed deer in the treatment area (x during the first survey season. In the second survey season we ¯ 5 0.88, s 5 1.49) white-tailed deer in the observed 44 (x treatment area. In the control area, we counted 53 (x ¯ 5 1.06, ¯ 5 2.58, s 5 2.86) white-tailed deer in s 5 1.61) and 129 (x the first and second survey seasons, respectively. In the treatment area, 26 axis deer were located within 67% of detection plots. Also in the treatment area, during the first survey season, 39% of detection plots were occupied by 9 and 12 white-tailed deer in the mornings and evenings, respectively. During the second survey season, 60% of detection plots in the treatment area were occupied by 5 and 15 white-tailed deer in the mornings and evenings, respectively. In the control area, we counted 4 and 13 whitetailed deer in mornings and evenings, respectively, in 62% of detection plots during the first survey season. In the second survey season, 48% of detection plots were occupied by 15 and 17 white-tailed deer in mornings and evenings, respectively. Between survey seasons, we removed 50 axis deer from the treatment area, 25 with the catch pen, 3 with the drop-net, and 9 by the helicopter crew. We spent approximately 70 man-hours in blinds and spotlighting at night, resulting in harvest of 10 axis deer. During the first 2 weeks of surveys, we observed 4 axis deer females on the survey route. After this initial observation 3 axis deer females were harvested within the first 2 months of surveys in the second survey season. We are confident that only one axis deer remained in the treatment area. Using the model with no covariate we estimated detection probability for axis deer to be 0.123 (Table 1). Based on activities to remove axis deer, minimum population size of axis deer was 51, close to the population size estimated from the ratio of the mean of survey counts and detection probability (5.17/0.123 5 42). The agreement in the 2 estimates of population size indicates that detection probabilities were not strongly biased. The model with no covariate was also selected for whitetailed deer (Table 1). From this model, we obtained a detection probability estimate of 0.069. From the detection probability and mean of survey counts, estimated population sizes were 21 white-tailed deer in the treatment area and 26 in the control area. The treatment area was 53.85% wooded habitat. Whitetailed deer used wooded habitat 34.6 6 4.2% during the first survey season and 79.7 6 4.5% during the second survey season, less and more than its availability, respectively (Fig. 2). The difference between percentage habitat use in the 2 survey seasons was 244.8 6 4.5%. Axis deer used wooded habitat 30.4 6 2.6% of the time in the first survey season, which was less than its availability. The control area consisted of 34.31% wooded habitat. White-tailed deer used wooded habitat 23.1 6 3.6% and The Journal of Wildlife Management N 74(4)
Table 1. Model selection analyses to examine whether detection probabilities of white-tailed deer and axis deer in Bandera County, Texas, USA, 2006– 2007, were influenced by diel period (diel), open or wooded habitat (hab), treatment or control area (site), survey season (yr), or not influenced by any of these variables (.). y 5 occupancy and p 5 detection probability. Model selection was based on Akaike Information Criterion coefficients (AICc), Akaike weights (wi), number of parameters (nPar), and 22 times the log likelihood (22LogLike). Modela White-tailed deer y(.),p(.) y(.),p(diel) y(.),p(hab) y(.),p(site) y(.),p(yr) Axis deer y(.),p(.) y(.),p(diel) y(.),p(hab) a
AICc
wi
nPar
22LogLike
631.30 631.69 632.80 633.13 635.15
0.33 0.30 0.17 0.15 0.05
2 3 3 3 3
627.06 625.21 626.32 626.65 628.67
171.15 173.72 173.74
0.51 0.20 0.20
2 3 3
166.48 166.31 166.33
For white-tailed and axis deer the y(.),p(.) model was selected.
25.1 6 3.1%, less than its availability, during survey seasons 1 and 2, respectively (Fig. 2). There was no difference in habitat use between the 2 survey seasons (22 6 4.8%). Because detection probabilities did not differ between diel periods, habitats, treatments, or survey seasons, estimates of wooded habitat use were similar for corrected or uncorrected counts. Hence, confidence intervals of the difference of percentages resulted in more conservative inferences, and a more robust analysis, when we used uncorrected counts, because sample sizes were much smaller. When we
calculated 95% confidence intervals of the difference of percentages from uncorrected counts our conclusions, however, were the same. Between survey seasons 1 and 2 there was an increase in use of wooded habitat by whitetailed deer (244 6 8%) in the treatment area but no change in use of wooded habitat in the control area (24 6 7%). In the treatment area selection by white-tailed deer for wooded habitat increased 2.2 (corrected counts) and 2.1 (uncorrected counts) times between survey seasons 1 and 2. In the control area selection by white-tailed deer for wooded habitat changed by (at most) 1.1 (corrected counts) and 1.2 (uncorrected counts) times between survey seasons 1 and 2. Displacements and Time Spent at Feeders We spent 58 hours observing feeders during 29 occasions. On 19 occasions we observed white-tailed deer or axis deer, but not both species (65.5%). Only during 8 occasions did we observe both white-tailed deer and axis deer at a feeder at the same time. During these sampling occasions, axis deer completely displaced white-tailed deer 17 times resulting in a mean rate of displacement of 0.140 (s 5 0.123). We did not observe complete displacement of axis deer by whitetailed deer. We observed incomplete displacements of white-tailed deer by axis deer 22 times (mean rate of displacement 5 0.086, s 5 0.046) and incomplete displacements of axis deer by white-tailed deer 2 times (mean rate of displacement 5 0.002, s 5 0.002). When we pooled both types of displacements, mean rate of axis deer displacing white-tailed deer was 0.226 (s 5 0.119) and mean rate of white-tailed deer displacing axis deer was 0.002 (s 5 0.002). Axis deer displaced white-tailed deer to a greater degree than white-tailed deer displaced axis deer for all displacement types (complete: ts 5 1.123, P , 0.001; incomplete: ts 5 0.666, P 5 0.005; combined: ts 5 1.789, P , 0.001). During all occasions when 1 animal of either species was present, we observed white-tailed deer at feeders alone ¯ 5 15.79, s 5 3.74), axis deer alone 300 minutes (x ¯ 5 11.21, s 5 3.85), and both species 213 minutes (x ¯ 5 5.26, s 5 1.85). Variance of each together 100 minutes (x species composition in each diel period differed (x25 5 29.89, P , 0.001) so we conducted the maximum likelihood ANOVA. There was no interaction between diel period and L
Figure 2. Proportion of wooded habitat used in survey seasons 1 and 2 by white-tailed deer compared to amount of wooded habitat available in treatment and control areas in Bandera County, Texas, USA, 2006–2007. Error bars represent 95% confidence intervals. Faas and Weckerly N Habitat Interference
703
species composition (t51 5 20.209, P 5 0.835). Amount of time axis and white-tailed deer each spent at feeders alone was longer than when both axis and white-tailed deer were present together (t51 5 21.89, P 5 0.064). When 1 animal of either species was present at a feeder, we recorded 1,542 deer-minutes for white-tailed deer (x ¯ 5 81.16, s 5 22.71) and 1,194 deer-minutes for axis deer (x ¯ 5 62.84, s 5 22.24). There were significant differences between variances in each diel period and species (x25 5 6.178, P 5 0.051) so we used the maximum likelihood ANOVA again. No interaction was present between diel period and species (t34 5 21.47, P 5 0.151), and number of deer-minutes whitetailed deer spent at feeders was not different from number of deer-minutes axis deer spent at feeders (t34 5 1.58, P 5 0.124). L
DISCUSSION Our study suggested that axis deer were superior at interference competition. First, we found that white-tailed deer shifted habitat use dramatically (36–52%) in the absence of axis deer, whereas white-tailed deer habitat use in the control area was constant over the same time span. Second, data from behavioral interactions clearly indicated that axis deer were socially dominant to white-tailed deer. We observed both male and female axis deer displacing male and female white-tailed deer. When we pooled both complete and incomplete displacements, 1 displacement of white-tailed deer by axis deer took place during every observation. In the 2 recorded observations of axis deer displacement by white-tailed deer, 2 male white-tailed deer made an aggressive advance toward 2 female axis deer resulting in those deer moving to a different location to feed. Under no circumstances was the rate of displacement of axis deer by white-tailed deer greater than that of the opposite. Third, at stations with a highly desired food source, each species occurred alone more frequently than both species together, which supports our hypothesis of interference competition of axis deer on white-tailed deer because the subordinate species should avoid areas used by the dominant species, or the subordinate species should change times it uses the resource to avoid interactions with the dominant species. Case and Gilpin (1974) and Carothers and Jaksic (1984) suggested that temporal partitioning between species may be one of the most recognizable symptoms of interspecific competition. Our study also suggests that the 2 species will not coexist because white-tailed deer were not superior at exploitative competition. Of the 3 conditions that should be met for white-tailed deer to be superior at exploitative competition, the 2 critical conditions were not met. There should have been no change in habitat use by white-tailed deer in the presence or absence of axis deer and white-tailed deer should have occurred more frequently at stations with a highly desired food source. If white-tailed deer had spent a greater amount of time at feeders than did axis deer, we would have expected white-tailed deer to be able to consume corn before or in a more efficient way than axis deer. Extending this finding to habitat use, white-tailed deer habitat use should L
704
be similar in the presence and absence of axis deer because foraging time in habitat is unaltered. To further refute the hypothesis that white-tailed deer are superior at exploitative competition, axis deer have an advantage over white-tailed deer in that axis deer are capable of using grasses, forbs, and browse whereas white-tailed deer mostly exploit only forbs and browse (Ables 1977, Putman 1988, Traweek and Welch 1992, Mungall and Sheffield 1994, Geist 1998). Food habits of axis deer vary according to the location in which the study is conducted. Whereas some have suggested grass is the primary food choice of axis deer, others have suggested that as much as half of axis deer diet may consist of forbs and browse (Ables 1977, Putman 1988, Mungall and Sheffield 1994, Geist 1998). The important conclusion is that axis deer are capable of using the same resources as white-tailed deer. Because white-tailed deer were not superior at exploitative competition, coexistence is unlikely to occur between whitetailed deer and axis deer (Amarasekare 2002, Stewart et al. 2002). Axis deer can interfere with resource selection of white-tailed deer and potentially exploit resources before white-tailed deer. Although we suggest that coexistence is unlikely, exploitative competition may not be as influential if densities of axis deer and white-tailed deer are kept low and habitat quality is high. Use of common resources manifests most when densities of animals are high and resources are limited (Stewart et al. 2002). When densities are low, there is greater potential for habitat partitioning, less interspecific competition, and fewer aggressive interactions, which should promote coexistence or niche partitioning to occur. When we designed this study, we thought densities of white-tailed deer and axis deer were approximately equal. However, we found that there were approximately twice as many axis deer as white-tailed deer in the treatment area, which possibly allowed axis deer to both interfere with and exploit resources before white-tailed deer (Amarasekare 2002, Stewart et al. 2002). If densities of white-tailed deer and axis deer were similar or if white-tailed deer were twice as abundant as axis deer in the treatment area, we suspect that axis deer would still be superior at interference competition but that white-tailed deer would be able to compete more strongly for available resources. Niche partitioning may take place to a greater degree, but over time, owing to interference from axis deer, white-tailed deer population sizes would decrease and axis deer would become more abundant. Our study is one of the first using both spatial and temporal controls to show habitat displacement of a native large mammal by an exotic large mammal (Berger 1985, Bartos et al. 2002, Stewart et al. 2002). Previous studies had only one type of control (usually spatial). Having both spatial and temporal controls was important because it allowed us to assess whether factors such as changes in the weather or differences in habitats in treatment and control areas affected results. For instance the first survey season was a below-average year for precipitation with only 61 cm recorded; however, the second survey season received .2.5 The Journal of Wildlife Management N 74(4)
times that amount (NOAA 2007). If the dramatic difference in habitat use by white-tailed deer in the treatment area could have been attributed to this difference in precipitation, a change in habitat use by white-tailed deer in the control area should have occurred, which did not happen. Although it was apparent that white-tailed deer shifted their habitat use because of axis deer, the underlying cause is unknown. The most likely conclusions were that axis deer displaced white-tailed deer from preferred habitat either by direct aggressive behavior resulting in interference competition or by increasing total deer density in wooded habitats to the point that white-tailed deer sought food in less preferred habitats. Our results suggest that interference competition took place; therefore, displacement by aggressive behaviors is most likely. Furthermore, given that axis deer are capable of consuming forbs and browse, their removal would decrease foraging pressure in wooded areas allowing white-tailed deer to use this habitat to a greater degree, assuming that white-tailed deer prefer and select for wooded habitat. However, this was not apparent in the control area where white-tailed deer used wooded habitat less than its availability. The difference in habitat selection of white-tailed deer between treatment and control areas may have been due to 2 possibilities. First, there could have been subtle differences in characteristics of habitat in the 2 areas (Fig. 1). For example, in the control area open habitat was characterized by more trees and shrubs scattered in the grassland and more broken terrain. More screening cover could have resulted in the deer spending greater amounts of foraging time away from the escape cover of wooded habitat. Second, densities of white-tailed deer probably differed by at least a factor of 2 between treatment (21 deer/130 ha 5 0.16) and control (25/ 360 5 0.07) areas. Most likely, habitat selection by deer is density dependent (Morris 1987). The dissimilar densities probably influenced habitat selection differently in the treatment and control areas.
MANAGEMENT IMPLICATIONS Due to their superiority at interference competition it is feasible to expect that axis deer densities will increase in central Texas. If axis deer densities actually increase throughout central Texas white-tailed deer will probably be displaced from higher quality to lesser quality habitats and abundance of white-tailed deer will decrease in this region because of their poor competitive abilities with axis deer. Whether axis deer densities will increase, however, is far from clear. How susceptible axis deer are to contemporary sources of predation in Texas has not been studied. Nor is it clear how spatial variation in habitat composition at different spatial scales and seasonal and annual variation in climatic conditions influences competitive interactions between the 2 species. The primary concern in the state of Texas is those areas with free-ranging axis deer. Currently, it is likely that there are fewer free-ranging axis deer than white-tailed deer throughout central Texas so negative effects are likely minimal. However, there are no population estimates or counts of axis deer throughout central Texas. Faas and Weckerly N Habitat Interference
Periodic monitoring of axis deer and white-tailed deer abundance in central Texas is warranted so that wildlife managers can determine whether white-tailed deer abundance is negatively associated with axis deer abundance.
ACKNOWLEDGMENTS We are grateful to C. Williams and the Flying A Ranch for access to the ranch and providing C. J. Faas with lodging and work during the field study. We also thank K. Lake, B. Trussell, and T. Reagan. Additional funding was provided from Dallas Safari Club, and C. J. Faas received 2 scholarships from the Houston Safari Club. R. Simpson, C. Green, and 2 anonymous reviewers commented on earlier drafts of the manuscript.
LITERATURE CITED Ables, E. 1977. The axis deer in Texas. Texas A&M Press, College Station, USA. Amarasekare, P. 2002. Interference competition and species coexistence. Proceedings of the Royal Society B: Biological Sciences 269:2541–2550. Bartos, L., D. Vankova, K. V. Miller, and J. Siler. 2002. Interspecific competition between white-tailed, fallow, red, and roe deer. Journal of Wildlife Management 66:522–527. Berger, J. 1985. Interspecific interactions and dominance among wild great basin ungulates. Journal of Mammalogy 66:571–573. Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Second edition. Springer Verlag, New York, New York, USA. Butts, G. L. 1979. The status of exotic big game in Texas. Rangelands 1:152–153. Carothers, J. H., and F. M. Jaksic. 1984. Time as a niche difference: the role of interference competition. Oikos 42:403–406. Case, T. J., and M. E. Gilpin. 1974. Interference competition and niche theory. Proceedings National Academy Sciences, USA 71:3073–3077. Clauss, M., R. R. Hofmann, J. Fickel, W. J. Streich, and J. Hummel. 2009. The intraruminal papillation gradient in wild ruminants of different feeding types: implications for rumen physiology. Journal of Morphology 270:929–942. Demarais, S., J. T. Baccus, and M. S. Traweek, Jr. 1998. Non-indigenous ungulates in Texas: long-term population trends and possible competitive mechanisms. Transactions North American Wildlife Natural Resource Conference 63:49–55. Demarais, S., K. V. Miller, and H. A. Jacobson. 2000. White-tailed deer. Pages 601–628 in S. Demarais and P. R. Krausman, editors. Ecology and management of large mammals in North America. Prentice Hall, Upper Saddle River, New Jersey, USA. Draper, N. R., and H. Smith. 1998. Applied regression analysis. Third edition. Wiley Interscience, New York, New York, USA. Fulbright, T. E., and J. A. Ortega-S, 2006. White-tailed deer habitat: ecology and management on rangelands. Texas A&M University Press, College Station, USA. Gabor, T. M., and E. C. Hellgren. 2000. Variation in peccary populations: landscape composition or competition by an invader? Ecology 81:2509– 2524. Geist, V. 1998. Deer of the world. Stackpole, Mechanicsburg, Pennsylvania, USA. Halls, L. K. 1984. White-tailed deer: ecology and management. Stackpole, Harrisburg, Pennsylvania, USA. Hedeker, D. R., and R. D. Gibbons. 2006. Longitudinal data analysis. Wiley Interscience, Hoboken, New Jersey, USA. Henke, S. E., S. Demarais, and J. A. Pfister. 1988. Digestive capacity and diets of white-tailed deer and exotic ruminants. Journal of Wildlife Management 52:595–598. Hirth, D. H. 1977. Social behavior of white-tailed deer in relation to habitat. Wildlife Monographs 53. Hofmann, R. R. 1985. Digestive physiology of deer—their morphological specialization and adaptation. Royal Society of New Zealand, Bulletin 22:393–407. 705
Kendall, W. L. 1999. Robustness of closed capture–recapture methods to violations of the closure assumption. Ecology 80:2517–2525. Lancia, R. A., C. S. Rosenberry, and M. C. Conner. 2000. Population parameters and their estimation. Pages 64–83 in S. Demarais and P. R. Krausman, editors. Ecology and management of large mammals in North America. Prentice Hall, Upper Saddle River, New Jersey, USA. Longoria, M. P., and F. W. Weckerly. 2007. Estimating detection probabilities of collared peccary from sign. Journal of Wildlife Management 71:652–655. MacKenzie, D. I. 2005. What are the issues with presence–absence data for wildlife managers? Journal of Wildlife Management 69:849–860. MacKenzie, D. I., J. D. Nichols, G. B. Lachman, S. Droege, J. A. Royle, and C. A. Langtimm. 2002. Estimating site occupancy rates when detection probabilities are less than one. Ecology 83:2248–2255. Manly, B. F. J. 1997. Randomization, bootstrap and Monte Carlo methods in biology. Chapman and Hall, London, United Kingdom. McGhee, J. D., and J. T. Baccus. 2006. Behavioral interactions between axis and fallow deer at high-value food patches. Southwestern Naturalist 51:358–367. Morris, D. W. 1987. Tests of density-dependent habitat selection in a patchy environment. Ecological Monographs 57:270–281. Morse, D. H. 1974. Niche breadth as a function of social dominance. American Naturalist 108:818–830. Mungall, E. C., and W. J. Sheffield. 1994. Exotics on the range: the Texas example. Texas A&M University, College Station, USA. National Oceanic and Atmospheric Administration [NOAA]. 2007. Climatological data annual summary, Texas. U.S. Department of Commerce, Washington, D.C., USA. Pinheiro, J. C., and D. M. Bates. 2000. Mixed-effects models in S and SPlus. Springer Verlag, New York, New York, USA. Putman, R. 1988. The natural history of deer. Cornell University Press, Ithaca, New York, USA. Rongstad, O. J., and R. A. McCabe. 1984. Capture techniques. Pages 655– 676 in L. K. Halls, editor. White-tailed deer ecology and management. Stackpole, Harrisburg, Pennsylvania, USA.
706
Sanders, C. L. 1963. Habitat preferences of white-tailed deer and several exotic ungulates in South Texas. Ecology 44:803–806. Sokal, R. R., and F. J. Rohlf. 1995. Biometry: the principles and practice of statistics in biological research. Third edition. W. H. Freeman, New York, New York, USA. Stewart, K. M., R. T. Bowyer, J. G. Kie, N. J. Cimon, and B. K. Johnson. 2002. Temporospatial distributions of elk, mule deer, and cattle: resource partitioning and competitive displacement. Journal of Mammalogy 83:229–244. Texas Natural Resources Information System. 2007. TNRIS data distribution system, Texas Natural Resources Information System, Austin, USA. ,http//www.tnris.state.tx.us.. Accessed 15 Feb 2007. Texas Parks and Wildlife Department. 2007. Hunting regulations. Texas Parks and Wildlife Department, Austin, USA. Thomas, J. W., R. M. Robinson, and R. G. Marburger. 1965. Social behavior in a white-tailed deer herd containing hypogonadal males. Journal of Mammalogy 46:314–327. Traweek, Jr., M. S., and R. Welch. 1992. Exotics in Texas. Texas Parks and Wildlife Department, Austin, USA. Van Soest, P. J. 1994. Nutritional ecology of the ruminant. Second edition. Comstock, Ithaca, New York, USA. Webb, S. L., J. S. Lewis, D. G. Hewitt, M. W. Hellickson, and F. C. Bryant. 2008. Assessing the helicopter and net gun as a capture technique for white-tailed deer. Journal of Wildlife Management 72:310–314. Weckerly, F. W. 1999. Social bonding and aggression in female Roosevelt elk. Canadian Journal of Zoology 77:1379–1384. Wheaton, C., and R. D. Brown. 1983. Feed intake and digestive efficiency of south Texas white-tailed deer. Journal of Wildlife Management 47:442–450. White, G. C., and R. A. Garrot. 1990. Analysis of wildlife radio-tracking data. Academic Press, San Diego, California, USA. Ziv, Y., Z. Abramsky, B. P. Kotler, and A. Subach. 1993. Interference competition and temporal and habitat partitioning in two gerbil species. Oikos 66:237–246. Associate Editor: Hudson.
The Journal of Wildlife Management N 74(4)
