| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
(American Journal of Botany. 2008;95:943-953.) doi: 10.3732/ajb.0800013 © 2008 Botanical Society of America, Inc. |
What's this? |
Ecology |
2 Harvard Forest, Harvard University, 324 North Main Street, Petersham, Massachusetts 01366-0068 USA 3 Florida International University 11200 SW 8th Street, Miami, Florida 33199 USA
Received for publication 13 January 2008. Accepted for publication 28 May 2008.
ABSTRACT
Comparative examination of a large sample of plant species can reveal important aspects of life history that influence the ecology and distribution of taxa and their vulnerability to local extinction. We investigated whether functional groups of 71 rare plant species with contrasting life history traits differed in terms of population losses over time, regional range contraction, and range-wide levels of imperilment. Using town-level occurrence data from herbaria and Natural Heritage Program databases, we characterized species’ extents of occurrence as 
-hulls encompassing the centroids of New England towns that contained well-documented populations of these rare taxa. Family affiliation was used as a covariate in analyses to reduce phylogenetic bias. Disparate functional groups of plants differed both in proportions of populations lost and declines in range areas over time, with insect-pollinated taxa, upland (vs. wetland) taxa, species with localized seed dispersal modes, and taxa reaching their northern range boundary in New England significantly more imperiled than other functional groups. These techniques permit a broad comparative assessment of the distribution of large numbers of plant taxa, so that we can identify several functional groups that warrant more concerted conservation attention.
Key Words: biogeography • functional groups • GIS • guilds • herbarium collections • rarity • species ranges
Plant ecologists have long theorized that the distribution and abundance of plants result in part from critical aspects of plant life history, such as dispersal capacity, pollination syndrome, and annual or perennial life span. Many studies have compared and contrasted rare and common species to understand whether particular traits tend to predispose plants toward rarity (Harper, 1979
; Kelly and Woodward, 1996; Kunin and Gaston, 1997; Hegde and Ellstrand, 1999; Cadotte and Lovett-Doust, 2002; Murray et al., 2002). Given that a large proportion of the world’s flora is in decline (Pitman and Jorgensen, 2002), it is critical to document the current status, distribution, and ecological traits of rare plant species to design efficient, large-scale conservation strategies that successfully target the most extinction-prone taxa and that potentially protect several species at once.
Species can be ranked quantitatively in terms of imperilment based on the proportion of the total number of constituent populations that have declined or disappeared over time. Data on the locations and temporal dynamics of rare plant populations frequently come from herbarium collections and information gathered during targeted botanical forays (Shaffer et al., 1998). These voucher or atlas data are often incomplete, however, and yield imprecise information on localities and population vigor. This vagueness can impede efforts to relocate historical populations and to ascertain whether populations are truly extinct or merely undetected. The quality and intensity of data collection and maintenance also vary across state and national boundaries, complicating the task of conducting regional or rangewide analyses (Rodrigues and Gaston, 2002). Likewise, botanical sampling can be biased toward known "hotspots" of biotic diversity or toward easily accessible areas near human population centers, and the intensity of botanical sampling can vary through time (Shaffer et al., 1998; Kadmon et al., 2004; Rich, 2006; Pautasso and McKinney, 2007). For most rare species, only scarce data are made publicly available; these are mostly in atlas form and reflect imperfect, sporadic, and spatially biased sampling that describes ranges only in a coarse-grained way.
However, spatial data analyzed with new statistical techniques and geographic information systems (GIS) can be used to assess and account for bias in historical sampling that is evident in museum and herbarium collections (Ponder et al., 2001; Stockwell and Peterson, 2002; Rich and Karran, 2006). Applying three statistical approaches adapted from Bayesian analysis and building on methods of Solow (1993), Burgman et al. (1995), McCarthy (1998), and Ungricht et al. (2005), we have previously assessed the level of uncertainty inherent in predicting whether historically documented wild populations ("occurrences" sensu NatureServe, 2007) of rare plant species are still extant in New England of the United States (Farnsworth and Ogurcak, 2006). Using these techniques, we have been able to identify and focus on a subset of 71 phylogenetically diverse species appropriate for reliably characterizing changes in the estimated distribution of occurrences over time.
Here, we compare the present biogeographic distribution of confirmed, extant occurrences to the distribution of "historic" occurrences (those not relocated since 1975, despite searches; Brumback et al., 1996) for each of these 71 species. The current study is one of only a few that have focused exclusively on rare species to determine if life history traits influence levels of imperilment (Rabinowitz et al., 1989; Lahti et al., 1991; Quinn et al., 1994; Thompson and Hodgson, 1996; Pocock et al., 2006). In contrast with prior studies, in which rarity has been inferred from a static measure of current range size, we investigate dynamic information on range contraction and numbers of populations lost over time.
Having access to this large sample of rare species enables us to detect commonalities in degree of rarity and causes for decline among species that are members of particular "functional groups" (sensu Gitay and Noble, 1997) based on apparent habitat affinities and life history characteristics. We subdivide our sample of 71 species into eight categories of ecological groupings based on aspects of life history, habitat affinity, and the spatial position of their range boundary relative to New England. The life history traits by which we chose to classify species are characters that have been shown previously to correlate with rarity (e.g., Kunin and Gaston, 1997; Pocock et al., 2006) and for which we have reliable data: pollination syndrome, dependency on species-specific symbioses, capacity for clonal as well as sexual reproduction, potential distance of seed dispersal, specialization on wetland vs. upland habitats, affinity for circumneutral substrates, tendency to occur in early-successional or disturbance-prone sites, and the position of the species’ range relative to the New England region. Using group-based comparisons, we test the hypothesis that species sharing particular ecological characteristics have undergone similar types of range shifts over time due to similar ecological constraints that determine their vulnerability to local extinction.
MATERIALS AND METHODS
The species
We analyzed 71 species (Table 1) for which we had unbiased presence/absence data (Farnsworth and Ogurcak, 2006) and for which a detailed, peer-reviewed Conservation and Research Plan had been published by the New England Wild Flower Society between 2001 and 2004 (Farnsworth, 2003). The taxa selected for conservation planning were chosen from the full group of 479 taxa identified in the Flora Conservanda: New England (Brumback et al., 1996) as either "State-Endangered/Threatened" (337 taxa) or "Regionally Rare" (recorded from <20 towns in New England and/or state-listed as "Special Concern" using NatureServe [2007] rarity criteria). Species were selected for concerted conservation planning based on several criteria, including the breadth of taxonomic representation, the absence of other plans such as U. S. Fish and Wildlife Service Endangered Species Recovery Plans already addressing them, the availability of knowledgeable botanists to assess the status of occurrences, and unambiguous evidence of the taxon’s rarity in the region. The state ranks of these taxa had not changed significantly in the decade since the Flora Conservanda was published; over half of the 44 changes—of 223 possible over the 71 species—involved upgraded rarity ranks (i.e., from Special Concern status to Endangered), and no taxa were delisted. Indeed, the up-to-date data provided in the Conservation and Research Plans have been used to inform state endangerment ranks since 1996.
|
criteria) in a mean of 8.2% of their total range or listed as critically imperiled (S1 or S2) in an average of 33.7 percent of their total range.
The 71 taxa spanned 35 plant families and comprised a representative subsample of the 479 Flora Conservanda taxa. Twenty-eight families (80%) contained only one or two species, indicating a fairly even distribution of taxa. The most speciose families analyzed included the Cyperaceae (with 10 species), the Asteraceae (7 spp.), and the Orchidaceae (7 spp.), broadly reflecting the distribution of species within families across all New England taxa. The distribution of the number of species per family in this subset of 71 taxa closely paralleled the number of species per family in the entire set of 479 Flora Conservanda taxa (regression of species/family of the Conservation Plan subset on the total set; r2 = 0.809; P < 0.001). To account for bias in the model due to phylogenetic relatedness, we entered the family level affiliation of each species as a factor in all analyses (Harvey and Pagel, 1993; see Statistical analyses section).
Classifying species into functional groups
We assembled trait and habitat information for each species from data in each Conservation and Research Plan, Natural Heritage element occurrence records, and the National PLANTS Database (USDA, NRCS, 2007). We then assigned each species to contrasting functional groups (Table 1). The first two traits we examined, pollination syndrome and dependency on species-specific symbioses, reflected the dependence of a particular plant species on other species needed to facilitate establishment, survival, and reproduction.
Pollination syndrome
We hypothesized that species dependent upon insect pollination would be more vulnerable to local extinction than wind- or self-pollinated species, given the fact that declines in plant species often parallel declines in pollinators (Biesmeijer et al., 2006; National Research Council, 2007) and that invasive plant species are increasingly disrupting reproductive mutualisms (Traveset and Richardson, 2006). Likewise, competition for pollinators appears to be intensifying in plant diversity "hotspots" (Vamosi et al., 2006) and concentrations of rare plants in New England also cluster in increasingly spatially isolated hotspots (Farnsworth and Ogurcak, 2006). Pollination syndromes were classified as "I" for insect-pollinated species (68%) and "W" for wind/water-pollinated or cleistogamous species (32%), based on flower morphology and/or observations of pollinator behavior in the field.
Dependency on species-specific symbioses
We posited that disruption of mutualisms caused by declines in mycorrhizal species (due to air pollution or substrate disturbance; Arnolds, 1991), host plants (for rare hemiparasites), and nodulating bacteria would place obligate mutualists at more risk than nonmutualist species. Species dependent upon symbioses with mycorrhizae, nodulating bacteria, or hemiparasitic relationships with host plants were designated "Y" for symbiotic interactions (primarily members of the Orchidaceae, Fabaceae, and Orobanchaceae; 18%); those without such interactions were termed "N" for nonspecialists (82%).
We next investigated characters that indicate the reproductive and dispersal flexibility of the taxa.
Capacity for clonal as well as sexual reproduction
We predicted that plant taxa capable of both clonal and sexual reproduction would lose fewer populations over time than species dependent solely on sexual reproduction because they can propagate flexibly when either pollinators or habitat are limiting. Species that reproduced by runners and rhizomes were given a score of "1" for clonality (35%); those with short vegetative buds were scored as "0.5" for intermediate clonality (19%), and those incapable of vegetative reproduction were scored as "0" (46%).
Potential seed dispersal distance
We hypothesized that species with seeds capable of dispersing long distances would maintain larger numbers of populations over time, given that long-distance dispersers tend to be more widespread than locally dispersed species (Kelly and Woodward, 1996; Gaston, 2003). We characterized dispersal strategy based on direct examination of seed morphology (for example, the presence of specialized structures to encourage animal dispersal) and information from existing literature on each species. Taxa with seeds dispersed by animals, wind, or water were categorized as "W" (for "widely dispersed"; 69%), while species with seeds dispersing by gravity or insects less than 1 m from the parent plant were designated "L" (for "local"; 31%).
We also categorized species based on their apparent habitat specialization, reasoning that congenitally "picky" species (sensu Rosenzweig and Lomolino, 1997) would be more likely to lose populations when their narrow habitats are annihilated (Pocock et al., 2006).
Wetland vs. upland specialists
Millions of hectares of wetlands have been permanently drained, converted, or destroyed over the past century (Bedford, 1999; Dahl, 2000), whereas upland habitats have undergone clearing but have rapidly afforested in the past century in the northeast United States (Hall et al., 2002). Thus, we expected that obligate or facultative wetland species, specializing on periodically inundated habitats, would decline more dramatically than would primarily terrestrial species. Categories of wetland status, determined from the U. S. Fish and Wildlife Service wetland indicator status (USDA, NRCS, 2007), were labeled "W" for species classified as facultative (FACW) or obligate (OBLW) wetland specialists (52% of the 71 species) and "U" for facultative (FACU) or obligate (OBLU) upland species (48%).
Affinity for circumneutral substrates
A number of species listed as rare in New England are specialists on calcium- or magnesium-rich bedrock substrates. This bedrock type occupies a restricted belt on the western fringe of the region, and calciphilic natural communities are considered rare (Motzkin, 1994; Anderson et al., 1999). Lahti et al. (1991) found that specialization on particular edaphic substrates such as calcareous bedrock tended to be associated with plant rarity. Thus, we hypothesized that species constrained to these relatively uncommon habitats (e.g., calcareous glades and fens) would be more vulnerable to decline than nonspecialists. Species designated as calciphiles ("C"; 49%) had the majority of their populations described from circumneutral substrates (bedrock or soils), whereas nonspecialists ("N") occupied a wider range of bedrock types (51%).
Early-successional species
Based on data from rare grassland plants (Farnsworth, 2007), we expected that species dependent on some forms of disturbance that reduce competition with later-successional species—such as flooding, fire, or gap formation—would undergo larger declines in New England, where many natural disturbance regimes have been suppressed in densely populated zones. Species classified as disturbance-dependent ("D"; 41%) had the majority of their populations in early-successional habitats (such as open fields, forest gaps, or roadside verges) or habitats subject to frequent disturbance (such as tidal wetlands or floodplains); successionally neutral species ("N"; 59%) occurred in multiple habitat types, from full shade to open sun.
Position of species’ range relative to New England states
Outlying populations occupy marginal habitats and may be subject to more stochastic variability in numbers and higher probabilities of extinction than populations at the heart of a species’ range. We compared species reaching the margin of their ranges with those for which New England formed the heart of the range. We tested the null hypothesis that rare species reaching either their northern or southern range boundary in New England would not lose more populations than species for which this region represents the center of their range. Species were classified by inspecting global range maps (NatureServe, 2007) as follows: southern ("S") species (38%) reaching the northern edge of their range in New England; central ("C") species (39%) in the heart of their range in New England; and northern ("N") species (23%) reaching the southern edge of their range in New England.
These eight categorization schemes obviously represented only a subset of the possible functional groupings into which these species could be classified. Previous authors, for example, found correlations between rarity status and other life-history traits such as self-incompatibility, seed size, capacity for seed dormancy, and annual vs. perennial lifespan (e.g., Kelly and Woodward, 1996; Kunin and Gaston, 1997; Hegde and Ellstrand, 1999; Cadotte and Lovett-Doust, 2002; Farnsworth, 2007). However, the highly uneven distribution of these latter character states among the 71 taxa studied here would have resulted in unbalanced sample sizes for comparison, and it was not possible to determine self-compatibility or seed dormancy status with certainty across all species. Thus, the eight functional groupings were the primary ones to which we could assign species with confidence.
These eight comparisons represented independent contrasts. That is, trait states were largely uncorrelated with each other and 
2 statistics did not indicate significant differences in the shape of the frequency distribution for other guild affiliation among members of contrasting guild types. However, in two cases, traits were disproportionately associated with each other: 95% of wind- or self-pollinated species had long-distance seed dispersal, and 92% of plant species with known mutualisms also had insect-pollinated flowers. However, statistical trends were not consistent among these functional groups, suggesting that the traits were largely independent in their effects on plant rarity.
Data on species’ occurrences
Data on each extant and historic occurrence of each species were made available by the Natural Heritage Programs of all six New England states for the purposes of the Conservation and Research Plan project. All occurrences documented as extant by Natural Heritage Programs were visited and confirmed by Conservation Plan authors at the time of writing or had been observed by New England Plant Conservation Program (NEPCoP) volunteers during the past 10 years. Natural Heritage Programs designated as "Historic" ("H") those occurrences for which targeted searches had been conducted but had not been observed since 1975; additional searches for some of these populations took place during the conservation planning period. Occurrences were ranked as "Extirpated" ("X") if they were judged to have poor prospects for being rediscovered due to destruction of existing habitat at a given site.
During 2001–2003, the Herbarium Recovery Project of the New England Wild Flower Society reviewed over 16 800 herbarium records from 42 major and minor herbaria in New England to confirm and update taxonomic identifications, to catalogue information on locations of 532 species of rare or poorly known plant species, and to collate this information with Natural Heritage records (Haines, 2003). This process significantly improved the accuracy of data on historical collections of plant populations and pinpointed taxa for which collections were scant or frequently misidentified. The project also demonstrated that temporal patterns of collection had been nearly identical in all six New England states (0.616 < r < 0.965, correlations among states of collections per decade from 1820 to 2000), indicating minimal spatial bias in sampling intensity over the history of botanical collection in the region.
Calculating probabilities of pseudoabsence
If sampling is imperfect and we erroneously consider populations to be "historical" when, in fact, we simply have not detected them, we refer to these populations as pseudoabsent (sensu Engler et al. [2004]). It is possible to overestimate the loss of populations if one simply subtracts the number of confirmed extant occurrences from the total number of observations of extant plus all presumed historical populations, when a proportion of the latter are actually pseudoabsent and should be treated as extant. Therefore, we took two steps to estimate more conservatively the maximum number of pseudoabsent occurrences (see methods of Farnsworth and Ogurcak [2006]). First, using the partial Solow equation (Solow, 1993; Burgman et al., 1995), we calculated the probability that each occurrence labeled as historical by Natural Heritage Programs was, in fact, pseudoabsent. The partial Solow equation estimates the probability that a taxon is still extant, based on an analysis of the frequency of museum or herbarium records collected between the first and most recent observations of a species, accounting for uneven collection effort over time. A high probability of extinction is inferred if collections are highly overdue relative to past collection patterns (McCarthy, 1998; Ungricht et al., 2005). While the partial Solow equation has typically been applied to estimate extinction probability of taxa, we used it to assess the probability of detecting a particular occurrence of a taxon given a known run of collection effort for the taxon as a whole in New England. This method enabled us to estimate the probability of any given historical occurrence being rediscovered eventually. For the 71 taxa examined here, the mean probability of relocating a historical population (that is, the probability that a so-called historical occurrence was actually pseudoabsent) was 0.17 ± 0.01 SE (Farnsworth and Ogurcak, 2006). Interestingly, Solow correction factors did not differ significantly among particular plant families (ANOVA; F33,37 = 1.39; P = 0.162), indicating little taxonomic bias in prior collection or the probability of rediscovering occurrences. The Solow correction factors computed for each species were then multiplied by the total subset of occurrences ranked as historical or extirpated by Natural Heritage Programs. We then added the resulting rounded number to the number of populations confirmed as extant from actual field surveys. Next, we lumped historical records that could not unambiguously be designated as distinct populations, either from each other or from known extant occurrences. This step eliminated an average of 15.3% ± 0.02 SD of historical populations from consideration. After applying these two methods, we used the resulting conservative figure to estimate the total number of potentially extant occurrences for each species.
Estimating range areas
Town-level occurrences were mapped in ArcGIS 8.3 software (ESRI, 2002) as centroids—point data given in Universal Transverse Mercator (UTM) coordinates—that corresponded to the center of mass of a polygon inscribing each town boundary. Next, a range-wide polygon corresponding to the -hull of the town centroid distribution (Burgman and Fox, 2003) was drawn to encompass all the town-level centroids that had positive occurrences of a particular taxon. Separate -hull polygons were developed to characterize the distributions of extant and historical populations for each species. Each polygon was constructed using a script in the program S-Plus (S-Plus 2000 Professional Release 3.0, Insightful Corp., Seattle, Washington, USA) kindly provided by M. A. Burgman and J. Fox. Alpha-hulls are a form of minimum convex hull (IUCN, 2001), which itself is the smallest polygon that contains all occurrence points in which no internal angle exceeds 180 degrees (Burgman and Fox, 2003). A Delauney triangulation is created by drawing lines between all points, subject to the constraint that lines do not intersect. An average line length is then calculated, indicating a mean distance among occurrences; lines longer than a value of multiplied by this average length are deleted from the triangulation, thus eliminating large empty areas of likely unsuitable "habitat" among distance occurrences. For example, lines connecting distant endpoints on a horseshoe-shaped point distribution would be deleted, yielding a more accurate representation of the range area than a coarse-grained rectangle. For the purposes of estimation here, lines were deleted when the edge length of a Delauney triangle was more than three times the average edge length of all triangles (Burgman and Fox, 2003). The areas of these -hull polygons, the sum of all remaining triangles, more conservatively estimate a species’ range than do conventional minimum convex polygons (e.g., Kaluzny et al., 1996). Occasionally, this method can exclude outlying occurrences and thus underestimate a range (Getz and Wilmers, 2004); however, by visually inspecting the overlap of the estimated -hull with the centroids of known occurrences for each species, we found that this was not the case with the taxa we studied. Because such polygons obviously overestimated the actual land area occupied by each species (populations did not occupy all points in space and habitat-specific information was insufficient to refine the exact range boundaries), they served as relative rather than absolute proxy measures for comparing extant and historical ranges ("areas of extent" sensu Gaston [2003]) within taxa.
Statistical analyses
All variables (Solow-corrected proportion of occurrences lost over time, proportion of historic range area lost, and area of extant range) were normally distributed and as such were analyzed using fixed-factor analyses of variance testing effects of functional group affiliation. Taxonomic family was entered as a factor in the model to estimate the contribution of phylogenetic relatedness among species to the variance. To account for the effects of starting (historic) range size, we included this variable as a covariate in ANCOVAs testing whether contrasting functional groups differed in the proportion of range area lost. Similarly, we incorporated the initial (historic) number of populations as a covariate in ANCOVAs comparing proportions of all populations lost over time among functional groups. Scheffé posthoc multiple comparisons were made among the three groups of taxa reaching their southern, central, and northern range margins and those with contrasting levels of clonal reproduction. Data were analyzed using S-Plus 2000 Professional Release 3.0 and plotted using SigmaPlot 2001 for Windows, version 7.101 (SPSS, Chicago, Illinois, USA). The resulting P values were interpreted subject to a Bonferroni correction for multiple comparisons, with = 0.006.
RESULTS
All species lost populations in New England over time (Fig. 1; see grand mean line showing a 44.5% loss of populations overall). For the most part, functional groups did not differ significantly in initial numbers of historic populations, except for upland species, which were recorded from a larger number of populations in the past than wetland species (Table 2).
|
|
|
|
Species forming obligate symbioses vs. nonspecialists
Mycorrhizal species, mycoheterotrophs, nodulating species, and hemiparasites did not differ significantly from nonsymbiotic species in terms of the percentage of range area lost or extant range area in New England, but had lost marginally higher proportions of historical populations than nonsymbiotic taxa (Table 2).
Species with asexual vs. exclusively sexual modes of reproduction
Species with a capacity for vegetative spread did not differ significantly from nonclonal species in any of the variables measured (Table 2).
Species with contrasting seed dispersal capacities
Species with localized seed dispersal lost proportionally more range area than animal-, wind-, or water-dispersed species (Fig. 2, Table 2), but did not differ from the contrasting functional group in any other variables measured.
Wetland vs. upland species
Upland species lost a significantly larger proportion of their former range areas (Fig. 2, Table 2) and a marginally higher proportion of historical populations (Table 2) than wetland species. Upland species also had been recorded from significantly more historical populations than wetland species (Table 2).
Species with affinities for circumneutral soils
Calciphilic species did not differ from nonspecialists in terms of losses of occurrences or the percentage of range area lost over time. Likewise, current extents of occurrence were similar among the two groups (Table 2).
Species of early-successional habitats vs. nonspecialists
Species with affinities for disturbance-prone or early-successional habitats did not differ significantly from nonspecialists in terms of proportions of populations lost, extant range area in New England, or the proportion of their range area lost over time than nonspecialists inhabiting a range of successional environments (Table 2).
Species at their range margins in New England
"Southern" species (namely, those reaching their northern range boundary in New England) had lost significantly more populations than "northern" species and those "central" taxa for which New England falls in the heart of their range (Fig. 1). Historic numbers of populations had originally been similar among southern and central taxa and larger than those of northern taxa (Table 2). Southern species lost marginally more range area over time than either northern or central species, while the last two groups did not differ from each other (Fig. 2). The extent of occurrence of extant populations in New England was also somewhat smaller among southern species than the areas of either northern or central taxa (Fig. 3, Table 2).
DISCUSSION
This biogeographic analysis of a large sample of taxa classified as regionally rare in New England has enabled us to compare range declines among disparate species that belong to contrasting functional groups. This exploration is of broad botanical interest because it identifies certain intrinsic life-history characteristics that may place certain species at risk for becoming rare. Utilizing a conservative estimate from collection and visitation records to estimate numbers of pseudoabsent and truly extinct populations, our methods permit a quantitative comparison of population loss and range contraction. Additionally, our methods for analyzing coarse-grained data from historical records should prove useful for all botanists seeking to derive information from natural history collections, whatever their research question or hypothesis. In so doing, we can begin to identify the most imperiled plant functional groups and to develop coherent strategies for managing taxa with similar ecological affinities and habitat requirements. Furthermore, because many of the rare taxa of New England hail from other biomes across North America (Crins, 1997; NatureServe, 2007), these results can inform ecological understanding and conservation practices throughout a broad geographical area.
Our findings accorded with the theory that certain suites of ecological and life-history traits predispose some species to rarity and identified characteristics of rare plants that may exacerbate their vulnerability. The group of insect-pollinated rare species, for example, has experienced a significant reduction in numbers of populations and range size relative to wind-, water-, or self-pollinated species (Figs. 1, 2, Table 2). Quinn et al. (1994) also found that obligately outcrossed species have smaller, more aggregated ranges than those taxa capable of self-pollination. Habitat fragmentation and population isolation demonstrably hinder plant–pollinator interactions (Honnay et al., 2005), making entomophilous species more vulnerable to the large-scale land-use changes that are taking place in New England and elsewhere. Habitat destruction has been identified by professional botanists as affecting nearly 10% of rare plant populations in the region, based on direct field observations (E. Farnsworth, unpublished data). Likewise, entomophilous plants may be responding to a cascade effect from local pollinator extinctions (Kearns et al., 1998; Memmott et al., 2004; Biesmeijer et al., 2006; National Research Council, 2007), competition with other plant species with which they share generalist pollinators (Gibson et al., 2006; Vamosi et al., 2006), and introduction of exotic species that compete for pollinator services (Traveset and Richardson, 2006).
Like pollination mutualisms, obligate symbioses with mycorrhizae, nodulating bacteria, and host plants also influence the persistence of rare plant occurrences, and their disruption results in poor recruitment and survivorship (McCormick et al., 2004). The present analysis did not find systematic differences between the species dependent upon these types of specialized mutualisms and nonspecialist species. However, mutualism-forming species have lost a marginally larger proportion of populations over time than nonspecialists, even though there were more records of their populations historically (Table 2). The species-specific and obligate vs. facultative nature of these mutualisms are poorly characterized for most rare plant taxa; a better understanding of the nature and importance of these mutualistic interactions would aid in designing recovery efforts.
Plant species producing seeds capable of being dispersed widely from the parent plant lost less range area over time than taxa with locally dispersed seeds (Fig. 2, Table 2), indicating that flexible, long-range seed dispersal may enable some plants to escape local threats and establish in a broader array of habitats.
In contrast, an inability to reproduce clonally did not consistently correlate with vulnerability to local extinction or the magnitude of range constriction (Table 2). This finding accorded with results obtained by Lahti et al. (1991) and Pocock et al. (2006), who found that current areas of occupancy by numerous rare plant species were not explained by the capacity for clonal reproduction. Rather, these studies suggested that habitat specialization was a more important determinant of range size.
Regarding habitat affinities, we might expect that rare plant species that are obligate or facultative inhabitants of wetlands also would have suffered drastic declines relative to terrestrial species because wetland systems have undergone precipitous rates of conversion and outright loss (Dahl, 2000). However, our analysis indicated that rare upland species have lost more populations (Fig. 1) and have undergone significantly larger range contractions (Fig. 2) than have rare wetland species (Table 2). Quinn et al. (1994) similarly found that species of upland habitats had smaller extant ranges than those inhabiting several types of wetlands. It is possible that rigorous federal and state legislation explicitly devised to protect wetlands has recently slowed habitat degradation, allowing for recovery of some obligate wetland species (Bedford, 1999), although wetlands are only rarely protected or managed solely for their value as habitat to rare species (Doust and Doust, 1995). Alternatively, the greater imperilment of upland species could be attributable to the large proportion of upland that has been subject to massive habitat conversion over time. About half (47%) of the rare upland species we examined typically inhabit understories of mature or old-growth forests (we discuss the other 53% of upland species that typically inhabit early-successional habitats later). Recalling the land-use history of New England, in which over 80% of upland forests were cleared for agriculture from 1700 to 1900, it is not surprising that these specialist upland species have declined due to past large-scale disruptions of their habitat (Bellemare et al., 2002). Even a century after afforestation recommenced in New England with the wholesale abandonment of agriculture, these species have not recovered; the signature of past land manipulation and the effects of fragmentation from contemporary development and tree harvesting still affect survivorship and establishment of many herbaceous upland plants (Bellemare et al., 2002, Cowell and Jackson 2002; Flinn and Vellend, 2005; Honnay et al., 2005; Kolb and Diekmann, 2005; Spyreas and Matthews, 2006). Thus, it is critical to protect and restore upland habitats and to devise forestry practices that are compatible with the recovery and long-term maintenance of rare plant populations (Foster et al., 2005).
Interestingly, rare plants specializing on circumneutral substrates in New England did not appear to be significantly more vulnerable to decline than other nonspecialist rare species (Table 2). This finding is somewhat surprising because circumneutral bedrock is localized in New England and habitat for strict calciphiles is apparently limiting. Such specialized natural community types, including calcareous fens and glades, have received much conservation attention throughout North America (Motzkin, 1994; Anderson et al., 1999; Mann et al., 1999), and they are recognized as hotspots for diversity of rare species in New England (Farnsworth and Ogurcak, 2006). Conservation efforts made to date in these areas may be successfully protecting these species and their habitats.
We found that the group of "disturbance-dependent" species (those found more commonly in early-successional or regularly disturbed upland habitats such as heathlands, coastal plain ponds, recent burns, and cliffs) did not differ significantly from nonspecialists in terms of populations or range area lost (Table 2). In contrast with this finding, competition with invasive exotic species and shading associated with succession to shrublands and woodlands frequently have been cited by botanists as severe threats to populations of New England’s rare species, affecting over 20% of populations (Farnsworth, 2004; E. Farnsworth, unpublished data). A great deal of current management effort in the region is expended on restoring or mimicking natural disturbance processes such as fire, flooding, and scour that can reset the successional clock and open new habitat for colonization by threatened species (Vickery and Dunwiddie, 1997; Litvaitis, 2003). The present analysis indicated that, while targeted management to reinstate controlled disturbance may be helpful to some populations (e.g., Lezberg et al., 2006), the alteration or dampening of disturbance regimes over the past century may constitute only part of a suite of larger-scale drivers of plant decline. Thus, management needs to be undertaken with caution, after a fuller understanding of landscape dynamics and species’ habitat requirements has been attained. Management trials need to be performed as controlled and replicable experiments, and the positive and negative impacts on the target rare species must be thoroughly and quantitatively described (Schemske et al., 1994; Falk et al., 1996).
Looking at the larger-scale, range-wide geographic context, we found that species hailing from regions south of New England have lost large proportions of populations (Fig. 1, Table 2) and a substantial proportion of their former range area (Fig. 2). Also, they currently occupy very small overall range areas in the region compared to northern and central taxa (Fig. 3). Thus, we conclude that southern species at their range margins are more at risk than northern species or those at the heart of their range in the region. These losses have occurred despite the fact that southern species were once recorded from a larger number of historic populations than northern species (Table 2). Farnsworth and Ogurcak (2006) similarly found that the overall centroid of the range of rare species generally has shifted north in New England due to the disproportionate loss of populations along the southern margin; this shift is primarily due to the elimination of southern populations, rather than a northward shift in occurrences that might reflect changing climate. These species reach their northern range margin in some of the densest population centers in North America—the human populations of Connecticut, Massachusetts, and Rhode Island are 
10 times denser than those of the three northern New England states (U.S. Census Bureau, http://www.census.gov/). Trampling by humans and outright habitat destruction to pave the way for housing and industry have been identified as primary threats to over a quarter of the populations of southern species now classified as extirpated in New England (E. Farnsworth, unpublished data). Likewise, many of the species reaching their northern range limit in New England are associated with relatively open, tall- and short-grass prairie habitats that are themselves severely imperiled in their midwestern U.S. strongholds (Table 1; Miller, 1989; Noss et al., 1995; Vickery and Dunwiddie, 1997; Mehrhoff, 2000). As discussed, southern species specializing on early-successional or disturbance-prone habitats may benefit from prudent management designed to mimic ecological processes common to prairies. However, given the prevalence of invasive species observed at certain rare plant sites in the southern New England states (Farnsworth, 2004), management to maintain early-successional habitats will have to be tailored so as not to encourage rapid invasion by exotic competitors.
Threats such as habitat alteration and invasive species typically act in concert to impact rare taxa (Gurevitch and Padilla, 2004; Didham et al., 2007). This regional analysis has demonstrated that concerted land protection and conservation efforts are sorely needed, particularly in the southern New England states. From an ecological perspective, however, it is important to keep in mind that many intrinsic features of rare plants, in concert with extrinsic threats, are synergistically influencing the persistence of populations at both local and regional scales.
FOOTNOTES
1 The authors are grateful for the collective work of the New England Plant Conservation Program (NEPCoP) and its 700 volunteers who monitor rare plants and for the 150+ professional botanists and peer reviewers that produced the Conservation and Research Plans. The authors thank the anonymous foundation that funded the NEPCoP planning project. The authors thank the Natural Heritage Programs of Maine, New Hampshire, Massachusetts, Connecticut, Rhode Island, and Vermont that made critical data available during the project. They extend their gratitude to M. A. Burgman and J. Fox for contributing script for computing -hulls. The comments of T. Sipe, K. Stinson, L. Knapp, A. M. Ellison, G. Motzkin, and two anonymous reviewers greatly improved the manuscript. A. M. Ellison and D. R. Foster provided excellent suggestions regarding data analyses. E.J.F. was supported by a Bullard Fellowship from Harvard University and NSF DGE0123490. 
4 Author for correspondence (e-mail: efarnswo{at}mtholyoke.edu)
LITERATURE CITED
Anderson, R. C., J. S. Fralish, AND J. M. Baskin [eds.]. 1999. Savannas, barrens, and rock outcrop plant communities of North America. Cambridge University Press, Cambridge, UK.
Arnolds, E. 1991. Decline of ectomycorrhizal fungi in Europe. Agriculture Ecosystems & Environment 35: 209–244.[CrossRef][Web of Science]
Bedford, B. L. 1999. Cumulative effects on wetland landscapes: Links to wetland restoration in the United States and southern Canada. Wetlands 19: 775–788.
Bellemare, J., G. Motzkin, AND D. R. Foster. 2002. Legacies of the agricultural past in the forested present: An assessment of historical land-use effects on rich mesic forests. Journal of Biogeography 29: 1401–1420.[CrossRef][Web of Science]
Biesmeijer, J. C., S. P. M. Roberts, M. Reemer, R. Ohlemüller, M. Edwards, T. Peeters, A. P. Schaffers, S. G. Potts, R. Kleukers, C. D. Thomas, J. Settele, AND W. E. Kunin. 2006. Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 313: 351–353.
Brumback, W. E., L. J. Mehrhoff, R. W. Enser, S. C. Gawler, R. G. Popp, P. Somers, D. D. Sperduto, W. D. Countryman, AND C. B. Hellquist. 1996. Flora Conservanda: New England. The New England Plant Conservation Program (NEPCoP) list of plants in need of conservation. Rhodora 98: 235–361.[Web of Science]
Burgman, M. A., AND J. C. Fox. 2003. Bias in species range estimates from minimum convex polygons: Implications for conservation and options for improved planning. Animal Conservation 6: 19–28.[CrossRef][Web of Science]
Burgman, M. A., R. C. Grimson, AND S. Ferson. 1995. Inferring threat from scientific collections. Conservation Biology 9: 923–928.[CrossRef][Web of Science]
Cadotte, M. W., AND J. Lovett-Doust. 2002. Ecological and taxonomic differences between rare and common plants of southwestern Ontario. Ecoscience 9: 397–406.[Web of Science]
Cowell, C. M., AND M. T. Jackson. 2002. Vegetation change in a forest remnant at the eastern presettlement prairie margin, USA. Natural Areas Journal 22: 53–60.[Web of Science]
Crins, W. J. 1997. Rare and endangered plants and their habitats in Canada. Canadian Field Naturalist 111: 509–519.
Dahl, T. E. 2000. Status and trends of wetlands in the conterminous United States 1986 to 1997: U. S. Department of Interior, Fish and Wildlife Service, Washington, D.C. USA.
Didham, R. K., J. M. Tylianakis, N. J. Gemmell, T. A. Rand, AND R. M. Ewers. 2007. Interactive effects of habitat modification and species invasion on native species decline. Trends in Ecology & Evolution 22: 489–496.[CrossRef][Web of Science][Medline]
Doust, L. L., AND J. L. Doust. 1995. Wetland management and conservation of rare species. Canadian Journal of Botany 73: 1019–1028.[CrossRef]
Engler, R., A. Guisan, AND L. Rechsteiner. 2004. An improved approach for predicting the distribution of rare and endangered species from occurrence and pseudo-absence data. Journal of Applied Ecology 41: 263–274.[CrossRef][Web of Science]
ESRI. 2002. ArcGIS: Editing in ArcMap. Redlands, California, USA.
Falk, D. A., C. I. Millar, AND M. Olwell. 1996. Restoring diversity: Strategies for reintroduction of endangered plants. Island Press, Washington, D.C. USA.
Farnsworth, E. J. 2003. Planning for plants: The New England plant conservation and research plan project. Plant Talk 34: 24–28.
Farnsworth, E. J. 2004. Patterns of plant invasions at sites with rare plant species throughout New England. Rhodora 106: 97–117.[Web of Science]
Farnsworth, E. J. 2007. Sandplain plant life history traits as guides to multi-species management strategies for rare taxa. Biological Conservation 136: 44–52.[CrossRef][Web of Science]
Farnsworth, E. J., AND D. E. Ogurcak. 2006. Biogeography and decline of rare plants in New England: Historical evidence and contemporary monitoring. Ecological Applications 16: 1327–1337.[CrossRef][Web of Science][Medline]
Flinn, K. M., AND M. Vellend. 2005. Recovery of forest plant communities in post-agricultural landscapes. Frontiers in Ecology and Evolution 3: 243–250.
Foster, D. R., D. Kittredge, B. Donahue, G. Motzkin, D. Orwig, A. M. Ellison, B. Hall, B. Colburn, AND A. D’Amato. 2005. Wildlands and woodlands: A vision for the forests of Massachusetts. Harvard Forest, Petersham, Massachusetts, USA.
Gaston, K. J. 2003. The structure and dynamics of geographic ranges. Oxford University Press, Oxford, UK.
Getz, W. M., AND C. C. Wilmers. 2004. A local nearest-neighbor convex-hull construction of home ranges and utilization distributions. Ecography 27: 489–505.
Gibson, R. H., I. L. Nelson, G. W. Hopkins, B. J. Hamlett, AND J. Memmott. 2006. Pollinator webs, plant communities and the conservation of rare plants: Arable weeds as a case study. Journal of Applied Ecology 43: 246–257.[CrossRef][Web of Science]
Gitay, H., AND L. R. Noble. 1997. What are functional types and how should we seek them? In T. M. Smith, H. H. Shugart, and F. I. Woodward [eds.], Plant functional types, 3–19. Cambridge University Press, Cambridge, UK.
Gurevitch, J., AND D. K. Padilla. 2004. Are invasive species a major cause of extinction?Trends in Ecology & Evolution 19: 470–474.[CrossRef][Web of Science][Medline]
IUCN. 2001. IUCN Red List categories, version 3. 1 International Union for the Conservation of Nature, Species Survival Commission, Gland, Switzerland.
Hall, B., G. Motzkin, D. R. Foster, M. Syfert, AND J. Burk. 2002. Three hundred years of forest and land-use change in Massachusetts, USA. Journal of Biogeography 29: 1319–1336.[CrossRef][Web of Science]
Harper, K. T. 1979. Some reproductive and life history characteristics of rare plants and implications for management. Great Basin Naturalist Memoirs 3: 129–137.
Haines, A. 2003. Herbarium Recovery Project. Report for the New England Wild Flower Society, Framingham, Massachusetts, USA.
Harvey, P. H., AND M. D. Pagel. 1993. The comparative method in evolutionary biology. Oxford University Press, Oxford, UK.
Hegde, S. G., AND N. C. Ellstrand. 1999. Life history differences between rare and common flowering plant species of California and the British Isles. International Journal of Plant Sciences 160: 1083–1091.[CrossRef][Web of Science][Medline]
Honnay, O., H. Jacquemyn, B. Bossuyt, AND M. Hermy. 2005. Forest fragmentation effects on patch occupancy and population viability of herbaceous plant species. New Phytologist 166: 723–736.[CrossRef][Web of Science][Medline]
Judd, W. S., C. S. Campbell, E. A. Kellogg, P. F. Stevens, AND M. J. Donoghue. 2008. Plant systematics: A phylogenetic approach, 3rd ed. Sinauer, Sunderland, Massachusetts, USA.
Kadmon, R., O. Farber, AND A. Danin. 2004. Effect of roadside bias on the accuracy of predictive maps produced by bioclimatic models. Ecological Applications 14: 401–413.[CrossRef][Web of Science]
Kaluzny, S. P., S. C. Vega, T. P. Cardoso, AND A. A. Shelly. 1996. S+ Spatialstats user’s manual. MathSoft, Seattle, Washington, USA.
Kearns, C. A., D. W. Inouye, AND N. M. Waser. 1998. Endangered mutualisms: The conservation of plant-pollinator interactions. Annual Review of Ecology and Systematics 29: 83–112.[CrossRef][Web of Science]
Kelly, C. K., AND F. I. Woodward. 1996. Ecological correlates of plant range size: Taxonomies and phylogenies in the study of plant commonness and rarity in Great Britain. Philosophical Transactions of the Royal Society of London, B, Biological Sciences 351: 1261–1269.[CrossRef]
Kolb, A., AND M. Diekmann. 2005. Effects of life-history traits on responses of plant species to forest fragmentation. Conservation Biology 19: 929–938.[CrossRef][Web of Science]
Kunin, W. E., AND K. J. Gaston. 1997. The biology of rarity: Causes and consequences of rare-common differences. Chapman and Hall, London, UK.
Lahti, T., E. Kemppainen, A. Kurtto, AND P. Uotila. 1991. Distribution and biological characteristics of threatened vascular plants in Finland. Biological Conservation 55: 299–314.[CrossRef][Web of Science]
Lezberg, A. L., K. Buresch, C. Neill, AND T. Chase. 2006. Mechanical land clearing to promote establishment of coastal sandplain grassland and shrubland communities. Restoration Ecology 14: 220–232.[CrossRef][Web of Science]
Litvaitis, J. A. 2003. Shrubland and early-successional forests: Critical habitats dependent on disturbance in the northeastern United States. Forest Ecology and Management 185: 1–4.[CrossRef][Web of Science]
Mann, L. K., A. W. King, V. H. Dale, W. W. Hargrove, R. Washington-Allen, L. R. Pounds, AND T. L. Ashwood. 1999. The role of soil classification in geographic information system modeling of habitat pattern: Threatened calcareous ecosystems. Ecosystems (New York, N.Y.) 2: 524–538.
McCarthy, M. A. 1998. Identifying declining and threatened species with museum data. Biological Conservation 83: 9–17.[CrossRef][Web of Science]
McCormick, M. K., D. F. Whigham, AND J. O’Neill. 2004. Mycorrhizal diversity in photosynthetic terrestrial orchids. New Phytologist 163: 425–438.[CrossRef][Web of Science]
Mehrhoff, L. J. 2000. Immigration and expansion of the New England flora. Rhodora 102: 280–298.[Web of Science]
Memmott, J., N. M. Waser, AND M. V. Price. 2004. Tolerance of pollination networks to species extinctions. Proceedings of the Royal Society of London, B, Biological Sciences 271: 2605–2611.
Miller, N. G. 1989. Pleistocene and Holocene floras of New England as a framework for interpreting aspects of plant rarity. Rhodora 91: 49–69.[Web of Science]
Motzkin, G. 1994. Calcareous fens of western New England and adjacent New York state. Rhodora 96: 44–68.[Web of Science]
Murray, B. R., P. H. Thrall, AND B. J. Lepschi. 2002. Relating species rarity to life history in plants of eastern Australia. Evolutionary Ecology Research 4: 937–950.[Web of Science]
National Research Council. 2007. Status of pollinators in North America. National Academies Press, Washington, D.C. USA.
NatureServe. 2007. NatureServe Explorer: An online encyclopedia of life [web application], version 4.4. NatureServe, Arlington, Virginia, USA. Website http://www.natureserve.org/explorer [accessed 15 December 2007].
Noss, R. F., E. T. Laroe III, AND J. M. Scott. 1995. Endangered ecosystems of the United States: A preliminary assessment of loss and degradation. Biological Report 28. U.S. Department of Interior, National Biological Service, Washington, D.C. USA.
Pautasso, M., AND M. L. McKinney. 2007. The botanist effect revisited: Plant species richness, county area, and human population size in the United States. Conservation Biology 21: 1333–1340.[CrossRef][Web of Science][Medline]
Pitman, N. C. A., AND P. M. Jorgensen. 2002. Estimating the size of the world’s threatened flora. Science 298: 989.
Pocock, M. J. O., S. Hartley, M. G. Telfer, C. D. Preston, AND W. E. Kunin. 2006. Ecological correlates of range structure in rare and scarce British plants. Journal of Ecology 94: 581–596.[CrossRef][Web of Science]
Ponder, W. F., G. A. Carter, P. Flemons, AND R. R. Chapman. 2001. Evaluation of museum collection data for use in biodiversity assessment. Conservation Biology 15: 648–657.[CrossRef][Web of Science]
Quinn, R. M., J. H. Lawton, B. C. Eversham, AND S. N. Wood. 1994. The biogeography of scarce vascular plants in Britain with respect to habitat preferences, dispersal ability and reproductive biology. Biological Conservation 70: 149–157.[CrossRef][Web of Science]
Rabinowitz, D., J. K. Rapp, S. Cairns, AND M. Mayer. 1989. The persistence of rare prairie grasses in Missouri: Environmental variation buffered by reproductive output of sparse species. American Naturalist 134: 525–544.[CrossRef][Web of Science]
Rich, T. C. G. 2006. Floristic changes in vascular plants in the British Isles: Geographical and temporal variation in botanical activity 1836–1988. Botanical Journal of the Linnean Society 152: 303–330.[CrossRef][Web of Science]
Rich, T. C. G., AND A. B. Karran. 2006. Floristic changes in vascular plants in the British Isles: Comparison of techniques for assessing changes in frequency of plants with time. Botanical Journal of the Linnean Society 152: 279–301.[CrossRef][Web of Science]
Rodrigues, A. S. L., AND K. J. Gaston. 2002. Rarity and conservation planning across geopolitical units. Conservation Biology 16: 674–682.[CrossRef][Web of Science]
Rosenzweig, M. L., AND M. V. Lomolino. 1997. Who gets the short bits of the broken stick? In W. E. Kunin, and K. J. Gaston [eds.], The biology of rarity, 63–90. Chapman and Hall, London, UK.
Schemske, D. W., B. C. Husband, M. H. Ruckelshaus, C. Goodwillie, I. M. Parker, AND J. G. Bishop. 1994. Evaluating approaches to the conservation of rare and endangered plants. Ecology 75: 584–606.[CrossRef][Web of Science]
Shaffer, H. B., R. N. Fisher, AND C. Davidson. 1998. The role of natural history collections in documenting species declines. Trends in Ecology & Evolution 13: 27–30.[CrossRef][Web of Science]
Solow, A. R. 1993. Inferring extinction from sighting data. Ecology 74: 962–964.[CrossRef][Web of Science]
Spyreas, G., AND J. W. Matthews. 2006. Floristic conservation value, nested understory floras, and the development of second-growth forest. Ecological Applications 16: 1351–1366.[CrossRef][Web of Science][Medline]
Stockwell, D. R. B., AND A. T. Peterson. 2002. Controlling bias during predictive modelling with museum data. In J. M. Scott, P. J. Heglund, M. Morrison, M. Raphael, J. Haufler, B. Wall, and F. Samson [eds.], Predicting species occurrences: Issues of scale and accuracy, 537–546. Island Press, Covelo, California, USA.
Thompson, K., AND J. G. Hodgson. 1996. More on the biogeography of scarce vascular plants. Biological Conservation 75: 299–302.[CrossRef][Web of Science]
Traveset, A., AND D. M. Richardson. 2006. Biological invasions as disruptors of plant reproductive mutualisms. Trends in Ecology & Evolution 21: 208–216.[CrossRef][Web of Science][Medline]
Ungricht, S., J.-Y. Rasplus, AND F. Kjellberg. 2005. Extinction threat evaluation of endemic fig trees of New Caledonia: Priority assessment for taxonomy and conservation with herbarium collections. Biodiversity and Conservation 14: 205–222.[CrossRef][Web of Science]
USDA, NRCS. 2007. The PLANTS Database: U.S. Fish and Wildlife Service wetland indicator status. Website http://plants.usda.gov/wetland.html [accessed 21 December 2007].
Vamosi, J. C., T. M. Knight, J. A. Steets, S. J. Mazer, M. Burd, AND T.-L. Ashman. 2006. Pollination decays in biodiversity hotspots. Proceedings of the National Academy of Sciences, USA 103: 956–961.
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, Massachusetts, USA.
CiteULike Complore Connotea Del.icio.us Digg Facebook Reddit Technorati Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
