Evaluation of a New Deer Repellent on Japanese Yews at Suburban Homesteads

Jersey,anexperimentaldeerrepellent,wasfieldtestedagainst2commercialrepellentsonJapaneseyews(Taxuscuspidata) near Ithaca, New York, during spring 1990. In Experiment 1, plots (n = 24) of 4 individually-potted yews were established, with 2 yews at each plot randomly treated with Jersey and 2 left as controls. Plots of 4 (1 x 4, n =12) and 16 (4 x 4, n = 2) plants were used in Experiment 2, with individual plants being treated with Jersey, Hinder", or Big Game RepellentR (BGRR) or left as controls. Photographs with a grid matrix placed behind each h plant were taken from 2 m at the beginning of the experiment and after 10 weeks. These photographs were analyzed to produce a cover index of plant size. Plots were monitored weekly to record browsing. In Experiment 1 more control (46/48) than treated (7/48)plants were browsed (P < 0.001). Controls were browsed earlier (x =1.7 wk) than treated yews (x = 4.4 wk, P < 0.01). At the end of 10 weeks, control plants were reduced in size more than Jersey-treated plants (P < 0.001). In Experiment 2, browsing rates did not differ among treatments in the 1 x 4 plots or 4 x 4 plots. However, controls were browsed more frequently than treated at both plot types (10/12 at 1 x 4, and 6/8 at 4 x 4 plots) (P < 0.05). Browsing reduced control plants by 56.8% (n =10) in 1 x 4 plots and 47.2% (n = 6) in 4 x 4 plots. These results suggest that Jersey reduced deer damage to a shrub preferred by deer. Moreover, Jersey was as effective as BGRR and Hinder" at reducing browsing. Experiments may need to be conducted under more severe conditions and over a longer time-period to separate efficacy of the 3 repellents. I'roc. East. Wildl. Damage Control Conf. 5:38-43.1992. Browsing damage to ornamental trees and shrubs by whitetailed deer (Odocoileus virginianus) is common in many suburban areas of the eastern United States, and some homeowners report high economic losses (Decker and Gavin 1987, Connelly et a1.1987, Sayre and Decker 1990). Homeowners use various methods to prevent deer damage, including physical barriers such as fencing or tree wraps, commercial repellents, soaps, human hair or animal blood attached to plants, and scare devices (Decker and Gavin 1987, Connelly et al. 1987, Sayre and Decker 1990). Despite their popularity, most of these methods have limited long-term success in deterring deer (Harris et al. 1983, Matschke et a1.1984, Swihart and Conover 1990, Andelt et al. 1991). However, fencing and commercial repellents appear to be the most effective of these methods. Some people are reluctant to use fences or physical barriers because they can be expensive (Palmer et a1. 1985), and many consider them to be unsightly (Decker and Gavin 1987). Commercial repellents, although not a cure for deer damage problems, have successfully reduced browsing on shrubs and trees (Conover 1984, 1987; Swihart and Conover 1991). The primary limitations with commercial repellents are the expense and need for repeated applications. An effective and longlasting deer repellent is needed to reduce deer damage to ornamental plants. Several studies of repellents have been conducted at commercial nurseries (Conover 1984, 1987; Swihart and Conover 1990), and with captive deer (Cambell and Bullard 1972, Palmer et al. 1983, Harris et a1.1983, Andelt et a1.1991), Present address: Dept. of Biology, P.O. Box 8238, University of North Dakota, Grand Forks, ND 58202. although to our knowledge, repellents have not been fieldtested at suburban settings. Foraging behavior of deer in a tree nursery may differ from foraging in a suburban landscape, where the plants are dispersed. Animal response to repellents also may differ between nurseries and homesites because of such differences in plant distribution. Controlled experiments with captive animals (e.g., Palmer et al. 1985, Andelt et al. 1991) provide valuable knowledge, but they have limitations because captive deer may behave differently from wild deer, which have alternative foods available. For example, Andelt et al. (1991) reported that hungry deer in captivity actually licked bars of soap suspended over apple twigs; this behavior would be unlikely in a wild setting. We tested the effectiveness of an experimental deer repellent called "Jersey" (Patent No. 45,965,070) on Japanese yews (Taxes cuspidata), a shrub highly preferred by deer. The experiments were conducted in or near suburban homesites in central New York, and were designed to simulate conditions where plants might be expected to be more scattered than at a nursery or orchard. In Experiment 1, Jersey was tested against a control, and in Experiment 2, Jersey was compared to the commercial repellents BGRR and Hinders. These repellents were chosen because they are currently the most effective commercial repellents available (Conover 1984). We make no endorsement of these products. Three variables were tested to determine differences between treatments: (1) evidence of browsing; (2) elapsed time before browsing was first detected; and (3) reduction in plant size due to browsing. We thank DeVisser's Nursery for donating the yews and are grateful to the homeowners who allowed us to conduct this Study on their property. We also thank T. Sayre for assistance n the field; K. Gerow and K. Newsome-Stuart for statistical advice and M. Fargione, N. Ingle, D. Jordan, and A.M. Wilkinson for comments and review of the manuscript. This project was funded by U. S. Department of Agriculture, Animal and Plant Health Inspection Service (USDA/APHIS). STUDY AREA The study was conducted east and southeast of Cayuga lake near the city of Ithaca, Tompkins County, New York. experiment 1 was conducted in Lansing Village (LV), 2.5 km north of Ithaca, and Experiment 2 in LV, and in Ellis Hollow EH), 5 km east of Ithaca. LV is a suburban community, with a mixture of residential and commercial developments, woodlots, abandoned fields, rid farmland still in production. Commercial and residential real constitute about one-half of the land area. Significant commercial developments include shopping malls, industrial search centers, and the Tompkins County Airport. EH is lore rural than LV, and consists of farmlands, woodlands, and single-unit houses along roads, which dissect the area at aproximately 1 km intervals. These areas were located on the Allegheny Plateau, a region of large hills (elevation 450-610 m), dissected by narrow nines, and broad valleys with steep upper slopes. The soils ere formed from shale and sandstone glacial till (Neeley X65). The area is within the Hemlock-White Pine-Northern ardwoods region, and is dominated by sugar maple (Acer Saccharum and beech (Fagus grandifolia) and their associa (Braun 1950). Before settlement by Europeans, the area was densely forested, but it was logged andclearedforagriculture firing the 18th and 19th centuries. Much of the farmland has been abandoned since the late 1800s, and forests have regenered. Two photographs of each plant were taken from orthogonal directions using a 35 mm camera with a SO mm lens using color slide film (ASA 100 or ASA 200). A density board with a 5 x 5 cm grid matrix was placed 10 cm behind the plant for each photo. The camera was held at a height of 1 m, and a distance of 2 m from the plant. Plots were monitored weekly for 10 weeks to determine when browsing occurred. After 10 weeks, each plant was rephotographed from the original positions. Experiment l.-Plots were established at 24 homesites in LV from 6 March through 2 April, and terminated from 15 May through 10 June 1990. All homesites contained ornamental trees and shrubs that had been browsed by deer during the months preceding the experiment. Most homesites were on medium-sized property lots (median = 0.71 ha; range = 0.2 to 9.2 ha). One plot, each with 4 plants spaced 20 m apart, was established at each homesite. Two plants at each homesite were randomly selected and sprayed once with Jersey while the remaining 2 plants were left as untreated controls. Experiment 2.-Two plot designs were used in Experiment 2, 1 x 4 plots (4 plants each) and 4 x 4 plots (16 plants each), with plants spaced 20 m apart in each plot. The 1 x 4 plots were established on 19 February in EH, and on 27 March in LV (n = 6 plots/study area). The 1 x 4 plots at EH were placed on properties ranging in size from 6 to 32 ha, while LV plots were placed on properties between 0.4 and 9.2 ha in size. Most plots were in fields, but there was 1 woodland plot selected in each study area. One plant in each plot was randomly selected for treatment with BGRR, Hinder', Jersey or designated a control. Repellents were applied once as in Experiment 1. Data collections were completed on 30 April in EH, and on 6 June in LV. DEER REPELLENT AT SUBURBAN HOMESITES • Sayre and Richmond 39