2007 NORTHEASTERN NATURALIST 14(1):51–60 Ice Damage to Trees on the Virginia Tech Campus from Ice Storms Richard W. Rhoades1,* and R. Jay Stipes2 Abstract - The purpose of this study was to analyze ice damage to 228 trees of 9 species on the Virginia Tech campus. Damage was caused by three severe ice storms in February and March 1994. There were significant differences among species in amount of damage. Four ways of expressing percent damage were compared (% individuals damaged, % basal area of damaged trees, % crown damage, and an average of the three). The average method yielded the most significant comparisons, followed by percent crown damage. The species, ranked in four groups by mean total damage, are as follows: most damage (Acer saccharum, Chamaecyparis nooktakensis, Ulmus americana, and Acer nigrum: 29.7–26.4%); less damage (Quercus alba and Platanus occidentalis: 18.4–13.7%); lesser damage (Cornus florida: 10.2%); and least damage (Quercus rubra and Quercus palustris: 6.7–0%). The differences among groups were significant at P 􀂔 0.05. Ice damage also caused a significant decrease in crown growth of four species. Comparisons with other studies revealed good correspondence, in general, with two or three exceptions. Our conclusions are that three factors were chiefly responsible for the relatively severe damage to trees: 1) the severity of the ice storms; 2) the open, exposed siting of all the trees, similar to trees growing at the edge of a forest; and 3) the high percentage of large trees with internal decay and asymmetric crowns. Introduction The effect of ice on trees has been studied since early in the 20th century (Harshberger 1904). Ice storms occur with irregular frequency over a large part of the deciduous forest of eastern North America. In the southern Appalachians, very severe ice storms occur at a given site about once in twenty years (Abell 1934). Severe ice storms have occurred in western Virginia more often than that (R.W. Rhoades, pers. observ.: Dec. 1969, amount unknown; 1979, > 1 cm, see Whitney and Johnson [1984]; 1994, see text; Jan. 1997, 0.2 cm; Jan. 1998, 1.27 cm; Jan. 1999, 1.1 cm; Feb. 2000, 1.27 cm; Dec. 2000, 0.04 cm; Jan. 2002, 0.81 cm; Dec. 2002, 1.07 cm; Feb. 2003, 1.27 cm; and Dec. 2005, 0.84 + 2.5 cm [data from 1997 to 2005 were recorded in a rain gauge at the residence of the senior author]). An ice burden can cause large losses of biomass from forests (Boerner et al. 1988, Bruederle and Sterns 1985, Elstner and Ware 2001, Rhoades 1999). In addition to ice accumulation, factors such as wind, site, and topography contribute to damage during ice storms. While wind amplifies the effect of ice (Lemon 1961), it was not a factor in this study. 1611 Rose Avenue, Blacksburg, VA 24060. 2Department of Plant Pathology, Physiology, and Weed Science, Virginia Tech, Blacksburg, VA 24061. *Corresponding author - rrhoades@blacksburg.net. 52 Northeastern Naturalist Vol. 14, No. 1 Various approaches have been used to study the effects of ice damage to trees. Three studies have been done in western Virginia. Whitney and Johnson (1984) compared damaged stands with undamaged stands on steep slopes. Rhoades (1999) compared density and basal area of a stand before and after the ice storms, and the effect of slope steepness. Warrillow and Mou (1999) analyzed the effect of topography on ice damage. One study was done in eastern Virginia (Elstner and Ware 2001). They described significant differences in ice damage between young and old stands and between small and large trees. All these studies of damaged areas have reported percent of individuals damaged, as did Siccama et al. (1976), Boerner et al. (1988), and Seischab et al. (1993). Hauer et al. (1993) assessed damage in an urban area, and Lemon (1961) assessed damage and species resistance. De Steven et al. (1991) studied long-term changes in forest composition, due in part to disturbance by ice storms. Four ice storms occurred in southwestern Virginia in 1994. The first was a large storm that spread over the southeastern US from February 9 to 13 (NOAA 1994a). Three storms occurred in March (NOAA 1994b). Ice deposits from the storms in Blacksburg were as follows: February 9–13, 2.5 cm; March 2–3, 2.9 cm; March 9, 2.65 cm; and March 18, 0.4 cm. Little or no wind was associated with any of these storms. The purpose of this study was to assess ice damage to landscape trees on the Virginia Tech campus caused by these storms. We postulated that ice damage would be affected by several variables, namely, size of the tree (diameter at breast height [dbh] and crown diameter), presence of heart rot or disease, or asymmetry of the crown. One independent variable, percent of paved area beneath the crown, was found to affect growth in a prior study (Rhoades and Stipes 1999); so we also tested the effect of this factor on percent ice damage. We also evaluated the effectiveness of four methods of expressing ice damage. These were % individuals damaged, % basal area of damaged trees, % crown damage, and an average of the three. The overall objective of this study was to test for significant differences among species, and to test the effects of several variables on percent ice damage. Methods We measured dbh and crown diameter of 228 trees of 9 species on the Virginia Tech campus each year from 1993 to 1995 for a study of growth of trees (Rhoades and Stipes 1999). Crown diameter was the mean of two measurements at right angles through the crown at the drip line. The trees were selected to represent a range of diameters, and to be evenly divided between “poor” and “good” sites. Sites were judged to be poor if more than 10% of the area beneath the crown was paved over. We also made notes on the condition of each tree, every year. These conditions included trunk or branch injury, presence of disease, evidence of heart rot, and crown structure, including degree of asymmetry. In 1995, we 2007 R.W. Rhoades and R.J. Stipes 53 estimated percent of crown missing due to ice damage. We assigned each tree to one of four classes of crown loss: 1) light, < 25%; 2) moderate, 25– 50%; 3) severe, 􀂕 51%; or 4) very severe, > 51% damage and tree removed. Degree of asymmetry was determined from a disparity between the two measurements of crown diameter. For most trees, the two diameters were similar, but in about 20% of trees the diameters were markedly different, and these were judged to be asymmetric. Percentage deviation from symmetry (larger diameter vs. smaller diameter) was light (10 to 50%), moderate (51 to 100 %), or maximal (101 to 200%). The preferred method for analyzing data was analysis of variance (ANOVA). However, if this did not indicate significant differences among groups for any variable, then we used the non-parametric Kruskal-Wallis test. All percentage data were transformed to arcsine for ANOVA. Paired comparisons after ANOVA were made with single classification ANOVA, with all possible pairs, namely 36 pairs among nine species. Similar comparisons after Kruskal-Wallis test were made with the Mann-Whitney U-test. All statistical procedures are from Sokal and Rohlf (1982). Nomenclature of trees is from Kartesz (1999). Results Ice storms damaged crowns of 47 of 228 trees and resulted in removal of 19 of the damaged trees because they were so badly damaged. Five of those removed were Ulmus americana L. (American elms) that had been recently infected with Dutch elm disease. One elm was removed principally because of severe ice damage. Other elms were removed due to a combination of disease and ice damage. Composition of the sample of trees, their basal area, crown diameter, and deviation from symmetry are in Table 1. Percent of crown damage by ice varied among species (Table 2). A one-way analysis Table 1. Mean basal area (m2), mean crown diameter (m), and mean deviation from crown symmetry (m) of species on the Virginia Tech campusA. Standard errors in parentheses. Mean Mean Mean crown deviation from Species N basal area diameter symmetry Ulmus americana (American elm) 18 8.0 (1.41) 22.8 (4.42) 6.3 a Quercus alba (white oak) 28 8.9 (0.80) 16.4 (0.85) 3.6 a Acer nigrum (black maple) 28 3.2 (0.65) 14.1 (2.74) 2.1 b Quercus rubra (northern red oak) 30 3.5 (0.34) 15.7 (0.68) 1.9 c Acer saccharum (sugar maple) 30 3.7 (0.83) 15.6 (2.87) 1.0 d Platanus occidentalis (sycamore) 15 4.5 (0.26) 14.4 (2.39) 1.2 d Quercus palustris (pin oak) 19 3.3 (0.80) 15.5 (3.59) 1.6 d Chamaecyparis nooktakensis 28 6.9 (0.50) 6.2 (1.24) 1.1 d (Alaskan white cedar) Cornus florida (flowering dogwood) 32 0.26 (0.14) 8.7 (1.07) 0.8 d Total 228 AResults of Kruskall-Wallis test; means followed by the same letter are not significantly different at P 􀂔 0.01. 54 Northeastern Naturalist Vol. 14, No. 1 of variance indicated that differences in % crown damage among species were significant and sorted the species into three groups. Percent damage to individuals and percent basal area of damaged trees also differed among species, but with less definite groups. Of 36 possible paired comparisons among nine species, % individuals damaged yielded two comparisons significant at P 􀂔 0.05 and two significant at P 􀂔 0.01. Percent basal area of damaged trees yielded five comparisons significant at P 􀂔 0.01. Percent crown damage yielded six comparisons significant at P 􀂔 0.01 and one significant at P 􀂔 0.05. Average damage yielded seven comparisons significant at P 􀂔 0.01 and five at P 􀂔 0.05 The ice storms had a pronounced effect on crown growth (Table 3), resulting in negative growth in six species, and significantly negative growth in the damaged trees of four species. Tables 4 to 8 are a summary of how various factors affected percent of ice damage. Asymmetry of crowns was an important determinant of ice damage. There were significant differences Table 2. Percent ice damage to trees on the Virginia Tech campus. % Mean % basal % total N N Trees Species individuals area crown damageA (1993) (1995) removed Chamaecyparis nooktakensis 32.1 a 24.7 a 32.1a 29.6 a 28 24 4 Acer saccharum 30.0 a 31.2 a 27.8 a 29.7 a 30 29 1 Ulmus americana 27.8 a 27.2 a 24.2 a 26.4 a 18 13 5 Acer nigrum 25.0 a 25.3 a 24.2 a 24.9 a 28 23 5 Platanus occidentalis 20.0 a 3.0 b 18.1 b 13.7 b 15 12 3 Quercus alba 17.0 a 22.7 a 15.4 b 18.4 b 28 28 0 Cornus florida 15.6 a 1.7 b 13.3 b 10.2 c 32 31 1 Quercus rubra 6.7 b 6.3 b 7.0 c 6.7 d 30 30 0 Quercus palustris 0.0 b 0.0 b 0.0 c 0.0 d 19 19 0 Means and totals 19.6 16.4 18.4 18.2 228 209 19 Standard errors (1.37) (1.34) (1.37) (1.32) AResults of a one-way ANOVA; means followed by the same letter are not significantly different at P 􀂔 0.05. Table 3. Mean crown growth and their standard errors (in parentheses) of damaged versus undamaged trees (m/yr, 1993 to 1995). Levels of significance (ANOVA) are indicated thus: ** = P 􀂔 0.01, * = non-significant. Two species are not included in this table: pin oak was undamaged and crowns of Alaskan white cedar were measured only once (in 1995). Damaged trees Undamaged trees Species N Mean crown growth (SE) N Mean crown growth (SE) Ulmus americana 5 -5.5 (3.4) 13 0.10 (0.24)** Acer nigrum 7 -5.2 (2.4) 23 -0.40 (0.24)** Platanus occidentalis 3 -5.8 (4.3) 12 0.48 (0.20)** Quercus alba 5 -1.9 (1.2) 23 0.28 (0.38)** Acer saccharum 9 -1.2 (1.0) 21 0.47 (0.20)* Cornus florida 7 -0.06 (0.11) 25 0.12 (0.07)* Quercus rubra 2 0.08 (0.28) 28 0.31 (0.14)* Grand means and totals 38 -2.7 (0.70) 145 0.18 (0.08)** 2007 R.W. Rhoades and R.J. Stipes 55 Table 6. Percent ice damage (% indiv. +% B.A. + % crown)/3 in three dbh classes. Small trees Medium trees Large trees (5–30 cm) (31–60 cm) (> 60 cm) Mean % damageA 2.4 a 16.5 b 53.5 c s.e. 0.19 0.03 1.20 N 55 86 87 AResults of ANOVA; means followed by the same letter are not significantly different at P 􀂔 0.01. Table 7. Crown diameter versus percent ice damage. Crown diameter < 7 m 8–15 m 16–25 m Mean % damageA 16.5 a 2.4 b 6.0 b (s.e.) 1.30 1.96 4.30 N 51 94 83 AResults of ANOVA; means followed by the same letter are not significantly different at P 􀂔 0.05. among species in asymmetry, and its effect was reflected in significant differences in percent ice damage between trees with low asymmetry and trees with a high degree of asymmetry (Table 4). Heart rot was also important in amount of ice damage. Trees with a large amount of heart rot were more severely damaged than trees with lesser amounts (Table 5). Ice damage was also significantly greater in large trees compared with smaller trees. Table 4. Asymmetry of crowns versus sum percent ice damage (% indiv. + % B.A. + % crown )/3. Deviation from symmetry 10–50 51–100 101–200 Mean % damageA 11.0 a 15.3 b 26.8 c s.e. 0.70 1.30 2.90 N 20 15 12 AResults of Kruskal-Wallis test; means followed by the same letter are not significantly different at P 􀂔 0.01. Table 5. Effect of heart rot on sum ice damage (% indiv. + % B.A. + % crown) /3. Percent heart rot 0 10–20 25–35 50 >50 Mean % damageA 2.2 a 5.0 b 30.0 b 50.0 c 65.0 c s.e. 0.80 0.07 4.80 10.00 5.40 N 189 8 7 7 17 AResults of ANOVA; means followed by the same letter are not significantly different at P 􀂔 0.01. 56 Northeastern Naturalist Vol. 14, No. 1 Diameter at breast height seemed to have a stronger effect than crown diameter (Tables 6 and 7). Seven Cornus florida L. (flowering dogwoods), three Platanus occidentalis L. (sycamores), and four Chamaecyparis nootkatensis (D. Don) Spach (Alaskan white cedars) received damage more typical of larger trees, resulting in the finding that there was significantly greater ice damage to trees with small crowns (Table 7). Percent of paved area beneath the crown also had a significant effect on ice damage (Table 8). Comparison of our study with others in is Table 9. There was fairly good correspondence among most studies; however, there was marked variation in percent crown damage. Discussion Severity of ice damage depends in part on the relative susceptibility of individual species. In a study in Illinois, Hauer et al. (1993) rated 34 species of trees commonly planted in urban areas. They rated American elm and Quercus palustris Muenchh. (pin oak) as susceptible to ice damage. Elm trees on the Virginia Tech campus sustained considerable damage, but pin oak suffered no visible damage. In a similar study (Hauer et al. 1993) in Urbana, IL, 8.4% of pin oaks were severely damaged by ice. Hauer et al. Table 9. Comparisons among studies by percent ice damage. Study (authors) % basal area damaged % indiv. damaged % crown damaged Boerner et al. (1988) Canopy 29.2 24.7 Subcanopy 15.5 13.1 Rhoades (1999) Canopy 29.8 28.4 Subcanopy 14.1 10.4 Warrillow and Mou (1999) Canopy 13.2 Subcanopy 41.6 Elstner and Ware (2001) Canopy 45.9 16.0 Subcanopy 27.0 11.0 Seischab et al. (1993) 20.0 20.0 Hauer et al. (1993) 26.0 Rhoades and Stipes (current) 19.6 16.4 18.4 Table 8. Effect of paved area beneath the crown on mean percent ice damage. Percent paved 0 10–20 21–40 41–60 Mean % damageA 0.16 a 3.50 b 1.80 b 4.70 b s.e. 0.12 1.20 0.44 6.00 N 181 6 14 27 AResults of ANOVA; means followed by the same letter are not significantly different at P 􀂔 0.05. 2007 R.W. Rhoades and R.J. Stipes 57 (1993) reported least damage to Quercus rubra L. (northern red oak; 0.6%), Quercus alba L. (white oak; 0.7%), and Acer saccharum Marshall (sugar maple; 1.9%). On the Virginia Tech campus, about as many sugar maple and Acer nigrum Michaux (black maple) were damaged as American elm, 30% and 25%, respectively. Sycamore also sustained about as many damaged trees (20%), as American elm, black maple, and sugar maple. Hauer et al. (1993) rated white oak and Thuja occidentalis L. (arborvitae) as resistant. On the Virginia Tech campus, 17% of white oaks were damaged, and 32.1% of Alaskan white cedar, a species with a growth habit similar to arborvitae. Hauer et al. (1993) also rated northern red oak as intermediate in susceptibility. On the campus, 7% of red oaks were damaged. Flowering dogwood was not rated by Hauer et al. (1993), but it was rated as resistant by Boerner et al. (1988). We rated flowering dogwood as moderately susceptible. On the Virginia Tech campus, trees that received the heaviest damage were mostly large trees with asymmetric crowns and internal decay: sugar maple with 56% internal decay, and black maple with 25%. Many white oaks had asymmetric crowns with many unsound limbs, with 18% internal decay of the trunk (Rhoades and Stipes 1999). American elm sustained heavy damage because it has many fine and brittle branches, similar to Ulmus pumila L.( Siberian elm) described by Hauer et al. (1993). On campus, several American elms were infected with Dutch elm disease. Dutch elm disease has an effect on ice damage similar to that of heart rot (Tattar 1989). Flowering dogwood on campus had considerable damage: 15.6% of individuals and 13.3% of crowns, but only 1.7% of basal area (Table 2). Five dogwoods had been injured by lawn mowers and had much rot. These received ice damage of 25–35%. Northern red oak, injured slightly, and pin oak, undamaged, were not subject to much internal decay (Rhoades and Stipes 1999). Sycamore is often infected by anthracnose that can result in increased production of small shoots. The increased surface area may increase susceptibility (Hauer et al. 1993). Alaskan white cedars on campus were mostly medium-sized trees, 96% over 20 cm dbh and 10 m tall. Their size probably accounted for the fairly heavy damage to this species compared with no damage to the arborvitae in Hauer’s study, which were small trees, less than 7 m tall (Hauer et al. (1993). Furthermore, the arborvitae were clustered in a windbreak. Trees in clusters tend to provide mutual support. Hauer et al. (1993) based their ratings on percent of trees damaged and percent of severely damaged trees. Similar to our findings, they also reported that percent of ice damage varied by tree dbh, i.e., 1.3% of small trees (􀂔 30 cm dbh), 6.5% of medium trees (31–60 cm dbh), and 17.1% of large trees (􀂕 61 cm dbh) were severely damaged. Seischab et al. (1993) recorded crown damage to four species in western New York that were also on the Virginia Tech campus. Compared to those of the same species we studied on campus, northern red oak had greater damage 58 Northeastern Naturalist Vol. 14, No. 1 (30%), sugar maple had less damage (20%), and white oak had equal damage (18%) (Table 2). Sugar maple on campus had a large amount of internal decay. The trees in New York occurred primarily on steep slopes. Crown damage to American elm (10%) in Seischab et al.’s (1993) study was lower than the 27.8% on campus because those elms were small, understory trees. The 20% of trees and 20% of crown damage reported by Seischab et al. (1993) were averages. Greater damage occurred at forest edges (50%) and in stands with large oaks, maples, and Tsuga canadensis (L.) Carr. (hemlocks) that had unsound limbs and internal decay (30%). Seischab et al. (1993) concluded that bottomland forests had less damage (15%) because of the small branch diameter of Acer saccharinum L. (silver maple), elm, and Fraxinus americana L. (white ash). Some other studies have also noted the effect of topography on ice damage. In a study by Boerner et al. (1988) in Neotoma Valley in southcentral Ohio, most damage occurred on the valley floor and the lower south-facing slope, where 15% of trees with 13% of the total basal area were severely damaged. Severity of ice damage there was positively correlated with tree height, basal area, and crown area. Warrillow and Mou (1999) reported on the effect of landform on species susceptibility. In four species, damage varied by topographic position. Most damage was on steep slopes and east-facing slopes, whereas least damage occurred on lower slopes and valley bottoms. Two species of pines, Pinus virginiana Mill. (Virginia pine) and Pinus rigida Mill. (pitch pine), were severely damaged. The most resistant species was Nyssa sylvatica Marsh. (black gum ). Damage to Quercus velutina Lam. (black oak), Acer rubrum L. (red maple), Pinus strobus L. (white pine), and Quercus montana Willd. (chestnut oak) varied greatly in different topographic sites. Rhoades (1999) found a significantly greater amount of ice damage on the steep south-facing slope compared with the gentle north-facing slope. Canopy trees (􀂕 25 cm dbh) that received the greatest damage were Quercus coccinea Muenchh (scarlet oak), white pine, and white oak. Understory trees (dbh 􀂕 10 cm and < 25 cm) that received severe damage were red maple and Amelanchier arborea (Michaux f.) Fernald (shadbush). Elstner and Ware (2001) noted that Virginia pine and P. taeda L. (loblolly pine) on roadsides were severely damaged. Trees on roadsides are susceptible to ice damage because they are like trees at the edges of a forest. That factor accounted for the severe damage to three sycamores on the Virginia Tech campus. Although these trees were small, they were growing on a roadside. Most severely damaged canopy trees reported by Elstner and Ware (2001) were loblolly pine (50%), Fagus grandifolia Erhrhart (beech, 38%), Liriodendron tulipifera L. (tuliptree, 35%), white oak (29%), and Oxydendron arboreum (L.) DC (sourwood, 20%). Most authors agree that gymnosperms are more susceptible to ice damage than angiosperms. Warrillow and Mou (1999) observed that conifers 2007 R.W. Rhoades and R.J. Stipes 59 were more susceptible than hardwoods, a phenomenon observed by others (Boerner et al. 1988, Elstner and Ware 2001, Whitney and Johnson 1984), but not by Hauer et al. (1993). We found that Alaskan white cedars on the Virginia Tech campus were very severely damaged. In the comparisons among studies (Table 9), the most variation was in % crown damage. However, this method of characterizing ice damage is probably more reliable than either % individuals damaged, or % basal area of damaged trees (Table 2). Better still is the average % ice damage used in Tables 4–8. It is calculated as (% individuals + % basal area + % crown)/3 (see also Table 2). Average % damage yielded 12 significant comparisons among species. Percent crown damage, % basal area, and % individuals yielded 7, 5, and 4 significant comparisons among species, respectively. From these results, it is apparent that, by itself, % individuals damaged is not a good indicator of ice damage. Variation among studies in percent crown damage is to be expected as the result of a chance event such as an ice storm. Several studies have emphasized that large trees are more susceptible due to large, asymmetric crowns and internal decay, conditions that predispose them to severe ice damage. Also trees on steep slopes or growing at the edge of a forest are quite susceptible. Our study has confirmed the idea that all of the above, except the effect of steep slopes, influence ice damage. There are few steep slopes on the Virginia Tech campus. We conclude from our study that three factors were largely responsible for the relatively severe ice damage to trees on the Virginia Tech campus: 1) the high percentage of large trees with internal decay and asymmetric crowns; 2) the severity of the ice storms; and 3) the open, exposed siting of all the trees, similar to trees growing at the edges of a forest. Acknowledgments We thank Jean Ratliff, Laboratory Specialist Senior in Plant Pathology at Virginia Polytechnic Institute and State University, for typing the first drafts of the manuscript. Literature Cited Abell, C.A. 1934. Influence of glaze storms upon hardwood forests in the southern Appalachians. Journal of Forestry 32:35–37. Boerner, R.E.D., S.D. Runge, C. DoSoon, and J.G. Kooser. 1988. Localized icestorm damage in an Appalachian Plateau Watershed. American Midland Naturalist 119:199–208. Bruederle, L.P., and F.W. Stearns. 1985. Ice storm damage to a southern mesic forest. Bulletin of the Torrey Botanical Club 112:167–175. De Steven, D.J., J. Kline, and P.E. Mattiae. 1991. Long-term changes in a Wisconsin Fagus-Acer forest in relation to glaze-storm disturbance. Journal of Vegetation Science 2: 201–208. Elstner, P., and S.A. Ware. 2001. Ice storm damage to Virginia coastal plain forests during the Christmas 1998 storm. Virginia Journal of Science 52:3–11. 60 Northeastern Naturalist Vol. 14, No. 1 Harshberger, J.W. 1904. The relation of ice storms to trees. Contributions of the Botany Laboratory, The University of Pennsylvania 2:345–349. Hauer, R.J., W. Wang, and J.O. Dawson. 1993. Ice storm damage to urban trees. Journal of Arboriculture 19:187–194. Kartesz, J.T. 1999. A synonymized checklist and atlas with biological attributes for the vascular flora of the United States, Canada and Greenland. First Edition. In J.T. Kartesz and C.A. Meacham (Eds.). Synthesis of the North American Flora, Version 1.0. North Carolina Botanical Garden, Chapel Hill, NC. Lemon, P.C. 1961. Forest ecology of ice storms. Bulletin of the Torrey Botanical Club 88:21–29. National Oceanic and Atmospheric Administration (NOAA). 1994a. 1994 weather in the Southeast. National Climatic Data Center Technical Report 94-03. Washington, DC. 20 pp. National Oceanic and Atmospheric Administration (NOAA). 1994b. Daily weather maps, weekly series. Feb.28–Mar. 20, 1994. Washington, DC. Rhoades, R.W. 1999. Ice damage in a small valley in southwestern Virginia. Castanea 64:243–251. Rhoades, R.W., and R.J. Stipes. 1999. Growth of trees on the Virginia Tech campus in response to various factors. Journal of Arboriculture 25:211–217. Seischab, R.K., J.M. Bernard, and M.D. Eberle. 1993. Glaze storm damage to western New York forest communities. Bulletin of the Torrey Botanical Club 120:64–72. Siccama, T.G., G. Weir, and K. Wallace. 1976. Ice damage in a mixed hardwood forest in Connecticut in relation to Vitis infestation. Bulletin of the Torrey Botanical Club 103:180–183. Sokal, R.R., and F.J. Rohlf. 1982. Biometry, 2nd Edition. W.H. Freeman and Company, San Francisco, CA. 776 pp. Tattar, T.A. 1989. Diseases of Shade Trees. Academic Press, San Diego, CA. 391 pp. Warrillow, M.P., and P. Mou. 1999. Ice storm damage to forest tree species in the ridge and valley region of southwestern Virginia. Journal of the Torrey Botanical Society 126:47–158. Whitney, H.E., and W.C. Johnson. 1984. Ice storms and forest succession in southwestern Virginia. Bulletin of the Torrey Botanical Club 111:429–437.