Northeastern Naturalist 45 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 Sheep Grazing as a Grassland Management Tool: Lessons Learned on Nantucket Island, Massachusetts Karen C. Beattie1,*, Jennifer M. Karberg1, Kelly A. Omand1, and Danielle I. O’Dell1 Abstract - On Nantucket Island, MA, the present range of coastal sandplain grasslands is primarily attributed to intense and prolonged historic sheep grazing following European settlement. The maintenance of this early successional habitat (ranked S-1 or “critically imperiled” in Massachusetts) relies on disturbance-based land-management tools. Habitat management efforts have focused primarily on mechanical and prescribed fire treatments, with limited emphasis on re-introducing sheep. This study examined and compared the impacts of repeated growing-season grazing, repeated growing-season mowing, and no management on vegetation community composition in a previously managed grassland. Sheep grazing effectively controlled and reduced clonal and vining woody plant coverage while increasing available bare ground for grassland species seed recruitment. However, grazing and mowing treatments resulted in an increase in weedy agricultural plant species, which may be an inherent side effect of management that results in soil disturbance. Given the long-term, variable nature of the ecological disturbances that created Nantucket’s sandplain grassland vegetation communities, one management technique alone will not likely result in successful habitat restoration over a short period of time. We recommend that sheep grazing be more widely considered as an addition to the existing sandplain grassland management “tool box”. Introduction The glacial outwash plain deposits along the coast of the northeastern United States, including parts of Long Island (NY), Block Island (RI), the Elizabeth Islands (MA), Cape Cod (MA), Martha’s Vineyard (MA), and Nantucket Island (MA), host the largest contiguous areas of coastal sandplain grassland (Neill 2007). This vegetation community occurs on acidic, nutrient-poor, drought-prone soils (Vickery and Dunwiddie 1997) and is dominated by native graminoids such as Schizachyrium scoparium (Little Bluestem), Carex pensylvanica (Pennsylvania Sedge), and Danthonia spicata (Poverty Oatgrass) intermixed with ericaceous shrubs, including Gaylussacia baccata (Black Huckleberry), Arctostaphylos uva-ursi (Bearberry), Vaccinium angustifolium (Lowbush Blueberry), and numerous forb species (Swain and Kearsley 2001). Prior to human settlement, these grasslands likely occurred in small, open areas impacted by wind, salt spray, and occasionally fire (Dunwiddie et al. 1996). Forest clearing by Native Americans and then European settlers, as well as agriculture and livestock grazing during the 1600s–late 1800s increased the natural expanse of this disturbance-driven community (Dunwiddie 1989). 1Science and Stewardship Department, Nantucket Conservation Foundation, PO Box 13, Nantucket, MA 02554-0013. *Corresponding author - kbeattie@nantucketconservation.org. Manuscript Editor: Daniel M. Keppie Natural History of Agricultural Landscapes 2017 Northeastern Naturalist 24(Special Issue 8):45–66 Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 46 Vol. 24, Special Issue 8 On Nantucket Island, intense and prolonged sheep grazing played a dominant role in the expansion and maintenance of sandplain grasslands (Dunwiddie 1989). A New England survey reported 25,000–30,000 sheep grazing on the island in 1704 (McManis 2010). Henry David Thoreau (1962;818) wrote in 1854, “This island must look exactly like a prairie, except that the view in clear weather is bounded by the sea.” Botanists soon noted the unique assemblage of plant species associated with these open grasslands on Nantucket and the role that sheep grazing played in their establishment and expansion (Harshberger 1917, Owen 1888, Rice 1946) and continue to study and document it today (Dunwiddie 1997, Sorrie and Dunwiddie 1996). Coastal sandplain grasslands require frequent disturbance to prevent vegetative succession into a community dominated by woody species (Foster and O’Keefe 2000). Compared to early European settlement, current land-use patterns on Nantucket and elsewhere in coastal New England include decreased agriculture and livestock grazing, increased fire suppression, and landscape development (Foster and Match 2003, Foster and O’Keefe 2000). This shift in land management practices has greatly diminished areas of intact coastal sandplain grassland to small, globally rare remnants (Dunwiddie et al. 1996) that now represent one of the highest priorities for conservation in the northeastern Unites States (Neill 2007). Numerous rare plant and animal species are recognized as strongly affiliated with sandplain grasslands, which are ranked S-1 (critically imperiled) in Massachusetts due to extreme rarity and vulnerability to extirpation (Barbour et al. 1998, Carlson et al. 1991, Steel 1999, Swain and Kearsley 2001). The current limited distribution of sandplain grasslands and the associated rare taxa they support are compelling ecological reasons for prioritizing their restoration, maintenance, and protection (Dunwiddie 1989). Nantucket Island contains some of the largest contiguous areas of sandplain grassland remaining in the northeastern United States (Dunwiddie 1989). Research and management efforts on Nantucket and elsewhere have concentrated on the use and effectiveness of mechanical vegetation treatments (mowing/brush-cutting), prescribed burning, herbicide use, and soil disturbance (plowing/harrowing) to maintain and increase sandplain grassland habitats (Dunwiddie et al. 1990). Little emphasis has been placed on re-introducing sheep as a potential restoration and management tool, despite the fact that sheep grazing was the primary historic land use associated with sandplain grassland establishment and expansion on the island (Dunwiddie 1997). Targeted grazing, a relatively new livestock management system, is the use of grazing ungulates to manipulate forage with a particular management goal or set of defined objectives (Chapman and Reid 2004). Grazing livestock are capable of modifying plant biomass, structure, and floral composition by removing vegetation through consumption and disturbing soil and ground cover with their movement patterns. These actions can create bare ground that facilitates colonization of vegetation through seed germination (Bullock et al. 1994, 1995; Silvertown and Smith 1988). The effects of grazing contrast with those of mowing, which is widely used as a grassland management technique. Mowing produces a spatially consistent disturbance treatment that does not typically create interstitial spaces for plant Northeastern Naturalist 47 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 recruitment (Huhta 2001). Thus, sheep grazing and mowing as grassland management techniques can potentially produce very different ecological results that have not been well studied within a paired treatment research design framework. This study examined the effects of repeated growing-season grazing, repeated growing-season mowing, and no management on vegetation communities in a previously managed open grassland. The goals of this research were to compare how mowing and sheep grazing performed as management tools for maintaining open grasslands, increasing sandplain grassland-associated plant species, and reducing woody shrub cover. Results of this study will help inform the management of sandplain grasslands and similar open grassland plant communities. Field-site Description Study site conditions Squam Farm (41°18'28"N, 69°59'42"W) is an 88.5-ha conservation property in the northeast portion of Nantucket Island, MA (Fig. 1), owned by the Nantucket Conservation Foundation, Inc. Nantucket is located 42 km south of Cape Cod and comprises ~116 km2 in land area. Average winter and summer temperatures are 1 °C and 22 °C, respectively; mean annual precipitation is 1070 mm (Langlois 1979). The study site contains a mosaic of upland and freshwater wetlands including coastal shrubland, open scrub oak shrubland, mowed grassland, wooded swamp, mixed deciduous forest, and shrub swamp. Study site land-use history Although no detailed records exist regarding historic land uses at Squam Farm, the vast majority of Nantucket was set aside as common grazing land for sheep and Figure 1. North Pasture location and research plot set-up. Squam Farm, Nantucket, MA. Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 48 Vol. 24, Special Issue 8 other livestock immediately after colonial settlement (Crevecoeur 1957). Records indicate that ~800 ha in the Squam area was planted with grain following a severe winter food shortage on the island in 1780; at least 3 farms operated within a 1.5 km radius of Squam Farm ca. 1850 (McManis 2010). Therefore, some or all of the site was likely historically utilized for agricultural purposes. By 1940, woody vegetation had reclaimed the majority of this area, as shown in aerial photographs (NHA 2016). More recently, ~30.5 ha of upland areas were cleared and maintained with mowing and small-scale livestock grazing by a previous owner in the 1980s. After acquiring the property in 2001, the Nantucket Conservation Foundation conducted annual dormant-season mowing in these upland areas to maintain early successional vegetation communities and provide suitable conditions for sandplain grassland-associated species. Research area This research project was initiated in 2005 in the North Pasture (0.69 ha) section of Squam Farm (Fig. 1). Initially started as a 1-year Master’s Thesis project (Schlimme 2006), the Nantucket Conservation Foundation continued the research for a total of 4 years. Elevation at North Pasture varies from 11 to 15 m above sea level (MassGIS 2012). Surficial geology consists of Pleistocene glacial end moraine (Oldale 1985). The dominant soil type is Plymouth-Evesboro complex (3–8% slopes) (Langlois 1979, MassGIS 2003). The North Pasture research area contained an assemblage of native sandplain grassland-associated species, including Little Bluestem, Pennsylvania Sedge, and Poverty Oatgrass, as well as non-native, cool-season grasses such as Holcus lanatus (Common Velvetgrass), Anthoxanthum odoratum (Sweet Vernalgrass), and Festuca ovina (Sheep Fescue). These non-native grasses are identified as a threat to sandplain grasslands because they form mats that change the character of the vegetation community (Swain and Kearsley 2001). Thus, the North Pasture research area provided an opportunity to test how both desirable and undesirable plant species responded to mowing and grazing treatments over multiple growing seasons. Methods Study design We divided the North Pasture area at Squam Farm into 9 research units (each 33 m x 21 m; Fig. 1) and randomly assigned each unit to a treatment for the length of the study: graze (n = 3), mow (n = 3), or control (n = 3). Prior to the initiation of this study, the entire North Pasture area was mowed annually for several years during the dormant season. Therefore, the control treatment provided an example of unmanaged vegetation during the course of this study. This study was initially designed to conduct 3 mow or graze treatments per growing season in each research unit for 3 years (2005–2007). However, drought conditions experienced in both 2005 and 2007 slowed post-grazing vegetation recovery, and a series of sheep flock management issues, including the unexpected Northeastern Naturalist 49 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 birth of lambs, caused further logistical complications in implementing the planned treatment schedule. As a result, the number of treatments varied each year: mowing and grazing treatments were applied either 2 or 3 times per growing season to all research units (Table 1). Additionally, the difference in number of treatments per growing season necessitated that the timing of treatments varied (Table 1). No treatments occurred in 2008, but we sampled vegetation that year to document the response during the first full growing season after ending treat ments. Grazing treatment. During each graze treatment, the sheep flock was contained in 1 randomly selected graze unit until most available forage was consumed; we then randomly rotated the flock through the other 2 graze units for the same number of days per unit (Table 1). Solar-powered, portable electro-net fencing (Wellscroft Fence Systems, LLC, Harrisville, NH) contained the sheep within the graze units. The size of the graze treatment sheep flock varied over the course of the study. In 2005, the Foundation borrowed 28 adult Cotswold sheep from the New England Heritage Breeds Conservancy (formerly in Great Barrington, MA; no longer in operation) for the duration of the summer season. The remainder of the research project utilized a reduced-size, mixed flock of Cotswolds (8 in 2006 and 3 in 2007) and Romneys (9 in 2006 and 4 in 2007). Cotswolds used in 2006–2007 were born and raised at the study site; Romneys were raised elsewhere on Nantucket and donated to the project by the Massachusetts Audubon Society. Although differences in flock size and breed composition were likely sources of variability in the graze-treatment effect, we believe that these differences are inherent to the use of livestock grazing as an ecological management regime. Mowing treatment. We mowed all mow units immediately following the end of each grazing treatment (Table 1) using a John Deere® tractor fitted with a rearmounted rotary cutter attachment. Table 1. Vegetation sampling times each year and corresponding sheep grazing and mowing treatment times over the course of the study. * the end of the sampling period indicates the late summer vegetation sampling time used throughout the analysis. Mowing treatments were performed at the end of each grazing treatment. Sampling dates Year Pre-treatment Post-treatment 1 Post-treatment 2 2005 July 14–Aug. 3 Aug. 8–12 Sept. 8–28* 2006 - Aug. 15–Oct. 13* 2007 - Sept. 25–Oct. 19* 2008 - July 24- Aug. 7 Sept. 9–24* Treatment dates Year Graze start Graze time Graze start Graze time Graze start Graze time 2005 Aug. 3 2 days/unit Sept. 7 1 day/unit - - 2006 June 13 2 days/unit July 11 2 days/unit Aug. 9 2 days/unit 2007 May 22 5 days/unit July 11 4 days/unit Aug. 14 6 days/unit 2008 - - - - - - Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 50 Vol. 24, Special Issue 8 Vegetation sampling Using ArcGIS (ESRI, Inc., Redlands, CA), we established 5 vegetation sampling plots within each of the 9 research units. Plots were separated from each other and the unit edge by at least 5 m to maximize spatial independence. Due to the size of the treatment unit and the need to buffer individual plots from each other and the unit edge, plot locations were not randomized. However, because plots were created in ArcGIS, no prior knowledge of site conditions or vegetation composition influenced plot location selection. To locate computer generated plots in the field, we exported plot coordinates to a Trimble® Geo XT™ Global Positioning System unit. Once we navigated to a plot point, we established a 1-m2 plot, permanently marking all plot corners with buried rebar to facilitate relocation with a metal detector. We established a total of 45 vegetation-sampling plots (5 plots per research unit; 15 plots per treatment type). We sampled vegetation community composition at each plot using a 1-m2 inclined point quadrat sampling frame containing 50 points per quadrat in a 10 x 5 grid with x-axis points spaced 10 cm apart and y-axis points spaced 20 cm apart. At each sampling point, a thin dowel (6.3 mm diameter) was inserted through the quadrat frame at a 32.5º angle. We recorded all plant species contacting the dowel as it slid to the ground, as well as the type of ground cover present (bare ground or litter) when the dowel touched the ground. Plants were identified to the species level whenever possible (Appendix 1). We recorded only once each plant species encountered per sampling point (dowel contact). From this data, we were able to calculate the percent foliar coverage of a species within each vegetation plot as a percent of total sampling points (dowel contacts) per 1-m2 plot (Olusuyi and Raguse 1968, Wilson 1963). For example, if Little Bluestem was encountered in 9 of the 50 sampling points within a plot, then percent foliar coverage was calculated to be 18%. Using an inclined point intercept with an angle of 32.5º dramatically reduced sampling error and over-estimation of shrubs and forbs (Wilson 1960, 1963; Winkworth 1954). Pre-treatment vegetation sampling occurred in mid-July 2005. Each year of the study, we conducted post-treatment vegetation sampling in mid-summer and late summer, ~1 month post treatment and thus varied annually based on the timing of treatments (Table 1). We sampled twice during 2008 to document vegetation response during the first full growing season after ending treatme nts. Analysis Broad functional groups. The low number of vegetation sampling plots in this study (n = 5 per unit and n = 15 total per treatment) meant that individual species could be encountered infrequently even when common in the research units, increasing the error associated with analyzing the data at the species level (Lavorel, et al. 1997, Sternberg, et al. 2000). To minimize this error, we analyzed vegetation data in broad functional groups (graminoid, forb, woody) as well as ground-cover categories (bare ground, litter). The response of species at the broad functional group level can be used as an indication of the direction and speed of vegetation community shifts, as these functional groups represent different stages of the successional process (Hellstrom et al. 2003, McIntyre et al. 1995). We chose Northeastern Naturalist 51 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 to examine broad plant growth forms and ground-cover categories because we were interested in the influence of these treatments on maintaining and increasing grassland-associated species, decreasing woody cover, and influencing conditions for successful grassland-associated species recruitment and establishment. All analyses were conducted using SPSS (IBM Corp. 2012) with an alpha of 0.05 to determine significance. Percent foliar coverage, representing the mean occurrence of each functional group, was calculated by dividing the total number of dowel contacts at a single sampling point by the number of sampling points (n = 50) in each vegetation sampling plot (Anderson 1986; Wilson 1960, 1963). Because multiple species were assigned to each functional group and dowel contacts were recorded by individual species, the percent foliar coverage of a functional group could exceed 100% within a sampling plot. Management interest functional groups. While changes in overall growth form and ground cover categories are useful in documenting general ecological successional trajectories, species composition is an important determinant of whether habitat restoration is successful. In order to examine the effects of our management treatments in more detail, we selectively classified key species in functional groups defined by habitat association and/or plant function that might influence grassland restoration goals. We created 4 management interest functional groups (Table 2). Agricultural species included non-native and weedy graminoids and forbs already present at the start of this study that might increase with disturbance. Sandplain grassland species included forbs and graminoids indicative of sandplain grassland communities as defined by the Massachusetts Natural Heritage and Endangered Species Program (MANHESP 2010). The clonal/vining woody functional group included clonal or vining woody species that are not reduced in cover by repeated disturbance (Dunwiddie 1997, Neiring and Driyer 1989) and therefore might decrease the success of grassland establishment and maintenance (Harper 1995). And last, we examined shrub woody species that form larger shrubs or trees in the absence of disturbance management. Only species that were deemed of management interest or defined as strongly indicative of habitat were included in a management interest functional group; hence, not all individual species encountered in the study were classified as part of a group. Percent foliar coverage, representing the mean occurrence of each management interest functional group, was calculated by dividing the total number of dowel contacts at a single sampling point by the number of sampling points (n = 50) in each vegetation sampling plot (Anderson 1986; Wilson 1960, 1963). Because multiple species were assigned to each functional group and dowel contacts were recorded by individual species, the percent foliar coverage of a functional group could exceed 100% within a sampling plot. Statistical analysis. To determine if vegetation communities differed prior to treatment, we conducted a 1-way analysis of variance (ANOVA) on the pre-treatment mean foliar coverage of each category (graminoid, forb, woody, bare ground, and litter) among all research units. Additionally, we conducted a 1-way analysis of variance (ANOVA) on pre-treatment mean foliar coverage of each management Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 52 Vol. 24, Special Issue 8 interest functional group (agricultural, sandplain grassland, clonal/vining woody, and shrub woody) between all research units. The vegetation composition did not differ among research units prior to the application of treatment (Table 3), which allowed us to conduct the following analysis on post-treatment data only. Table 2. Species classified into management interest functional groups based on community definitions and native species guides for Massachusetts as well as literature review and personal observations on Nantucket. Note that not all species encountered within this study were assigned to a management interest functional group. Management interest Broad functional group/ functional Scientific name Common name (status) group Agricultural species Anthoxanthum odoratum Sweet Vernalgrass (non-native) Graminoid Dactylis glomerata Orchard Grass (non-native) Graminoid Festuca ovina Sheep Fescue (non-native) Graminoid Hieracium spp. Hawkweeds (native, disturbance) Forb Holcus lanatus Common Velvetgrass (non-native, invasive) Graminoid Leucanthemum vulgare Oxeye Daisy (non-native) Forb Potentilla canadensis Dwarf Cinquefoil (native, disturbance) Forb Rumex acetosella Red Sorrel (non-native) Forb Vicia sativa Common Vetch (non-native) Forb Sandplain grassland species Agrostis hyemalis Winter Bentgrass (native) Graminoid Carex pensylvanica Pennsylvania Sedge (native) Graminoid Crocanthemum dumosum Bushy Frostweed (native, special concern) Forb Danthonia spicata Poverty Oatgrass (native) Graminoid Euthamia graminifolia Grass-leaved Goldenrod (native) Forb Hypericum stragulum St. Andrew's Cross (native, endangered) Forb Juncus greenei Greene's Rush (native) Graminoid Panicum virgatum Switchgrass (native) Graminoid Polygala polygama Bitter Milkwort (native) Forb Schizachyrium scoparium Little Bluestem (native) Graminoid Sisyrinchium fuscatum Coastal Plain Blue-eyed Grass (native, special concern) Forb Solidago rugosa Wrinkleleaf Goldenrod (native) Forb Symphyotrichum spp. American-asters (native) Forb Clonal/vining woody species Corylus cornuta Beaked Hazelnut (native) Woody Gaultheria procumbens Eastern Wintergreen (native) Woody Gaylussacia baccata Black Huckleberry (native) Woody Lonicera japonica Japanese Honeysuckle (non-native, invasive) Woody Parthenocissus quinquefolia Virginia Creeper (native) Woody Rubus spp. Dewberries (native) Woody Smilax spp. Greenbriers (native) Woody Vaccinium spp. Blueberries (native) Woody Vitis labrusca Fox Grape (native) Woody Shrub/tree woody species Morella caroliniensis Bayberry (native) Woody Quercus spp. Scrub Oaks (native) Woody Sassafras albidum Sassafras (native) Woody Northeastern Naturalist 53 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 Table 3. Analysis of variance (ANOVA) and post-hoc Tukey’s HSD results (C = control, G = graze, M = mow) for all broad functional groups over the 4-year study. All pre-treatment sampling occured in 2005. * indicates significant P values. Columns for each treatment indicate the mean (M) and standard deviation (SD) of each treatment. [Table continued on next page.] Broad functional groups/ Graze Control Mow Tukey’s sampling year M SD M SD M SD F 2,42 P HSD Graminoid Pre-treatment 143.20 61.12 135.60 59.67 142.80 58.81 0.08 0.93 2005 41.47 29.80 118.67 54.78 57.60 26.75 16.21 0.01* C > G,M 2006 151.33 53.51 174.27 76.54 191.87 37.14 1.84 0.17 2007 111.87 79.27 46.27 10.55 88.93 20.81 7.31 0.02* C < G,M 2008 164.80 42.87 142.13 92.75 196.93 47.03 2.70 0.08 Forb Pre-treatment 72.40 47.68 70.00 47.68 92.53 66.51 0.68 0.52 2005 4.80 7.12 40.93 42.55 6.80 7.85 9.65 0.01* C > G,M 2006 54.40 25.60 46.80 30.32 52.40 33.26 0.26 0.77 2007 54.93 57.03 16.00 10.17 20.00 18.73 5.57 0.07* G > C,M 2008 71.10 39.69 63.87 53.56 65.10 53.15 0.92 0.91 Woody Pre-treatment 113.87 59.86 120.00 70.71 102.93 36.65 0.34 0.72 2005 42.93 27.31 86.27 53.72 23.87 10.73 12.28 0.01* C > G,M 2006 78.67 46.63 161.07 91.58 52.27 20.91 13.18 0.01 C > G,M 2007 122.53 73.06 26.27 31.35 12.80 7.81 25.26 0.01 G > C,M 2008 84.40 51.65 103.33 62.69 38.80 17.08 7.19 0.02 M < C,G Bare ground Pre-treatment 12.67 13.22 13.87 11.30 14.00 17.05 0.04 0.96 2005 19.60 14.37 18.67 12.27 7.73 7.87 4.67 0.02* M < C,G 2006 24.13 13.45 13.73 13.22 17.60 16.23 2.01 0.15 2007 2.40 2.60 12.67 10.07 7.20 5.22 8.74 0.01* G < C 2008 19.87 9.93 8.40 8.21 10.13 5.73 8.64 0.01* G > C,M Litter Pre-treatment 80.40 15.82 82.67 16.38 82.67 14.67 0.11 0.91 2005 88.27 6.28 94.93 4.13 89.47 9.27 3.99 0.03* G < C 2006 94.53 5.97 96.27 12.85 97.73 3.01 0.55 0.58 2007 90.80 6.36 76.80 18.48 86.80 9.85 4.88 0.01* G < C 2008 80.40 9.45 89.60 5.82 91.87 8.09 8.80 0.01* G < C,M Agricultural Pre-treatment 77.20 25.06 66.80 35.69 85.60 37.36 1.21 0.31 2005 8.27 9.47 45.73 33.32 11.73 14.73 13.61 0.01* C > G,M 2006 94.93 43.75 63.60 36.75 85.87 43.96 2.25 0.12 2007 24.00 13.37 35.33 35.57 33.87 36.22 0.62 0.54 2008 102.00 48.90 38.67 23.56 96.40 60.43 8.38 0.01* C < G,M Sandplain Pre-treatment 118.53 34.62 125.07 35.15 132.93 32.54 0.67 0.52 2005 32.13 21.85 103.73 28.95 31.20 21.43 43.90 0.01* C > G,M 2006 39.07 27.73 77.73 29.11 51.33 30.58 6.89 0.03* C > G,M 2007 35.20 13.26 124.67 33.33 65.73 18.23 57.49 0.01* C > M > G 2008 105.47 36.49 157.07 42.96 141.07 27.70 7.96 0.01* G < C,M Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 54 Vol. 24, Special Issue 8 A repeated measures analysis of variance (RMANOVA) was used to examine changes in the mean foliar coverage in each broad functional group and groundcover category in response to multiple annual mowing and grazing treatments over the 4-year study. The same method examined changes in mean foliar coverage in each management interest functional group. Where the RMANOVAs proved significant, we used a post hoc Tukey HSD test to examine how each treatment affected broad functional group, ground-cover group, and management interest functional group coverage. Results Broad functional groups Prior to the start of the study, vegetation sampling plots showed no significant difference in percent foliar coverage of broad vegetation functional groups (graminoid, forb, woody) and ground-cover categories (bare ground, litter) between the research units (Fig. 2). Post-treatment sampling in the first year of the study (2005) showed a significant reduction in the foliar coverage of each functional group in both the grazing and mowing treatments compared to the lack of treatment in the controls (graminoid [F2,42 = 16.21, P = 0.01], forb [F2, 42 = 9.65, P = 0.01], woody [F2, 42 = 12.28, P = 0.01]). Additionally, this first year of the mowing treatment (2005) resulted in significantly less bare ground compared to the grazing and control treatments (F2, 24 = 4.67, P = 0.02), while the grazing treatment had significantly less litter compared to the mowing and control treatments (F2, 24 = 3.99, P = 0.03). Post-treatment sampling over the next 3 years of the study allowed us to examine the trajectory of restoration as the vegetation community responded to repeated management. Foliar coverage of functional and ground-cover groups exhibited varied responses to each treatment over the 4 years of the study (Table 3). These responses may have also been influenced by precipitation levels. Average precipitation during the growing season on Nantucket (May–September) is 335.3 mm (NOAA 2016). Table 3, continued. Broad functional groups/ Graze Control Mow Tukey’s sampling year M SD M SD M SD F 2,42 P HSD Clonal/vining woody Pre-treatment 87.73 44.54 91.87 45.48 82.13 27.90 0.22 0.08 2005 23.07 19.08 64.80 35.09 14.80 10.30 18.10 0.01* C > G,M 2006 62.93 39.49 123.60 62.44 43.60 18.18 13.55 0.01* C > G,M 2007 18.80 23.84 97.73 60.57 10.40 6.77 24.39 0.01* C > G,M 2008 73.30 50.42 117.87 66.11 48.67 23.82 7.40 0.02* C > G,M Shrub woody Pre-treatment 11.73 15.76 14.80 27.05 0.80 3.09 2.46 0.10 2005 8.53 12.13 13.73 24.26 0.13 0.52 2.97 0.06 2006 9.07 14.12 16.27 27.53 0.40 0.82 2.44 0.10 2007 12.93 23.93 16.93 29.15 0.13 0.52 2.88 0.07 2008 18.67 32.53 16.93 28.65 0.00 0.00 2.55 0.09 Northeastern Naturalist 55 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 Growing-season precipitation over the study varied each year (2005 = 106.7 mm, 2006 = 577.9 mm, 2007 = 195.8 mm, 2008 = 404.1 mm) and was greatly reduced in 2005 and 2007, which likely influenced vegetation sampling results. Vegetation sampled in the 2008 growing season experienced near average precipitation and was therefore likely representative of an initial year of recovery from the cumulative influence of each treatment type. At the end of this study, foliar coverage of the woody functional group was significantly lower in the mowing treatment as compared to the grazing and control treatments (F2, 42 = 7.19, P = 0.01; Fig. 3). Additionally, bare ground was significantly higher and litter significantly lower in the grazing treatment as compared to the control and mowing treatments (bare ground: F2, 42 = 8.64, P = 0.01; litter: F2, 42 = 8.80, P = 0.01; Fig. 3). Graminoid foliar coverage was not significantly different between the grazing and mowing treatments in 2008, although mean coverage was higher in both treatments as compared to the control (F2, 42 = 2.69, P = 0.08). Overall, forb species Figure 2. Foliar percent coverage (total hits/total samples per plot *100) of each treatment (control, graze, mow) by management interest functional group before the start of this project (2005). Note that because more than 1 species representing a functional group may be present in a research unit, percent foliar coverage by a functional group can be larger than 100%. Bars are standard errors. Species composition of agricultural, sandplain grassland, and clonal/vining woody management interest functional groups for this study is presented in Table 2. Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 56 Vol. 24, Special Issue 8 increased in both the grazing and mowing treatments, although there was no significant difference between the treatments at the end of the study . Management interest functional groups Prior to the start of the study, vegetation sampling plots showed no significant difference in percent foliar coverage of the management interest functional groups’ species (Table 2) between the research units (Table 3) (Fig. 2). Post-treatment sampling in the first year of the study (2005) showed a significant reduction in the percent foliar coverage of each functional group in both the grazing and mowing treatments compared to the controls (agricultural: F2, 42 = 13.61, P = 0.01; sandplain: F2, 42 = 43.89, P = 0.01; clonal/vining woody: F2, 42 = 18.10, P = 0.01]. The shrub woody functional group showed no difference between treatments. The coverage of this functional group was negligible over the entire study site and did not significantly increase over the course of the study . Figure 3. Foliar percent coverage (total hits/total samples per plot *100) of each treatment (control, graze, mow) by management interest functional group in the recovery year of this project (2008). Note that because more than one species representing a functional group may be present in a research unit, percent foliar coverage by a functional group can be larger than 100%. Bars are standard errors; letters indicate significant differences based on ANOVA and Tukey’s HSD. Species composition of agricultural, sandplain grassland, and clonal/vining woody management interest functional groups for this study is presented in Table 2. Northeastern Naturalist 57 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 After 3 consecutive years of treatment (2008), we were able to detect a significant change in coverage of all species of management interest functional groups, with the exception of the shrub woody group. Percent foliar coverage of agricultural species significantly increased (F2, 42 = 8.38, P = 0.01) and clonal/vining woody species coverage significantly decreased (F2, 42 = 7.40, P = 0.02) in the grazing and mowing treatments as compared to controls (Fig. 3). The reduction in clonal/ vining woody species was primarily responsible for the reduction of the broader woody functional group within the mowing treatments, as discussed in the previous section. Sandplain grassland associate species were significantly reduced in the grazing treatments compared to the control and mowing treatments (F2, 42 = 7.96, P = 0.0; Fig. 3). Discussion Current grassland management practices on Nantucket focus on slowing establishment and decreasing cover of woody species that, through succession, will outcompete native grassland species (Dunwiddie 1989). Three consecutive years of sheep grazing or mowing, conducted multiple times during the summer growing season, resulted in significant shifts in vegetation community composition in a grassland previously managed with only dormant-season mowing. Overall, woody species cover was significantly reduced in the mow units, and clonal/vining woody species, which actively compete with native grasses and forbs, were reduced in both the mow and graze units. Although not statistically significant, both forb and graminoid species increased in the grazing and mowing treatments. This finding suggests that both grazing and mowing treatments resulted in shifts in the ecological trajectory towards a vegetation community containing more cover of early successional species and less cover of competitive woody species in just 3 years. Coastal sandplain grassland communities require periodic natural and/or anthropogenic disturbances in order to persist (Dunwiddie 1989). Prescribed fire and mowing are the most common techniques currently used, although sheep grazing has been identified as the primary historic post-colonial disturbance on Nantucket (Dunwiddie 1989). Both the type and seasonality of management are important factors influencing management outcomes. On Nantucket, mowing and prescribed fire management have typically been conducted in the spring and late autumn dormant seasons to limit smoke impacts and avoid affecting nesting birds and wildlife (K.C. Beattie, pers. observ.). In contrast, sheep grazing needs to occur during the growing season when live, palatable forage is present (late spring–late fall). Research conducted in grasslands on Nantucket suggests that growing-season mowing or burning can be more effective in reducing shrub cover and increasing frequency of graminoids and forbs compared to these same treatments applied during the dormant season (Dunwiddie 1998, Dunwiddie and Caljouw 1990, Dunwiddie et al. 1995). Our research results provide additional evidence that growing-season treatments may be more effective than dormant-season treatments in restoring sandplain grassland habitat. Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 58 Vol. 24, Special Issue 8 Sheep grazing differs from mowing in the way that vegetation is removed from a site and how the soil surface is affected. Mowing reduces all vegetation, regardless of species, to a uniform height, including new leaf and shoot growth as well as older woody stems (Huhta 2001). In contrast, grazing is more selective, as sheep show preferences or aversions to particular species based on their varying nutritional needs (Provenza 2003). Sheep tend to prefer newer growth because it is more digestible and nutritious (Launchbaugh and Walker 2006). Sheep hooves rework the litter and soil surface, opening up areas of bare soil and providing opportunities for increased graminoid seed germination (Silvertown and Smith 1988). Grazing removes standing vegetation through consumption, whereas mowing removes standing vegetation by dropping it to the soil surface to decompose. At Squam Farm, plant litter on the soil surface was significantly reduced and bare ground available for seedling germination was significantly increased within grazing treatments as compared to mowing treatments. While examining response to management at a broad functional group level can help illustrate overall successional trends (Hellstrom et al. 2003), analyzing results at a finer scale provides additional insight into the ecology of localized succession. Pre-treatment data show that prior to this study, the research units at Squam Farm contained a mosaic of desirable native grassland species, undesirable grasses and forbs associated with agriculture, and undesirable woody species capable of outcompeting sandplain grassland species in the absence of disturbance (Table 2). Clonal woody species such as Black Huckleberry and vining woody species such as Vitis labrusca (Fox Grape) can inhibit grassland restoration through competition with desired grasses and forbs (Harper 1995; Reiners 1965; J.M. Karberg, pers. observ.). Clonal and vining woody species typically respond positively to disturbance management, potentially creating conflicting management outcomes when promoting early successional vegetation communities (Dunwiddie et al. 1990, Harper 1995, Reiners 1965). At Squam Farm, we saw a reduction in the coverage of clonal/vining woody species in both the mowing and grazing treatments over the course of the study. This result suggests that mowing and grazing, either separately or potentially together, may be effective tools at reducing competitive woody cover in early successional community management. On the other hand, this study documented a decrease in sandplain grassland species in the grazing treatment and an increase in undesirable agricultural species in both the grazing and mowing treatment. Ecosystems that have a history of past agricultural use often have plant communities that remain altered for many decades by persistent non-native, agricultural plant species (Neill et al. 2007, Von Holle and Motzkin 2007), which can hinder establishment of desirable native sandplain grassland species (Neill et al. 2015). The agricultural species present at the site prior to treatments likely increased due to increased disturbance from both the grazing and mowing treatments. Previous studies have documented that grazing can result in the creation of gaps (patches of bare soil exposed to increased light) as the dominant vegetation is suppressed by consumption. These gaps are more likely to be colonized by stress-tolerant, Northeastern Naturalist 59 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 competitive ruderal species (Bullock et al. 2001, Smith and Rushton 1994). While bare ground provided suitable space for seed germination in the grazing treatments, it appears that agricultural species were able to establish and possibly outcompete sandplain species over the 4 years of the study. Reconstruction of palynological post-glacial vegetation communities using radiocarbon-dated sediment cores from Nantucket and elsewhere in New England show a rapid increase in pollen indicative of both native grass species and agricultural weeds immediately post-European settlement, coincident with increases in sheep grazing and other agricultural activities (Dunwiddie 1990, Foster and Motzkin 2003, Motzkin and Foster 2002). This trend continued through the late 19th century and then declined with reductions in historic grazing activity. These findings suggest that weedy agricultural species may have been a component of the successional trajectory that resulted in the establishment of current sandplain grassland habitats. During the height of sheep grazing activity on Nantucket, Foster and Motzkin (2003:139) reported that overgrazing led to a loss in soil and vegetation cover followed by wind erosion, resulting in a landscape “described as degraded, barren, rutted, eroded, and wasteland.” Thus, while soil disturbance was likely an integral component of the historic disturbance regimes, ultimately resulting in sandplain grassland establishment and expansion, palynological data suggest that these highly eroded and degraded soils may have initially been colonized by weedy agricultural plant species. Results from our study suggest that the use of sheep grazing alone may not be effective in restoring sandplain grassland habitat at sites where weedy agricultural species are already present or become introduced during management. Native seed availability in the soil seed bank has been identified as a limiting factor in the establishment of native species in sandplain grassland restoration efforts (Omand et al. 2014). In a former agricultural field on Martha’s Vineyard, MA, soil disturbance in the form of tilling combined with native seed addition increased native species presence and cover (Wheeler et al. 2015), although this technique did not eliminate the non-native agricultural species formerly dominant at the site. In our study, sheep grazing was effective in creating soil disturbance, which appears to be an integral component of grassland restoration. Combining sheep grazing with native seed addition could be an important next step towards understanding how grazing can be utilized as an effective grassland restoration tool. Additional research is needed to determine if grazing can successfully maintain this community type once restored. In this study, the amount of time that the sheep were grazed in each treatment unit was controlled by the study design, and grazing occurred during the later portion of the growing season. Neither of these conditions likely mimicked the historic grazing pressures that resulted in Nantucket’s current grassland communities, which were continuously grazed year-to-year over many decades and throughout the growing and non-growing seasons. While the trends seen in this study provide an indication of how grazing pressure can alter these early successional communities, our project was designed as a replicable study that controlled Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 60 Vol. 24, Special Issue 8 for as many variables as possible and therefore was not a simulation of historic, unmanaged grazing patterns. Future research is needed to examine the effects of an intensive, year-round grazing regime. Grazing that results in a mosaic of treatment and intensity may be more effective at maintaining the suite of native forb and graminoid species desirable within grassland communities. Responsible animal husbandry would not allow the use of animals today the way they might have been grazed historically. The restoration and maintenance of early successional grasslands is complicated and involves factors related to pre- and post-restoration vegetation composition, weather variability, and management technique and timing. In this study, sheep grazing was effective at decreasing coverage of competitive woody species and creating soil disturbance, which appear to be important first steps in grassland restoration. The observed increase in undesirable agricultural species within grazing and mowing treatments may be an inherent side effect of management that results in soil disturbance. An important next step towards overcoming weedy-species establishment and expansion could be the combined use of sheep grazing and native seed addition. Given the long-term, variable nature of the ecological disturbances that created these early successional communities, one management technique alone likely will not result in successful grassland restoration over a short period of time. We recommend that sheep grazing be more widely considered as an addition to the sandplain grassland restoration “tool box”. Acknowledgments We would not have been able to complete this project without the help and support of numerous individuals. The Nantucket Conservation Foundation, Lucy Dillon, the Boston Foundation, Mr. and Mrs. Ian MacKenzie, Gretchen Penrose, Robert McDaugh, Patty Gibian, the Good Samaritan Foundation, the Wolff Family Foundation, and an anonymous donor provided funding. Rachael Freeman, Kurt Schlimme, and Brooke Brewer contributed invaluable project development expertise. The following provided field work oversight and/ or sheep flock care and management: Rachael Freeman, Brooke Brewer, Nicole DuPont, Jessica Pykosz, Constance Helstosky, Pam Buckley, Tom Lennon, Chris Iller, Richard Mack, Donnie Mack, Tom Larrabee, Billy Coffin, Sarah Bois, Sarah Hinman, Offshore Animal Hospital, Dr. Maia Howard, Dr. Constance Breese, Dr. Scott White, Sarah Oktay, Sarah Flack, Phil Lindsay, and Andy Rice. Kurt Schlimme oversaw the initial year of data collection as part of his Master’s thesis research project. The New England Heritage Breeds Conservancy and the Massachusetts Audubon Society donated sheep to the project. Many thanks to numerous Nantucket Conservation Foundation board members, Science and Stewardship Department field assistants, and volunteers who provided project support and data collection assistance. Literature Cited Anderson, E.W. 1986. A guide for estimating cover. Rangelands 8:236–238. Barbour, H., T. Simmons, P.C. Swain, and H. Woolsey. 1998. Our irreplaceable heritage: Protecting biodiversity in Massachusetts. Natural Heritage and Endangered Species Program and Masschusetts Chapter of the Nature Conservancy, Boston, MA. Northeastern Naturalist 61 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 Bullock, J.M., B. Clear Hill, M.P. Dale, and J. Silvertown. 1994. An experimental study on the effects of sheep grazing on vegetation change in a species-poor grassland and the role of seedling recruitment into gaps. Journal of Applied Ecology 31:493–507. Bullock, J.M., B. Clear Hill, J. Silvertown, and M. Sutton. 1995. Gap colonization as the source of grassland community change: Effects of gap size and grazing on the rate and mode of colonization by different species. Oikos 72:273–282. Bullock, J.M., J. Franklin, M.J. Stevenson, J. Silvertown, S.J. Coulson, S.J. Gregory, and R. Tofts. 2001. A plant-trait analysis of responses to grazing in a long-term experiment. Journal of Applied Ecology 38:253–267. Carlson, L., M. Babione, P.J. Godfrey, and A. Fowler. 1991. Ecological survey of heathlands in Cape Cod National Seashore, MA. Botany Department, University of Massachusetts, Amherst, MA. Chapman, C.K., and C.R. Reid. 2004. Sheep and goats: Ecological tools for the 21st century. Fact Sheet. Utah State University Extension, Logan, Utah. Crevecoeur, J.H.S. 1957. Letters from an American Farmer. E.P. Dutton and Company, New York, NY. Dunwiddie, P.W. 1989. Forest and heath: The shaping of the vegetation on Nantucket Island. Journal of Forest History 33:126–133. Dunwiddie, P.W. 1990. Postglacial vegetation history of coastal islands in southeastern New England. National Geographic Research 6:178–195. Dunwiddie, P.W. 1997. Long-term effects of sheep grazing on coastal sandplain vegetation. Natural Areas Journal 17:261–264. Dunwiddie, P.W. 1998. Ecological management of sandplain grasslands and coastal heathlands in southeastern Massachusetts. Pp. 83–93, In T.L. Pruden and L.A. Brennan (Eds.). Fire in Ecosystem Management: Shifting the Paradigm from Suppression to Prescription. Tall Timbers Research Station Tallahassee, FL. Dunwiddie, P.W., and C. Caljouw. 1990. Prescribed burning and mowing of coastal heathlands and grasslands in Massachusetts. New York State Museum Bulletin 471:271–275. Dunwiddie, P.W., W.A. Patterson III, and R.E. Zaremba. 1995. Evaluating changes in vegetation from permanent plots: An example from sandplain grasslands in Massachusetts. Pp. 245–250, In T.B. Herman, S. Bondrup-Nielsen, J.H. Martin Willison and N.W.P. Munro (Eds.). Ecosystem Monitoring and Protected Areas. Science and Management of Protected Areas Association Wolfville, NS, Canada. Dunwiddie, P.W., R.E. Zaremba, and K.A. Harper. 1996. Classification of coastal heathlands and sandplain grasslands in Massachusetts. Rhodora 98:117–145. Foster, D.R., and G. Motzkin. 2003. Interpreting and conserving the openland habitats of coastal New England: Insights from landscape history. Forest Ecology and Management 185:127–150. Foster, D.R., and J.F. O’Keefe. 2000. New England Forests Through Time: Insights from the Harvard Forest Dioramas. Harvard Forest, Petersham, MA. Harper, K.A. 1995. Effect of expanding clones of Gaylussacia baccata (Black Huckleberry) on species composition in sandplain grassland on Nantucket Island, Massachusetts. Bulletin of the Torrey Botanical Club 122:124–133. Harshberger, J.W. 1917. American heaths and pine heaths. Brooklyn Botanic Garden Memoirs 1. Hellstrom, K., A.-P. Huhta, P. Rautio, J. Tuomi, J. Oksanen, and K. Laine. 2003. Use of sheep grazing in the restoration of semi-natural meadows in northern Finland. Applied Vegetation Science 6:45–52. Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 62 Vol. 24, Special Issue 8 Huhta, A. 2001. Restorative mowing on semi-natural grasslands: Community-level changes and species-level responses. University of Oulu, Oulu, Finland. 40 pp. IBM Corp. 2012. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY. Langlois, K.H., Jr. 1979. Soil survey of Nantucket County, Massachusetts. USDA Soil Conservation Service, Massachusetts Agricultural Experiment Station, Amherst, MA. Launchbaugh, K., and J. Walker. 2006. Targeted Grazing: A Natural Approach to Vegetation Management and Landscape Enhancement. Cottrell Printing, Centennial, CO. Lavorel, S., S. McIntyre, J. Landsberg, and T.D.A. Forbes. 1997. Plant functional classifications: From general groups to specific groups based on response to disturbance. TREE 12:474–478. Massachusetts Natural Heritage and Endangered Species Program (MANHESP). 2010. Sandplain grassland factsheet. Available online at http://www.mass.gov/eea/docs/dfg/nhesp/ natural-communities-facts/sandplain-grassland-factsheet.pdf. Accessed May 2016. MassGIS. 2003. MassGIS data download: NRCS-SSURGO-Certified soils. Available online at http://www.mass.gov/anf/research-and-tech/it-serv-and-support/applicationserv/ office-of-geographic-information-massgis/datalayers/download-elevation-contours- 15000.html. Accessed May 2016. MassGIS. 2012. MassGIS data download: Elevation contours (1:5,000). Available online at http://www.mass.gov/anf/research-and-tech/it-serv-and-support/application-serv/ office-of-geographic-information-massgis/datalayers/ftpsoi.html. Accessed May 2016. McIntyre, S., S. Lavorel, and R.M. Tremont. 1995. Plant life-history attributes: Their relationship to disturbance response in herbaceous vegetation. Journal of Ecology 83:31–44. McManis, D.A. 2010. Town Farms and Country Commons: Farming on Nantucket. Mill Hill Press, Nantucket MA. Motzkin, G., and D.R. Foster. 2002. Grasslands, heathlands, and shrublands in coastal New England: Historical interpretations and approaches to conservation. Journal of Biogeography 29:1569–1590. Neill, C. 2007. The challenge of managing disturbance regimes, terrestrial communities, and rare species in a suburbanized region: The northeastern US coastal sandplain. Biological Conservation 136:1–3. Neill, C., W.A. Patterson, and D.W. Crary. 2007. Responses of soil carbon, nitrogen, and cations to the frequency and seasonality of prescribed burning in a Cape Cod oak–pine forest. Forest Ecology and Management 250:234–243. Neill, C., M.M. Wheeler, E. Loucks, A. Weiler, B. Von Holle, M. Pelikan, and T. Chase. 2015. Influence of soil properties on coastal sandplain grassland establishment on former agricultural fields. Restoration Ecology 23:531–538. Neiring, W.A., and G.D. Dreyer. 1989. Effects of prescribed burning on Andropogon scoparius in postagricultural grasslands in Connecticut. American Midland Naturalist 122:88. Nantucket Historical Association (NHA). 2016. Aerial views of Nantucket from the collections of the Nantucket Historical Association. Available online at http://www.nha.org/ digitalexhibits/aerialviews/AP1940Web/pages/ap1940_142.htm. Accessed May 2016. National Oceanographic and Atmospheric Administration (NOAA). 2016. National Centers for Environmental Information: Climate data online. Available online at http://www. ncdc.noaa.gov/cdo-web/. Accessed 2016. Oldale, R.N. 1985. Geologic map of Nantucket and Nearby Islands, Massachusetts. Department of the Interior, US Geological Survey, Washington, DC. Olusuyi, S.A., and C.A. Raguse. 1968. Inclined-point estimation of species contributions to dry pasture matter production. Agron 60:441–442. Northeastern Naturalist 63 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 Omand, K.A., J.M. Karberg, K.C. Beattie, D.I. O’Dell, and R.S. Freeman. 2014. Soil seed bank in Nantucket’s early successional communitites: Implications for management. Natural Areas Journal 34:188–199. Owen, M.L. 1888. A Catalogue of Plants Growing without Cultivation in the County of Nantucket, Mass. Gazette Printing Company, Northampton, MA. Provenza, F.D. 2003. Foraging Behavior: Managing to Survive in a World of Change. USDANRCS Grazing Lands Technology Institute; Utah State University Department of Forest, Range, and Wildlife Servies; Utah Agricultural Experiment Station, Logan, UT. Reiners, W.A. 1965. Ecology of a heath-shrub synusia in the pine barrens of Long Island, New York. Bulletin of the Torrey Botanical Club 92:448–464. Rice, M.A. 1946. Trees and shrubs of Nantucket. Maria Mitchell Association, Nantucket, MA. 77 pp. Schlimme, K.S. 2006. Returning sheep to the island of Nantucket: A conservation grazing experiment at Squam Farm. Duke University, Durham, NC. Silvertown, J., and B.A. Smith. 1988. Gaps in the canopy: The missing dimension in vegetation dynamics. Vegetatio 77:57–60. Smith, R.S., and S.P. Rushton. 1994. The effects of grazing management on the vegetation of mesotrophic (meadow) grassland in northern England. Journal of Applied Ecology 31:13–24. Sorrie, B.A., and P.W. Dunwiddie. 1996. The vascular and non-vascular flora of Nantucket, Tuckernuck, and Muskeget Islands. Massachusetts Audubon Society, Massachusetts Natural Hertiage and Endangered Species Program, Nantucket Maria Mitchell Association, The Nature Conservancy, Nantucket, MA. Steel, J. 1999. Losing Ground: An Analysis of Recent Rates and Patterns of Development and Their Effects on Open Space in Massachusetts. Massachusetts Audubon Society, Lincoln, MA. Sternberg, M., M. Gutman, A. Pereolotsky, E.D. Ungar, and J. Kigel. 2000. Vegetation response to grazing management in a Mediterranean herbaceous community: A functional group approach Journal of Applied Ecology 37:224–237. Swain, P.C., and J.B. Kearsley. 2001. Classification of the natural communities of Massachusetts. Natural Heritage and Endangered Species Program; Massachusetts Division of Fisheries and Wildlife, Westborough, MA. Thoreau, H.D. 1962. The Journal of Henry D. Thoreau, Volumes I–VII (1837–October 1855). B. Torrey and F.H. Allen (Eds.). Dover Publications, Inc. New York, NY. 926 pp. Vickery, P.D., and P.W. Dunwiddie. 1997. Grasslands of Northeastern North America: Ecology and Conservation of Native and Agricultural Landscapes. Massachusetts Audubon Society, Lincoln, MA. Von Holle, B., and G. Motzkin. 2007. Historical land use and environmental determinants of nonnative plant distribution in coastal southern New England. Biological Conservation 136:33–43. Wheeler, M.M., C. Neill, E. Loucks, A. Weiler, B. Von Holle, M. Pelikan, and T. Chase. 2015. Vegetation removal and seed addition contribute to coastal sandplain grassland establishment on former agricultural fields. Restoration Ecology 23:539–547. Wilson, J.W. 1960. Inclined point quadrats. New Phytologist 59:1–8. Wilson, J.W. 1963. Estimation of foliage denseness and foliage angle by inclined point quadrats. Australian Journal of Botany 11:95–101. Winkworth, R.E. 1954. The use of point quadrats for the analysis of heathland. Australian Journal of Botany 3:68–81. Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 64 Vol. 24, Special Issue 8 Appendix 1. Scientific name and authority, common name, broad functional group designation, and family of plant species encountered in this project. Species name and authority Common name Functional group Family Achillea millefolium L. Common Yarrow Forb Asteraceae Antennaria spp. Pussytoes Forb Asteraceae Cirsium spp. Thistle Forb Asteraceae Crocanthemum dumosum Bickn. Bushy Frostweed Forb Cistaceae Euthamia caroliniana (L.) Greene ex Porter & Britton Slender Goldentop Forb Asteraceae Euthamia graminifolia (L.) Nutt. Grass-leaved Goldenrod Forb Asteraceae Fragaria virginiana Duchesne Wild Strawberry Forb Rosaceae Hieracium spp. Hawkweed Forb Asteraceae Hypericum perforatum L. Common St. John's-wort Forb Clusiaceae Hypochaeris radicata L. Hairy Cat’s ear Forb Asteraceae Lechea maritima Legg. ex Britton, Sterns & Poggenb. Beach Pinweed Forb Cistaceae Lechea spp. Pinweed Forb Cistaceae Lespedeza capitata Michx. Roundheaded Lespedeza Forb Fabaceae Leucanthemum vulgare Lam. Oxeye Daisy Forb Asteraceae Oxalis spp. Oxalis Forb Oxalidaceae Polygala polygama Walter Bitter Milkwort Forb Polygalaceae Potentilla canadensis L. Dwarf Cinquefoil Forb Rosaceae Potentilla simplex Michx. Spreading Cinquefoil Forb Rosaceae Pseudognaphalium obtusifolium (L.) Hilliard & B.L. Burtt Fragrant Everlasting Forb Asteraceae Rumex acetosella L. Red Sorrel Forb Polygonaceae Sisyrinchium fuscatum E.P. Bicknell Coastal Plain Blue-eyed Grass Forb Iridaceae Solidago rugosa Mill. Wrinkleleaf Goldenrod Forb Asteraceae Symphyotrichum dumosum (L.) G.L. Nesom Bushy Aster Forb Asteraceae Symphyotrichum patens (Aiton) G.L. Nesom Clasping-leaved Aster Forb Asteraceae Symphyotrichum undulatum (L.) G.L. Nesom Wavyleaf Aster Forb Asteraceae Viola sagittata Aiton Arrow-leaved Violet Forb Violaceae Seriocarpus asteroides (L.) Britton, Sterns & Poggenb. Toothed Whitetop Aster Ford Asteraceae Northeastern Naturalist 65 K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 Vol. 24, Special Issue 8 Species name and authority Common name Functional group Family Agrostis perennans (Walter) Tuck. Upland Bentgrass Graminoid Poaceae Agrostis stolonifera L. Creeping Bentgrass Graminoid Poaceae Anthoxanthum odoratum L. Sweet Vernalgrass Graminoid Poaceae Carex pensylvanica Lam. Pennsylvania Sedge Graminoid Cyperaceae Cyperus lupulinus (Spreng.) Marcks Great Plains flatsedge Graminoid Cyperaceae Danthonia spicata (L.) P.Beauv. Ex Roem. & Schult. Poverty Oatgrass Graminoid Poaceae Dichanthelium spp. (Hitchc. & Chase) Gould Panic Grass Graminoid Poaceae Festuca ovina L. Sheep Fescue Graminoid Poaceae Holcus lanatus L. Common Velvetgrass Graminoid Poaceae Juncus greenei Oakes & Tuck. Greene's Rush Graminoid Juncaceae Panicum virgatum L. Switchgrass Graminoid Poaceae Poa spp. Bluegrass Graminoid Poaceae Schizachyrium scoparium (Michx.) Nash Little Bluestem Graminoid Poaceae Cladonia spp. Hill ex P. Browne Reindeer Lichen Lichen Cladoniaceae Corylus cornuta Marshall Beaked Hazelnut Woody Corylaceae Gaylussacia baccata (Wangenh.) K. Koch Black Huckleberry Woody Ericaceae Ilex verticillata (L.) A. Gray Common Winterberry Woody Aquifoliaceae Lonicera japonica Thunb. Japanese Honeysuckle Woody Caprifoliaceae Morella caroliniensis (Mill.) Small Bayberry Woody Myricaceae Parthenocissus quinquefolia (L.) Planch. Virginia Creeper Woody Vitaceae Prunus serotina Ehrh. Black Cherry Woody Rosaceae Quercus ilicifolia Wangenh. Scrub Oak Woody Fagaceae Rhus copallinum L. Winged Sumac Woody Anacardiaceae Rosa virginiana Mill. Virginia Rose Woody Rosaceae Rubus flagellaris Willd. Northern Dewberry Woody Rosaceae Rubus hispidus L. Bristly Dewberry Woody Rosaceae Sassafras albidum (Nutt.) Nees Sassafras Woody Lauraceae Smilax glauca Walter Sawbrier Woody Smilacaceae Smilax rotundifolia L. Greenbrier Woody Smilacaceae Toxicodendron radicans (L.) Kuntze Poison Ivy Woody Anacardiaceae Northeastern Naturalist K.C. Beattie, J.M. Karberg, K.A. Omand, and D.I. O’Dell 2017 66 Vol. 24, Special Issue 8 Species name and authority Common name Functional group Family Vaccinium angustifolium Aiton Lowbush Blueberry Woody Ericaceae Vaccinium corymbosum L. Highbush Blueberry Woody Ericaceae Viburnum dentatum L. Arrowwood Woody Caprifoliaceae Vitis labrusca L. Fox Grape Woody Vitaceae