2012 NORTHEASTERN NATURALIST 19(1):49–66 Deer Impacts on Seed Banks and Saplings in Eastern New York Carrie R. Levine1,2,*, Raymond J. Winchcombe3, Charles D. Canham3, Lynn M. Christenson1, and Margaret L. Ronsheim1 Abstract - Selective browsing by Odocoileus virginianus (White-tailed Deer) has shifted plant communities in the Northeast, but the effects of seed dispersal by deer on forest seed banks are unknown. We used data from deer exclosures in hunted and unhunted properties in southeastern New York to determine whether deer browsing and different deer management histories have altered composition and/or abundance of forest saplings and seed banks. Results indicate that deer did not alter species richness, abundance, or composition in seed banks in either hunted or unhunted areas. Deer did, however, decrease sapling density and richness at the unhunted site but not at the hunted site. We conclude that impacts of unmanaged deer populations are greater on sapling recruitment than on seed banks. Introduction Densities of Odocoileus virginianus Zimmermann (White-tailed Deer) have increased dramatically in recent decades in the northeastern United States, accompanied by shifts in forest vegetation (Côté et al. 2004, Rooney and Waller 2003, Russell et al. 2001). Deer directly impact forest structure and composition by selectively grazing understory forbs and browsing trees and shrubs. Intensive deer herbivory can increase the susceptibility of forests to non-native plant invasions and can result in the local extirpation of preferred plant species (Augustine and Jordon 1998, Eschtruth and Battles 2008, Ruhren and Handel 2003). These deer-mediated shifts in plant communities have negatively affected a variety of other taxa, including invertebrates, herpetofauna, and songbirds (Allombert et al. 2005, Bergstrom and Edenius 2003, DeCalesta 1994). In addition to browsing impacts, deer may also shape plant community composition by endozoochorous and epizoochorous seed dispersal (Myers et al. 2004, Schmidt et al. 2004). White-tailed Deer ingest seeds as they forage on leaves and twigs and transport a large number and diversity of seeds in their gut across their home range in the Northeast (≈4 ha in summer, and >5 ha in winter in a suburban setting; Porter et al. 2004, Williams et al. 2008). It has been estimated that 30% of seeds ingested by deer travel over 1 km (Vellend et al. 2003), indicating that deer may play an important role in long-distance seed dispersal in forested areas for some plant species. 1Biology Department, Vassar College, 124 Raymond Avenue, Poughkeepsie, NY 12604. 2Current address - Department of Forest and Natural Resource Management, State University of New York College of Environmental Science and Forestry, 1 Forestry Drive, Syracuse NY 13210. 3Cary Institute of Ecosystem Studies, Box AB, Millbrook, NY 12545. *Corresponding author - crlevi01@syr.edu. 50 Northeastern Naturalist Vol. 19, No. 1 While seeds of some plant species germinate immediately after dispersal into forest interiors (Vellend 2002), other seeds may be incorporated into a persistent seed bank (Baskin and Baskin 1998). Forest seed banks are not typically as diverse as those found in other ecosystems, and a significant component of the seeds present are often of early successional and shade-intolerant species (Leckie et al. 2000, Pickett and McDonnell 1989). The resulting low similarity between the species present in the seed bank and the current vegetation of the area can play an important role in gap dynamics and response to disturbance (Hopfensperger 2007, Pickett and McDonnell 1989, Tanentzap and Bazely 2009). Heinken et al. (2006) demonstrated that at least one species of large mammal (Sus scrofa L. [Wild Boar]) can both increase seed density and alter species composition of localized forest seed banks, but the impact of White-tailed Deer on seed bank dynamics is not well known. Seed herbivory by deer may alter seed bank composition by reducing the seed input by forest herbs such as the rare Panax quinquefolius L. (American Ginseng; Furedi and McGraw 2004). The combined effects of herbivory and seed dispersal by deer on deciduous forests are of particular importance in forest fragments at the interface of residential and woodland areas found in many northeastern landscapes (Williams et al. 2008). Forest edges often act as conduits for non-native species and thus have the potential to influence community composition within forests (Cadenasso and Pickett 2001, Tanentzap et al. 2010). Deer facilitate this process by consuming seeds in residential areas and transporting seeds into forested sites (Myers et al. 2004, Williams et al. 2008). High numbers of viable seeds of non-native, invasive species such as Lythrum salicaria L. (Purple Loosestrife) and Lonicera spp. (honeysuckle) have been found in deer feces, indicating that deer may play an important role in the long-distance dispersal of some invasive species (Myers et al. 2004, Vellend 2002). However, it is not clear what effect seed dispersal by deer has on aboveground plant communities in forests. In this study, we used experimental exclosures to examine the impact of deer on saplings and seed banks in two landscapes with different deer management histories in Dutchess County, NY. Our objective was to determine the effect of deer on the diversity, composition, and abundance of sapling recruitment and seed banks in northeastern forests. Field-site Description This study was conducted at two sites in Millbrook, NY: the Rockefeller University Field Research Center (Rockefeller; 41.76ºN, 73.75ºW) and the Cary Institute of Ecosystem Studies (CIES; 41.79ºN, 73.74ºW). The Rockefeller site is 423 ha and the CIES site is 835 ha, and they are located approximately 5 km apart. Forests at both sites are second-growth stands with a history of past use as either pastures or woodlots. Stands at both sites are estimated at >50 yrs old. The CIES site has had an extensively monitored deer management regime, including a controlled annual hunt, since 1976 (Winchcombe 1993). Based on harvest data, current deer density at CIES is estimated to be 12 deer km-2. In contrast, the neighboring Rockefeller site has not been hunted since the late 1960s, and a well-defined browse line is present throughout the forest. 2012 C.R. Levine, R.J. Winchcombe, C.D. Canham, L.M. Christenson, and M.L. Ronsheim 51 The forest canopy at Rockefeller is dominated by Quercus prinus L. (Chestnut Oak) and Q. rubra L. (Northern Red Oak), with Q. velutina Lam. (Black Oak) and Betula lenta L. (Sweet Birch) also present. The forest canopy at CIES is dominated by Q. rubra, with Tsuga canadensis L. (Eastern Hemlock), Acer rubrum L. (Red Maple), Q. prinus, and Q. velutina (Christenson et al. 2002, Katz et al. 2010). Between 1992 and 2008, the average annual temperature of the study area was 9.6 °C, the average annual precipitation was 113 cm, and there was an average of 144 frost-free days between mid-May and late September (Cary Institute of Ecosystem Studies, unpubl. data). Both study sites are within the Nassau-Cardigan complex of soils. These are shallow, somewhat excessively drained Nassau soils and moderately deep, well-drained Cardigan soils formed in glacial till deposits. Rock outcrop covers 2–10% of the surface at the two sites (USDA 2001). Methods Exclosure design and vegetation sampling In 1992, deer exclosures were established at CIES and Rockefeller using a paired-plot design. Four sets of paired 10- x 10-m plots were installed within an approximately 20-ha area at each site, and the paired plots were located 10–20 m apart from one another. All plots were selected to maximize similarity of light availability, canopy composition, and topography. Within each pair, one control plot remained unfenced, and the other plot was enclosed with high-tensile woven wire fencing at the border of the plot edge. The fencing was 3.5 m tall with 12-cm mesh. The sampling area used for saplings was 10 x 10 m. Soil samples and seed bank samples were taken within the inner 8- x 8-m area of each plot. In each plot, all saplings 0.5–2 m in height were identified to species. Pretreatment measurements at Rockefeller and CIES were collected in the spring of 1992. The plots were routinely resampled from 1992–2004, and the data used in this analysis were collected in spring 2004. Soil samples Field methods. To determine seed bank composition, soil samples were collected in October 2008 in the control and fenced plots using a bulb corer. Two transects were randomly located in each plot, and six soil samples (5 cm in depth, excluding the leaf litter) were taken along transects at 1-m intervals, beginning 1 m from the edge of the plot. Soils were stored at 4 °C, and pre-processed by hand sorting to remove large roots, twigs, and stones and to allow for homogenization of samples. The samples from each plot were then pooled, as high spatial variability within the soil seed bank is known to exist (Baskin and Baskin 1998). Soil processing. Soil pH was measured using a 1:2 slurry of soil and Nanopure ® water. Mixtures were stirred intermittently for 30 min and allowed to settle for 90 min. The supernatant was decanted, and the pH was determined using an electrode (Orion Model 420A, Orion Laboratories, Oak Ridge, TN). Loss on ignition was used to determine soil organic matter content (SOM), using 20 g of oven-dried soil muffled at 375 °C for 12 hr. Seed germination. Seed bank samples were cold-stratified in an incubator at 5 °C for three months to simulate winter conditions. After the cold-stratification 52 Northeastern Naturalist Vol. 19, No. 1 period, the density of viable seeds in the soil seed bank was estimated using the seed-emergence method (Brown 1992). The samples were spread in 30- x 70-cm trays over a 3-cm-deep layer of sterile potting soil (Metro-Mix 360, Sun Gro Horticulture, Bellevue, WA). The trays were placed in growth chambers at temperature and light regimes consistent with the June 1 averages for Poughkeepsie, NY, characterized by a 15-hour photoperiod with a daytime temperature of 23 °C and a nighttime temperature of 15 °C (NOAA National Climatic Data Center 2009). Each tray was watered with 450 ml tap water daily and fertilized with 50 g fertilizer (Professional Formula Water Soluble All Purpose Plant Food, Scotts Miracle-Gro Co., Marysville, OH) every two weeks. The emergent seedlings were harvested every two weeks. After six weeks, the top centimeter of soil in each tray was gently mixed with a metal stirring rod (1 mm diameter) to bring ungerminated viable seeds to the surface. Germinants of unknown species were repotted to prevent crowding in trays and grown until they could be identified to species. When germination in all trays had ceased for a total of two weeks, watering was discontinued and the remaining seedlings were harvested. Seed banks were grown for 12 weeks in total. Seedlings were categorized in three groups: trees and shrubs (hereafter woody), grasses and sedges (hereafter graminoids), and herbaceous plants. All woody and herbaceous seedlings were identified to the species level, and graminoids were identified to the family level. Species were also categorized as either native or non-native, and as annual, biennial, or perennial based on the USDA NRCS Plants Database (2011). Grasses, sedges, and unidentified species were included in estimates of seed density, but not in estimates of species richness. Data analyses. All data were analyzed using non-parametric tests. Wilcoxon signed-rank tests were used when a paired comparison was necessary, and Mann- Whitney U tests were used when samples were not paired. Paired tests were used to compare plots within sites (fenced vs. control), and unpaired tests were used to compare differences between fenced and control plots between the two sites. Response variables for the seed bank component of the study included total seed density (seeds m-2), species richness (species per 10- x 10-m plot), and the proportion of species and seeds of non-native species, herbaceous, graminoid, and woody species. We also compared soil pH and SOM between sites. To examine pretreatment sapling density (saplings ha-1) and species richness (number of species per 10- x 10-m plot) between sites, we used an unpaired test to compare the average density and richness of control and fenced plot pairs. We also used a paired test to compare density and richness between control and fenced plots within each site in 1992 (pretreatment) and 2004 (post-treatment). We also used an unpaired test to compare the fenced plots at Rockefeller and CIES in 1992 to determine whether there were preexisting site differences, and used the same test on the fenced plots at both sites in 2004 to assess whether sapling regeneration in the absence of deer differed between the two sites after 12 years of exclusion. Sørenson’s index of similarity was used to calculate the similarity between aboveground sapling/shrub species composition and belowground seed bank composition for the fenced and control pairs at each site, where Sørenson 2012 C.R. Levine, R.J. Winchcombe, C.D. Canham, L.M. Christenson, and M.L. Ronsheim 53 similarity = 2w/(A/B), with A being the number of species aboveground, B the number of species in the seed bank, and w the number of species found both above and belowground (Hopfensperger 2007). Sørenson’s index of similarity ranges from 0 (no species in common) to 1 (all species in common). It was also used to calculate the similarity between aboveground sapling/shrub species composition in the fenced and control plots within each site. All statistical tests were done in R (R Development Core Team 2010). Results Site conditions The two sites did not differ in soil pH (W = 22, P = 0.32) or SOM (W = 39.5, P = 0.46). Rockefeller soils had a mean pH (± s.e.) of 4.5 ± 0.1 and an average SOM content of 92.0 ± 0.3%. CIES soils had an average pH of 4.6 ± 0.1 and an average SOM content of 91.9 ± 0.2%. There were no significant differences in pH or SOM between the fenced and control sites at Rockefeller and CIES (Rockefeller: pH [V = 6.5, P = 0.72], SOM [V = 3, P = 0.99]; CIES: pH [V = 4, P = 0.85], SOM [V = 0, P = 0.10]). At Rockefeller, the mean SOM of the fenced plots was 91.9 ± 0.5%, and the average SOM of the control plots was 92.0 ± 0.2%. At CIES, the average SOM of the fenced sites was 91.7 ± 0.2%, and the average SOM of the control plots was 92.0 ± 0.3%. Seed bank composition A total of 281 seeds germinated from the soil seed bank samples. The seed bank included woody species (101 seeds of 10 species), herbaceous dicot species (65 seeds of 21 species, with 5 seeds of 4 species remaining unidentified to species), and graminoid species (115 seeds total) (Table 1). The seed bank at both sites consisted predominately of native perennials. Herbaceous species that grow in open or disturbed habitats were also common, with only 5 seeds from 4 shadetolerant species present across all samples (Eupatorium rugosum Houtt. [White Snakeroot], Hamamelis virginiana L. [American Witchhazel], Solidago caesia L. [Wreath Goldenrod], and Thalictrum thalictroides (L.) Eames & B. Boivin [[Rue Anemone]). Betula lenta was the primary tree species found in the seed bank at both sites and represented 29% of the total number of seeds that germinated. Nearly 84% of species were represented by 5 or fewer seedlings (Table 1). Sites did not differ in the density of seeds (V = 10, P = 0.69) or the number of species present (V = 8, P = 0.99). Mean seed density (± s.e.) at Rockefeller was 422 ± 66 seeds m-2, with a mean of 4 ± 1 species per plot. Mean seed density in the CIES seed banks was 441 ± 94 seeds m-2, with a mean of 3 ± 1 species per plot (Fig. 1). The fenced treatment had no detectable effect on the number of seeds (Rockefeller W = 6, P = 0.88; CIES: W = 7, P = 0.63) or the number of species at either site (Rockefeller: W = 6, P = 0.88; CIES: W = 1.5, P = 0.99) (Fig. 1). While the similarity in species composition between the fenced and control plots was higher at CIES (Sørenson’s index of similarity mean ± s.e.: 0.46 ± 0.85) than at Rockefeller (0.22 ± 0.83), the difference was not significant (W = 4.5 P = 0.36). At Rockefeller, the seed bank was composed of 28% herbaceous seeds, 51% graminoid seeds, and 21% woody seeds. The CIES seed bank was composed of 54 Northeastern Naturalist Vol. 19, No. 1 Table 1. Seedlings germinated from the Rockefeller and CIES seed banks in fall 2008. The species name is followed by the number of seeds germinated. * indicates non-native species. # = number of seeds. Herbaceous Annual Perennial Woody Site Treatment Graminoid # Species # Species # Species # Rockefeller Fenced Cyperaceae spp. 12 Lobelia inflata L. 4 Oxalis stricta L. 10 Betula lenta L. 16 Poaceae spp. 16 Erigeron annuus (L.) Pers. 1 *Hypericum perforatum L. 2 Rubus flagellaris Willd. 1 Hieracium canadense Michx. 1 *Lythrum salicaria L. 1 Potentilla canadensis L. 1 Solidago bicolor L. 1 Solidago odora Aiton 1 Thalictrum thalictroides (L.) 1 Eames & B. Boivin Unknown sp. A 1 Unknown sp. B 1 Total seeds: 28 Total seeds: 5 Total seeds: 20 Total seeds: 17 Total species: 2 Total species: 10 Total species: 2 Control Cyperaceae spp. 33 Erechtites hieraciifolia (L.) 7 Solidago bicolor L. 3 Betula lenta L. 8 Raf. ex DC. Poaceae spp. 8 Lobelia inflata L. 1 Eupatorium rugosum Houtt. 1 *Ailanthus altissima (Mill.) 1 Swingle * Galium cruciata (L.) Scop. 1 *Celastrus orbiculatus Thunb. 1 Phytolacca americana L. 1 Vitis aestivalis Michx. 1 Potentilla canadensis L. 1 Unknown sp. C 1 Unknown sp. D 1 Total seeds: 41 Total seeds: 8 Total seeds: 9 Total seeds: 11 Total species: 2 Total species: 7 Total species: 4 2012 C.R. Levine, R.J. Winchcombe, C.D. Canham, L.M. Christenson, and M.L. Ronsheim 55 Table 1, continued. Herbaceous Annual Perennial Woody Site Treatment Graminoid # Species # Species # Species # CIES Fenced Cyperaceae spp. 20 Lobelia inflata L. 11 *Hypericum perforatum L. 1 Betula lenta L. 37 Poaceae spp. 1 Amaranthus retroflexus L. 1 Solidago caesia L. 1 Rubus occidentalis L. 3 Unknown sp. B 1 Toxicodendron radicans (L.) 3 Kuntze Total seeds: 21 Total seeds: 12 Total seeds: 3 Total seeds: 43 Total species: 2 Total species: 3 Total species: 3 Control Cyperaceae spp. 15 Mollugo verticillata L. 1 Hypericum perforatum L. 4 Betula lenta L. 22 Poaceae spp. 10 Solidago caesia L. 2 Toxicodendron radicans (L.) 4 Kuntze Potentilla canadensis L. 1 Acer rubrum L. 1 Gleditsia triacanthos L. 1 Hamamelis virginiana L. 1 Vitis aestivalis Michx. 1 Total seeds: 25 Total seeds: 1 Total seeds: 7 Total seeds: 30 Total species: 1 Total species: 3 Total species: 6 56 Northeastern Naturalist Vol. 19, No. 1 16% herbaceous seeds, 33% graminoid seeds, and 52% woody seeds (Fig. 2). The proportion of herbaceous, graminoid, and woody seeds in the seed bank did not differ between sites (herbaceous: W = 8, P = 0.99; graminoids: W = 9, P = 0.89; woody: W = 10, P = 0.69) or between fenced and control plots at either site (Fig. 2). A total of 18 species were identified in the seed bank at Rockefeller, of which 28% were non-native, comprising 5% of all seeds. A total of 13 species were identified at CIES, of which 8% were non-native, also comprising 5% of all seeds (Table 1, Fig. 3). There was no difference between the proportion of non-native species at each site (W = 8, P = 0.99) or between fenced and control plots at Rockefeller (V = 3, P = 0.99) or CIES (V = 1, P = 0.99). Similarly, there was no Figure 1. Seed density (A) and seed species richness (B) of the seed bank in control and fenced plots at Rockefeller and CIES in 2008; n = 4 paired plots at each site. Error bars represent standard error. Differences between treatments were not significant. 2012 C.R. Levine, R.J. Winchcombe, C.D. Canham, L.M. Christenson, and M.L. Ronsheim 57 difference between the proportion of non-native seeds at each site (W = 8, P = 0.99) or between fenced and control plots at Rockefeller (V = 3, P = 0.99) and CIES (V = 1, P = 0.99) (Fig. 3) . The Sørenson’s index of similarity for above- and belowground tree/ shrub species composition was low overall and not different for fenced vs. control plots at either site (Rock: Sexclosure mean = 0.1 ± 0.05, Scontrol mean = Figure 2. Density and species richness of graminoid, herbaceous, and woody germinants from the seed banks in fenced and control plots at Rockefeller and CIES in 2008; n = 4 paired plots at each site. Error bars represent standard error. Differences between treatments and sites were not significant. 58 Northeastern Naturalist Vol. 19, No. 1 0.0 ± 0.0, V = 3, P = 0.37; CIES: Sexclosure mean = 0.06 ± 0.05, Scontrol mean = 0.12 ± 0.07, V = 0, P = 0.37). Sapling response to protection from deer The two study sites differed in numbers of saplings and species richness before the fences were constructed in 1992. The average sapling density was Figure 3. The number of native and non-native seeds (A) and species (B) germinated from the seed banks in fenced and control plots at Rockefeller and CIES in 2008; n = 4 paired plots at each site. Error bars represent standard error. Differences between treatments and sites were not significant. 2012 C.R. Levine, R.J. Winchcombe, C.D. Canham, L.M. Christenson, and M.L. Ronsheim 59 about 5.4 times higher at CIES than at Rockefeller (W = 0, P = 0.03), and the species richness was twice as high at CIES than at Rockefeller (W = 1.5, P = 0.08) (Fig. 4, Table 2). There were no pre-treatment differences in sapling density Figure 4. Sapling (0.5–2 m) density (A) and species richness (B) in control and fenced plots at the Rockefeller and CIES sites. Bars show pre-treatment data (1992) and data after 12 years of deer exclusion (2004); n = 4 paired plots at each site. Error bars represent standard error. In 1992, pre-treatment differences between sites were significant, but differences between treatments within each site were not significant. In 2004, differences between treatments were significant at Rockefeller but not at CIES. 60 Northeastern Naturalist Vol. 19, No. 1 Table 2. Species composition of saplings (0.5–2 m) in the fenced and control plots (10 x 10 m) at Rockefeller (unhunted) and CIES (hunted). # = number of saplings. 1992 2004 Site Treatment Species # Species # Rockefeller Fenced Amelanchier arborea (Michx. f.) Fernald 4 Quercus prinus L. 21 Quercus prinus L. 4 Quercus rubra L. 7 Betula lenta L. 3 Amelanchier arborea (Michx. f.) Fernald 5 Quercus alba L. 2 Carya cordiformis (Wangenh.) K. Koch 5 Carya glabra (Mill.) Sweet 1 Quercus velutina Lam. 5 Fraxinus americana L. 1 Betula lenta L. 4 Quercus rubra L. 1 Ostrya virginiana (Mill.) K. Koch 3 Quercus velutina Lam. 1 Pinus strobus L. 3 Acer rubrum L. 2 Fraxinus americana L. 2 Prunus serotina Ehrh. 2 Quercus alba L. 2 Acer saccharum Marsh. 1 Cornus florida L. 1 Total: 17 Total: 63 Control Quercus prinus L. 29 Quercus prinus L. 3 Amelanchier arborea (Michx. f.) Fernald 10 Ostrya virginiana (Mill.) K. Koch 2 Quercus rubra L. 6 Quercus alba L. 2 Betula lenta L. 5 Quercus rubra L. 2 Ostrya virginiana (Mill.) K. Koch 4 Amelanchier arborea (Michx. f.) Fernald 1 Quercus velutina Lam. 4 Quercus velutina Lam. 1 Acer rubrum L. 2 Total: 60 Total: 11 2012 C.R. Levine, R.J. Winchcombe, C.D. Canham, L.M. Christenson, and M.L. Ronsheim 61 Table 2, continued. 1992 2004 Site Treatment Species # Species # CIES Fenced Ostrya virginiana (Mill.) K. Koch 96 Ostrya virginiana (Mill.) K. Koch 81 Acer pensylvanicum L. 29 Acer saccharum Marsh. 17 Amelanchier arborea (Michx. f.) Fernald 27 Acer pensylvanicum L. 12 Acer saccharum Marsh. 22 Amelanchier arborea (Michx. f.) Fernald 8 Quercus rubra L. 7 Carya cordiformis (Wangenh.) K. Koch 8 Fraxinus americana L. 4 Quercus prinus L. 7 Quercus prinus L. 4 Pinus strobus L. 4 Salix alba L. 4 Fraxinus americana L. 3 Carya glabra (Mill.) Sweet 3 Quercus rubra L. 3 Acer rubrum L. 2 Betula lenta L. 2 Carya cordiformis (Wangenh.) Koch 2 Acer rubrum L. 1 Betula lenta L. 1 Quercus alba L. 1 Robinia pseudoacacia L. 1 Sassafras albidum (Nutt.) Nees 1 Total: 201 Total: 148 Control Acer saccharum Marsh. 52 Ostrya virginiana (Mill.) K. Koch 32 Amelanchier arborea (Michx. f.) Fernald 33 Acer pensylvanicum L. 15 Acer pensylvanicum L. 32 Amelanchier arborea (Michx. f.) Fernald 8 Ostrya virginiana (Mill.) K. Koch 28 Acer saccharum Marsh. 6 Quercus prinus L. 28 Betula lenta L. 4 Quercus rubra L. 16 Quercus prinus L. 4 Carya glabra (Mill.) Sweet 7 Carya cordiformis (Wangenh.) K. Koch 3 Betula lenta L. 3 Quercus rubra L. 3 Quercus velutina Lam. 2 Fagus grandifolia Ehrh. 2 Acer rubrum L. 1 Fraxinus americana L. 2 Amelanchier canadensis (L.) Medik. 1 Acer rubrum L. 1 Fraxinus americana L. 1 Pinus strobus L. 1 Total: 204 Total: 81 62 Northeastern Naturalist Vol. 19, No. 1 (Rockefeller: V = 12, P = 0.28; CIES: V = 4, P = 0.28) or species richness (Rockefeller: V = 12, P = 0.28; CIES: V = 8, P = 0.99) between the control and fenced plots within either site. After 12 years of protection from deer at the Rockefeller site, stem density in the fenced plots was approximately 16 times higher than the stem density in the control plots (V = 0, P = 0.02). Species richness was about 4.5 times higher in the fenced plots than the control plots (V = 0, P = 0.02; Fig. 1). At CIES, there were no significant differences between the fenced and control plots for either stem density (V = 8, P = 0.99) or species richness (V = 6, P = 0.61) ( Fig. 4). In the fenced plots, where sapling density had been significantly higher at CIES than at Rockefeller in 1992 (W = 0, P = 0.03), we saw no difference in sapling density after 12 years of exclusion (W = 9, P = 0.89). Based on the Sørenson’s index of similarity comparing the species composition of woody species in fenced and control plots, species similarity was higher at CIES (0.68 ± 0.09) than at Rockefeller (0.34 ± 0.08) (W = 1, P = 0.06). Discussion We found that the seed bank composition did not differ between the two sites or between treatments within sites. Seed banks were dominated by open-habitat, perennial herbs and shade-intolerant tree species, with relatively low species diversity and low seed density. This finding is consistent with the seed bank composition reported from other Northeast second-growth, deciduous forests (Pickett and McDonnell 1989, Roberts and Vankat 1991). Although there was a higher proportion of graminoid seeds and a correspondingly lower proportion of tree seeds at the site without active deer management (Rockefeller), these proportions were not significantly different from those found at CIES. Consistent with other forest seed bank studies (e.g., Leckie et al. 2000), spring ephemerals were nearly absent, with only 1 Thalictrum thalictroides seedling germinating amongst all our samples. The proportion of non-native species in the seed bank (8% at CIES, 28% at Rockefeller) was relatively low in these formerly pastured and wooded sites as compared to nearby sites that were intensively farmed (51% non-native species at the nearby Vassar College Ecological Preserve; C.R. Levine, unpubl. data). Land-use history can play an important role in the species composition of seed banks (Plue et al. 2009). The pattern of germinable tree seeds in the seed bank was also consistent with other studies of northeast forests. Shade-tolerant tree species such as Acer saccharum Marsh. (Sugar Maple) were absent from the seed bank, despite the presence of both sapling and adult Sugar Maples in the plots. Seeds of large-seeded tree genera such as Quercus and Carya were also absent from the seed bank. Seeds of these tree types often germinate soon after dispersal, remaining in the understory as seedling and sapling banks rather than entering into a persistant seed bank (Pickett and McDonell 1989). The most common tree species in the seed bank was the shade-intolerant B. lenta. Shade-intolerant tree genera such as Betula are often a dominant component of seed banks in northeast forests and are able to germinate and establish in light gaps (Leckie et al. 2000, Yorks et al. 2000). 2012 C.R. Levine, R.J. Winchcombe, C.D. Canham, L.M. Christenson, and M.L. Ronsheim 63 There was little overlap in aboveground saplings/shrubs and belowground woody species composition at either site. Land-use history plays an important role in shaping the species composition of the seed bank of secondary forests (Plue et al. 2010). For example, the seeds of weedy species brought in via agriculture can persist for many decades in the seed bank, leading to reduced similarity between seed bank composition and aboveground woody plants with successional age (Plue et al. 2010). The low similiarity of above and belowground woody species composition found in these forests is not atypical of forests in general and reflects the lack of shade-tolerant species in the seed bank (Hopsfenperger 2007). It is possible that greater similiarity would have been found if the aboveground vegetation had been sampled in 2008 when the seed bank component of the study was conducted, or if we had been able to include herbaceous species in the similarity analysis. We think the latter is unlikely, as the vast majority of herbaceous species found in the seed bank were of open habitat and are rare in >50-year-old forest stands. In addition, vegetation surveys conducted in 2004 near our study sites identified 36 herbaceous species; but only one species, Eupatorium rugosum, overlapped with those we found in the seed bank (Katz et al. 2010). While deer can play an important role in seed dispersal via endozoochory of both native and non-native seeds (Myers et al. 2004, Williams et al. 2008), our data indicate that, over a 16-year period, excluding deer did not affect species composition, seed density, or proportion of non-native species present in the seed bank at either of our sites (Fig. 3). Thus, the effect of endozoochory by deer on forest composition at these sites has more likely consisted of rare, long-distance dispersal events into microsites suitable for germination and establishment (Nathan et al. 2008, Vellend 2002), rather than the incorporation of new seeds into the persistant seed bank. There are many possible reasons why the persistent seed bank was unaffected by deer exclusion (see Baskin and Baskin 1998, Leck et al. 1989). These may include the possibility that ingestion by deer may break innate dormancy mechanisms so that the seeds germinate and do not enter the persistent seed bank. Alternatively, as the seeds of many plant species lack physiological mechanisms for long-term dormancy, the particular plant species that are dispersed by deer may not be physiologically able to persist in seed banks. Also, seeds dispersed in deer feces may have higher rates of pathogen attack and so may not enter the seed bank. It is also possible that there is an indirect effect on the behavior of seed predators, where deer feces attract these animals and so seeds are eaten before they can enter the seed bank. In contrast to their lack of impact on the forest seed bank, deer did have a significant impact on both the number and diversity of saplings present in the understory at the unhunted site (Rockefeller). The fenced plots at Rockefeller were the only plots in which there was an increase in the number of species over the 12-year period, leading to a lower species similarity between the fenced and contol plots at that site. These vegetational shifts were not seen in either the fenced or the control plots at CIES, where deer have been hunted since 1976, as both the number of stems and number of species remained similar over time. While a number of studies have found deer effects on tree regeneration at ≥8 deer km-2 64 Northeastern Naturalist Vol. 19, No. 1 (DeCalesta 1997, Horsley et al. 2003), in this study we found no evidence for deer limiting sapling regeneration at a density of 12 deer km-2. Current vegetation, seed dispersal, and seed bank dynamics can all play important roles in forest dynamics in areas that are heavily browsed (Leckie et al. 2000, Peterson and Pickett 1995, Pickett and McDonnell 1989). With limited tree regeneration in the forest understory, more growing space becomes available for other vegetation types present in the seed bank. For example, Horsley et al. (2003) and Rooney (2009) found an increase in ferns and graminoids with high deer densities. A similar dynamic may be occurring at Rockefeller, where graminoids and herbaceous plants represented over 80% of the seed bank. Our results demonstrate that nearby sites (≈5 km) with different deer-hunting histories exhibited different responses in the sapling layer to the exclusion of deer over a 12-year period. In contrast, different deer-hunting histories and deer exclusion did not alter seed availability in the soil seed banks over a 16-year period. Thus, intensive deer herbivory appears to be a stronger driver of sapling dynamics than deer-mediated effects on seed bank dynamics over the time periods measured in this study. Understanding the potential synergistic relationship of deer herbivory on seedlings, saplings, and recruitment from seed banks is important for assessing the future trajectory of forest regeneration. Acknowledgments We thank K. Van Camp for assisting with the seed bank portion of the study, S. Stehman for helpful comments on the statistical analyses, and R. Yanai, A. Harrison, and S. Stout for helpful reviews of the manuscript. We would also like to thank E. Faison, A.J. Tanentzap, and an anonymous reviewer for additional comments that have contributed to this manuscript. Literature Cited Allombert, S., S. Stockton, and J.L. Martin. 2005. A natural experiment on the impact of overabundant deer on forest invertebrates. Conservation Biology 19:1917–1929. Augustine, D.J., and P.A. Jordan. 1998. Predictors of White-tailed Deer grazing intensity in fragmented deciduous forests. The Journal of Wildlife Management 62:1076–1085. Baskin, C.C., and J.M. Baskin. 1998. Seeds: Ecology, Biogeography, and Evolution of Dormancy and Germination. Academic Press, San Diego, CA. 667 pp. Bergström, R., and L. Edenius. 2003. From twigs to landscapes: Methods for studying ecological effects of forest ungulates. Journal for Nature Conservation 10:203–211. Brown, D. 1992. Estimating the composition of a forest seed bank: A comparison of the seed-extraction and seedling-emergence methods. Canadian Journal of Botany 70:1603–1612. Cadenasso, M.L., and S.T.A. Pickett. 2001. Effect of edge structure on the flux of species into forest interiors. Conservation Biology 15:91–97. Christenson, L.M., G.M. Lovett, M.J. Mitchell, and P.M. Groffman. 2002. The fate of nitrogen in Gypsy Moth frass deposited to an oak forest floor. Oecologia 131:444–452. Côté, S.D., T.P. Rooney, J. Tremblay, C. Dussault, and D.M. Waller. 2004. Ecological impacts of deer overabundance. Annual Review of Ecology, Evolution, and Systematics 35:113–147. 2012 C.R. Levine, R.J. Winchcombe, C.D. Canham, L.M. Christenson, and M.L. Ronsheim 65 DeCalesta, D.S. 1994. Effect of White-tailed Deer on songbirds within managed forests in Pennsylvania. The Journal of Wildlife Management 58:711–718. DeCalesta, D.S. 1997. Deer and ecosystem management. Pp. 267–280, In W.J. McShea, H.B. Underwood, and J.H. Rappole (Eds.). The Science of Overabundance: Deer Ecology and Population Management. Smithsonian Institute Press, Washington, DC. 402 pp. Eschtruth, A.K., and J.J. Battles. 2008. Acceleration of exotic plant invasion in a forested ecosystem by a generalist herbivore. Conservation Biology 23:388–399. Furedi, M.A., and J.B. McGraw. 2004. White-tailed Deer: Dispersers or predators of American ginseng seeds? American Midland Naturalist 152:268–276. Heinken, T., M. Schmidt, G. von Oheimb, W. Kriebitzsch, and H. Ellenberg. 2006. Soil seed banks near rubbing trees indicate dispersal of plant species into forests by Wild Boar. Basic and Applied Ecology 7:31–44. Hopfensperger, K.N. 2007. A review of similarity between seed bank and standing vegetation across ecosystems. Oikos 116:1438–1448. Horsley, S.B., S.L. Stout, and D.S. deCalesta. 2003. White-tailed Deer impact on the vegetation dynamics of a northern hardwood forest. Ecological Applications 13:98–118. Katz, D.S.W., G.M. Lovett, C.D. Canham, and C.M. O’Reilly. 2010. Legacies of land-use history diminish over 22 years in a forest in southeastern New York. Journal of the Torrey Botanical Society 137:236–251. Leck M.A., V.T. Parker, and R.L. Simpson (Eds.). 1989. Ecology of Soil Seed Banks. Academic Press, San Diego, CA. 462 pp. Leckie, S., M. Vellend, G. Bell, M.J. Waterway, and M.J. Lechowicz. 2000. The seed bank in an old-growth, temperate deciduous forest. Canadian Journal of Botany 78:181–192. Myers, J.A., M. Vellend, S. Gardescu, and P.L. Marks. 2004. Seed dispersal by Whitetailed Deer: Implications for long-distance dispersal, invasion, and migration of plants in eastern North America. Oecologia 139:35–44. Nathan, R., F.M. Schurr, O. Spiegel, O. Steinitz, A. Trakhtenbrot, and A. Tsoar. 2008. Mechanisms of long-distance seed dispersal. Trends in Ecology and Evolution 23:638–647. National Oceanic and Atmospheric Administration (NOAA) National Climatic Data Center. 2009. Available online at http://www.ncdc.noaa.gov/oa/ncdc.html. Accessed 17 February 2009. Peterson, C.J., and S.T.A. Pickett. 1995. Forest reorganization: A case study in an oldgrowth forest catastrophic blowdown. Ecology 76:763–774. Pickett, S.T.A., and M.J. McDonnell. 1989. Seed bank dynamics in temperate deciduous forest. Pp. 123–147, In M.A. Leck, V.T. Parker, and R.L. Simpson (Eds.). Ecology of Soil Seed Banks. Academic Press, San Diego, CA. 462 pp. Plue, J., J.L. Dupouey, K. Verheyen, and M. Hermy. 2009. Forest seed banks along an intensity gradient of ancient agriculture. Seed Science Research 19:103–114. Plue, J., K. Verheyen, H. Van Calster, D. Marage, K. Thompson, R. Kalamees, M. Jankowska-Blaszczuk, B. Bossuyt, and M. Hermy. 2010. Seed banks of temperate deciduous forests during secondary succession. Journal of Vegetation Science 21:965–978. Porter, W.F., H.B. Underwood, and J.L. Woodard. 2004. Movement behavior, dispersal, and the potential for localized management of deer in a suburban environment. Journal of Wildlife Management 68:247–256. R Development Core Team. 2010. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online at http://www.R-project.org/. 66 Northeastern Naturalist Vol. 19, No. 1 Roberts, T.L., and J.L.Vankat. 1991. Floristics of a chronosequence corresponding to old-field deciduous forest succession in Southwestern Ohio. II. Seed banks. Bulletin of the Torrey Botanical Club 118:377–384. Rooney, T.P. 2009. High White-tailed Deer densities benefit graminoids and contribute to biotic homogenization of forest ground-layer vegetation. Plant Ecology 202:103–111. Rooney, T.P., and D.M. Waller. 2003. Direct and indirect effects of White-tailed Deer in forest ecosystems. Forest Ecology and Management 181:165–176. Ruhren, S., and S.N. Handel. 2003. Herbivory constrains survival, reproduction, and mutualisms when restoring nine temperate forest herbs. Journal of the Torrey Botanical Society 130:34–42. Russell, F.L., D.B. Zippin, and N.L. Fowler. 2001. Effects of White-tailed Deer (Odocoileus virginianus) on plants, plant populations, and communities: A review. The American Midland Naturalist 146:1–26. Schmidt, M., K. Sommer, W. Kriebitzsch, H. Ellenberg, and G. von Oheimb. 2004. Dispersal of vascular plants by game in northern Germany. Part I: Roe Deer (Capreolus capreolus) and Wild Boar (Sus scrofa). European Journal of Forest Research 123:167–176. Tanentzap, A.J., and D.R. Bazely. 2009. Propagule pressure and resource availability determine plant community invisibility in a temperate forest understorey. Oikos 118:300–308. Tanentzap, A.J., D.R. Bazely, and R. Lafortezza. 2010. Diversity-invasibility relationships across multiple scales in disturbed forest understoreys. Biological Invasions 12:2105–2116. United States Department of Agriculture (USDA) Natural Resources Conservation Service (NRCS). 2001. Soil Survey of Dutchess County, New York. A publication of the National Cooperative Soil Survey. 357 pp. USDA, NRCS. 2011. Plants Database. Available online at http://plants.usda.gov/java/. Accessed 1 March 2011. Vellend, M. 2002. A pest and an invader: White-tailed Deer (Odocoileus virginianus Zimm.) as a seed dispersal agent for honeysuckle shrubs (Lonicera L.). Natural Areas Journal 22:230–234. Vellend, M., J.A. Myers, S. Gardescu, and P.L. Marks. 2003. Dispersal of trillium seeds by deer: Implications for long-distance migration of forest herbs. Ecology 84:1067–1072. Williams, S.C., J.S. Ward, and U. Ramakrishnan. 2008. Endozoochory by White-tailed Deer (Odocoileus virginianus) across a suburban/woodland interface. Forest Ecology and Management 255:940–947. Winchcombe, R.J. 1993. Controlled access hunting for deer population control: A case study. Northeast Wildlife 50:1–9. Yorks, T.E., D.J. Leopold, and D.J. Raynal. 2000. Vascular plant propagule banks of six Eastern Hemlock stands in the Catskill Mountains of New York. Journal of the Torrey Botanical Society 127:87–93.