Differential Effects of Understory and Overstory Gaps on Tree Regeneration1

By Beckage, Brian Kloeppel, Brian D; Yeakley, J Alan; Taylor, Sharon F; Coleman, David C

BECKAGE, B. (Department of Plant Biology, University of Vermont, Marsh Life Science Building, Burlington, VT, USA 05405), B. D. KLOEPPEL (Odum School of Ecology, University of Georgia, Coweeta Hydrologic Laboratory, 3160 Coweeta Lab Road, Otto, North Carolina USA 28763), J. A. YEAKLEY (Environmental Science and Management, Portland State University, Portland, Oregon, USA 97207), S. F. TAYLOR (Odum School of Ecology, University of Georgia, Coweeta Hydrologic Laboratory, 3160 Coweeta Lab Road, Otto, North Carolina USA 28763), AND D. C. COLEMAN (Odum School of Ecology, University of Georgia, Athens, Georgia, USA 30602). Differential effects of understory and overstory gaps on tree regeneration. J. Torrey Bot. Soc. 135: 1-11. 2008.-Gaps in the forest canopy can increase the diversity of tree regeneration. Understory shrubs also compete with tree seedlings for limited resources and may depress tree recruitment. We compared effects of shrub removal and canopy windthrow gaps on seedling recruitment and understory resource levels. Shrub removal, with the canopy left intact, was associated with increased levels of understory light and soil moisture and coincided with increased species richness and diversity of tree regeneration compared to both control plots and canopy gaps. Canopy windthrow gaps, however, resulted in a more than 500 fold increase in soil nitrate concentrations, and seedling growth rates that were twice as high as that observed with shrub removal. Our results suggest that gaps in the understory shrub layer and the overstory canopy may have complementary effects on resource availability with corresponding benefits to seedling establishment and growth. Key words: canopy gaps, Rhododendron, shrubs, species richness, tree seedling, understory gaps.

Low light levels beneath the forest canopy can limit tree regeneration to species that are tolerant of deep shade (e.g., Canham 1988, Pacala et al. 1994). The diversity of tree species observed in the forest overstory, which includes shade intolerant species, may result from elevated levels of light and other resources that are found beneath gaps in the forest canopy (e.g., Pickett and White 1985, Whitmore 1989). Canopy gaps result from disturbances such as windstorms, drought, insect outbreaks, or pathogens that remove overstory trees, creating openings in the forest canopy. Canopy gaps can provide recruitment opportunities for tree seedlings, increasing the diversity of tree regeneration, and have figured prominently in empirical and theoretical investigations of mechanisms that promote forest diversity (e.g., Shugart 1984, Platt and Strong 1989, Busing and White 1997). The empirical evidence supporting the role of gaps in forest regeneration, however, has been equivocal with studies both supporting (Barden 1980, Runkle 1981, Kneeshaw and Bergeron 1998, Lusk and Smith 1998) and questioning the importance of canopy gaps in promoting the diversity of the forest overstory (Brewer and Merritt 1978, Hibbs 1982, Cho and Boerner 1991, Hubbell et al. 1999, Webb and Scanga 2001). Ambiguous evidence in support of the role of gaps in forest regeneration may result from differences in the density of understory vegetation as well as from variability in gap characteristics such as their size and mode of creation (Putz et al. 1983, Canham et al. 1990, Beckage et al. 2000).

Understory vegetation can mediate tree regeneration in forests through their influence on resource levels at the forest floor (Connell et al. 1997). Dense forest understories compete with tree seedlings for resources, limiting tree recruitment (Lorimer et al. 1994, Baker and Van Lear 1998, George and Bazzaz 1999, Beckage et al. 2000). Understory shrubs also usurp resources made available by canopy gaps, reducing tree regeneration (Phillips and Murdy 1985, Nakashizuka 1989, Clinton et al. 1994, Beckage et al. 2000). Understory gaps may be functionally similar to canopy gaps, elevating resource levels at the forest floor and increasing tree regeneration (Connell et al. 1997), but we are aware of few studies that compare canopy and understory gap effects on both abiotic conditions and tree regeneration (e.g., Pecot et al. 2007). We used the serendipitous occurrence of windthrown canopy trees within an existing shrub removal experiment to compare the effects of understory and canopy gaps on the species richness and diversity of tree regeneration in a temperate deciduous forest. The windthrow occurred within a control plot adjacent to a treatment plot where the understory shrub Rhododendron maximum L. had just been removed. Soil moisture and nutrient levels had been monitored for the prior year in both plots providing a unique opportunity to relate seedling recruitment to environmental responses. We subsequently established two additional control plots and monitored seedling recruitment over four years while continuing to monitor nutrients and soil moisture and also measuring light levels. While the unplanned nature of this experiment limited us to an unreplicated study, our results are suggestive of differential responses to these two modes of disturbance (overstory windthrow vs. shrub removal) and are reported in order to be subject to further exploration and testing.

Materials and Methods. STUDY AREA. Our study was conducted at Coweeta Hydrologie Laboratory, near Otto, North Carolina in the southern Appalachian Mountains (35[degrees]03′ N, 83[degrees]25′ W). Elevations in the Coweeta Basin range from 675 m to 1592 m and encompass a drainage area of 1626 ha. Mean annual precipitation is 1770 mm at the Coweeta base climate station. Vegetation in the lower elevations at Coweeta consists of secondgrowth mixed oak-hickory (Quercus-Carya) forest often with an understory of the evergreen ericaceous shrub Rhododendron maximum (Swank and Crossley 1988). Rhododendron can form a dense subcanopy (Baker and Van Lear 1998) with leaf area indices in the range of 4.8 to 6.6 and can have strong effects on the understory light environment (Beckage et al. 2000). Low light levels beneath Rhododendron subcanopies can have a dramatic impact on seedling regeneration, precluding nearly all seedling establishment (Beckage et al. 2000, Lei et al. 2002).

Our study resulted from the serendipitous occurrence of windthrown canopy trees in a study of the effect of Rhododendron removal on biogeochemical cycling. The original experiment examined watershed differences in nutrient fluxes due to shrub removal. We employed a watershed level experimental design (i.e., one treatment vs. one reference watershed), which is commonly used because of the high cost of watershed level studies, that relies on pre- and post- treatment data in a randomized intervention analysis that compares changes in temporal signals (Yeakley et al. 2003). Originally, two study sites, located within 100 m of each other were instrumented for the collection of data on nutrient fluxes and one year of pre- treatment data was collected prior to the removal of the Rhododendron understory layer in one of the sites (Yeakley et al. 2003). At the end of August 1995, the aboveground portions of all Rhododendron stems within a 10 m by 30 m area of one site were manually removed and the herbicide Roundup (Monsanto, Luling, LA) was applied once to the top of cut stumps. Sixty-five stems of Rhododendron were removed, representing approximately 30% of total above-ground woody biomass. This plot is referred to as the shrub removal treatment. On 4 October 1995, Hurricane Opal blew down 9 canopy trees in the adjacent site, which was originally intended as a control plot, while not damaging the shrub removal plot. We subsequently designated a 10 m by 30 m area within this blowdown site as the canopy removal plot. There was minimal damage to the shrub layer in the canopy removal plot, with forty-five intact stems of Rhododendron present. Undisturbed reference plots were located adjacent to both the shrub and canopy removal treatments, to maximize similarity between paired treatment and controls, and are referred to as the shrub and canopy controls, respectively. All four plots had similar pre-treatment overstory composition and Rhododendron density as well as aspect, topography, and soils (Yeakley et al. 2003).

We subsequently began monitoring the species richness and diversity of tree regeneration as well as the growth and survivorship of individual seedlings in all four of these plots. Our nomenclature follows Wofford (1989). Seedling censuses were conducted in ten l m by 1 m quadrats that were randomly located in each of the four plots and that were permanently marked with PVC corner posts. The height and species of all tree seedlings present in the quadrats were recorded during censuses conducted in the growing season of four consecutive years (1997-2000), and each seedling that was recorded in a prior census was scored as alive or dead in subsequent censuses. Species richness and the ShannonWiener index of species diversity (H’) were calculated for each quadrat in every census year using the seedling counts. We calculated the relative growth rates (RGR) for species with seedlings that survived three years; i.e., from the first census in 1997 to the final census in 2000, in two or more treatment plots. RGR was calculated as log(Ht/H0) where H^sub T^ was the final height of the seedling in the 2000 census, H^sub 0^ was the initial height of the seedling in the 1997 census, and T was the elapsed time in years (i.e., 3 years; Beckage and Clark 2003). The availability of light, soil moisture, and soil nutrients was measured across sites. Photosynthetically active radiation (PAR; [mu]mol m^sup -2^ s^sup -1^) was measured at a height of 1 m over each quadrat during overcast days in the 1998 growing season using a sunfleck ceptometer (Decagon Devices, Pullman, Washington). Soil moisture in each of the seedling quadrats was measured six times from early June to late September 1998 using time domain reflectometry (TDR), with 3 mm diameter stainless steel rods inserted vertically 5 cm apart to a depth of 20 cm, following established methods (Topp et al. 1985, Yeakley et al. 1998). Eight lysimeters were installed at a depth of 20-25 cm (BA horizon) in both the canopy removal and shrub removal plots (16 lysimeters total), with four lysimeters installed in the BA horizon in the shrub control plot (Yeakley et al. 2003). Soil water samples were taken weekly and composited monthly for laboratory analysis of nutrient concentrations, including NO^sub 3^, NH^sub 4^, PO^sub 4^, SO^sub 4^, K, Na, Ca, and Mg (Yeakley et al. 2003). Soil water samples were measured beginning one year prior to the removal of the canopy and shrub vegetation in 1994 and continuing through 2000. We present results for NO^sub 3^ (nitrate), averaged over the growing season (May-Sep) of each year, as nitrate had the largest response to disturbance (Yeakley et al. 2003).

We sampled forest floor litter components in February 2000 using ten 30 x 30 cm (0.09 m^sup 2^) quadrats in each of our four treatment plots. Material within each quadrat was separated into three components: litter (Oi), a combined fermentation and humus component (Oe + Oa), and the total organic layer (Oi + Oe + Oa). Small wood (< 7.5 cm diameter) within the 30 x 30 cm sampling frame was cut using pruning shears, and forest floor material was removed by component (i.e., Oi, Oe + Oa, Oi + Oe + Oa) after cutting along the inside of the sampling frame. Forest floor materials were placed in a paper bag and transported to the laboratory where they were dried at 60[degrees]C to a constant mass.

Statistical Analysis. Our analyses of species richness, diversity, seedling survival, and abiotic responses are intended to be exploratory in nature because of the limitations of our data. Our study lacked true replication across spatial units because of its opportunistic nature. Sampling units were located within a single site that experienced the same canopy disturbance or shrub removal treatment (Hurlbert 1984). The seedling recruitment portion of our study benefited from four years of sampling as well as from measurements of abiotic resources including nutrient responses to disturbance.

We estimate seedling regeneration and abiotic responses in our plots using Bayesian methods (Gelman et al. 2003): the posterior distributions calculated in this study will provide prior distributions for more extensive studies of the relative importance of Rhododendron removal vs. canopy removal. The resulting posterior distributions also provide an estimate of the within-plot variability ‘ relative to plot effects. We report if the estimated posterior distributions of treatment effects differ with a probability greater than 0.95. We constructed likelihood functions to estimate the underlying responses while accounting for autocorrelation among repeated sampling periods and year effects, while also using the most appropriate error distribution. seedling survival was modeled as a Bernoulli process with an annual survival probability that is a function of treatment, and year, and also includes a random effect that captures seedling to seedling variability (e.g., Lavine et al. 2002, Beckage et al. 2005). Species richness was estimated using a Poisson likelihood that adjusted for year effects and for serial autocorrelation across repeated censuses of seedling quadrats using an autoregressive (AR) term of order 1 (Beckage and Stout 2000, Beckage and Platt 2003). We used a similar model for the species diversity data except that the errors were normally distributed (Beckage and Stout 2000). The PAR measurements and soil moisture data were analyzed using a normal likelihood. Our likelihood function accounted for sequential measurement periods in the soil moisture sampling intervals. Models were fit using Bayesian methods and either the winBugs (www. mrc-bsu.cam.ac.uk) or R (www.r- project.org) software. We include the complete descriptions of our statistical models, posterior parameter estimates, and associated statistics in an appendix. Actual code for these procedures can be found at www.uvm.edu/~bbeckage/ reprints.html.

Results. Light levels in the forest understory were extremely variable across quadrats (Fig. 1a), with median PAR values of 150, 42, 25, and 3 [mu]mol m 2 s^sup -1^ in the shrub removal, canopy removal, shrub control, and canopy control plots, respectively. PAR levels were significantly higher in the canopy and shrub removal plots compared to controls (P >/= 0.95) as well as in the shrub removal compared to the canopy removal plots (P >/= 0.95). Soil moisture was also higher in the shrub removal plot compared to the canopy removal or shrub control plots (P >/= 0.95) (Fig. 1b). Canopy removal was not associated with significantly greater soil moisture compared to its control (P < 0.95). Concentrations of several nutrients changed in response to the blowdown (Yeakley et al. 2003), but the most pronounced soil water nutrient change occurred for NO^sub 3^. Nitrate concentrations increased by a factor of > 500 at depths of both 20-25 cm and 40-45 cm in the canopy removal plot in the post-hurricane period relative to the pre-treatment period, compared to increases 6 y a factor of 1.3 and 3.3 in the shrub removal plot over the same period. The concentration of nitrate was 770 times greater in the canopy removal treatment compared to shrub removal in the 1997 growing season (Fig. 1c). Increases in nitrate lasted approximately 5 years in our study.

The highest species richness and diversity of tree regeneration occurred in the shrub removal plot followed by canopy removal and controls (Fig. 2). Species richness and diversity were significantly greater in the shrub removal plot compared to the canopy removal plot (P >/= 0.95), and both treatment plots had higher species richness and diversity compared to their controls (P >/= 0.95). Mean species richness was 3.1 (SE = 0.13) and 2.4 (SE = 0.18) species m^sup -2^ with shrub and canopy removal, respectively, compared to 0.98 (SE = 0.084) and 1.4 (SE = 0.17) species m^sup -2^ for their controls. Similarly, mean species diversity was 0.77 (SE = 0.036) and 0.48 (SE = 0.059) with shrub and canopy removal, compared to 0.05 (SE = 0.022) and 0.31 (SE = 0.072) in the shrub and canopy controls, respectively. The maximum species richness of an individual quadrat was 5 and occurred in the shrub removal treatment, whereas the minimum species richness of a quadrat was 0 and occurred in the canopy removal, canopy control, and shrub control plots, but not in the shrub removal plot (Table 1). Species richness increased by a factor of 2.3 and 1.5 in the shrub and canopy removal plots compared to their controls when calculated within plots rather than within quadrats. Recruitment responses to canopy and shrub removal treatments resulted both from the occurrence of novel species, such as the shade intolerant species Betula lenta and Liriodendron tulipifera, and from increased abundance of species that occurred across all treatments (Table 1).

FIG. 1. Abiotic responses to shrub and canopy removal. A) Photosynthetically active radiation (PAR) measured over each quadrat during the 1998 growing season using a sunfleck ceptometer. One measurement was made at a height of l m over each quadrat on an overacast day for a total of 10 measurements per plot. SC corresponds to shrub control, SR to shrub removal, CC to canopy control, and CR to canopy removal. B) Volumetric soil moisture in the shrub removal, canopy removal, and control plots across the 1998 growing season measured using time domain reflectometry. One measurement was made for each quadrat six times across the 1998 growing season for a total of 10 measurements per plot. We show means and standard errors. C) Nitrate concentrations in the BA horizon (20-25 cm in depth) for the 1994-2000 growing seasons (May- Sep) in the shrub control, shrub removal, and canopy removal plots. Removal of the canopy and understory layers is indicated by the arrow. Data are based on eight lysimeters in both the canopy removal and shrub removal plots and four lysimeters in the shrub control plot. We report means and standard errors.

FIG. 2. Species richness and diversity of tree regeneration (mean and s.e.) following removal of shrub understory (shrub removal) in August 1995, or formation of windthrow gaps (canopy removal) in October 1995. Censuses of all quadrats were made in May and October 1997, May, June, and September 1998, June and September 1999, and June 2000, and were combined into a single measurement for each quadrat within a year. The means and s.e.’s are based on the combined yearly censuses for the 10 quadrats in each plot. The location of points along the x-axis have been offset slightly to better distinguish points.

Comparisons of relative growth rates across treatments were limited to two species, Acer rubrum and Liriodendron tulipifera, and to the canopy and shrub removal plots, because of low seedling numbers of other species across treatments. More seedlings of the shade tolerant A. rubrum occurred in the canopy removal plot than did seedlings of the shade intolerant L. tulipifera, likely because A. rubrum could better tolerate the lower light levels in the understory of the canopy removal plot (Fig. 3). The seedlings that established in the canopy removal plot, however, had higher growth rates, which were associated with elevated nitrogen levels, compared to other plots (Fig. 3). Relative growth rates were nearly twice as great in the canopy removal compared to the shrub removal plots for both of these species (P >/= 0.95) (Fig. 3), although the precision of these estimates was limited by low numbers of seedlings in some treatments. Betula lenta seedlings also had very high growth rates in the canopy removal plot (mean RGR = 0.67) in areas of highly disturbed soils in windthrow tip-up mounds. We estimated the effect of treatments on seedling survival of Acer rubrum using 823 newly germinated seedlings that were widely distributed across plots, i.e., 237 (shrub removal), 305 (canopy removal), 171 (shrub control), and 110 (canopy control) seedlings. seedling survival was highest in the shrub removal treatment compared to the canopy removal (P >/= 0.95) or its control plot (P >/= 0.95). Seedling survival was also higher in the canopy removal plot compared to its control plot (P >/= 0.95), but the shrub and canopy control plots were not significantly different (P < 0.95). The mean annual survival probabilities of A. rubrum seedlings were 0.86 (shrub removal), 0.82 (canopy removal), 0.59 (shrub control), and 0.60 (canopy control). Year to year variation in seedling survival was estimated at 3.6 (median on logit scale), which was much larger than the variance associated with individual seedlings (0.095, median on logit scale).

Table 1. Mean density (SE) per m^sup 2^ and frequency of tree seedlings over four years(a).

The soil organic layer was reduced in the shrub removal plot compared to the other treatments, primarily due to decreased amounts of humus (Table 2).

FIG. 3. Relative growth rate in height for Acer rubrum and Liriodendron tulipifera seedlings over the period from 1997 to 2000. The sample means are indicated by horizontal lines and are based on sample sizes of 17, 42, 24, and 4 seedlings from left to right in the panel, respectively.

Discussion. Our limited study suggests that canopy and understory gaps may have complementary effects on resource availability and patterns of seedling recruitment. While both canopy and shrub removal were associated with increased seedling recruitment compared to controls, the greatest diversity and highest survivorship of seedlings coincided with shrub removal while the highest seedling growth rates occurred with canopy removal. The initial establishment of tree seedlings may benefit from the environmental conditions more strongly associated with the removal of understory shrubs, including higher levels of light and soil moisture: shrub removal coincided with seedling regeneration that was 1.3 times more species rich and 1.6 times more diverse than in the canopy removal plot. The understory shrub Rhododendron does not inhibit tree regeneration through allelopathy, modification of seed rain, or through increased predation (Nilsen et al. 1999, Lei et al. 2002, Beckage and Clark 2005), suggesting that the higher diversity of tree seedlings in the shrub removal plot was likely related to elevated levels of light and soil moisture. In addition, seedling establishment may have been facilitated by a > 43% reduction in the soil organic layer that occurred following shrub removal (Table 2). Seedling growth rates, in contrast to seedling diversity, were highest in the canopy removal plot, which corresponded with very large increases (> 700 times greater) in soil nitrate levels. Elevated nitrate levels may be related to the high level of soil disturbance caused by the uprooting of trees in the canopy removal site (Schaetzl et al. 1989, Aber et al. 1998, Greenberg and McNab 1998), which did not occur in any of the other plots. Shrub removal that leaves the canopy relatively intact can occur naturally with low intensity fire (McGee and Smith 1967, Thaxton and Platt 2006, Waldrop et al. 2007) for instance, resulting in a much lower level of soil disturbance than canopy windthrows. Our results are consistent with previous studies finding that seedling establishment is sensitive to light conditions and soil moisture (Haeussler et al. 1995, Negi et al. 1996, Weisberg and Baker 1995) while seedling growth rates respond strongly to nutrient additions especially in conjunction with elevated light levels (Phares 1971, Walters and Reich 2000, Beckage and Clark 2003). Our limited study suggests the need for more extensive field studies of the influence of understory and overstory gaps on tree regeneration, including the potential for complementary effects on light, moisture, and nutrient resources with corresponding effects on seedling establishment, survival and growth (e.g., Beckage and Clark 2003).

Table 2. Litter, humus, and total organic layer components of the forest floor(a).

Our results indicate that equivocal support for the role of canopy gaps in tree regeneration may stem, in part, from both the variable density of understory vegetation across forest stands and the propensity for different modes of canopy gap formation to create understory gaps. Canopy gaps created by drought or insect outbreaks produce standing dead trees with little damage to the forest understory (Clinton et al. 1993, Beckage et al. 2000), and catastrophic forest disturbance, such as results from large blowdowns or landslides, can lead to the elimination of both overstory and understory layers (Veblen and Ashton 1978, Rebertus et al. 1997). Forests with dense understories may require severe disturbance for successful tree regeneration (Veblen 1982, Nakashizuka 1989, Veblen 1989). Forests of the southern Appalachians, for example, have areas with extensive understories of Rhododendron, Kalmia, and Gaylussacia, and canopy gaps created by standing dead trees have failed to increase tree regeneration (Clinton et al. 1994, Beckage et al. 2000). Dense understory layers occur in forests worldwide (e.g., Niering and Egler 1955, Nakashizuka 1989, Veblen 1989, Dolling 1996) and the differential effects of disturbance on overerstory and understory gaps could play a significant role in determining patterns of forest regeneration.

1 This research was partially supported by National Science Foundation Grants DEB 9632854 and DEB 0218001 to the Coweeta LTER program.

Literature Cited

ABER, J., W. MCDOWELL, K. NADELHOFFER, A. MAGILL, G. BERNSTON, M. KAMAKEA, S. MCNULTY, W. CURRIE, L. RUSTAD, AND I. FERNANDEZ. 1998. Nitrogen saturation in temperate forest ecosystems-hypotheses revisited. BioScience 48: 921-934.

BAKER, T. T. AND D. H. VAN LEAR. 1998. Relations between density of rhododendron thickets and diversity of riparian forests. Forest Ecol. Manag. 109: 21-32.

BARDEN, L. S. 1980. Tree replacement in a cove hardwood forest of the southern Appalachians. Oikos 35: 16-19.

BECKAOE, B. AND J. S. CLARK. 2003. Seedling survival and growth of three southern Appalachian Forest tree species: the role of spatial heterogeneity. Ecology 8: 1849-1861.

BECKAGE, B. AND J. S. CLARK. 2005. Does predation contribute to tree diversity? Oecologia 143: 458-469.

BECKAGE, B., J. S. CLARK, B. D. CLINTON, AND B. L. HAINES. 2000. A long-term study of tree seedling recruitment in southern Appalachian forests: the effects of canopy gaps and shrub understories. Can. J. Forest Res. 30: 1617-1631.

BECKAGE, B., M. LAVINE, AND J. S. CLARK. 2005. Survival of tree seedlings across space and time: estimates from long-term count data. J. Ecol. 93: 1177-1184.

BECKAGE, B. AND W. J. PLATT. 2003. Predicting severe wildfire years in the Florida Everglades. Front. Ecol. Environ. 1: 235-239.

BECKAOE, B. AND I. J. STOUT. 2000. The effects of repeated burning on species richness in Florida sandhills: a test of the intermediate disturbance hypothesis. J. Veg. Sci. 11: 113-122.

BREWER, R. AND P. G. MERRITT. 1978. Windthrow and tree replacement in a climax beech-maple forest. Oikos 30: 149-152.

BUSING, R. T. AND P. S. WHITE. 1997. Species diversity and small- scale disturbance in an oldgrowth temperate forest: a consideration of gap partitioning concepts. Oikos 78: 562-568.

CANHAM, C. D. 1988. Growth and canopy architecture of shade tolerant trees: the response of Acer saccharum and Fagus grandifolia to canopy gaps. Ecology 69: 786-795.

CANHAM, C. D., J. S. DENSLOW, W. J. PLATT, J. R. RUNKLE, T. A. SPIES, AND P. S. WHITE. 1990. Light regimes beneath closed canopies and treefall gaps in temperate and tropical forests. Can. J. Forest Res. 20: 620-631.

CHO, D. AND R. E. J. BOERNER. 1991. Canopy disturbance patterns and regeneration of Quercus species in two Ohio old-growth forests. Vegetatio 93: 9-18.

CLINTON, B. D., L. R. BORING, AND W. T. SWANK. 1993. Canopy gap characteristics and drought influences of the Coweeta basin. Ecology 74: 1551-1558.

CLINTON, B. D., L. R. BORING, AND W. T. SWANK. 1994. Regeneration patterns in canopy gaps, mixed-oak forests of the southern Appalachians: influences of topographic position and evergreen understory. Am. Midl. Nat. 132: 308-319.

CONNELL, J. H., M. D. LOWMAN, AND I. R. NOBLE. 1997. Subcanopy gaps in temperate and tropical forests. Aust. J. Ecol. 22: 163-168.

DOLLING, A. H. U. 1996. Interference of bracken with Scots pine and Norway spruce seedling establishment. Forest Ecol. Manag. 88: 227-235.

GELMAN, A., J. B. CARLIN, H. S. STERN, AND D. B. RUBIN. 2003. Bayesian Data Analysis, 2nd edition. Chapman and Hall, Boca Raton, FL. 668 p.

GEORGE, L. O. AND F. A. BAZZAZ. 1999. The fern understory as an ecological filter: emergence and establishment of canopy-tree seedlings. Ecology 80: 833-845.

GREENBERG, C. H. AND W. H. MCNAB. 1998. Forest disturbance in hurricane-related downbursts in the Appalachian mountains of North Carolina. Forest Ecol. Manag. 104: 179-181.

HAEUSSLER, S., J. C. TAPPEINER II, AND B. J. GREBER. 1995. Germination, survival, and early growth of red alder seedlings in the central Coast Range of Oregon. Can. J. Forest Res. 25: 1639- 1651. HIBBS, D. E. 1982. Gap dynamics in a hemlockhardwood forest. Can. J. Forest Res. 12: 522-527.

HUBBELL, S. P., R. B. FOSTER, S. T. O’BRIEN, K. E. HARMS, R. CONDIT, B. WECHSLER, S. J. WRIGHT, AND S. Loo dE LAO. 1999. Light- gap disturbances, recruitment limitation, and tree diversity in a neotropical Forest. Science 283: 554-557.

HURLBERT, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monogr. 54: 187-211.

KNEESHAW, D. D. AND Y. BERGERON. 1998. Canopy gap characteristics and tree replacement in the southeastern boreal forest. Ecology 79: 783-794.

LAVINE, M., B. BECKAGE, AND J. S. CLARK. 2002. Statistical modeling of seedling mortality. J. Agr. Biol. Envir. St. 7: 21-41.

LEI, T. T., S. W. SEMONES, J. F. WALKER, B. D. CLINTON, AND E. T. NILSEN. 2002. Effects of Rhododendron maximum thickets on tree seed dispersal, seedling morphology, and survivorship. Int. J. Plant Sci. 163: 991-1000.

LORIMER, C. G., J. W. CHAPMAN, AND W. D. LAMBERT. 1994. Tall understory vegetation as a factor in the poor development of oak seedlings beneath mature stands. J. Ecol. 82: 227-237.

LUSK, C. H. AND B. SMITH. 1998. Life history differences and tree species coexistence in an oldgrowth New Zealand rain forest. Ecology 79: 795-806.

MCGEE, C. E. AND R. C. SMITH. 1967. Undisturbed Rhododendron thickets are not spreading. J. Forest. 65: 334-336.

NAKASHIZUKA, T. 1989. Role of uprooting in composition and dynamics of an old-growth forest in Japan. Ecology 70: 1273-1278.

NEGI, A. S., G. C. S. NEGI, AND S. P. SINGH. 1996. Establishment and growth of Quercus floribunda seedlings after a mast year. J. Veg. Sci. 7: 559-564.

NIERING, W. A. AND F. E. EGLER. 1955. A shrub community of Viburnum lentago, stable for twenty-five years. Ecology 36: 356- 360.

NILSEN, E. T., J. O. WALKER, O. K. MILLER, S. H. SEMONES, T. T. LEI, AND B. D. CLINTON. 1999. Inhibition of seedling survival under Rhododendron maximum: could allelopathy be a cause? Am. J. Bot. 86: 1597 1605.

PACALA, S. W., C. D. CANHAM, J. A. SILANDER JR., AND R. K. KOBE. 1994. Sapling growth as a function of resources in a north temperate forest. Can. J. Forest Res. 24: 2172-2183.

PECOT, S. D., R. J. MITCHELL, B. J. PALIK, E. B. MOSER, AND J. K. HIERS. 2007. Competitive responses of seedlings and understory plants in longleaf pine woodlands: separating canopy influences above and below ground. Can. J. Forest Res. 37: 634-648.

PHARES, R. E. 1971. Growth of red oak (Quercus rubra L.) seedlings in relation to light and nutrients. Ecology 52: 669-672.

PHILLIPS, D. L. AND W. H. MURDY. 1985. Effects of Rhododendron (Rhododendron maximum L.) on regeneration of southern Appalachian hardwoods. Forest Sci. 31: 226-233.

PICKETT, S. T. A. AND P. S. WHITE. 1985. The Ecology of Natural Disturbances and Patch Dynamics. Academic Press, Orlando, FL. 472 p.

PLATT, W. J. AND D. R. STRONG. 1989. Gaps in forest ecology. Ecology 70: 535.

PUTZ, F. E., P. D. COLEY, K. Lu, A. MONTALVO, AND A. AIELLO. 1983. Uprooting and snapping of trees: structural determinants and ecological consequences. Can. J. Forest Res. 13: 1011-1020.

REBERTUS, A. J., T. KITZBERGER, T. T. VEBLEN, AND L. M. ROOVERS. 1997. Blowdown history and landscape patterns in the Andes of Tierra Del Fuego, Argentina. Ecology 78: 678-692.

RUNKLE, J. R. 1981. Gap regeneration in some oldgrowth forests of the eastern United States. Ecology 62: 1041-1051.

SCHAETZL, R. J., S. F. BURNS, D. L. JOHNSON, AND T. W. SMALL. 1989. Tree uprooting: review of impacts on forest ecology. Vegetatio 79: 165-176.

SHUGART, H. H. 1984. A Theory of Forest Dynamics: The Ecological Implications of Forest Succession Models. Springer-Verlag, Berlin. 278 p.

SWANK, W. T. AND D. A. CROSSLEY. 1988. Introduction and site description, p. 3-16. In W. T. Swank and D. A. Crossley [eds.], Forest hydrology and ecology at Coweeta. Springer-Verlag, New York, NY.

THAXTON, J. M. AND W. J. PLATT. 2006. Small-scale fuel variation alters fire intensity and shrub abundance in a pine savanna. Ecology 87: 1331-1337.

TOPP, G. C., J. L. DAVIS, AND A. P. ANNAN. 1985. Measurement of soil water content using timedomain reflectometry. Soil Sci. Soc. Am. J. 49: 19-24.

VEBLEN, T. T. 1982. Growth patterns of Chusquea bamboos in the understory of Chilean Nothofagus forests and their influences in forest dynamics. Bull. Torrey Bot. Club 109: 474-487.

VEBLEN, T. T. 1989. Tree regeneration responses to gaps along a transandean gradient. Ecology 70: 541-543.

VEBLEN, T. T. AND D. H. ASHTON. 1978. Catastrophic influences on the vegetation of the Valdivian Andes, Chile. Vegetatio 36: 149- 167.

WALDROP, T. A., L. BRUDNAK, AND S. RIDEOUTHANZAK. 2007. Fuels on disturbed and undisturbed sites in the southern Appalachian Mountains, USA. Can. J. Forest Res. 37: 1134-1141.

WALTERS, M. B. AND P. B. REICH. 2000. Seed size, nitrogen supply, and growth rate affect tree seedling survival in deep shade. Ecology 81: 1887-1901.

WEBB, S. L. AND S. E. SCANGA. 2001. Windstorm disturbance without patch dynamics: twelve years of change in a Minnesota forests. Ecology 82: 893-897.

WEISBERG, P. J. AND W. L. BAKER. 1995. Spatial variation in tree regeneration in the forest-tundra ecotone, Rocky Mountain National Park, Colorado. Can. J. Forest Res. 25: 1326-1339.

WHITMORE, T. C. 1989. Canopy gaps and the two major groups of forest trees. Ecology 70: 536-538.

WOFFORD, B. E. 1989. Guide to the vascular plants of the Blue Ridge. University of Georgia Press, Athens, GA. 384 p.

YEAKLEY, J. A., D. C. COLEMAN, B. L. HAINES, B. D. KLOEPPEL, J. L. MEYER, W. T. SWANK, B. W. ARGO, J. M. DEAL, AND S. F. TAYLOR. 2003. Hillslope nutrient dynamics following upland riparian vegetation dynamics. Ecosystems 6: 154-167.

YEAKLEY, J. A., W. T. SWANK, L. W. SWIFT, G. M. HORNBERGER, AND H. H. SHUGART. 1998. Soil moisture gradients and controls on a southern Appalachian hillslope from drought through recharge. Hydrol. Earth Syst. Sci. 2: 41-49.

Brian Beckage2

Department of Plant Biology, University of Vermont, Marsh Life Science Building, Burlington, VT, USA 05405

Brian D. Kloeppel

Odum School of Ecology, University of Georgia, Coweeta Hydrologie Laboratory, 3160 Coweeta Lab Road, Otto, North Carolina, USA 28763

J. Alan Yeakley

Environmental Science and Management, Portland State University, Portland, Oregon, USA 97207

Sharon F. Taylor3

Odum School of Ecology, University of Georgia, Coweeta Hydrologie Laboratory, 3160 Coweeta Lab Road, Otto, North Carolina, USA 28763

David C. Coleman

Odum School of Ecology, University of Georgia, Athens, Georgia, USA 30602

2 Author for correspondence. E-mail: Brian. Beckage@uvm.edu

3 Current address: Land Trust for the Little Tennessee River, Franklin, North Carolina, USA. Received for publication May 17, 2007, and in revised form October 8, 2007.

Appendix

Appendix

Appendix

Copyright Torrey Botanical Society Jan/Feb 2008

(c) 2008 Journal of the Torrey Botanical Society. Provided by ProQuest Information and Learning. All rights Reserved.