D. TOLLESON, Texas Agricultural Experiment Station, Vernon, TX 76384
D. ROLLINS, Texas Agricultural Extension Service, San Angelo, TX 76901
W. PINCHAK, Texas Agricultural Experiment Station, Vernon, TX 76384
M. IVY, USDA-Soil Conservation Service, Albany, TX 76430
A. HIERMAN, USDA-Soil Conservation Service, Albany, TX 76430
Feral hogs (Sus scrofa) are attracting a lot of attention in Texas. Some people view the feral hog as a disease-carrying nuisance, rooting up range and cropland, or damaging fences and equipment. Yet, to others they are an economic asset. Feral hogs are quite prolific and can provide meat or added income from trapping and selling or from leasing the rights to hunt them. To some managers of white-tail deer (Odocoileus virginianus) and upland gamebirds however, feral hogs are a competitor for mast and forage or a potential predator (Kroll 1986). Like them or not, the feral hog is most probably here to stay, thus their impact on wildlife and agriculture in Texas should be addressed.
The purpose of this paper is to examine the interrelationship of feral hogs and bobwhite quail (Colinus virginianus). The bobwhite is a very important game species in Texas for several reasons. In the Rolling Plains prices charged for quail hunting range from $1-$2.50/ac for season-long leases and from $50-$100 per day for short-term hunts. Quail hunters purchase food, lodging and supplies making a significant contribution to rural economies. Bobwhites also provide enjoyment to non-consumptive interests such as photographers and ornithologists. Thus, aesthetically, recreationally and economically, the bobwhite is an integral part of the Texas wildlife community.
Feral hogs and bobwhite quail co-inhabit many areas in Texas. Feral hogs have been reported to prey on the eggs of ground nesting birds such as quail and turkeys (Synatzske 1979). In response to the recent proliferation of feral hogs in the Rolling Plains, a study was initiated in the spring of 1992 to examine the impact of feral hogs on the survival of simulated quail nests at 2 different locations within the region.
Experiment 1 was conducted on the Y Experimental Ranch (YER) located in Foard, Cottle, King and Knox counties. Much of the YER consists of expansive mesquite (Prosopis glandulosa) flats broken by numerous juniper (Juniperus pinchotii) ridges. Experiment 2 was carried out on the Nail, Cook and Newell Ranches in Shackelford county. This area is in a transition zone between the western Cross Timbers and the southeastern Rolling Plains. As compared to Experiment 1, the ranches in Experiment 2 contained more open grassland, interspersed with mottes of mesquite and oak (Quercus spp.).
An aerial census conducted on the YER in March of 1992 indicated that at least 171 feral hogs occupied the 38,000ac ranch (Rollins et al. 1992). Aerial surveys were not available on the Shackelford county ranches, however, each ranch is reported to have “moderate to high” hog populations.
Each nest consisted of 3 small unwashed chicken eggs. An ideal bobwhite nest site has been described as a bunch grass, 10-12 inches in diameter, 6-8 inches tall, with dead material from the previous year, located near some type of escape cover (Bidwell, et al 1992). An attempt was made to simulate these conditions when locating nest sites. Plastic gloves were worn while handling the eggs and care was taken to minimize disturbance to the immediate area while constructing the nests.
A transect consisted of 12 points (Experiment 1) or 20 points (Experiment 2) approximately 25 yards apart. Nest sites were chosen at least 10 yards perpendicular to each point, alternating left and right.
In Experiment 1, 192 simulated quail nests were constructed. Four transects of 12 nests each were placed in 4 different pastures. Two different vegetation types (mesquite vs. juniper) were located approximately 1 mile apart in each pasture, and 2 different topographic locations (upland vs. lowland) separated by approximately .25 miles within each vegetation type were utilized. This arrangement resulted in 4 distinct habitat types (mesquite-upland, mesquite-lowland, juniper-upland and juniper- lowland) being represented in each pasture.
Three hundred sixty nests were constructed in Experiment 2. Six transects of 20 nests each were located on each ranch. Three located in upland and 3 in lowland areas. Again, each topographic location was separated by approximately .25 miles and each pair was approximately 1 mile apart.
Nests were inspected once a week for 6 weeks from mid-May through early July. Each transect was traversed on foot, the nests located and their status recorded. If a nest was intact, it was left undisturbed. If the nest had been destroyed, it was examined more closely in an effort to identify the predator. Tracks, droppings, “rooting” and the condition of any eggshell fragments were used as a means of identification. Notes describing the signs and condition of the nest were made and representative pictures were taken.
Differences in probabilities of survival between habitat types were determined using General Linear Model Split-Split Plot Analysis of Variance procedures (Statistical Analysis System). Main effects in the model were week, vegetation type and location (Experiment 1), or week and location (Experiment 2). Descriptive analyses of cumulative depredation rates and classification of predators were conducted.
Most (56%) of the depredation in Experiment 1 was attributed to unknown causes (Fig. 1A). Feral hogs, at 28%, were the most often identified predator. Their activity was distributed across vegetation types and locations. Opposums and raccoons however, numerically preferred upland (16%) over lowland (7%) areas.
Of the original 192 nests, 84% were destroyed after the first 3 weeks and only an additional 14% destruction had occurred after 6 weeks (Fig. 2). Cumulative depredation by week 5 had reached 100% as compared to 94% in juniper vs. mesquite sites respectively (Fig. 2A). Upland locations had experienced 100% depredation after 6 weeks whereas lowland areas had suffered a 96% loss (Fig. 2B).
Percent weekly depredation was greater during the first 3 weeks than the final 3 (p<.02, 38 vs 19% respectively). Weekly depredation was also greater in upland than in lowland sites (p< .02, 32 vs 25% respectively). There was no difference in weekly depredation between mesquite (31%) and juniper (26%) vegetation types. There was however, a significant week x vegetation type interaction (p<.05).
Coyotes (32%), skunks (23%) and snakes (16%) were the most often implicated predators in this experiment (Fig. 1B). Coyote activity appeared to be more prevalent in upland (37%) than in lowland (26%) areas. Feral hogs (8%) were not a major factor in either location.
After 3 weeks, 25% of the original 360 nests were destroyed (Fig. 3). Nest destruction had increased over 2.5 times to 67% by week 6. These results contrast sharply with those from the YER (Fig. 2).
Percent weekly depredation was greatest in week 6 (41%, p<.05), followed by week 5 (18%, p<.05). Weeks 2,3 and 4 were intermediate (7, 15 and 14% respectively), while the least amount of depredation occurred in the first week (5%, p<.05). There was no significant difference in weekly depredation between lowland (15%) and upland (18%) locations.
Most depredation on the YER occurred within the first 3 weeks of the experiment while in Shackelford county, nest loss was least in week 1 and greatest in week 6. One original hypothesis of this research was that as the eggs rotted, they would become more vulnerable to depredation. The rate of depredation in Experiment 2 supports this hypothesis whereas the results of Experiment 1 do not. Another interesting observation is that feral hogs were associated with more nest loss in Experiment 1 than Experiment 2. The reasons for these inconsistencies are unknown. Observer bias is one probable factor as there was a different technician responsible for each experiment. Additionally, differences in predator populations and habitat types between the 2 experimental locations could certainly affect the rate of nest destruction.
Another hypothesis was that depredation would be greater in lowland than upland areas because feral hogs tend to favor wet habitats. The results presented here disagree with that hypothesis (Figs. 2 and 3). At both experimental locations, feral hogs are most often sighted around water, though it is not uncommon to encounter them in upland habitats. However, 1992 had a very wet spring, so water was not a limiting factor. The distribution of feral hogs could change if precipitation patterns differ.
The results of these 2 experiments indicate feral hogs could have a detrimental effect upon nesting success of bobwhite quail. Even though depredation of simulated nests was high on the YER; feeding, nesting and brooding cover were quite adequate and the quail population is up this year. As conditions change during future replications of this experiment, and more information is obtained; more definitive conclusions may be drawn. Information gathered under various conditions will be more useful since the density of quail relative to predators (including hogs), the quality of nesting cover for the quail and the amount of other foods available to the hogs will all affect the interrelationship between the two species.
Bidewell, T.G., S.R. Tully, A.D. Peoples and R.E. Masters. 1992. Habitat Appraisal Guide for Bobwhite Quail. Cooperative Extension Service, Oklahoma State University.
Kroll, J.C. 1986. Interspecific Competition Between Feral Hogs and White-tailed Deer in the Post-Oak Savannah Region of Texas. Texas Parks Wildl. Dep. Fed. Aid Proj. W-109-R-8. Job No. 44.
Rollins, D., W.E. Pinchak and D.R. Tolleson. 1992. Wildlife Research. Y Experimental Ranch 1992 Progress Report. 24 pp.
Synatzske, D.R. 1979. Status of the Feral Hog in Texas. Unpubl. Rep. 9 pp.