Southeastern Naturalist
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
396
2014 SOUTHEASTERN NATURALIST 13(2):396–406
Spider (O: Araneae) Responses to Fire and Fire Surrogate Fuel
Reduction in a Piedmont Forest in Upstate South Carolina
Michael E. Vickers1,* and Joseph D. Culin2
Abstract - The two forest-management practices of prescribed burning and thinning are
techniques used to reduce heavy fuel loads that have resulted from years of fire suppression.
Therefore, the National Fire and Fire Surrogate (NFFS) study was conducted to determine
the effects of prescribed burning and thinning on different environmental factors. This study
was conducted in the Clemson University Experimental Forest in South Carolina, one of
the 13 NFFS sites in the United States, and examined the impacts of these management
practices on spider populations. We used pitfall traps to sample ground-dwelling spiders
to determine if changes in population levels had occurred one year following implementation
of these practices. We collected a total of 1220 specimens of Araneae, representing 13
families. Results indicated that by 1-year post-treatment, spider populations had recovered
following the initial (2001) burning and thinning. However, in 2002, when we compared
the first post-burn samples to pre-burn samples in thin+burn plots, we found a significant
decrease in the mean abundance of Agelenidae and Linyphiidae after the prescribed burn.
Introduction
During the 20th century, fire-exclusion policies were put into practice by federal
land managers, resulting in the accumulation of an extraordinary amount of fuel
(i.e., dead trees, logs, leaves, and litter; Agee and Lolley 2006, Lasko 2010, Schwilk
et al. 2009, Stephens 1998). High levels of fuel accumulation results in fires that
burn hotter, move slower, and have more profound ecological effects than fires in
areas with lower fuel accumulation (McCullough et al. 1998). Such high-fuel fires
create unnaturally severe wildfires and lead to an overall deterioration of forest
ecosystem integrity (Stephens 1998).
Management strategies have been developed to reduce fuel loads and are practiced
across the United States. The two primary practices are prescribed burning
and thinning. A prescribed burn has characteristics similar to naturally occurring
fires but is controlled by forest managers so it burns material mainly at ground
level and moves slowly through a forest. Thinning is used to remove smaller trees
that have the potential to spread fire into forest canopies (Converse et al. 2006).
However, it is not clearly understood if controlled burning, thinning, or the combination
of thinning and burning harm (decrease shelter and food resources, cause
heightened disturbance to the area) or benefit (increase niche diversity and habitat
complexity) forest ecosystems.
1Department of Biology, Millikin University, 1185 West Main Street, Decatur, IL 62521.
2School of Agricultural, Forest and Environmental Sciences, Clemson University, 101 Barre
Hall, Clemson, SC 29634. *Corresponding author - mvickers-alum@millikin.edu.
Manuscript Editor: Vincent A. Cobb
Southeastern Naturalist
397
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
Therefore, the National Fire and Fire Surrogate (NFFS) study was established to
quantify consequences that fuel-load–reduction treatments have both on economics
and on ecological factors such as, vegetation, wildlife, arthropods, and pathology
(Youngblood et al. 2005).
Our study assessed the effects of forest-management strategies on the spider
community. The few studies conducted in the United States that have examined
the effects on spiders of prescribed burning, thinning, or thinning combined with
burning have reported highly variable results ranging from spider populations
being: unaffected (Haskins and Shaddy 1986), decreasing dramatically (Coyle
1981, Merrott 1976, Reichert and Reeder 1972, Willett 2001), or increasing (Vogl
1993). The goal of our study was to determine if the fuel-load–management practices
of prescribed burning and thinning has an effect on spider populations in a
southern piedmont forest.
Methods
During 2002, we used pitfall traps to sample spider populations (Araneae) after
three forest-management treatments: prescribed burning (burn-only), thinning
(thin-only), and thinning followed by prescribed burning (thin+burn). Control areas
were established as a fourth treatment to allow comparisons with un-impacted
areas. Replicated plots of the 4 treatments were located in the Clemson Experimental
Forest in Anderson, Oconee, and Pickens counties, SC (Fig. 1). Burning of
burn-only plots was conducted in the spring 2001. Thinning in the thin-only plots
was conducted in the winter of 2001. Thinning and burning in the thin+burn
plots were conducted in the winter of 2001 and spring 2002, respectively.
Initial design setup
Because the Clemson Experimental Forest was one of 13 sites participating in the
National Fire and Fire Surrogate (NFFS) study, the initial plot design and establishment
were completed by USDA Forest Service researchers prior to our sampling.
In that design, the burn-only, thin+burn, and control treatments were replicated 3
times, and thin-only treatments were replicated 5 times. Thin-only treatments were
replicated 5 times due to the herbicide for two established thin+herbicide treatments
not being applied prior to our sampling; therefore, those two treatments were
designated as thin-only in this study. Within each treatment area, 40 permanent grid
points were established with 50-m spacing between each grid point in one of the
cardinal directions. At randomly selected grid points, four 20-m x 50-m plots were
established. Each 20-m x 50-m plot was divided in half, and a coin was tossed to
determine which 10-m x 50-m plot was used during each of the 6 sampling periods.
The 10-m x 50-m plot was further divided into five 10-m x 10-m subplots, and a
Figure 1 (following page). Replicated treatments, shaded areas within highlight boxes, located
in the Clemson Experimental Forest in Anderson, Oconee, and Pickens counties, SC.
Inset: Expanded view of one treatment area with replicated plots showing the 10-m x 50-m
area containing the five subplots that were sampled.
Southeastern Naturalist
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
398
Southeastern Naturalist
399
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
pitfall trap was placed in the center of each subplot (Fig. 1). A total of 280 pitfall
traps were deployed on each sampling date.
Sampling methods
We conducted pitfall-trap sampling every 2 months for 1 year beginning in
January 2002. Pitfall traps were active for approximately 48 hrs during each
sampling period. Traps were constructed using a 473-ml (16-oz) plastic cup
placed into the ground with a 266-ml (9-oz) cup, placed inside. The 473-ml cups
remained in the ground during the entire study, while the 266-ml cups were only
used during the trapping periods. To kill and preserve organisms that fell into the
traps, we placed approximately 80 ml of 70% EtOH into the 266-ml cup. After we
brought samples to the laboratory, we rinsed and stored them in 70% EtOH. Using
published taxonomic keys (Dondale and Redner 1990, Kaston 1972, Roth 1993),
we then sorted spiders to family and placed them into separate containers labeled
with date, treatment, and plot. We deposited voucher specimens in the Clemson
University Arthropod Collection (CUAC), Clemson, SC.
Statistical analysis
To determine spider family diversity and richness in the 4 treatments, we
plotted the proportional abundance against rank in the spider community for the
4 treatments. We used the plotted rank-abundance curves to determine which
treatments supported more-diverse spider communities, so as to focus on the
most-abundant families.
To determine if spider populations were affected by the management practices, we
used a general linear model (P = 0.05) to analyze numbers of individuals per family
collected at the plot level (SAS 1999) to obtain least square means. We made comparisons
1) between individual treatments (burn-only, thin-only, and thin+burn) and the
control, and 2) among treatments (burn-only, thin-only, and thin+burn). Also, to exclude
seasonality effects on spider families, we analyzed sampling dates versus
treatment interactions. Sampling dates were January/February, March/April, June,
August, October, and December. All statistical analyses were performed using SAS
(SAS 1999). Also, we compared pre-burn samples (January/February) to post-burn
samples (March/April, June, August, October, and December) in thin+burn plots to
determine if burning had an immediate effect on spider populations in these plots.
Results
Plot rank-abundance
We collected a total of 1220 specimens of Araneae, representing 13 families
(Table 1). After comparing each family’s proportional abundance versus rank, we
considered 7 families to be abundant enough for analyses (Fig. 2).
Effects of treatment on mean abundance of spider populations
Four of the 7 families identified—Agelenidae, Gnaphosidae, Pisauridae, and
Thomisidae—were significantly affected by one or more treatment. The other 3
Southeastern Naturalist
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
400
families—Clubionidae, Linyphiidae, and Lycosidae—were not significantly affected
by any treatment (Table 2).
Effect of sampling period versus treatment on mean abundance of spider
populations
The families Agelenidae and Linyphiidae exhibited significant effects related to
treatment over the 6 sampling periods, while the other 5 families analyzed did not
Table 1. Total number of individuals collected for each of the 13 spider families during the six sampling
periods in the Clemson Experimental Forest.
Family # collected Control Burn-only Thin-only Thin+burn
Agelenidae 144 59 33 23 29
Araneidae 5 1 1 1 2
Atypidae 3 1 1 0 1
Clubionidae 44 17 9 14 4
Gnaphosidae 212 61 59 68 24
Hahniidae 34 6 3 22 3
Linyphiidae 283 66 46 108 63
Lycosidae 321 33 64 166 58
Oxyopidae 16 2 2 10 2
Pisauridae 73 22 4 34 13
Salticidae 33 6 11 8 8
Theridiidae 6 1 1 3 1
Thomisidae 46 20 10 14 2
Figure 2. Rank-plot abundance of spiders collected during the six sampling periods in the
four treatment plots. A total of 13 families were collected. Families are ranked from highest
to lowest in proportional abundance.
Southeastern Naturalist
401
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
(Table 3). In addition, there was a significant difference in pre- versus post-burn
mean numbers of agelenids (Table 3, Fig. 3) and linyphiids (Table 3, Fig. 4) following
the March 2002 burn in the thin+burn treatment.
Table 2. Total number of individuals collected for each of the 7 abundant spider families during the 6
sampling periods in the Clemson Experimental Forest and LSmean totals per treatment. Superscripted
letters indicate families that exhibited a significant difference (general linear model: df = 3, P = 0.05):
Abetween control plots and either burn-only, thin-only, or thin+burn plots; and Bbetween burn-only compared
to thin-only or thin+burn plots.
LSmean
Family # collected Control Burn-only Thin-only Thin+burn
Lycosidae 321 0.5322 0.9411 1.4310 0.8055
Linyphiidae 283 1.0645 0.6764 0.9310 0.8750
Gnaphosidae 212 0.9838 0.8676 0.5862A 0.3333A,B
Agelenidae 144 0.8709 0.6176 0.1637A,B 0.4166A
Pisauridae 73 0.3548 0.0588A 0.2931B 0.1805
Thomisidae 46 0.3225 0.1470A 0.1206A 0.0273A
Clubionidae 44 0.2741 0.1323 0.1206 0.0555
Figure 3. LSmean number of Agelenidae (O: Araneae) collected during each of the six
sampling periods in the four treatment plots. The burn indicates when the burn occurred in
thin+burn plots in 2002.
Southeastern Naturalist
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
402
Discussion
Burn-only plots
Plots burned in April 2001 had significantly lower mean numbers of Pisauridae
and Thomisidae compared to control plots, suggesting that populations of
Table 3. Total number of individuals collected for each of the 7 abundant spider families during the
six sampling periods in the Clemson Experimental Forest and LSmean totals per sampling period by
treatment. Superscripted letters indicate families that exhibited a significant difference (general linear
model: df = 3, P = 0.05): Abetween control plots and either burn-only, thin-only, or thin+burn plots;
Bbetween burn-only compared to thin-only or thin+burn plots; and Cbetween thin-only compared to
burn-only and thin+burn plots.
LSmean
Family # collected Jan/Feb Mar/Apr June Aug Oct Dec
Lycosidae 321
Control 0.0909 0.2857 1.3636 0.8181 0.4545 0.0909
Burn-only 0.5000 0.8333 0.8333 1.8333 1.1666 0.2500
Thin-only 1.3888 0.6842 1.2500 1.7368 3.3000 0.2000
Thin+burn 1.4166 0.4166 1.0000 0.9166 0.8333 0.2500
Linyphiidae 283
Control 3.2727 3.7142 0.2727 0.0909 0.0000 0.0000
Burn-only 2.0000A, B 0.7500A 0.9166 0.1666 0.0000 0.0000
Thin-only 4.8888A 0.5263A 0.2500 0.1578 0.0000 0.1000
Thin+burn 4.1666A, B 0.6666A 0.0833 0.2500 0.0000 0.0833
Gnaphosidae 212
Control 0.5454 0.8571 2.1818 0.5454 1.7272 0.0000
Burn-only 0.0833 1.5000 1.7500 0.7500 0.6666 0.2500
Thin-only 0.2777 0.2631 1.6500 0.3684 0.9000 0.0000
Thin+burn 0.0000 0.5000 0.8333 0.2500 0.4166 0.0000
Agelenidae 144
Control 3.5454 0.0000 0.4545 0.5454 0.0909 0.2727
Burn-only 1.0833A 1.5000A 0.3333 0.2500 0.1666 0.2500
Thin-only 0.4444A, C 0.3684B 0.0500 0.0526 0.0000 0.1000
Thin+burn 1.7500A 0.2500B 0.3333 0.0000 0.0000 0.1666
Pisauridae 73
Control 0.0000 0.2857 0.3636 0.7272 0.6363 0.0909
Burn-only 0.0833 0.0000 0.0833 0.1666 0.0000 0.0000
Thin-only 0.0555 0.1052 0.4500 0.7894 0.3500 0.0000
Thin+burn 0.0000 0.0000 0.5833 0.0833 0.3333 0.0833
Thomisidae 46
Control 0.1818 0.0000 1.2727 0.1818 0.0909 0.0909
Burn-only 0.1666 0.0833 0.5000 0.0833 0.0000 0.0000
Thin-only 0.1111 0.2105 0.3000 0.0526 0.0000 0.0500
Thin+burn 0.0000 0.0000 0.0000 0.0769 0.0833 0.0000
Clubionidae 44
Control 0.0000 0.1428 0.1818 0.9090 0.3636 0.0000
Burn-only 0.0000 0.0000 0.0000 0.2500 0.5000 0.0000
Thin-only 0.0000 0.1052 0.0000 0.4736 0.1500 0.0000
Thin+burn 0.0000 0.1666 0.0000 0.0833 0.0833 0.0000
Southeastern Naturalist
403
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
these two families had been reduced by burning and had remained impacted for
at least 1 year (Table 2). Both pisaurids and thomisids are wandering spiders that
actively hunt prey. Although thomisids are generally considered to be sit-and-wait
predators, Uetz (1975) has observed them running and pouncing on prey. The decrease
in numbers of these 2 families could be due to either direct mortality, prey
reduction, or reduction in habitat complexity (Phillips et al. 2004, Reichert and
Reeder 1972). The results for the five families not having significantly different
mean numbers in burn-only versus control plots (Table 2) may be due to those spiders’
ability to escape the fire by burrowing into the ground, hiding in protected
locations, or rapidly recolonizing impacted habitats (Merrott 1976, Reichert and
Reeder 1972, Vogl 1993).
Thin-only plots
The families Agelenidae, Gnaphosidae, Linyphiidae, and Thomisidae were
negatively affected by thinning (Table 2). Agelenids and linyphiids are commonly
found on vegetation, and a reduction in understory habitat may have caused this
reduction. Correspondingly, Coyle (1981) has suggested that a decrease in number
of ground-level web-building spiders was a result of decreased forest canopy and
Figure 4. LSmean number of Linyphiidae (O: Araneae) collected during each of the six
sampling periods in the four treatment plots. The burn indicates when the burn occurred in
thin+burn plots in 2002.
Southeastern Naturalist
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
404
its corresponding effects on ground-level microclimate and reduction of litter. Both
gnaphosids and thomisids are wandering spiders, and a reduction in litter depth,
habitat complexity, and prey abundance could result in higher mortality (Uetz
1979), or in a high number of spiders being captured in pitfall traps due decreased
vegetation (Phillips et al. 2004).
Thin+burn plots
The families Agelenidae, Gnaphosidae, and Thomisidae were negatively affected
by the combination of prescribed burning and thinning (Table 2). For these
three families, the drastic change in environment or potential changes in prey density
could have caused increased mortality (Coyle 1981, Huhta 1971, Phillips et al.
2004, Uetz 1979).
Agelenidae and Linyphiidae were the only families to show a significant reduction
immediately following prescribed burning in thin+burn plots when pre-burn
and post-burn mean numbers were compared (Fig. 4). These results suggests that
agelenids and linyphiids were negatively impacted due to fire mo rtality.
Burn-only versus thin-only plots
Mean population numbers of Agelenidae were significantly lower in thinonly
compared to burn-only plots (Table 2), suggesting that thinning had a greater
negative impact than burning on this family. Agelenids in thin-only plots may have
moved to other areas to find more suitable environments.
Pisaurids had significantly higher numbers in thin-only plots than in burnonly
plots (Table 2), suggesting that this family was able to tolerate the impacts
of thinning. Pisaurids are commonly found on ground surfaces where they actively
search for prey and potential mates. High pisaurid collection numbers in
thin-only plots may have resulted from increased movement of spiders of those
species within the pitfall-trapping areas due to habitat disturbance and an increase
in habitat complexity. Habitat complexity potentially was increased because of
an increase in the amount of woody debris throughout thinned areas compared to
burned areas (Waldrop et al. 2004).
Sampling period versus treatment
When we analyzed treatments over the 6 sampling periods, we found significant
seasonal changes in mean population numbers for the families Agelenidae
and Linyphiidae (Table 3). When the three treatments burn-only, thin-only, and
thin+burn were compared to control plots, we found significant differences in
mean population numbers during specific sampling periods. For instance, agelenids
and linyphiids had higher numbers during January/February, March/April,
June, and August. Additionally, it is important to note that seasonal fluctuations
in mean spider populations could have affected number of spiders collected in
the 4 treatments during sampling (Coyle 1981, Greenberg and Forrest 2003, Uetz
1979). Multiple-year sampling would have provided a better estimate of familylevel
phenology.
Southeastern Naturalist
405
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
Conclusion
By one year following fuel-load–reduction treatments in the Clemson Experimental
Forest, individual spider families exhibited different patterns of impact.
Spider families were either not affected, or had recovered quickly from the various
fuel-load–reduction treatments. There was an immediate impact following burning
of the thin+burn plots on the families Agelenidae and Linyphiidae, which had a
significant decrease following the controlled burn. However, by the second postburn
(Agelenidae) and third post-burn (Linyphiidae) samples, mean agelenid and
linyphiid numbers were not significantly different from those in the control plots.
This information corroborates other studies that have examined the impacts of
the management practices of prescribed burning and thinning (Haskins and Shaddy
1986, Merrott 1976, New and Hanula 1998). It is important to note, our study was
only conducted over 1 year, and it is yet to be determined what the long-term effects
of prescribed burning and thinning are on spider populations.
Acknowledgments
This research was funded in part by the US Joint Fire Science Program, USDA Forest
Service (Agreement # SRS 01-CA-11330136-490). In addition, we would like to thank the
College of Agriculture, Forestry and Life Sciences, the Clemson Experiment Station, and
the Clemson Experimental Forest.
Literature Cited
Agee, J.K., and M.R. Lolley. 2006. Thinning and prescribed-fire effects on fuels and potential
fire behavior in an eastern Cascades forest, Washington, USA. Fire Ecology
2(2):142–158.
Converse, S.J., G.C. White, K.L. Farris, and S. Zack. 2006. Small mammals and forest-fuel
reduction: National-scale responses to fire and fire surrogates. Ecological Applications
16(5):1717–1729.
Coyle, F.A. 1981. Effects of clearcutting on the spider community of a southern Appalachian
forest. Journal of Arachnology 9:285–298.
Dondale, C.D., and J.H. Redner. 1990. The insects and arachnids of Canada: The wolf spiders,
nurseryweb spiders, and lynx spiders of Canada and Alaska. Part 17. Pp.15–121.
Research Branch, Agriculture Canada, Ottawa, ON, Canada.
Greenberg, C.H., and T.G. Forrest. 2003. Seasonal abundance of ground-occurring macroarthropods
in forest and canopy gaps in the southeastern Appalachians. Southeastern
Naturalist 2(4):591–608.
Haskins, M.F., and J.H. Shaddy. 1986. The ecological effects of burning, mowing, and
plowing on ground-inhabiting spiders (Araneae) in an old-field ecosystem. Journal of
Arachnology 14:1–13.
Huhta, V. 1971. Succession in the spider communities of the forest floor after clearcutting
and prescribed burning. Annales Zoologici Fennici 8:483–542.
Kaston, B.J. 1972. How to Know the Spiders. 2nd Edition.Wm. C. Brown Company Publishers,
Dubuque, IA. 290 pp.
Lasko, R. 2010. Implementing Federal wildland fire policy: Responding to change. Fire
Management Today 70:5–7.
Southeastern Naturalist
M.E. Vickers and J.D. Culin
2014 Vol. 13, No. 2
406
McCullough, D.G., R.A. Werner, and D. Neumann. 1998. Fire and insects in northern and
boreal forest ecosystems of North America. Annual Review of Entomology 43:107–127.
Merrott, P. 1976. Changes in the ground-living spider fauna after heathland fires in Dorset.
Bulletin of the British Arachnological Society 3:214–221.
New, K.C., and J.L. Hanula. 1998. Effect of time elapsed after prescribed burning in Longleaf
Pine stands on potential prey of the Red-cockaded Woodpecker. Journal of Applied
Forestry 22:175–183.
Phillips, R.J., T.A. Waldrop, G.L. Chapman, H.H. Mohr, M.A. Callaham, and C.T. Flint,
Jr. 2004. Effects of fuel-reduction techniques on vegetative composition of Piedmont
Loblolly-Shortleaf Pine communities: Preliminary results of the National Fire and
Fire Surrogate Study. Proceedings of the 12th Biennial Southern Silvicultural Research
Conference. General Technical Report SRS-71. US Department of Agriculture, Forest
Service, Southern Research Station, Asheville, NC. 594 pp.
Reichert, S.E., and W.G. Reeder. 1972. Effects of fire on spider distribution in southwestern
Wisconsin prairies. Proceedings of the Second Midwest Prairie Conference,
1970:75–90.
Roth, V.D. 1993. Spider Genera of North America. 3rd Edition. Sold on behalf of the American
Arachnological Society, Portal, AZ. 203 pp.
SAS Institute, Inc. 1999. SAS Version 8. Cary, NC.
Schwilk, D.W., J.E. Keeley, E.E. Knapp, J. McIver, J.D. Bailey, C.J. Fettig, C.E. Fiedler,
R.J. Harrod, J.J. Moghaddas, K.W. Outcalt, C.N. Skinner, S.L. Stephens, T.A. Waldrop,
D.A. Yaussy, and A. Youngblood. 2009. The national Fire and Fire Surrogate study:
Effects of fuel reduction methods on forest vegetation structure and fuels. Ecological
Society of America 19(2):285–304.
Stephens, S.L. 1998. Evaluation of the effects of silvicultural and fuels treatments on potential
fire behavior in Sierra Nevada mixed-conifer forests. Forest Ecology and Management
105:21–35.
Uetz, G.W. 1975. Temporal and spatial variation in species diversity of wandering spiders
(Araneae) in deciduous forest litter. Environmental Entomology 4:719–724.
Uetz, G.W. 1979. The influence of variation in litter habitats on spider communities. Oecologia
(Berl.) 40:29–42.
Vogl, R.J. 1973. Effects of fire on the plants and animals of a Florida wetland. American
Midland Naturalist 89:334–347.
Waldrop, T.A., D.W. Glass, S. Rideout, V.B. Shelburne, H.H. Mohr, and R.J. Phillips. 2004.
An evaluation of fuel-reduction treatments across a landscape gradient in Piedmont forests:
Preliminary results of the National Fire and Fire Surrogate study. Proceedings of
the 12th Biennial Southern Silvicultural Research Conference. General Technical Report
SRS-71. US Department of Agriculture, Forest Service, Southern Research Station,
Asheville, NC. 594 pp.
Willett, T.R. 2001. Spiders and other arthropods as indicators in old growth versus logged
Coastal Redwood forests. Restoration Ecology 9:410–420.
Youngblood, A., K.L. Metlen, E.E. Knapp, K.W. Outcalt, S.L. Stephens, T.A. Waldrop, and
D. Yaussy. 2005. Implementation of the Fire and Fire Surrogate study: A national research
effort to evaluate the consequences of fuel reduction treatments. Pp. 315–321, In
C.E. Peterson and D.A. Maquire (Eds.). Balancing Ecosystem Values. General Technical
Report PNW-GTR-635. USDA Forest Service, Pacific Northwest Research Station,
Portland, OR.