WILDFIRE DAMAGE ASSESSMENT OF A YOUNG OAK FOREST IN PENNSYLVANIA

J. APPLIED FIRE SCIENCE, Vol. 23(1) 91-14, 213-214 WILDFIRE DAMAGE ASSESSMENT OF A YOUNG OAK FOREST IN PENNSYLVANIA MARC D. ABRAMS SARAH E. JOHNSON Penn State University, University Park ABSTRACT A large area of mature mixed-oak (Quercus spp.) forest was clearcut in the early 199s on a dry mountain plateau in south-central Pennsylvania. The majority of the developing forest was burned from an intense wildfire in 25. Our 29 vegetation survey revealed that the unburned areas contained a high density (3579 stems per acre) of sapling and pole-sized trees dominated by mixed-oaks, red maple (Acer rubrum), sweet birch (Betula lenta), and black cherry (Prunus serotina). The burned areas had 43% lower tree density, including fewer oaks and black cherry, but a higher proportion of lower value trees (e.g., black locust; Robinia pseudoacacia). In the burned areas, 91% of surviving oaks had basal fire scars averaging 39 in length, were overwhelmingly multiple-stemmed after being top-killed by fire, and shorter in height and smaller in diameter than the oaks in the unburned units. The burned units had higher cover of shrub and herbaceous species. The results of this study suggest that intense wildfire can significantly damage young oak forests, and that the negative direct and indirect effects of this will persist long into the future. INTRODUCTION It is now generally accepted that recurring fire played a major ecological role in the historical development of oak (Quercus) forests in the eastern United States 213, Baywood Publishing Co., Inc. doi: http://dx.doi.org/1.219/af.23.1.f http://baywood.com 91

92 / ABRAMS AND JOHNSON (Abrams, 1992; Lorimer, 1985; Nowacki & Abrams, 28). High fire frequency in oak forests dates back to the early Holocene and suggests that Native Americans actively used fire to manage these forests (Day, 1953; Delcourt & Delcourt, 1997; Gill, Williams, Jackson, Lininger, & Robinson, 29; Guyette, Muzika, & Dey, ). The use of fire in this manner created more open forests for oak regeneration and increased mast production for the Native American diet as well as a food source for the animals they hunted (Abrams & Nowacki, 28). Fire frequency remained high during the early and middle European settlement period of the eastern forest and peaked during the clearcut and catastrophic fire era of 188-193 (Abrams, 1992; Brose, Schuler, van Lear, & Berst, 21). Following the fire exclusion policy of the 193s, much of the eastern oak forest experienced significantly less fire which, over time, dramatically altered their composition and structure (Dey, Royo, Brose, Hutchinson, Spetich, & Stoleson, 21; Nowacki & Abrams, 28). Forest density has increased, oak regeneration and recruitment has often failed, and later successional species, especially red maple (Acer rubrum), have increased in importance (Abrams, 1998; Fei & Steiner, 29). Prescribed fire is now considered an important management tool in oak forests because of the many ecological benefits that it provides (Brose, Gottschalk, Horsley, Knopp, Kochenderfer, McGuinness, et al., 28; Lorimer, 1989). However, it also creates at least one major problem basal scarring and the high potential for rot developing in the butt logs of trees (Toole, 1959; Wagener & Davidson, 1954). The negative impact on butt log wood quality has been one of the major deterrents to using prescribed fire by timber managers, despite the fact that oaks are capable of compartmentalizing wounds from fire (Smith & Kennedy Sutherland, 1999). Fire is also known to cause significant mortality to smaller oak trees in mature stands (Huddle & Pallardy, 1996; Maslen, 1989; Ward & Stephens, 1989). However, there is little scientific information on the effects of fire at various stages in oak forest development, particularly in young forests (Hepting, 1935). In 25, a severe wildfire burned through a large, approximately 15-year-old, post-clearcut, mixed-oak forest in south-central Pennsylvania. This represented a unique opportunity to conduct a detailed wildfire damage assessment in the early stage of oak forest development. STUDY AREA The study area is a 986 acre tract on private land near the town of Fairhope in Somerset County, south-central Pennsylvania (39.84 N, 78.79 W; Figure 1). It is comprised of a steep, southwest facing slope adjacent to railroad tracks (at 9 feet elevation; not used in this study), leading to more gently sloping terrain to the upper plateau portion of Savage Mountain (at 25 feet elevation). The soils on the property are typical of ridges in southern Pennsylvania and are mainly the Hazelton, Leck, and Albrights series (Yaworski, 1983; Table 1). The Hazelton soil is classified as very stony or very bouldery sandy loam. The Leck

WILDFIRE DAMAGE ASSESSMENT / 93 Figure 1. The study area in Fairhope, Pennsylvania. The inner boundaries demarcate the north, west, central, and east 25 burned survey areas. The northwest, northeast, and east-central survey areas (exterior to the burned areas) are young oak forests that did not burn in 25. All seven burned and unburned units were clearcut during 199-91.

94 / ABRAMS AND JOHNSON Table 1. Physical Attributes of Unburned and Burned Sampling Units Near Fairhope, Somerset County, Pennsylvania Unit Average elevation (feet) Average slope (degrees) Topographic position Aspect Soil texture Unburned Northeast East Central Northwest 2316 277 2323 19 23 23 Upper slope Mid slope Upper slope E-SE SE SE Very stony sandy loam Very stony sandy loam Very stony silt loam Burned North Central East West 7 1989 1768 1918 32 37 35 38 Upper slope Mid slope Lower slope Mid to lower slope SE and SW E-SE SE S-SW-W Very stony silt and sandy loam Very stony silt and sandy loam Very stony sandy loam Very stony silt loam

WILDFIRE DAMAGE ASSESSMENT / 95 Figure 2. Four photographs taken in 29 of the burned (25) and unburned units in the study area at Fairhope, Pennsylvania. Top-left: high density oak trees in the northwest unburned (background) and brushy vegetation in the north burned unit (foreground). The Philadelphia rod is 12 feet tall. Top-right: large fire scar on an oak that survived the 25 fire. Bottom-left: multiple-stem chestnut oak that resprouted after being top-killed by fire with extensive deer browse. Bottom right: Brushy vegetation and low value trees developing in the east burned area. and Albrights soils are very stony silt loams. A mature mixed-oak forest on the study area was clearcut in the early 199s (mainly 199-91). The former composition was confirmed by aerial photographs taken in the 197s and a

96 / ABRAMS AND JOHNSON timber evaluation report from 1982. This evidence and eye witness accounts indicate that prior to the 25 fire, the burned and unburned areas were similar in composition and structure before and after the 199s logging. After harvesting, the site produced a healthy, dense forest of mostly sapling to pole-sized oak and mixed-hardwood species throughout the cut area. We consider the unburned forest to be in the stem exclusion stage of development (Figure 2; Oliver, 1981). A wildfire on October 1, 25 burned approximately 626 acres of the property and was thought to have been produced by a railroad ignition. A logging road prevented the fire from spreading to the unburned areas. A wildfire investigation report produced by the Pennsylvania Bureau of Forestry (number 4(5)2-3) reported a midflame windspeed of 8.1 mph, a Behave fuel model 9 (hardwood litter), a dead fuel moisture of 9%, a rate of spread of 16.6 chains/h, a heat per unit area of 335 Btu/ft 2, a fireline intensity of 12 Btu/ft/s, and a flame length of 3.8 ft. We examined photographs taken by the State of Pennsylvania Fire Marshall a few days after the fire that indicate the wildfire was quite intense (including extensive crown fire that top killed much of the vegetation) throughout most of the study areas. Therefore, we attribute the major vegetation differences between the burned and unburned units to the 25 fire rather than other physical or biological factors (see Methods). METHODS During September 3-5, 29, we conducted a vegetation survey within four burned clearcut areas and three unburned clearcut areas within the 199-91 harvested portion of the study area (Figure 1). In 29, the unburned study area was approximately 19 years old, whereas the burned study areas was 4 years old. The burned units ranged in size from 96-167 acres and the unburned units from to 25-6 acres. The burned units were located in the north, east, central, and west portions of the property (excluding the steep slope to the south leading to railroad tracks). The three unburned units were located exterior to the burned units along the northeast, east, and northwest boundaries of the property. The individual units were defined based on differences in topographic position, elevation and aspect (Table 1). Slope varied between the burned (mean 35% slope) and unburned (mean 21% slope), but not among the units in each category. Soil texture was similar (very stony sandy or silt loam) among all the units. The burned units had an average elevation of 1986, compared with 39 for the unburned units. In general, the units occurred on middle or upper slopes with a southeast aspect (with the west burned unit being an exception). Dividing of the burned and unburned areas into multiple units better enabled sampling in a large and somewhat disjunctive study area, and allowed for statistical comparisons between the burned and unburned treatments.

WILDFIRE DAMAGE ASSESSMENT / 97 Twenty fixed-area plots were used for vegetation sampling, located at random intervals along transects through the interior of each of the seven units. Saplings and pole-sized trees (combined) and seedlings were counted in nested circular plots of 1 ft 2 and 5 ft 2, respectively. Trees were classified as woody stems 4.5 feet in height, whereas seedlings were < 4.5 feet in height (first year germinants were not counted). Multiple stem trees and seedlings were noted along with the number of stems each unit contained, but each unit (regardless of the number of stems) was considered as a single individual for the density calculations. Shrub cover by species and total herbaceous cover were estimated into cover classes (-5%, 5-25%, 25-5%, 5-75%, 75-95%, and 95-1%) in the 5 ft 2 plots. In each plot, the total height and diameter at breast height (4.5 feet) were measured on the tallest two or three oaks with a Philadelphia (or leveling) rod and a diameter tape. The presence and length of fire scars on all oak saplings were recorded in the burned plots. T-tests (assuming unequal variance) were used to determine statistical differences between the burned and unburned units at p <.5. RESULTS AND DISCUSSION The three unburned units were dominated by saplings and poles of red oak (Q. rubra), white oak (Q. alba), chestnut oak (Q. montana), red maple, sweet birch (Betula lenta), witch-hazel (Hamamelis virginiana), and black cherry (Prunus serotina; Table 2). The east-central unit had the highest over-all tree density, with particularly high values for red oak, red maple, witch-hazel, and black cherry, but lacked sweet birch. In contrast, sweet birch led species density in the northeast and west units. The burned units were dominated by saplings of black locust (Robinia pseudoacacia), red maple, sweet birch, and chestnut oak (Table 3). The north and west units had higher stem density than the other two units. This was due to the high density of sweet birch in the west unit and the high density of red maple, chestnut oak, and black oak (Q. velutina) in the north unit. Oak sapling density was higher in the north unit and lower in the central and east units. The unburned units had a mean sapling density of 3579 per acre versus 247 per acre in the burned area (Tables 2 and 3). The unburned units contained significantly (p <.5) more oaks (1619 per acre) than the burned units (594 per acre). Most of this difference was due to increased red oak and white oak. The unburned units also had a high density of red maple, sweet birch, witch hazel, and black cherry trees. The burned units had a higher proportion of less desirable hardwoods (e.g., black locust, red maple, and sweet birch), despite the fact that their densities were generally lower than in the unburned units. The proportion of single-stemmed trees was 83.7% on the unburned units compared to only 32.8% on the burned units. We do not attribute the major vegetation and structural

98 / ABRAMS AND JOHNSON Table 2. Species Density (per Acre) and Mean (± Standard Error) of Sapling and Pole-Sized Trees in Three Unburned Units Surveyed in 29 a Young Oak Forest Near Fairhope, Pennsylvania Species Northeast East Central Northwest Mean Red oak 545 137 35 719 ± 32 White oak 348 566 436 45±63 Chestnut oak 392 414 24 349±55 Black oak 19 51±32 Scarlet oak 152 51±51 Red maple 261 98 51 581 ± 211 Sweet birch 741 828 523 ± 263 Witch-hazel 174 741 24 385 ± 179 Black cherry 675 218 35 ± 194 Black locust 87 36±26 Blackgum 29±19 American chestnut ±13 Striped maple ± Tulip poplar ± Pignut hickory 15±15 Sassafras 15±15 Bigtooth aspen 7±7 Total 294 4813 2984 3579 ± 617 differences to the variation in elevation and slope or simply the age difference between the 19 years unburned and 4-year-old burned units. The unburned units had a seedling density of 3688 per acre compared with 5663 per acre in the burned units (Table 4). We attribute this difference to the burned units being more open and to fire top-killing most of the preexisting sapling/pole trees, which transformed these individuals to seedling-sized sprouts. The latter also explains the scarcity of saplings in the burned areas. In addition, we believe the unburned units are in the stem exclusion stage of forest development, which normally has few seedlings (Oliver, 1981). The seedlings in the burned were primarily red maple, sweet birch, black locust, witch-hazel, and devil s walking stick (Aralia spinosa). The unburned units contained a high

WILDFIRE DAMAGE ASSESSMENT / 99 Table 3. Species Density (per Acre) and Mean (± Standard Error) Density of Sapling and Pole-Sized Trees Four Burned Units Surveyed in 29 After a 25 Wildfire in a Young Oak Forest Near Fairhope, Pennsylvania Species North Central East West Mean Chestnut oak 51 35 21 ± 123 Red oak 196 87 261 147±5 Black oak 218 174 131±39 White oak 152 87 131 93±34 Bear oak 16±16 Scarlet oak 5±5 Black locust 414 414 283 327 359±33 Red maple 937 87 35 338 ± 29 Sweet birch 87 167 316 ± 251 Devil s walking stick 152 414 147±95 Witch-hazel 196 218 19 131±5 Black cherry 87 19 82±1 Pignut hickory 19 ±27 Sassafras ±9 American chestnut 11±11 Striped maple 5±5 Total 2831 1416 1285 27 247 ± 45 proportion of red maple, sassafras (Sassafras albidum), and witch hazel seedlings. The dominant oak seedling in both the unburned and burned units was chestnut oak. The unburned units contained significantly more (1147 per acre) and a higher percentage (31%) of oak seedlings (all species) compared with the burned unit (3 per acre and 11.5% of the total). The number of pignut hickory (Carya glabra) seedlings was higher in the burned than unburned units. Mean shrub cover was slightly higher in the burned (14.2%) than the unburned units (9.5%). Both areas averaged 5-8% blueberry (Vaccinium spp.), but the burned units had 7.7% Rubus spp. not found in the unburned units. Mean total herbaceous cover on the burned units (5.8%) was significantly higher than that of the unburned units (6.8%). Table 5 summarizes the major differences between the burned and unburned units in relation to a fire damage assessment. The burned units contained 43%

1 / ABRAMS AND JOHNSON Table 4. The Mean (± Standard Error) Density of Tree Seedlings in Three Unburned and Four Burned Units Surveyed in 29 After a 25 Wildfire in a Young Oak Forest Near Fairhope, Pennsylvania. Mean Followed by Different Letters are Significantly Different at p <.5. An Asterisk () Indicates Mean Values Less than.5 Seedlings/acre Species Unburned Burned Red maple Sassafras Chestnut oak Red oak White oak Black oak Bear oak Witch-hazel Sweet birch Blackgum Striped maple American chestnut Black locust Mockernut hickory Pignut hickory Sugar maple Devil s walking stick Black cherry Bigtooth aspen 1147.1 ± 453.6 638.9 ± 574.8 493.7 ± 34.2 246.8 ± 11.6 217.8 ± 115.2 188.8 ± 167.5 392. ± 75.4 a 11.6 ± 14.5 a 11.6 ± 16.6 58.1 ± 58.1 43.6 ± 43.6 14.5 ± 14.5 14.5 ± 14.5 14.5 ± 14.5 14.5 ± 14.5 1361.3 ± 386.9 283.1 ± 168.2 348.5 ± 28.1 141.6 ± 48.3 43.6 ± 3.8 98. ± 98. 21.8 ± 21.8 773.2 ± 113. b 1241.5 ± 348.7 b 457.4 ± 217.4 239.6 ± 239.6 588.1 ± 421. 43.6 ± 17.8 21.8 ± 12.6 Total 3688.1 ± 932.3 5662.8 ± 218.5 fewer trees per acre and 63% fewer high value oak and cherry trees. The burned units had 29% more lower value (less desirable) trees, such as red maple, black locust, and sweet birch. Only 24% of the high value species in the burned units were single stemmed, compared with 73% in the unburned units. Fire scars were recorded in 91.4% of the oak trees and had an average length of 38.9 inches in the burned units, and none in the unburned unit (Figure 2). Fire scars of this size on

WILDFIRE DAMAGE ASSESSMENT / 11 Table 5. Components of Damage Estimates on the Burned Versus Unburned Sites as a Result of the 25 Fire in Fairhope, Pennsylvania Forest attribute Burned Unburned Burned difference Trees/acre 247 3579 43% Oak, hickory, cherry/acre 72 1938 63% % less desirable trees % 46% +29% % single stem oak, hickory, cherry 24% 73% 67% % fire scarred oaks 91.4% % +91.4% Fire scar length 38.9 +38.9 Mean height of tallest oaks 12.4 21.8 43% Oak diameter 2.5 2..6% Rotation loss 15+ years 15+ years trees typically represent a significant loss in future timber grade due to the likely development of stem rot in the highly valued butt log (Loomis, 1977). The mean height and diameter of the tallest oaks in the burned and unburned units was 12.4 versus 21.8 feet and 2.5 versus 2. inches, respectively. The burned units suffered a minimum rotation loss of 15 years (based on the number years between the 199 harvest and the 25 fire). However, we believe the negative impacts of fire in the burned units will exist far beyond 15 years. The 25 fire top-killed many of the oak stems, after which multiple stem sprouts formed (Table 3; Figure 2). This potentially represents another significant loss in future value as multiple stem trees tend to have poorer form (more lean and sweep), as well as smaller crowns and smaller diameters (Palmer, 24). The impact of this may be lessened somewhat in the future as multiple stems will self-thin over time. However, we frequently observed deer browse on the multiple stemmed sprouts (Figure 2), which we believe will further retard their development into quality timber (Moore & Johnson, 1967). In extreme cases, intense and persistent deer browsing causes the formation of tree seedling grubs with little or no timber potential. The Fairhope fire significantly reduced the number and quality of the high value oak and black cherry saplings and poles at the site, while increasing the proportion of less desirable tree species (e.g., black locust, red maple, and sweet birch), shrubs, and herbs. A 1932 wildfire in a mature mixed-oak forest in Connecticut initially caused 84% mortality to saplings and poles (Ward & Stephens, 1986). By 1987, however, the burned portion had higher density and

12 / ABRAMS AND JOHNSON basal area of oak and larger mean stem diameter than did the unburned portion. This suggests that the negative impact of fire may be temporary in our study area, but there are several important differences between the Ward and Stephens (1989) study area and ours and many negative factors impacting present-day oak forests that did not exist to same extent in 1932. The Fairhope fire occurred in a 19-year-old clearcut, not in the understory of a mature forest with residual oak trees. Hepting (1935) reported that all but the smallest fire scars will remain open during the life of the tree that results in subsequent radial and horizontal spread of decay from fungus and insect damage. Therefore, we believe that future oak development and economic value in the Fairhope burn will be greatly limited by a lack of seed source, intense deer browsing, wood decay, and competition from herbs, shrubs, and other tree species (Figure 2). In contrast to the burned areas, the composition, size, and quality of young oak and other trees in the unburned units were exceptional. In our opinion, vegetation in the burned units will not develop the composition, growth, and structural qualities possessed in the unburned units in the next 15 years or more. Indeed, a major problem in managing oak in eastern forests is the lack of oak seedlings and saplings before or after harvest (Dey et al., 21). Once this situation develops it is very difficult to overcome (Abrams, 1998). We believe in the important benefits of using prescribed fire for the ecology and management of eastern oak forests. Research suggests that fire is most appropriate as an understory treatment in maturing oak forests (e.g., after 5 years of age) to create advanced regeneration and/or to release seedlings following harvest (Brose et al., 28; Dey et al., 21). While the intensity of wildfires are often greater than that of prescribed fire (Brewer & Rogers, 26), based on the results of this study we believe that both types of fire should be avoided, whenever possible, in young oak forests in the sapling or pole-sized stages. REFERENCES Abrams, M. D. (1992). Fire and the development of oak forests. BioScience, 42, 346-353. Abrams, M. D. (1998). The red maple paradox. BioScience, 48, 355-364. Abrams, M. D., & Nowacki, G. J. (28). Native Americans as active and passive promoters of mast and fruit trees in the eastern USA. The Holocene, 18, 1123-1137. Brewer, S., & Rogers, C. (26). Relationships between prescribed burning and wildfire occurrence and intensity in pine-hardwood forests in north Mississippi, USA. International Journal of Wildland Fire, 15, 23-211. Brose, P., Schuler, T., Van Lear, D., & Berst, J. (21). Bringing fire back: The changing regimes of the Appalachian mixed-oak forest. Journal of Forestry, 99, 3-35. Brose, P. H., Gottschalk, K. W., Horsley, S. B., Knopp, P. D., Kochenderfer, J. N., McGuinness, B. J., et al. (28). Prescribing regeneration treatments for mixed-oak forests in the Mid-Atlantic region. U.S. Forest Service General Technical Report NRS-33, 84 pp.

WILDFIRE DAMAGE ASSESSMENT / 13 Day, G. M. (1953). The Indian as an ecological factor. Ecology, 34, 329-346. Delcourt, H. R., & Delcourt, P. A. (1997). Pre-Columbian Native American use of fire on southern Appalachian landscapes. Conservation Biology, 11, 11-114. Dey, D. C., Royo, A. A., Brose, P. H., Hutchinson, T. F., Spetich, M. A., & Stoleson, S. H. (21). An ecologically based approach to oak silviculture: A synthesis of 5 years of oak ecosystem research in North America. Revista Columbia Forestal, 13, 21-2. Fei, S., & Steiner, K. C. (29). Rapid capture of growing space by red maple. Canadian Journal of Forest Research, 39, 14-1452. Gill, J. L., Williams, J. W., Jackson, S. T., Lininger, K. B., & Robinson, G. S. (29). Pleistocene megafaunal collapse, novel plant communities, and enhanced fire regimes in North America. Science, 326, 11-113. Guyette, R. P., Muzika, R. M., & Dey, D. C. (). Dynamics of an anthropogenic fire regime. Ecosystems, 5, 472-486. Hepting, G. H. (1935). Decay following fire in young Mississippi delta hardwoods. USDA Technical Bulletin No. 494, 32 pp. Huddle, J. A., & Pallardy, S. G. (1996). Effects of long-term annual and prescribed burning on tree survival and growth in a Missouri Ozark oak-hickory forest. Forest Ecology and Management, 82, 1-9. Loomis, R. M. (1977). Wildlife effects on an oak-hickory forest in southeast Missouri. Research Note NC-219. USDA Forest Service, North Central Research Station, St. Paul, MN. 4 pp. Lorimer, C. G. (1985). The role of fire in the perpetuation of oak forests. In J. E. Johnson (Ed.), Challenges in oak management and utilization (pp. 8-25). Madison, WI: Cooperative Extension Service, University of Wisconsin. Lorimer, C. G. (1989). The oak regeneration problem: New evidence on causes and possible solutions. Forest Resource Analyses, No. 8. Madison, WI: University of Wisconsin, Department of Forestry. 31 pp. Maslen, P. (1989). Response of immature oaks to prescribed fire in North Carolina Piedmont. USDA Forest Service General Technical Report, SO-74 (pp. 259-266). New Orleans, LA: USDA Forest Service, Southern Forest Experiment Station. Moore, W. H., & Johnson, F. M. (1967). Nature of deer browsing on hardwood seedlings and sprouts. Journal of Wildlife Management, 31, 351-353. Nowacki, G. J., & Abrams, M. D. (28). Demise of fire and mesophication of eastern U.S. forests. BioScience, 58, 123-138. Oliver, C. D. (1981). Forest development in North America following major disturbances. Forest Ecology and Management, 3, 153-168. Palmer, B. (24). Timber stand improvement. Missouri Department of Conservation Extension Bulletin F35, 7 pp. Smith, K. T., & Kennedy Sutherland, E. (1999). Fire-scar formation and compartmentalization in oak. Canadian Journal of Forest Research, 29, 166-171. Toole, E. R. (1959). Decay after fire injury to southern bottomland hardwood. USDA Forest Service Technical Bulletin 1189, 25 pp. Wagener, W. W., & Davidson, R. W. (1954). Heart rot in living trees. Botanical Review, 2, 61-134. Ward, J. S., & Stephens, G. R. (1989). Long-term effects of a 1932 surface fire on stand structure in a Connecticut mixed hardwood forest. In G. Rink & C. A. Budelsky

14 / ABRAMS AND JOHNSON (Eds.), Proceedings of the 7th Central Hardwoods Forest Conference, U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. General Technical Report NE-274, pp. 267-273. Yaworski, M. (1983). Soil survey of Somerset County, Pennsylvania. USDA Soil Conservation Service, 148 pp. Direct reprint requests to: Marc D. Abrams 37 Forest Resources Building School of Forest Resources Penn State University University Park, PA 1682 e-mail: agl@psu.edu