1.  Introduction

 Despite widespread alteration of the New England landscape since European settlement, there have been few major shifts in species distributions, but several notable changes in the relative abundance of tree taxa [1]. Various studies have collected data on the overstory throughout the region [2-4]. In one recent study conducted at the University of New Hampshire by [5], the overstory species reported were similar to those of the three islands of our study [6]. The species included Acer rubrum, Acer saccharum, Betula alleghaniensis, Betula lenta, Fagus grandifolia, Fraxinus americana, Pinus strobus, Quercus rubra, Tilia americana, and Tsuga canadensis. A few of these species, including Tsuga canadensis, have been important in the natural history of New Hampshire [5]. In this study, we continue to assess the influence of environmental variation and land-use history on forest composition and distribution on islands of central New Hampshire [6]. We focused on the vegetation of three of the 253 islands in Lake Winnipesaukee: Blueberry (BI), Hawk’s Nest (HNI), and Three Mile (TMI) islands. These three islands are either owned or managed by the Appalachian Mountain Club (AMC), which allowed the senior author, who is a member of the Club, easy access for sampling. Detailed site descriptions of the three islands are available in [6]. Residents of the Lake Winnipesaukee area began tracking “ice-out” on the Lake in 1887. The ice-out date has been observed and recorded for nearly 130 years. The setting of the exact date and time is non-scientific and is now determined by an observer in a small plane that flies over the lake several times a day. Records show the latest date for ice-out on May 12, 1888, while the earliest date occurred on March 18, 2016. Thus, these records provide documentation for an overall warming trend in central New Hampshire. 

In 1900, the AMC established a permanent camp on Three Mile Island [7]. TMI soil is strongly acidic, dominated by gravelly sand and gravelly muddy sand, low in essential nutrients and fairly homogenous in terms of moisture and bulk density [8]. This island has an elevation gradient of 50 feet, increasing as one moves to the interior. Shore areas have elevations of 470 feet above sea level, while the island’s center is 520 feet above sea level [9]. By the early 1970’s the Camp had grown and the managers’ realization that certain land uses needed to be restricted led to the implementation of a land use plan [38,39]. Following an adaptation from [12] four types of environments needed by humans, the Camp was divided into Protective, Productive, Urban, and Compromise zones [13]. Protective and Productive zones are exposed to relatively little human use, whereas Urban and Compromise zones support the majority of human activity. Lake islands are not typically considered “closed” ecosystems as compared to oceanic islands [14]. Lake islands, such as those of Lake Winnipesaukee, are generally closer to the mainland than are oceanic islands. As such their floras may be comparable to the surrounding mainland due to their close proximity and lack of definitive physical barriers separating them from the mainland flora. Despite this, they may still be subject to broad ecological principles that apply to oceanic islands [15]. Humans have often endangered certain plant species through different activities such as habitat destruction and unintentional species movement. In particular, these activities often lead to the introduction of nonnative plant species, which sometimes may threaten to outcompete or eliminate native vegetation. The introductions of such exotic species have led to concern about the impact that these nonnative species may have on the native flora and fauna of ecosystems [16,17]. This study was part of a continuing effort to document the flora of three islands which started in 1901 when the camp on TMI was founded, and the earliest floristic survey was made by landscaper [18]. The most extensive early compilation of vascular plant species on the island was made by [19] and included many of the notations and collections made by J.H. Emerton (in May 1906) and R.A. Ware (in July 1906), as well as by Pease. Other collections on the island included those of [20] and [13] undertook an extensive study that compared their collections with those of [18,19,20]. Many of these collections are preserved in two Massachusetts herbaria (NEBC at Harvard University and SCHN at Smith College). Floristic studies were also conducted on other islands in Lake Winnipesaukee, including Rattlesnake [21], Bear [22] and Timber Islands [14,23]. Sizes of these islands and the most recent number of species found on each are compared below (Table 1). We also looked at the relationships between the native and nonnative plant species on the three smallest islands. “Nonnative” species are those that had been introduced to the islands since 1901, even those that are native to the New England area. In 1901 Harlan Kelsey donated 481 plants of 60 different species to the TMI Camp, only 9 of which were native to the island. However, most of these introductions did not survive over the years, and by the mid-1980’s only 20% of Kelsey’s introductions were observed [13]. Species reported by [13] which had not been recorded on Pease’s list include: Aster nemoralis, Chenopodium album, Festuca elatior, Oxalis europaea, Phalaris arundinacea, and Stellaria media, all of which were found in the Urban Zone. The current investigation looked at the composition and abundance of the native and nonnative plant species historically since 1978 when ecological studies began [39]. Any invasive, nonnative, or threatened species was considered a “species of concern” [24]. This manuscript follows up on previous work described by [18,19] [13] and [6], and underscores the utility of long term studies in identifying trends in vegetation change and dynamics especially in the context of climate change. Distributions of tree and shrub species were generated for each plot in each of the four land-use zones. To provide a manageable dataset, we report here on numbers of adult woody species (trees and shrubs), and will report on seedlings and herbs in a separate manuscript. Each year of measurement used the bootstrap, sampling with replacement [25,26]. The resampling procedure involved pooling the data for each year and zone; for example, one data pool contained data from the protective plots assessed in 1978. Then N samples were randomly selected from and replaced to a given data pool, where N equals the number of circular plots within a given land use zone. The mean of the N samples was then calculated and recorded and the procedure repeated for each measurement year and land-use zone combination through 10,000 iterations using @Risk software (Palisade, Inc., Ithaca, NY), within Excel (Microsoft Inc., Redmond, WA). Distributions of tree and shrub species were modeled by bootstrapping data for each sampling period. The increase in total overstory species in the protective and compromise zones between 1978 and 1991 was significant, and was very likely the result of a major storm in December 1980 which opened up the canopy [6]. No other significant differences in the frequency distribution of total overstory species in any of the other land-use zones, over the periods 1978-1991 and 1991-2001, were observed [11] and 2001-2011 [27]. The 1980 storm affected eight of the sampling plots on the north end of TMI, one plot on the north end of HNI, and one plot on the northeastern end of BI. While individual trees have been struck by lightning or toppled by strong winds, no storm events as massive as the December 1980 storm have been experienced since then.

The following are the “species of concern” on TMI, HNI, and BI. The species listed as invasive in various databases are Berberis thunbergii [17], Poa compressa [17], and Robinia hispida [16]. The species that are not native to the islands but were introduced by Harlan Kelsey in 1901 are Halesia carolina and Rhododendron calendulaceum [52]. The rare, threatened and endangered species are Cypripedium arietinum [28], Rhododendron maximum [28], and Rhododendron viscosum [52]. The species that are of concern to island campers, but do not fall under the above categories are Apios americana and Desmodium perplexum [17,52]. We present results from plot-based ecological studies that quantify the frequency, distribution, density, and dominance of woody species listed in [13] floristic survey. Based on the theory of island biogeography [29,30,31] we predicted total species richness would be highest on the largest island (Three Mile), lowest on the smallest (Blueberry), and that species richness would increase over time [32]. Second, we expected to see a decrease over time in the importance of Fagus grandifolia and Tsuga canadensis, which were threatened in other places by two non-native invasive species: the hemlock wooly adelgid beetle to the south [33,34,35] and beech bark disease on the mainland [36,37]. Third, we predicted species richness would be higher in the Protective and Productive plots, and lower in the Urban and Compromise plots where most trails are located.

2.  Materials and Methods 

This study was a part of a regular ecological sampling of the vascular flora that has occurred on TMI, HNI and BI islands since 1978 [10]. Samplings occurred on the islands at roughly ten year intervals in 1978, 1991, and 2001 [6,32]. The fourth sampling occurred in the summer of 2011 from June 11 to July 2 and followed the protocol of the other samplings. In 1978, 25 permanently marked circular plots were randomly distributed across the three islands, and their numbers were assigned based on coordinates in a number grid [10]. Each circular plot was 34 m (111.5 ft) in diameter and was approximately 908 m² [39], and the overstory plants were sampled in the circular plots [10, 11], while 1m x 1m square quadrats nested within each larger plot were used for sampling understory plants [39]. Nomenclature followed [40] and [52]. In the summer of 2011, a team of volunteers was gathered to assist with the sampling. They were campers at TMI and had prior knowledge of the islands’ flora because of their experiences as professional foresters or volunteers with the New England Wildflower Society. Sampling consistency since 1978 has been provided by the senior author. The team sampled the understory first and then the overstory for each plot. Then the team surveyed the islands for any plant species of concern, and made notes on abundance and location for each species. Within the 25 larger plots, woody plants over 2.0 m in height were recorded as part of the “overstory,” which included both tree and shrub layers [6]. The number of individuals (density), the number of plots (frequency), and the size and/or canopy position (dominance) of each species were recorded [6]. Two different methods were used to calculate relative dominance. The first method used a rating system of canopy position developed by Smith [6,41,42] and the second measured diameter at breast height (dbh) [43]. Individual trees and shrubs were assigned to one of four categories descriptive of their relative vertical position in the forest, including: overtopped (formerly “suppressed” in Smith 1962), intermediate, codominant and dominant, respectively [32,39,42]. 

Dominant trees were those having crowns extending above the level of general crown cover receiving full light from above and partly from the sides. Co-dominant trees had medium-sized crowns and received little light from the sides. Intermediate trees were shorter than those in the two preceding classes; receiving a little direct light from above but none from the sides. Overtopped (e.g. short trees, saplings, and tall shrubs) had crowns entirely below the general level of the crown cover [6,42]. From these data, a mean was calculated by dividing the four summed ratings for each species by the number of individuals of the species in each plot. Knowing the mean canopy class has proven to be biologically meaningful in studying natural forests where competition is fierce for trees growing on gravelly sandy substrates such as these islands. For the 2011 sampling, diameter at breast height (dbh) was also recorded for the trees on TMI and HNI that were larger than 2.5 cm dbh. but not at BI due to the presence of bald eagles nesting on the island. After each tree was identified and recorded, a piece of visible tape was placed on the tree to ensure that it was not recorded again. As had been done in the 1978, 1991, and 2001 samplings, importance values were calculated for each species. The importance value of a species is a relative quantitative measurement that combines density, dominance, and frequency. These parameters’ influence can vary in calculating importance values, where one parameter such as frequency may be higher than density. Using [44] method, importance values are also influenced by the number of species, so if only one species is present, its importance value would be 300. In general, an importance value represents the prominence of a species in an ecosystem [45]. Importance values of overstory species were based on relative frequency, relative dominance and relative density [44]. Note that in the following formulas, “q” represents one species. Relative dominance was calculated using the following formula: (rating species q/ ratings of all species) x 100 or (basal area of species q/ basal area of all species) x 100. In an effort to obtain the most accurate data for species dominance, the 2011 sampling plan had been to measure the dbh of each woody specimen. However, NH Audubon would only allow us to sample on BI if we agreed to move quickly on and off the island. Thus, in consideration of the bald eagle eaglets, the team agreed to use the faster “rating” approach. Relative density was calculated using the following formula: (number of individual species q)/ (total number of individuals of all species) x 100. Relative frequency was calculated using the following formula: ((number of plots species q was found in/total frequency of all species) * 100). Then, the importance values were calculated from the sum of the relative density, relative dominance and relative frequency. Lastly, the prominent species in the overstory of each island were determined from the importance value data. The dbhs recorded on Three Mile and Hawk’s Nest islands were converted into basal areas using the following formula: Basal area = (DBH²*0.7458)/10000 (cm²/m²) [43], and the basal areas were then summed for each species in the plots. The dbhs from 2011 were summed for HNI and TMI and these data were used to calculate relative dominance. Each calculation of relative dominance was used to determine a second set of importance values. These importance values were compared with the set of 2011 importance values calculated from ratings to demonstrate any differences in the overstory sampling methods. Shannon’s Diversity index was calculated to determine the diversity of the plots and the islands. The Shannon’s Diversity index took into account both the species richness and evenness in the sampling year. The index was calculated using the following formula: H’ = -Σ^S_i=1(p_i*ln(p_i)) where p_i is the proportion of individuals of a species (number of individuals/total # of individuals in the sampling) and S is the total number of species in the sampling. The closer the index is to ln(S), the more even the sampling. Species evenness was calculated using the formula E = H’/ln(S) where H’ is the Shannon’s Diversity Index and S is the number of species from that sample. From these values, the diversity and evenness were compared across the four years of sampling for the islands’ overstory.

3.  Results

3.1.  Three Mile Island

Since 1978, the species with the highest importance values (IV) on TMI were (Table 3, Figure 1): Acer pensylvanicum, Acer rubrum, Fagus grandifolia, Hamamelis virginiana, Pinus strobus, Quercus rubra, and Tsuga canadensis. Whether ratings or basal areas were used to calculate dominance in 2011, these same seven species are most important (Table 2). The density of Acer pensylvanicum, Fagus grandifolia, Hamamelis virginiana, and Tsuga canadensis increased over the 33year sampling (Figure 2). By 2011, Acer pensylvanicum, Acer rubrum, Pinus strobus, Quercus rubra, and Tsuga canadensis appeared in all 19 plots (Figure 3). In 2011, Quercus rubra had the highest canopy position and size in the overstory (Figure 4). Species of concern were only encountered on TMI (Tables 2,3). While Robinia hispida has been found in the plots during three samplings, its mean importance value is 4. Halesia Carolina was sampled in the overstory in 1991, but not since (Table 3), and Rhododendron maximum was only sampled in 2011. The understory species of concern will be discussed in a separate manuscript.

3.2.  Hawk’s Nest Island

Since 1978, the species with the highest importance values on Hawk’s Nest Island were (Table 5, Figure 5): Alnus incana ssp. rugosa, Betula populifolia, Pinus resinosa, Pinus strobus, Quercus rubra, and Tsuga canadensis. Whether ratings or basal areas were used to calculate dominance in 2011, these same six species are most important (Table 4). The density of Betula populifolia, Quercus rubra, and Tsuga canadensis increased over the 33year sampling (Figure 6). By 2011, Pinus resinosa, Pinus strobus, Quercus rubra, and Tsuga canadensis appeared in all three plots (Figure 7). In 2011, the two pines had the highest canopy position and size in the overstory (Figure 8).

3.3.  Blueberry Island

Since 1978, the species with high importance values on BI were (Table 7, Figure 9): Acer rubrum, Betula papyrifera, Betula populifolia, Ilex mucronata, Nyssa sylvatica, Picea rubens, Pinus strobus, Tsuga canadensis, and blueberry bushes. Undoubtedly, both species of blueberry found in 2011 (Vaccinium corymbosum and Vaccinium fuscatum) had been present throughout the 33year sampling, but they had not previously produced flowers and fruits at the same time. In 2011 the most dominant species on BI were Ilex mucronata, Vaccinium corymbosum, and Vaccinium fuscatum (Table 6). The density of Ilex mucronata, Vaccinium corymbosum, and Vaccinium fuscatum were each over 800 individuals in 2011 (Figure 10). In 2011, Acer rubrum, Nyssa sylvatica, Pinus strobus, Tsuga canadensis, and Vaccinium fuscatum appeared in all 3 plots (Figure 11). In 2011, Picea rubens had the highest canopy position and size in the overstory (Figure 12).

3.4.  Richness, Evenness, and Diversity

Over the 33 year sampling period, 39 species were sampled in the overstory (Table 8). Throughout the sampling, evenness and diversity remained relatively constant, while species richness generally increased (Table 8). On TMI, species richness increased significantly from 1978 to 1991 and remained relatively constant in the later samplings (Figure 13). Species evenness decreased from 1978 to 1991 and remained constant in the later samplings (Figure 14). However, the Shannon diversity index remained constant throughout the four samplings (Figure 15). On HNI, species richness increased from 12 species to 18 from 1978 to 2011 (Figure 13). Species evenness and Shannon’s diversity index increased through the sampling years (Figures 14,15). On BI, species richness increased through the years (Figure 13). However, species evenness decreased through the sampling years (Figure 14). Similarly to TMI, the BI Shannon diversity index remained relatively constant through the years, but it was highest during the 2001 sampling (Figure 15). Overall, the three islands saw a significant increase in total species richness from 1978 to 1991, though it remained constant in 2001 and 2011 (Figure 13).

When mean species richness is examined by zone, it was highest in the Urban zone in 1978, but increased in all four zones over the 33year sampling period, and was highest in the Protective zone in 2011 (Table 11). Thus, our third hypothesis was not supported. Species evenness did not change much among the four samplings (Figure 14). Lastly, Shannon’s diversity on TMI and BI increased from 1978 to 1991 but decreased slightly from 2001 to 2011, while it increased during each sampling on HNI (Figure 15). On TMI, the following species were absent from the plots in 2011 but not 2001: Ilex mucronata, Lyonia ligustrina, Populus tremuloides, Quercus alba, and Viburnum nudum var. cassinoides. The following species were present in the plots in 2011 but not 2001: Rhododendron maximum, Vaccinium fuscatum, and Viburnum lentago. On HNI, Gaylussacia baccata was absent in 2011 but not 2001. The following species were present in 2011 but not 2001: Nyssa sylvatica, Rhododendron canadense, Rosa palustris, and Vaccinium corymbosum. On BI, the following species were absent in 2011 but not 2001: Myrica gale, Populus grandidentata, Prunus pensylvanica, and P. serotina. The following species were present in 2011 but not 2001: Fraxinus nigra, Populus tremuloides, Quercus rubra, Rosa palustris, Vaccinium fuscatum, and Viburnum lentago. Even though some species were not located in the plots in 2011, that does not mean that they were not present in other locations on these islands. Throughout the years of sampling, the overstory increased in the number of individuals and in total ratings on all three islands (See Tables 8-10). In particular, the overstory on BI increased significantly from 1978 to 1991 in both the number of individuals and total ratings (Tables 9 and 10). Also, the BI increases in overstory individuals from 1991 to 2001 were relatively the same as from 2001 to 2011 (Table 9). It should be noted that while numbers of species in various plots may have increased over the 33 years of sampling, that no new overstory species have been found since the collections in the early 1900s [18,19].

3.5.  Synthesis

Throughout the sampling period, the three islands can be characterized as dominated by red maple (Acer rubrum), paper birch (Betula papyrifera), American witchhazel (Hamamelis virginiana), eastern white pine (Pinus strobus), and eastern hemlock (Tsuga canadensis). Over time, Hawk’s Nest Island increased in species richness, species evenness, and Shannon’s Diversity (Tables 8-10), which may be attributed to this island’s designation as falling entirely with in a Protective zone. Shrubs in the genera Gaylussacia (huckleberry), Ilex (holly), Lyonia (maleberry), Vaccinium (blueberry), and Viburnum (arrowwood) have become established along each island’s shoreline, exhibiting higher importance values today than at the start of the ecological sampling in 1978.

4.  Discussion

4.1.  Succession

Since the earliest records for BI, HNI, and TMI date from the 1890s, we assume that the seeds of trees noted [46] in presettlement New Hampshire (Fagus grandifolia, Tsuga canadensis, and Pinus spp.) may have been moved by wind or mammals onto the islands. Seeds of maple, birch, and conifers are wind dispersed and with the mainland less than 0.5 miles (1 km) away from the islands, the mainland could have provided the seed source for these species. Whereas records from the 1890s [7] indicate that TMI was covered by Betula spp. and Populus spp., and we assume the same was true for HNI and BI, by the early 2000s the most frequently encountered tree species were A. rubrum, F. grandifolia, P. strobus, Q. rubra, and T. canadensis. By 2011, both B. papyrifera and B. populifolia were present on TMI, HNI and BI, but were not as frequent as in the past [6]. The sampling of the three islands on Lake Winnipesaukee is part of ongoing floristic and ecological studies [6,18,19,20,27,39]. As of this writing, healthy populations of Tsuga canadensis have been found on all three islands, undoubtedly due to the fact that hemlock wooly adelgid has not moved north to Lake Winnipesaukee. In spite of the infestation of beech bark disease on TMI, density of Fagus grandifolia has increased over the study period; F. grandifolia has not been found on HNI, but the one lone F. grandifolia on BI continues to survive. In 2011, many of TMI’s dominant species overlapped with those found in the study of [5]. These species include Acer rubrum, Fagus grandifolia, Pinus strobus, Quercus rubra and Tsuga canadensis. Only a few dominant species on HNI and BI overlap with the Asahina study. These species are Pinus strobus, Quercus rubra, and Tsuga canadensis. Similarly, several woody species were found on both TMI and Timber Island [14]. These species include Acer pensylvanicum, Acer rubrum, Acer saccharum, Betula alleghaniensis, Betula papyrifera, Fagus grandifolia, Gaylussacia baccata, Hamamelis virginiana, Ostrya virginiana, Pinus resinosa, Pinus strobus, Quercus rubra, Tsuga canadensis, Vaccinium angustifolium, and Vaccinium corymbosum.

4.2.  Role of Large Grazing Herbivores

Application of the Simple Matching Index to the island floras of TMI and Timber Island produced a value of 65.59% [14]. This high similarity value is striking given the difference in human traffic on the two islands. Timber Island (135 acres) is the largest undeveloped island in Lake Winnipesaukee and supports a large deer population, while TMI (43 acres) experiences regular human disturbance from early June through late September when various AMC events are underway. [6] suggested that this heavy human activity had protected TMI from deer grazing throughout the first 23 years of vegetation sampling. However, a pregnant doe and a yearling either walked on the ice or swam over to TMI during winter/spring 2015, and two fawns were sighted with their mother and the yearling throughout that summer. The doe was not threatened by TMI campers, and at least two does were sighted on numerous occasions throughout the Urban and Compromise zones during summer 2016. The role of large grazing herbivores on TMI is already of interest to biologists, staff, and campers at the AMC camp, and the extent of deer browse is being documented.

4.3.  Role of Shrubs and Short Trees

All three islands had shrubby understory species become more important throughout the course of the study. Five out of the 25 plots sampled on BI, HNI, and TMI are shoreline plots, bordered by the Lake. It is in these plots that woody shrubs have proliferated, and the preponderance of Gaylussacia baccata, Ilex mucronata, Ilex verticillata, Vaccinium corymbosum, and Vaccinium fuscatum have been documented. Undoubtedly the easy access to sunlight has allowed these shrubs to dominate the shorelines of the three islands. In the center of the island, two TMI “overtopped” tree species, Acer pensylvanicum and Hamamelis virginiana, increased in importance throughout the study. One explanation is that the lowered density of large woody trees like Pinus strobus, Quercus rubra, and Tsuga canadensis after the 1980 blowdown opened the canopy for smaller tree and shrub species. As had been predicted [6] the shorter woody species (both trees and shrubs) exhibiting high importance values in 2001 on the islands’ edges have continued their importance there.

4.4.  Land Use on Each Island

This study emphasized the value of an ecological land use plan for monitoring the natural habitats of three small islands. After the extensive floristic survey which occurred in the late 1970s and early to mid 1980s [13] the land use plan was made more restrictive with the re-discovery of Cypripedium arietinum in the middle of the Productive Zone. When this rediscovery was brought to the attention of the volunteer management committee, the area was designated Protective, a designation which continues today. Unfortunately, C. arietinum has not been sighted since 1985. Three TMI plots were established in wetland areas: these support woody hydrophytes such as Cephalanthus occidentalis and Alnus rugosa. All three of these plots are designated as part of the Protective zone, and have been accumulating organic matter in the three lowest inland areas on TMI over at least the last 100 years. [47] studied over 20 wetlands across the globe and found that most wetlands are net carbon sinks. Thus, Mitsch et al. suggest that humans should endeavor to protect existing wetlands, since the older they become, the more carbon they can sequester. TMI is the only island that is large enough to support all four (Urban, Compromise, Productive, and Protective) land use zones [13]. HNI is completely zoned Protective, while BI is zoned Protective on the north and south ends of the island, but is zoned Compromise in the island’s center. Mean overstory species richness in plots of the Protective zone increased by 4.2 species during the 33year sampling period (Table 11). Enacting long-term land use plans may play useful roles in assisting managers to protect natural habitats in other areas throughout the world.

4.5.  Biodiversity

The earth is in the midst of a biodiversity crisis, and projections indicate continuing and accelerating rates of global changes [48]. [49] estimate that land use and related pressures have already reduced local biodiversity intactness across 58.1% of the world’s land surface, where 71.4% of the human population live. Many ecosystem services are underpinned by biodiversity, but globally there is currently a lack of coordinated action to halt biodiversity declines [50]. In general, the compositions in the permanent plots of TMI, HNI, and BI demonstrated several changes in vegetation over time. During the 33year sampling period, overstory species richness increased on each island. By 2011, TMI supported more species than BI and HNI probably due to its larger size. However, the small BI supported more species than HNI undoubtedly because the density of shrub growth on each end of BI allowed protection for dense tree growth. At this time, no changes in overstory species can be attributed directly to climate change. However, as temperatures warm, fires associated with lightning strikes may intensify. The three islands studied represent a microcosm of species and natural habitats found in the Lake Winnipesaukee, NH, region. The main difference between the islands and the mainland is the intensity of development on the mainland, especially along the shore. By establishing a camp on TMI in 1900, the AMC demonstrated its continuing commitment to sustaining and protecting the natural vegetation of the Lakes/White Mountain region. [51] examine global sustainability of marine fisheries, and note the importance of human rewards for maintaining biodiversity. They report that getting incentives right matters, and suggest that the ways in which these incentives can shift specific feedbacks in social ecological systems hold promise for conservation and management efforts in the ocean. A similar examination of incentives for management of northern forested ecosystems could prove beneficial.

5.   Acknowledgments

We thank the TMI Camp Committee for the opportunity to assist with the ongoing floristic research on the Lake Winnipesaukee islands. We would like to thank all the following volunteers who helped us sample the plots on TMI, HNI, and BNI in 2011: Jeanne Albert, Betsy Atkins, Carolynn Ernst, Philip Fitz, Marjorie Freeman, Nancy Grant, Joan Gulovsen, Lesley Harrington, Betsy Johnson, Carolyn Kasper, Anne Latchis, Joan Liehe, Tersh Palmer, Heather Pembrook, and Holly Tanguay. We also would like to thank the Three Mile Island Camp Staff and the Appalachian Mountain Club for providing us food and shelter during the samplings. C. John Burk, William Clapham, Heather Pembrook, and two anonymous reviewers provided helpful comments on various versions of the manuscript.

Figure 9: Importance values for ten trees and shrub species on Blueberry Island.

Figure 15: Overstory Shannon’s Diversity Index of TMI, HNI and BI in the four years of plot sampling.

Table 1: Lake Winnipesaukee islands where flora was surveyed. This table includes the most recent number of species and sizes of the islands in hectares. Bear, Timber, and Rattlesnake islands’ data were summarized from [14]. Three Mile, Hawk’s Nest (HNI) and Blueberry Islands (BI) were summarized from the 2011 sampling.

Table 2: 2011 Sampling Data for trees and shrub species of Three Mile Island.

Table 3: Importance values (IV) for tree and shrub species on Three Mile Island.

Table 4: 2011 sampling data for tree and shrub species of Hawk's Nest Island.

Table 5: Importance values (IV) for woody tree and shrub species on Hawk's Nest Island.

Table 6: 2011 sampling data for woody tree and shrub species of Blueberry Island.

Table 7: Importance values (IV) for woody trees and shrub species on Blueberry Island.

Table 11: Mean species richness in zones of overstory plots on TMI, HNI, and BI.

1. Hall B, Motzkin G,  Foster D, Syfert M, Burk JS (2002) Three hundred years of forest and land-use change in Massachusetts, USA. J Biogeogr 29: 1319-1335.
2. McIntosh RP (1972) Forests of the Catskill Mountains, New York. Ecol Monogr 2: 143-161.
3. Siccama TG (1971) Presettlement and present forest vegetation in Northern Vermont with special reference to Chittenden County. Amer Midl Naturalist 85: 153-172.
4. Whitney GG (1990) The history and status of the hemlock hardwood forests of the Allegheny Plateau. J Ecol 78: 443-458.
5. Asahina HTD, Lee TD, Eckert RT (2014) Disturbance-mediated dynamics of mid-tolerant hardwoods in an old pine-hemlock-hardwood forest, New Hampshire, USA. Rhodora 116: 125-147.
6. Holland MM, Clapham WM (2012) Vegetation changes in temperate forested island communities in Lake Winnipesaukee, New Hampshire, USA. Rhodora 114: 383-405.
7. Rogovin LS (1983) A history of Three Mile Island Camp. Published by Three Mile Island Camp, Meredith, NH.
8. O’Sullivan D (1981) Chemical and physical analyses of soil from Three Mile Island in Lake Winnipesaukee, NH. Senior seminar thesis, Department of Biology, College of New Rochelle, New Rochelle, NY.
9. Stone L (1972) Map of Three Mile Island. TMI Appalachian Mountain Club Camp. Laconia, NH.
10. Holland, MM, WM Clapham, and Maciejowski JD (1982) Volunteer success in natural resource management at an Appalachian Mountain Club Camp in New Hampshire, pp. 162-170. In: A. A. Merrill, ed, Proc. Volunteers in the Backcountry Conference, U.S. Forest Service, Durham, NH.
11. Clapham W, Holland MM, Fedders DJ (2009) Relationship of biodiversity and land use in forested plots in New Hampshire, USA. Poster #52-121, session: Temperate Forest Habitats, ESA Annual meeting, Albuquerque, NM. Ecological Society of America, Washington, DC.
12. Odum EP (1969) Strategy of ecosystem development. Science 164: 262-270.
13. Holland MM, Sorrie B (1989) Floristic dynamics of a small island complex in Lake Winnipesaukee, New Hampshire. Rhodora 91: 315-338.
14. Bradley AF, Crow G (2010) The Flora and Vegetation of Timber Island, Lake Winnipesaukee, New Hampshire, USA. Rhodora 112: 156-190.
15. Powledge F (2003) Island Biogeography’s lasting impact. BioScience 53: 1032-1038.
16. Invasive Plant Atlas (2010) Invasive plant atlas of the United States. Center for Invasive Species and Ecosystem Health, University of Georgia, Tifton, GA.
17. IPANE: Invasive Plant Atlas of New England (2011) Dept. Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT.
18. Kelsey HP (1902) Reports of the Councilors for the autumn of 1901: Natural History. Appalachia 10: 75-79.
19. Pease AS (1911) List of plants on Three Mile Island: Pteridophyta and Spermatophyta. Appalachia 12: 266-276.
20. Hartmann E (1941) The flora of Three Mile Island. Appalachia 23: 560-563.
21. Berry WH (1966) A Floristic Study of Rattlesnake Island, Lake Winnipesaukee, New Hampshire. M.S. Thesis, University of New Hampshire, Durham, NH.
22. Jackson, N. J. 1969. A Floristic Study of Bear Island, Lake Winnipesaukee, New Hampshire. M.S. Thesis, University of New Hampshire, Durham, NH.
23. Bradley AF (2005) The Flora and Vegetation of Timber Island, Lake Winnipesaukee, New Hampshire. M.S. Thesis, University of New Hampshire, Durham, NH.
24. Winkler MG (2012) Survey of the native and nonnative vascular plant species of three islands in Lake Winnipesaukee, New Hampshire. M.S. Thesis, University of Mississippi, University, MS.
25. Efron B, Tibshirani RJ (1994) An introduction to the bootstrap. Chapman and Hall/CRC, New York, NY: 456.
26. Efron B, Tibshirani RJ (1998) An introduction to the bootstrap. Monographs on Statistics and Applied Probability 57. Chapman and Hall/CRC, Boca Raton, FL.
27. Holland, MM, Clapham, WM, Winkler M (2013) Maturation of forested island vegetation in Lake Winnipesaukee, NH. Poster #52-121, session: Temperate Forest Habitats, ESA Annual Meeting, Minneapolis, MN. Ecological Society of America, Washington, DC.
28. New Hampshire Heritage (2011) Rare plant list for New Hampshire. New Hampshire Natural Heritage Bureau, Division of Forests and Lands, Concord, NH.
29. MacArthur R, Wilson EO (1967) The Importance of Islands. The Theory of Island Biogeography. Princeton Univ. Press, Princeton, NJ.
30. Adserson H (1995) Research on Islands: Classic, Recent, and Prospective Approaches Islands: 7-21.
31. Vitousek PM, Loope LL, Adserson H (1995) Islands: Biological Diversity and Ecosystem Function. New York: Springer-Verlag. 238.
32. Briggs F, Davis, T, Holland, MM (2008) Assessment of Vegetation Changes at Lake Winnipesaukee in New Hampshire. Sigma Xi Annual Poster Conference at Univ. Mississippi, University, MS. Sigma Xi, The Scientific Research Society, Research Triangle Park, NC.
33. Orwig DA, Foster DR (2002) Stand, landscape, and ecosystem analyses of Hemlock Woolly Adelgid outbreaks in southern New England: an overview. J Biogeogr 10: 1475-1487.
34. USDA Forest Service. (2005) a. USDA Forest Service, Northeastern area, Newtown Square, PA.
35. USDA Forest Service (2005) b. Pest alert: Hemlock Wooly Adelgid. USDA Forest Service, Northeastern area, Newtown Square, PA.
36. Gilman EF, Watson DG (1993) Fagus grandifolia American Beech. Fact Sheet ST-243. Environmental Horticulture Dept. Florida Cooperative Extension Service, Inst. Food and Agricultural Sciences. Univ. Florida, Gainesville, FL.
37. Houston DR (1994) Major new tree epidemics: Beech bark disease. Annual Rev Phytopathol 32: 75-87.
38. Clapham, W, Holland, MM, Maciejowski JD (2010) Botanical surveys reveal effects of land-use plans at Three Mile Island Camp. Appalachia 62: 138-140.
39. Maciejowski JW, Clapham, Holland MM (1981) Environmental assessment begins at Three Mile Island Camp. Appalachia 43: 137-142.
40. Haines A (2011) Flora Novae Angliae. Yale University Press, New Haven, CT.
41. Smith D (1962) The Practice of Silviculture. John Wiley and Sons, Inc, New York, NY.
42. Smith D (1986) The Practice of Silviculture. 8th ed. John Wiley and Sons, Inc, New York, NY.
43. Brewer R, McCann MT (1982) Laboratory and Field Manual of Ecology. Saunders College Publishing, Philadelphia, PA.
44. Curtis JT (1959) The vegetation of Wisconsin. An ordination of plant communities. University of Wisconsin Press, Madison, WI.
45. Mueller-Dombois D, Ellenberg H (2002) Aims and Methods of Vegetation Ecology. The Blackburn Press, Caldwell, NJ.
46. Cogbill CV, Burk JS, Motzkin G (2002) The forests of presettlement New England, USA: Spatial and compositional patterns based on town proprietor surveys. J Biogeogr 29: 1279-1304.
47. Mitsch WJ, Bernal B, Nahlik AM, Mander U, Zhang L, et al. (2013) Wetlands, carbon, and climate change. Landscape Ecology 28: 583-597.
48. Sorte CJB, Davidson VE, Franklin MC, Benes KM, Doellman MM, et al. (2016) Long-term declines in an intertidal foundation species parallel shifts in community composition. Global Change Biology. Doi: 10.1111/gcb.13425.
49. Newbold T, Hudson LN, Arnell AP, Contu S, De Palma A, et al. (2016) Has land use pushed terrestrial biodiversity beyond the planetary boundary? A global assessment. Science 353: 288-291.
50. Tittensor DP, Walpole M, Hill SLL, Boyce DG, Britten GL, et al. (2014) A mid-term analysis of progress toward international biodiversity targets. Science 345: 241-244.
51. Lubchenco J, Cerny-Chipman EB, Reimer JN, Levin SA (2016) The right incentives enable ocean sustainability successes and provide hope for the future. Pnas.1604982113. 113: 14507-14514.
52. USDA NRCS (2001-2016) The Plants Database. National Plant Data Center, Baton Rouge, LA. Website Accessed Sep 2016.