2009 NORTHEASTERN NATURALIST 16(1):67–84 Relationship of Bryophyte Occurrence to Rock Type in Upstate New York and Coastal Maine Natalie L. Cleavitt1,*, Susan A. Williams2, and Nancy G. Slack3 Abstract - Bryophytes are often noted for their growth on specific rock types and their value as indicator species. However, some evidence suggests that restriction of a species to specific rock types may be less rigid and could vary under different environmental conditions. We assessed richness and distribution patterns of bryophytes at 22 rock outcrop locations in upstate New York (NY) and coastal Maine (ACAD). At each location, detailed surveys were done in five replicate 5-m by 2-m plots on vertical rock faces. We report on the 194 bryophyte species found in these surveys and present detailed analyses for the 137 species that occurred at two or more of the 22 locations. In general, liverworts were less likely to be dominant within a plot than mosses. Within-site dominance and frequency of liverworts were less well correlated to larger-scale frequency (number of locations and regions of occurrence). In NY, although there was no significant difference in bryophyte richness by rock type, rare species were more often found on calcium-containing rock types. Rock type, soil influence, number of liverwort species, and region were significant correlates with bryophyte species composition patterns. The importance of rock type in explaining species composition patterns was still significant though weaker when ACAD locations were included in the analysis. This difference resulted from a higher prevalence of leafy liverwort species on calcium-containing rock types in ACAD. Our results present further evidence that apparent restrictions to specific rock types may shift depending on environmental conditions such as increased humidity and narrower temperature extremes that occur along the north Atlantic coast. Introduction Bryophytes represent a unique component of the richness, structural complexity, and functioning of many community types in the northeast. Rock-outcrop communities are notable upland assemblages in which bryophytes can dominate as the main vegetation. The role of rock type in determining bryophyte assemblages is well known from Europe (e.g., Bates 1978, Smith 1982), but is less well studied in North America, and then mainly in western North America (e.g., Cleavitt 2001, Horton 1988, Shaw 1981). Although many bryophyte floristic surveys have been conducted in the regions of upstate New York and coastal Maine (e.g., Allen 2005, Ketchledge 1980, Patterson 1930, Schuster 1949), none have focused specifically on rock-outcrop communities through a standardized sampling protocol, and the communities remain under-documented. Remarkably different bryophyte assemblages can be present within close proximity when calcareous and non-calcareous rock types are juxtaposed 1Department of Natural Resources, 8F Fernow Hall, Cornell University, Ithaca, NY 14853. 2340 Maple Street, Hinsdale, MA 01235. 3Biology Department, The Russell Sage Colleges, Troy, NY 12180. *Corresponding author - nlc4@cornell.edu. 68 Northeastern Naturalist Vol. 16, No. 1 on the landscape (Bates 1978, Miller and McDaniel 2004). Smith (1982) presented results from several European studies demonstrating the relationship between bryophyte species assemblages and rock type. In particular, limestone had a more specific flora than the non-calcareous rock types (Bates 1978, Smith 1982). Wide tolerance for substrate chemistry has been correlated with common or widespread species, and pH was the strongest explanatory variable separating the habitats of species (see results of Shaw [1981] for Pohlia and Horton [1988] for Encalypta). Jenkins (2004) observed that six rare bryophytes found in the Adirondack Park of New York all occur on “limy” ledges. Recent work on bryophyte communities in eastern Europe has shown that calcium-containing rock types contain higher overall species richness; however, liverwort species richness was higher on noncalcareous rock types (Kubešová and Chytrý 2005). The variation in substrate specificity throughout the range of a given bryophyte species may be underappreciated. For instance, species have been documented to shift from mainly tree-dwelling to rock-dwelling at their range limits (Piippo 1982, Smith 1982) and species can also demonstrate regional preference for calcium depending on pollution influence (Bates 1993). In the Gulf of St. Lawrence region, influence of rock type (areal extent of calcareous rock outcrops) was secondary to climatic factors such as growing-season temperature and ocean influence for explaining species distributions (Belland 2005). In this study, we examined patterns in bryophyte species occurrences on a range of rock types in upstate New York (NY), and then tested to see if species exhibited the same patterns on these rock types in coastal Maine (Acadia National Park; ACAD). We had four main questions (Q) and hypotheses (H). Q1) What are the main correlates to bryophyte species composition at 19 NY locations? H1) Sites would separate most strongly along a rock-type gradient by degree of calcium content. Q2) Do patterns of bryophyte species composition found in NY hold for similar rock types in ACAD? H2) Maine sites would follow the bryophyte composition patterns found in NY. Q3) How do species richness and rare species occurrences compare between sites and rock types in the three regions of NY? H3) More rare species and higher bryophyte species richness would occur on calcium-containing rock types. Q4) Do mosses and liverworts differ in their prevalence on different rock types or in their frequency and abundance patterns? H4) Liverworts would be comparatively more prevalent on non-calcareous rock types. Their frequencies would be similar to those of mosses, but their abundances would be lower than mosses given the generally smaller size of liverworts. Methods Site selection In NY, six rock outcrops were surveyed in each of the three regions: the Adirondack Mountain region in the northeast (Adirondacks), the Catskill Mountain and Shawangunk Mountain region in the southeast (Catskill-Shawangunk), 2009 N.L. Cleavitt, S.A. Williams, and N.G. Slack 69 and the Finger Lakes region in the center (Finger Lakes) (Fig. 1, Table 1). One additional site from the mid-eastern section of the state was also included for a total of 19 sites in NY. Rock outcrops in the Adirondacks were all at relatively lower elevations for that region and were comparable with the elevations of rock outcrops in the other two regions. Three locations were surveyed in ACAD, with two sites on Isle au Haut and the third on Mount Desert Island (Fig. 1, Table 1). N.L. Cleavitt and S.A. Williams surveyed NY locations from June to September 2005 and ACAD locations in August 2006. Site-selection protocol required that: 1) rock outcrops were in a wooded landscape such that all rock outcrops received some shading from the tree canopy, and 2) the rock-outcrop complex was large enough to accommodate our plot-selection methods (five rock outcrops with a vertical face at least 2 m high by 5 m wide). In addition, site selection was designed to include a range of rock types with equal representation of calcareous and non-calcareous rocks. Initial rock-type designations for site selection were made using information provided by the NY Heritage Element Occurrence Records for “Cliff Communities” (on file with NY State Heritage Program, Albany, NY) for NY locations, and Gilman and Chapman (1988) for ACAD locations. Final rock-type descriptions included in Table 1 were made from rock samples taken from each site. Samples were examined by Professor William Bassett (Geologist, Cornell University, Ithaca, NY) using HCl tests for calcium carbonates and a dissecting scope to ascertain mineral composition and degree of weathering. Surveyed rock types in NY varied from base-poor quartzite and conglomerates (Catskill-Shawangunk) to calcareous limestone, shales, and sandstones (Finger Lakes). The Adirondack rock outcrops included Figure 1. Locations of 22 sites where rock outcrops were surveyed for bryophytes in New York (19 sites) and Maine (Acadia National Park, 3 sites). In NY, sites were selected in three regions: Adirondack Mountains (4R, DL, HM, LD, PL, WH); Catskill-Shawangunk Mountains (BC, BLT, MRTP, NL, OM, TR); and Finger Lakes (BF, CF, LB, LSP, SB, TEG). TP was an additional limestone site in mid-eastern NY just north of the Catskill Mountain sites. 70 Northeastern Naturalist Vol. 16, No. 1 the widest diversity of rock types from limestone to granite (Table 1). The ACAD rock outcrops included granite, volcanic tuffs, and gabbro-diorite. Plot surveys Within each site, five replicate plots 2 m high by 5 m wide were surveyed on vertical rock outcrops. Selection for plot locations was "restricted random," with the restrictions being that the first five plots encountered that met the following criteria were surveyed: 1) greater than 1-m2 cover of bryophytes, Table 1. Sites, grouped by region, where rock outcrop surveys were conducted in New York and Maine. Site locations are given in Figure 1. Site abbreviations are used in Figures 1 and 2. Rock types were determined from rock samples taken at each site. Site (abbreviation) Aspect Rock type Adirondack Mountains (A), NY Deer Leap (DL) E Gneiss containing quartz, unaltered feldspars, magnitite and iron Huckleberry Mt. (HM) NNE Well-crystalized granitic gneiss Little Diameter (LD) SSW Gneiss containing amphibolite bands Peltigera Rock outcrop (PL) ESE Calcite inclusions and high calcium carbonate content, formation also containing muscovite, sheet-silicates and mica crystals River Road, RR (4R) SE Well-crystalized granitic gneiss Warner Hill (WH) W Limestone Catskill-Shawangunk (C) Blue Rock outcrop Trail (BLT) NNW Arkose sandstone derived from granite and encompassing different degrees of weathering of feldspars to clay Bonticou Crag (BC) N Quartzite Mt. Road - Taconic Pkwy. (MRTP) W Mudstone (shale with some schist components), hematite stained sheet silicates containing chlorite, quartz and mica, no carbonates North Lake (NL) W Arkose sandstone derived from granite and encompassing different degrees of weathering of feldspars to clay Overlook Mt. (OM) W Arkose sandstone, highly weathered and containing talc, biotite and iron stained silicates Table Rocks (TR) NNE Quartzite Finger Lakes (F) Buttermilk Falls (BF) NNE Calcareous shale and mudstone Chittenango Falls (CF) NNE Limestone Letchworth State Park (LSP) E, N Calcareous mudstone Lick Brook (LB) NNE Calcareous shale and mudstone Stony Brook (SB) ESE Calcareous shale and mudstone gorge; band of schist lacking carbonates Treman/Enfield Glen (TEG) NE Calcareous shale and mudstone gorge Additional NY site: Thacher State Park (TP) NNE Limestone of the Helderberg Escarpment Acadia National Park, Maine (ACAD) Long Pond (MELP) E Gabbro-diorite (calcium containing igneous rock) Duck Harbor Mt. (MEDH) NW Volcanic tuffs mixed with some calcite crystals Mansell Mt. (MEMM) N, NE Coarse granite with feldspars (non-calcareous) 2009 N.L. Cleavitt, S.A. Williams, and N.G. Slack 71 2) at least 2 m high, 3) no soil areas with vascular plant cover, and 4) at least 5 m away from the last surveyed plot. Per plot, a species tally was gathered and an asterisk was placed by all species that achieved greater than 100 cm2 cover (a 10-cm x 10-cm transparent grid divided into 1-cm2 cells was used for reference) within an individual plot. For a given site, a species could then be “dominant” at 0–5 of the plots in the site. While not a detailed cover estimate, this method is repeatable and not overly time demanding while providing a useful quantification of species’ relative importance on the rock outcrops. Each plot was surveyed by both S.A. Williams and N.L. Cleavitt, with each person starting at opposite ends and crossing roughly in the middle to finish where the other person began. Complete sets of vouchers were collected for each site, with additional collections within a site for species that lacked good field characters. This work is based on the collection of over 1200 voucher specimens deposited at New York State Museum (NYS), New York Botanical Garden (NYBG), Bailey Hortorium at Cornell University (BH), and College of the Atlantic Herbarium (HCOA; designated ACAD repository). Data analyses The relationship between species composition of sites was analyzed using non-metric multidimensional scaling (NMS) in PC-ORD (ver. 4.27). The NMS analysis is an ordination method that uses an iterative search for rankings and placement of the analyzed variables to find the solution that minimizes stress (McCune and Mefford 1999). Because NMS uses ranks, it can be considered a non-parametric form of ordination with relaxed assumptions on data structure that are usually more applicable to ecological data (McCune and Mefford 1999). All NMS analyses were run using Sorensen’s distance measure with 40 runs using real data and 50 runs of randomized data. The instability criterion was 0.00001, with 400 as the maximum number of iterations. The final solution was chosen based on the dimensionality with the lowest mean stress from a run comparing randomized to real data (McCune and Mefford 1999). Two ordinations were run: 1) a dataset including 134 bryophyte species that occurred at more than one site was used for NY sites only, and 2) a dataset of 137 species occurring at two or more of all 22 sites (Appendix A). The frequency of a species within a site (number of the five plots where the species was present) was used as the abundance value in the ordination data set. The correlation of available site variables was examined by overlaying joint plot vectors onto the ordination calculated from the species composition data. The explanatory values are expressed as correlation (r2) to the ordination axes. Visually, vectors represent the hypotenuse of a right triangle with adjacent and opposite sides being the r2 values to the two axes displayed (McCune and Grace 2002). Therefore, both the length and angle of the vector express information about the strength of the variable’s correlation to the axes. For the vector joint plot, rock type, region, degree of soil influence (cover classes on surveyed areas), total richness of bryophytes, and number of liverwort species were entered. The rock types were coded as: quartzite, 72 Northeastern Naturalist Vol. 16, No. 1 weathered sandstone and gneiss, sandstone and gneiss with feldspars still present, igneous calcium-containing rocks, calcium-rich mudstones and shales, and limestones. Regions were coded as: Adirondacks, Catskills- Shawangunks, Finger Lakes, or ACAD. Differences in species richness, number of rare species, and number of liverworts between calcareous and non-calcareous rock types was tested by one-way ANOVA. One site (Warner Hill [WH]) was excluded from these analyses because it had very high variability in species richness, which resulted because one of the five restricted-random rock outcrop sections had 100% dominance of a single species. Using histograms and Mann Whitney U-tests, differences between liverworts and mosses were examined for frequency between and within sites and dominance within sites . All non-ordination analyses were preformed in SPSS ver. 14.0. All means are presented with ± 1 SE of the mean. A list of species that occurred at only one NY site is given in Appendix B. Twelve species present at ACAD were absent from NY sites (Appendix B). Because of the potential that they are coastally restricted, these species were left out of the ordination data set, so that they would not cause the Maine sites to cluster together. Rare species designations for NY are based on Cleavitt et al. (2006). Nomenclature follows Crosby et al. (1999) for mosses and Schuster (1966–1992) for liverworts. Results Correlates to bryophyte composition at NY sites Although only the ordination from the full 22-site analysis is shown, the patterns in Figure 2 are also representative of patterns for NY sites alone. Tables 2 and 3 give details on the minor changes between the ordinations. Two main patterns emerged from the NMS ordinations: 1) species composition separated along a rock-type and soil-influence gradient (Fig. 2), and 2) liverworts separated out significantly toward sites with non-calcareous rock types (Fig. 2). In the NY site ordination, both axes relate to rock types, with the ends of the gradient in species composition as three limestone (WH, TP, and CF) and the two quartzite sites (BC and TR) (Table 2a, Fig. 2). There were a total of 182 bryophyte species verified from the NY sites, with 141 mosses and 41 liverworts. Of these species, 134 (101 mosses, 33 liverworts) occurred at two or more of the 19 survey sites (Appendix A), and these were used in the ordination analysis. For the NMS ordination, a two-dimensional solution was best, with relatively low stress (9.51) and instability (p less than 0.0001) indicating a robust ordination appropriate for ecological interpretation (McCune and Grace 2002). Species with high indicator value (i.e., Pearson’s r value indicated a significant correlation) for non-calcareous rock outcrops included: Andreaea rothii (0.636), Blepharostoma trichophyllum (0.625), Dicranum fulvum (0.919), Dicranum montanum (0.831), Jamesoniella autumnalis (0.676), Leucobryum glaucum (0.716), Plagiothecium laetum 2009 N.L. Cleavitt, S.A. Williams, and N.G. Slack 73 (0.618), Pohlia nutans (0.729), Pseudotaxiphyllum distichaceum (0.659), Pylaisiadelphus tenuirostris (0.749), and Scapania nemorea (0.800). Species with higher indicator value (r value) for calcium-containing rock outcrops included: Fissidens bryoides (0.723), Brachythecium oxycladon (0.618), Campylium chrysophyllum (0.676), Encalypta procera (0.666), and Myurella sibirica (0.680). Differences between NY and ACAD composition patterns We found a total of 67 bryophyte species at ACAD, of which 22 were leafy liverworts. A total of 194 bryophyte species were present in the combined site checklists (148 mosses, 46 liverworts) (Appendices A and B). Of these, 137 species (103 mosses, 34 liverworts) occurred at two or Figure 2. Ordination (calculated by non-metric multidimensional scaling) showing the relative similarities and differences in bryophyte species composition for 22 sites in New York and Maine and joint bi-plot with site characteristics. Site abbreviations are those from Table 1. Site abbreviations for Maine sites begin with ME. Sites are coded here by rock-type: ▲ = limestones, ♦ = calcareous mudstones and shales, ● = igneous calcium-containing rocks, ° = sandstones and gneiss with feldspars, Δ = weathered sandstones, gneiss and granites, ◊ = quartzite. Vectors represent the correlation of variables to the ordination axes. Significantly correlated variables were rock type (six rock type classes as coded above), soil (degree of soil influence), region (four study regions), and liverworts (the number of liverworts). Table 3 contains further details of vector correlation significance. 74 Northeastern Naturalist Vol. 16, No. 1 more of the 22 survey sites, and these species were used in the correlation and ordination analyses. The NMS ordination was robust (10.345 = final stress for 2-dimensional solution, 0.00001 = final instability) and explained 91.6% of the variation in bryophyte species composition (Fig. 2). The most notable changes from the ordination with NY sites only were the increased importance of region in explaining variation along axis 2, and the coincident lack of correlation for rock type and soil influence with the second axis (Table 2b). The pattern of liverwort prevalence toward the positive end of axis 1 (non-calcareous rock types) is equally strong in both ordinations (Table 2). As in NY, liverworts accounted for a higher proportion of the bryophyte flora on non-calcareous rock outcrops in ACAD. Mansell Mountain (granite) rock outcrop flora was 39.5% liverworts, Duck Harbor (volcanic tuffs) was 32.5%, and Long Pond (gabbro-diorite) was 26.5%. However, liverworts contributed two-times more to the flora on calcium-containing rock Table 2. Correlation of site traits to the two ordination axes from non-metric multidimensionsal analyses. a) For NY sites (n = 19). Axis 1 Axis 2 Trait Pearson’s r Significance Pearson’s r Significance Rock type -0.930 <0.001 -0.724 <0.001 Region 0.026 ns -0.553 <0.020 Soil influence -0.668 <0.01 -0.807 <0.001 Total species -0.397 ns -0.161 ns No. liverwort spp. 0.668 <0.01 0.326 ns b) For NY and ACAD sites combined (n = 22). Significant relationships are displayed as vectors on Figure 2. Axis 1 Axis 2 Trait Pearson’s r Significance Pearson’s r Significance Rock type -0.915 <0.001 0.110 ns Region 0.136 ns -0.576 <0.01 Soil influence -0.827 <0.001 -0.107 ns Total species -0.331 ns 0.182 ns No. liverwort spp. 0.684 <0.001 -0.359 ns Table 3. Correlations between averages for number of regions, number of sites, and frequency and dominance within sites for mosses (below diagonal; n = 103) and liverworts (above diagonal; n = 35). Values are given in Appendix A. Histogram summaries are shown in Figure 3. Significance values are given as superscripts: ns = not significant at P = 0.05; * denotes P < 0.05; ** P < 0.01; and *** P < 0.001. Regions Sites Frequency Dominance Regions 0.777*** 0.326ns 0.196ns Sites 0.751*** 0.560*** 0.368* Frequency 0.444** 0.627*** 0.620*** Dominance 0.201* 0.356** 0.550*** 2009 N.L. Cleavitt, S.A. Williams, and N.G. Slack 75 outcrops in ACAD than in NY (for NY: N = 10, % liverworts = 13.5 ± 6.3), and this trend is also evident in the placement of ACAD sites on the ordination (Fig. 2). All three Maine sites clustered with the non-calcareous rock floras of upstate NY regardless of rock type (Fig. 2). Patterns of species richness in NY Species richness did not differ significantly between rock types (noncalcareous: 34.7± 3.21, calcareous: 39.7 ± 3.12; F1,17 = 1.25, P = 0.281), although there were significant differences in richness between sites (P < 0.001; range = 23–54 species). Mean species richness within the 2-m by 5-m plot area ranged from 9.2 ± 1.1 to 26.6 ± 2.4 species. The three NY regions (each with six survey sites) had comparable species richness on rock outcrops, with the Adirondacks (109 spp.) and Finger Lakes (107 spp.) regions slightly higher than the Catskill-Shawangunk (97 spp.) region. The Finger Lakes had the highest number of species found only in that region (24 spp.) and also had the greatest number of rare species (8 spp.) based on the updated NY rare moss list (Cleavitt et al. 2006). Although the Finger Lakes has the most unique species on a regional scale, within the region there is a higher number of shared species (41 spp. occurring at four or more of the six sites) than within the other two regions (Adirondacks: 14 spp., and Catskill-Shawangunk: 19 spp.). The high number of shared species in the Finger Lakes relates at least in part to the greater homogeneity in rock types compared to the other two regions. The number of rare species was significantly higher on calcareous (2.13 ± 0.44) relative to non-calcareous rock outcrops (0.78 ± 0.32; F1,16 = 6.26, P = 0.024). Fifteen moss species regarded as rare in NY were found during the rock outcrop surveys and two of these species, Tortula pagorum and Fabronia ciliaris, were only recently reported for NY (Cleavitt et al. 2006, Trigoboff 2005) (Appendices A and B). Three of the other 13 species were collected at three or more of the sites and they may be more widespread than herbarium records indicate (Appendix A). The other ten species are split evenly between S1 (1–5 occurrences in NY) and S2 (6–20 occurrences in NY) ranks (Appendices A and B). Only three of the rare mosses— Pseudotaxiphyllum distichaceum, Schwetschkeopsis fabronia, and Sematophyllum demissum— were collected from non-calcareous rock types; therefore, 80% of the rare species were on calcium-containing rocks. Differences between mosses and liverworts (all sites) Liverworts accounted for a significantly higher percentage of species richness at non-calcareous rock outcrops (10.2 ± 1.04%) than at calcareous rock outcrops (5.89 ± 0.87%) (F1,17 = 10.2, P = 0.006). Mosses were dominant more often than liverworts within a site (Mann Whitney U: z = - 2.245, P = 0.025; Fig. 3). There were no significant differences between mosses and liverworts for the within-site frequency or the number of sites of occurrence (Fig. 3). Frequency and dominance within a site were significantly correlated with higher frequency at both the site and regional scales for mosses. For 76 Northeastern Naturalist Vol. 16, No. 1 liverworts, within-site frequency and dominance were not correlated with the number of regions of occurrence (Table 3). Figure 3. Histograms of moss (left) and liverwort (right) species occurrences with average values (± standard errors), for all sites in New York and Maine. A. Species distribution by the number of sites of occurrence (possible values are 2–22). B. Species distribution by their frequency within sites as the average value across sites (for each site the possible values are 1–5 plots). C. Species distributions by dominance within sites as their average values across sites (for each site the possible values are 0–5, where 5 would mean greater than 100 cm2 cover at all five plots within a site). 2009 N.L. Cleavitt, S.A. Williams, and N.G. Slack 77 Discussion Most of our original hypotheses were supported by the results of this study, with the exceptions that the ACAD sites did not ordinate with the NY sites that had similar rock types and the number of liverworts was a stronger correlate to the composition of bryophyte rock outcrop assemblages than initially anticipated. Although rock type was the strongest correlate to axis 1, which explained 80.5% of the variation in the data, the two calcium-containing rock types in ACAD were closest in bryophyte species composition to the non-calcareous rock types from New York. This pattern resulted from the greater importance of leafy liverworts on all three of the ACAD sites. The higher humidity, including frequent fogs along the coast, may promote leafy liverwort growth and allow these species to compete well on all rock types in ACAD. This hypothesis is partially supported by the restriction of Isothecium stoloniferum and Aulocomnium androgynum to the coastal rock cliffs (species included in Appendix B). These species were found by Hedderson and Brassard (1990) to be mainly limited by moisture availability. The exclusion of most leafy liverwort species from calcareous rock types in NY may be due at least in part to differences in the water-holding capacities or water-retention times of the rocks rather than wholly to differences in rock chemistry (Aho and Weaver 2006). In contrast to leafy liverworts, the thalloid liverworts only occurred on calcium-containing rock types in this study, but their contribution to species composition patterns was small (Appendices A and B). This unanticipated pattern in the occurrence of leafy liverworts demonstrates the complexity of species ecology and suggests that apparent substratum specificity or restrictions may change under different climate regimes. Our finding of a lower proportion of liverwort species relative to mosses on calcareous rocks agrees with a recent paper by Kubešová and Chytrý (2005) and is the first quantification of this pattern for northeastern North America. Our results differed from that study in that we did not find higher species richness on calcium-rich rocks. This discrepancy may partly be due to a difference in methodology. Kubešová and Chytrý (2005) conducted plotless site surveys and found the number of species to increase with cliff area, while we restricted ourselves to the same amount of rock outcrop area at each site. Therefore, while we often noted additional species between surveyed areas, these were not included in the data set presented here. In addition, Kubešová and Chytrý (2005) note that base-rich rock may have a larger range of pH when sampling includes the more base-poor organic soil components. Our methodology limited the amount of such area included to more closely reflect rock chemistry and to target areas without vascular plants, since in such areas bryophyte richness may be lowered by competition (Kuntz and Larson 2006). We interpret soil influence in our study as representing differences in the amount of fracturing and small cracks. The calcareous mudstones and shales were among the most friable rock types and tended to have mineral soil coating the rock surfaces as well 78 Northeastern Naturalist Vol. 16, No. 1 as filling the many small cracks of the rock layers. Rock texture and friability are additional rock characters, other than rock chemistry, that may account in part for some species occurrences on specific rock types. Previous studies comparing bryophyte assemblages across rock types have suggested that limestone had the most unique flora (Bates 1978, Smith 1982). In particular, Bates (1978) conducted work most similar to ours by investigating the bryoflora and chemical signatures of four rock types in Scotland (limestone, acid sandstone, basalt, and ultrabasic rock). While the limestone flora was quite distinct from the flora on the other three rock types, Bates (1978) noted that inclusion of other base-rich sedimentary rocks might make limestone flora appear less differentiated. Our ordination clearly shows this caveat to be correct, with species composition showing a continuous gradient. Without the calcareous mudstone and shale floras, there would have been a large gap in species composition between igneous calcium-containing rocks and limestone, which still forms the end of the calcium-rich gradient (Fig. 2). Our results were similar to Bates (1978) in that no taxon was abundant on all rock types. The most ubiquitous species that we found were Anomodon rostratus, A. attenuatus, Dicranum fulvum, Plagiothecium cavifolium, Pseudotaxiphyllum elegans, and Scapania nemorea. The two species of Anomodon were absent from the most base-poor rocks (quartzite, granites, and weathered calcium-poor sandstone and gneiss) and tended to increase in abundance on calcium-rich rocks. In contrast, Dicranum fulvum and Scapania nemorea were absent from the most calcium-rich rocks (limestones and calcareous mudstones and shales) and increased in abundance on calciumpoor rock types. Both are indicator species for base-poor rock. Plagiothecium cavifolium and Pseudotaxiphyllum elegans were absent from the most basic and the most base-poor rock types, but were abundant elsewhere. While bryophyte assemblages in the northeastern US do show strong separation along a rock-type gradient, the reliability of certain species as indicators of calcium status should be treated with caution for several reasons. Firstly, calcium content of the rock may not be the sole reason for species presence on a rock outcrop. Other important factors to consider are: 1) influence of local climate, 2) importance of rock water relations, and 3) modification of the rock environment by micro-topography (e.g., overhangs, underhangs, crevices) or soil presence. Some species do appear to be robust indicators, if not specifically of calcium status, then more generally of rock pH. For instance, the resolution of Encalypta procera as an indicator species for base-rich rock agrees with the findings of Hedderson and Brassard (1990) for this species in Newfoundland. Andreaea spp. are also well documented as indicators of acid rock types (Heegaard 1997). As suggested by Jenkins (2004), calcareous rock types supported a larger number of rare moss species in NY than non-calcareous rock types. Rare liverworts were not included in this analysis since they are not yet assessed for NY. Many of the rare species for this study (only present at one NY site; Appendix B) were found only on calcium-containing rock types. 2009 N.L. Cleavitt, S.A. Williams, and N.G. Slack 79 The NY species with S1 or S2 ranks from calcareous rocks included: Conardia compacta, Cyrto-hypnum pygmaeum, Didymodon ferrugineus, Mnium ambiguum, Platydictya jungermannioides, Seligeria calcarea, Seligeria donniana, and Taxiphyllum taxirameum (Appendices A and B). Conservation of bryophytes on rock outcrops in the northeast US may be guided in part by the following general conclusions: 1) calcareous rock types support more rare moss species than non-calcareous rock types, 2) non-calcareous rock outcrops are critical for leafy liverwort species richness, and therefore, 3) protection of habitat that includes the full range of rock types present in a region would more fully protect rock bryophyte assemblages than focusing on a single rock type. Acknowledgments Funding for this research was provided through the New York State Biodiversity Research Initiative to N.L. Cleavitt, S.A. Williams, and N.G. Slack, and an L.L. Bean Acadia Research Fellowship to N.L. Cleavitt and S.A. Williams. We are grateful to: David Werier for preparing Figure 1 and helping with site visits around the Finger Lakes region; Professor William A. Bassett of the Department of Earth and Atmospheric Sciences, Cornell University, for identifying rock samples for the summaries provided in Table 1; Alison C. Dibble for assisting with fieldwork on Isle au Haut and providing comments on a previous version of the paper; Troy Weldy and Greg Edinger, New York Natural Heritage, for providing access to information for selection of NY locations; Wayne Barder, Stuart West, and David Manski of Acadia National Park for making our visit to Isle au Haut and survey work in Acadia both pleasant and possible; Jerry Jenkins for assisting with several site visits in the Adirondacks region, and particularly for canoe transport to WH; Steve Rice, Union College, for acting as guest editor on this paper; and two anonymous reviewers for helpful suggestions. Literature Cited Aho, K., and T. Weaver. 2006. Measuring water relations and pH of cryptogam rocksurface environments. The Bryologist 109:348–357. Allen, B. 2005. Maine mosses: Sphagnaceae–Timmiaceae. Memoirs of The New York Botanical Garden 93:1–420. The New York Botanical Garden Press, Bronx, NY. Bates, J.W. 1978. The influence of metal availability on the bryophyte and macrolichen vegetation of four rock types on Skye and Rhum. Journal of Ecology 66:457–482. Bates, J.W. 1993. Regional calcicoly in the moss Rhytidiadelphus triquetrus: Survival and chemistry of transplants at a formerly SO2-polluted site with acid soil. Annals of Botany 72:449–455. Belland, R.J. 2005. A multivariate study of moss distributions in relationship to environment in the Gulf of St. Lawrence region, Canada. Canadian Journal of Botany 83:243–263. Cleavitt, N.L. 2001. Disentangling moss species limitations: The role of substrate specificity for six species occurring on substrates with varying pH and percent organic matter. The Bryologist 104:59–68. 80 Northeastern Naturalist Vol. 16, No. 1 Cleavitt, N.L., S.A. Williams, and N.G. Slack. 2006. Updating the rare moss list for New York State: Ecological community and species-centered approaches. Final report for the Biodiversity Research Initiative. NY State Museum, Albany, NY. Crosby, M.R., R.E. Magill, B. Allen, and S. He. 1999. A checklist of the Mosses. Available online at http://www.mobot.org/MOBOT/tropicos/most/checklist. Accessed in 2001. Gilman, R.A., and C.A. Chapman. 1988. Bedrock Geology of Mount Desert Island. Bulletin 38, Maine Geological Survey, Augusta, ME. Hedderson, T.A., and G.R. Brassard. 1990. Microhabitat relationships of five cooccurring saxicolous mosses on cliffs and scree slopes in eastern Newfoundland. Holarctic Ecology 13:134–142. Heegaard, E. 1997. Ecology of Andreaea in western Norway. Journal of Bryology 19:527–636. Horton, D.G. 1988. Microhabitats of new world Encalyptaceae (Bryopsida): Distribution along edaphic gradients. Beiheft zur Nova Hedwigia 90:261–282. Jenkins, J. 2004. The Adirondack Atlas: A geographic portrait of the Adirondack Park. Wildlife Conservation Society, Bronx, NY. Ketchledge, E.H. 1980. Revised checklist of the mosses of New York State. New York State Museum Bulletin 440. Albany, NY. 19 pp. Kubešová S., and M. Chytrý. 2005. Diversity of bryophytes on treeless cliffs and talus slopes in a forested central European landscape. Journal of Bryology 27:35–46. Kuntz, K.L., and D.W. Larson. 2006. Microtopographic control of vascular plant, bryophyte, and lichen communities on cliff faces. Plant Ecology 185:239–253. McCune, B., and J.B. Grace. 2002. Analysis of ecological communities. MjM Software, Gleneden Beach, OR. 300 pp. McCune, B., and M.J. Mefford. 1999. PC-ORD. Multivariate analysis of ecological data, version 4. MjM Software Design, Gleneden Beach, OR. Miller, N.G., and S.F. McDaniel. 2004. Bryophyte dispersal inferred from colonization of an introduced substratum on Whiteface Mountain, New York. American Journal of Botany 91:1173–1182. Patterson, P.M. 1930. The mosses of Mt. Desert Island, Maine. The Bryologist 33:83–89. Piippo, S. 1982. Epiphytic bryophytes as climatic indicators in Eastern Fennoscandia. Acta Botanica Fennica 119:1–39. Schuster, R.M. 1949. The ecology and distribution of hepaticae in central and western New York. American Midland Naturalist 42:513–712. Schuster, R.M. 1966–1992. The Hepaticae and Anthocerotae of North America. Volumes I–VI. Columbia University Press, New York, NY. Shaw, A.J. 1981. Ecological diversification among nine species of Pohlia (Musci) in western North America. Canadian Journal of Botany 59:2359–2378. Smith, A.J.E. 1982. Epiphytes and epiliths. Pp. 191–227, In A.J.E. Smith (Ed.). Bryophyte Ecology. Chapman and Hall, New York, NY. Trigoboff, N. 2005. Tortula papillosa and Tortula pagorum (Pottiaceae) in New York State. Evansia 22:85–89. 2009 N.L. Cleavitt, S.A. Williams, and N.G. Slack 81 Appendix A. Summary table of the 137 bryophyte species included in the ordination for 22 sites occurring on rock outcrops. Rare species are indicated by a superscript with their S-rank for NY. R = regions of occurence, which are abbreviated as: A = Adirondacks, C = Catskill-Shawangunk, F = Finger Lakes, and M = ACAD, Maine. Frequency is the number of cliff sections (maximum of 5) within a location in which the species occurred. Dominance is the number of sections (maximum of 5) within a location where the species cover was greater than 100 cm2. No. Average Average Species sites frequency dominance R Mosses Amblystegium serpens (Hedw.) Schimp. 3 1.0 0.5 AC A. varium (Hedw.) Lindb. 3 2.3 0.3 AF Amphidium lapponicum (Hedw.) Schimp. 2 1.0 0.0 CF A. mougeotii (Bruch. & Schimp.) Schimp. 4 1.8 0.0 CM Andreaea rothii F. Weber & D. Mohr. 7 3.4 1.8 ACM A. rupestris Hedw. 5 3.4 0.7 CM Anomodon attenuatus (Hedw.) Huebener 13 3.1 1.0 ACF A. rostratus (Hedw.) Schimp. 16 3.2 1.1 ACFM A. viticulosus (Hedw.) Hook. & Taylor 5 3.0 1.8 AF Atrichum angustatum Bruch & Schimp. 2 1.0 0.0 F A. oerstedianum (Müll. Hal.) Mitt. 2 1.5 0.0 F Aulacomnium palustre (Hedw.) Schwägr. 2 1.0 0.5 A Bartramia pomiformis Hedw. 10 2.5 0.4 ACFM Brachythecium oxycladon (Brid.) A. Jaeger 10 3.4 2.0 ACF B. plumosum (Hedw.) Schimp. 2 1.0 0.5 F B. rivulare Schimp. 3 2.0 1.3 AF B. rutabulum (Hedw.)Schimp. 2 1.5 0.0 F B. salebrosum (Hoffm.) Schimp. 5 1.2 0.6 AF B. velutinum (Hedw.) Schimp. 6 1.0 0.0 ACF Brotherella recurvans (Michx.) M. Fleisch. 3 1.6 0.0 CM Bryhnia graminicolor (Brid.) Grout 10 2.4 0.1 ACF Bryoerythrophyllum recurvirostrum (Hedw.) P.C. Chen 6 2.7 0.0 AF Bryum caespiticium Hedw. 3 1.7 0.3 A B. flaccidum Brid. 7 2.4 0.0 ACF B. lisae var cuspidatum (Bruch. & Schimp.) Margad. 3 1.3 0.0 AF B. pseudotriquetrum (Hedw.) P. Gaertn. 9 2.3 0.0 ACFM Callicladium haldanianum (Grev.) H.A. Crum 2 1.0 0.0 F Campylium chrysophyllum (Brid.) Lange 12 2.5 0.5 ACF C. hispidulum (Brid.) Mitt. 2 1.0 0.0 AF Ceratodon purpureus (Hedw.) Brid. 4 1.5 0.0 AM Conardia compactaS1 (Müll. Hal.) H. Rob. 2 1.5 0.0 F Cratoneuron filicinum (Hedw.) Spruce 6 2.0 0.7 AF Dichodontium pellucidum (Hedw.) Schimp. 3 1.3 0.0 AF Dicranella heteromalla (Hedw.) Schimp. 5 1.6 0.4 CF D. varia (Hedw.) Schimp. 3 2.3 0.0 F Dicranum fulvum Hook. 15 3.6 2.2 ACFM D. montanum Hedw. 13 2.7 0.2 ACM D. scoparium Hedw. 5 2.2 0.3 ACM Didymodon rigidulus Hedw. 5 1.8 0.4 AF D. ferrugineusS1/S2 (Schimp.) M.O. Hill 2 1.0 0.0 F Diphyscium foliosum (Hedw.) D. Mohr 7 1.3 0.0 ACM Encalypta procera Bruch 7 3.3 0.7 AF Eurhynchium hians (Hedw.) Sande Lac. 2 3.0 0.0 F Fabronia ciliarisS1 (Brid.) Brid. 2 1.0 0.0 AF Fissidens adianthoides Hedw. 3 1.3 0.3 F F. bryoides Hedw. 9 3.3 0.0 AF F. dubius P. Beauv. 9 2.0 0.2 ACFM F. taxifolius Hedw. 5 2.2 0.0 F 82 Northeastern Naturalist Vol. 16, No. 1 No. Average Average Species sites frequency dominance R Gymnostomum aeruginosum Sm. 8 3.3 1.6 ACF Hedwigia ciliata (Hedw.) P. Beauv. 4 1.7 0.2 ACM Herzogiella striatella (Brid.) Z. Iwats. 2 2.0 0.5 CM Homalia trichomanoides (Hedw.) Schimp. 3 1.3 0.0 C Homomallium adnatum (Hedw.) Broth. 5 1.4 0.0 ACF Hygroamblystegium tenax (Hedw.) Jenn. 8 1.8 0.8 AF Hygrohypnum luridum (Hedw.) Jenn. 2 1.0 0.5 F Hymenostylium recurvirostrum (Hedw.) Dixon 6 2.8 1.8 AF Hypnum cupressiforme Hedw. 4 3.0 3.0 AM H. imponens Hedw. 4 2.3 1.0 CM Isopterygiopsis muelleriana (Schimp.) Z. Iwats. 5 1.8 0.2 CF Leskeella nervosa (Brid.) Loeske 2 1.0 0.0 A Leucobryum glaucum (Hedw.) Ångström 8 2.7 0.0 ACFM Mnium marginatum (Dicks.) P. Beauv. 10 2.8 0.4 AF M. thomsonii Schimp. 3 1.0 0.0 F Myurella julaceaS2 (Schwägr.) Schimp. 2 1.0 0.0 A M. sibirica (Müll. Hal.) Reimers 10 2.8 0.0 AF Orthotrichum anomalum Hedw. 3 2.0 0.0 A Oxystegus tenuirostris (Hook. & Taylor) A.J.E. Sm. 9 1.9 0.0 ACFM Paraleucobryum longifolium (Ehrh.) Loeske 5 1.8 0.2 CM Philonotis fontana (Hedw.) Brid. 3 1.3 0.7 A P. marchica (Hedw.) Brid. 3 2.0 0.0 ACF Plagiomnium ciliare (Müll. Hal.) T.J. Kop. 2 2.5 0.0 F P. cuspidatum (Hedw.) T.J. Kop. 10 2.1 0.2 ACF P. rostratum (Schrad.) T.J. Kop. 6 2.8 0.5 F Plagiopus oederiana (Brid.) Limpr. 4 2.0 0.8 AF Plagiothecium cavifolium (Brid.) Z. Iwats. 12 3.5 1.1 ACF P. denticulatum (Hedw.) Schimp. 4 1.8 0.0 AC P. laetum Schimp. 7 1.8 0.3 ACM Platydictya confervoides (Brid.) H.A. Crum 3 2.7 0.0 AF P. jungermannioidesS2 (Brid.) H.A. Crum 3 2.0 0.0 AF Platygyrium repens (Brid.) Schimp. 6 1.5 0.0 AC Platyhypnidium riparioides (Hedw.) Dixon 3 1.0 0.0 F Pohlia cruda (Hedw.) Lindb. 5 2.8 0.0 ACFM P. nutans (Hedw.) Lindb. 10 3.8 0.4 ACFM Polytrichastrum alpinum (Hedw.) G.L. Sm. 6 3.0 0.7 ACM Polytrichum pallidisetum Funck 3 1.0 0.0 AC Pseudotaxiphyllum distichaceumS2/S3 (Mitt.) Z. Iwats. 4 2.0 0.3 AC P. elegans (Brid.) Z. Iwats. 13 3.7 0.6 ACFM Pterigynandrum filiforme Hedw. 4 1.8 0.3 AC Pylaisiadelpha tenuirostris (Bruch & Schimp.) W.R. Buck 6 2.0 0.3 AC Racomitrium heterostichum Cardot 5 1.8 0.8 AC Rhabdoweisia crispata (Dicks.) Lindb. 5 2.8 1.5 CM Rhizomnium punctatum (Hedw.) T.J. Kop. 2 1.0 0.0 F Rhodobryum ontariense (Kindb.) Paris 2 1.0 0.0 AF Rosulabryum capillare (Hedw.) J.R. Spence 5 2.6 0.0 AC Schistdium apocarpum (Hedw.) Bruch & Schimp. 9 2.0 0.3 ACFM S. lancilifolium (Kindb.) H.H. Blom 3 2.3 0.3 CF Steerecleus serrulatus (Hedw.) H. Rob. 4 1.7 0.0 FM Taxiphyllum deplanatum (Schimp.) M. Fleisch. 3 1.0 0.0 F Tetraphis pellucida Hedw. 4 1.0 0.0 CM Thuidium delicatulum (Hedw.) Schimp. 13 1.4 0.4 ACFM Tortella tortuosa (Hedw.) Limpr. 9 2.2 0.8 ACF Tortula mucronifolia Schwägr. 2 1.0 0.0 F Ulota hutchinsiae (Sm.) Hammar 6 2.0 0.2 ACM 2009 N.L. Cleavitt, S.A. Williams, and N.G. Slack 83 No. Average Average Species sites frequency dominance R Liverworts Anastrophyllum michauxii (F. Weber) H. Buch. 3 1.3 0.0 ACM A. minutum (Schreb.) R.M. Schust. 6 1.3 0.0 ACM Barbilophozia attenuata (Nees) Loeske 6 1.8 0.3 ACM Bazzania denudata (Torr.) Trevis. 3 2.0 0.0 C B. trilobata (L.) Gray. 6 2.3 0.7 CFM Blepharostoma trichophyllum (L.) Dumort. 3 1.3 0.0 AC Calypogeia fissa (L.) Raddi 2 1.5 0.0 C C. muelleriana (Schiffner) K. Müller 2 1.0 0.0 AC Cephaloziella rubella (Nees) Warnst. 7 1.9 0.0 ACF C. stellulifera (Taylor) Schiffner 2 1.0 0.0 F Cololejeunea biddlecomiae (Austin) A. Evans 5 1.8 0.0 ACF Conocephalum conicum (L.) Underw. 7 2.1 1.1 AF Diplophyllum apiculatum (A. Evans) Stephani 5 2.4 0.3 CFM Frullania eboracensis Gottsche 4 1.5 0.0 AC Jamesoniella autumnalis (DC.) Stephani 10 2.3 0.1 ACFM Lejeunea cavifolia (Ehrh.) Lindb. 6 1.2 0.0 CF Lepidozia reptans (L.) Dumort 5 1.4 0.0 CM Lophocolea heterophylla (Schrad.) Dumort. 4 1.5 0.0 CFM L. minor Nees 2 2.0 0.0 A Lophozia ventricosa (Dicks.) Dumort. 5 2.2 0.0 ACM Marchantia polymorpha L. 2 1.0 0.0 F Marsupella emarginata (Ehrh.) Dumort. 2 2.5 1.0 AM Metzgeria conjugata Lindb. 6 2.2 0.5 CF M. furcata (L.) Corda 2 1.5 0.0 C Plagiochila asplenioides (L.) Dumort. 7 1.4 0.3 CFM Porella platyphylla (L.) Pfeiff. 5 1.4 0.3 ACM Preissia quadrata (Scop.) Nees 4 2.3 0.5 F Ptilidium pulcherrimum (Weber) Hampe 9 2.1 0.0 ACM Radula complanata (L.) Dumort. 8 2.1 0.0 ACFM R. tenax Lindb. 3 2.0 0.0 C Reboulia hemisphaerica (L.) Raddi 3 1.7 0.0 AF Scapania mucronata H. Buch. 3 2.3 0.0 A S. nemorea (L.) Grolle 12 3.7 1.0 ACFM Tritomaria exsecta (Schmider) Schiffner 3 1.3 0.0 AC T. exsectiformis (Breidl.) Schiffner 3 1.0 0.0 AC Appendix B. List of the species not included in the ordination. a) Species that were found at only one of the 19 sites surveyed in NY, and b) species that were found only at the ACAD (Maine) sites. For NY, rare species are indicated by a superscript with their S-rank. a) Species that were found at only one of the NY sites: Mosses Abietinella abietina (Hedw.) M. Fleisch. Amblystegium riparium (Hedw.) Schimp. Aulacomnium heterostichum (Hedw.) Bruch & Schimp. Barbula unguiculata Hedw. Brachythecium acuminatum (Hedw.) Austin B. populeum (Hedw.) Schimp. Bryhnia novae-angliae (Sull. & Lesq.) Grout Campylium polygamum (Schimp.) C.O.E. Jensen Coscinodon cribrosusS1 (Hedw.) Spruce Cyrto-hypnum pygmaeumS2 (Schimp.) W.R. Buck & H.A. Crum Ditrichum lineare (Sm.) Lindb. 84 Northeastern Naturalist Vol. 16, No. 1 Encalypta ciliata Hedw. Fissidens subbasilaris Hedw. Funaria hygrometrica Hedw. Grimmia pilifera P. Beauv. Hygroamblystegium fluviatile (Hedw.) Loeske Loeskeobryum brevirostre (Brid.) M. Fleisch. Hypnum lindbergii Mitt. Leucodon andrewsianus (H.A. Crum & L.E. Anderson) W.D. Reese & L.E. Anderson Mnium ambiguumS2 H. Müll. M. spinulosum Bruch & Schimp. M. stellare Reichard Orthotrichum stellatum Brid. O. strangulatum P. Beauv. Pohlia annotina (Hedw.) Lindb. Pohlia bulbifera (Warnst.) Warnst. Racomitrium aciculare (Hedw.) Brid. Schistidium rivulare (Brid.) Podp. Schwetschkeopsis fabroniaS1 (Schwägr.) Broth. Seligeria calcareaS1 (Hedw.) Bruch & Schimp. S. donnianaS2 (Sm.) Müll. Hal. Sematophyllum demissumS1 (Wilson) Mitt. Taxiphyllum taxirameumS2 (Mitt.) M. Fleisch. Thamnobryum subserratum (Hook.) Nog. & Z. Iwats. Thuidium scitum (P. Beauv.) Austin Timmia megapolitana Hedw. Tortella fragilis (Hook. & Wilson) Limpr. Tortula pagorumS1 (Milde) De Not. Warnstorfia fluitans (Hedw.) Loeske Weissia controversa Hedw. Liverworts Aneura pinguis (L.) Dumort. . Bazzania tricrenata (Wahlenb.) Trevis. Calypogeia neesiana (C. Massal. & Carestia) K. Müll. Frullania tamarisci var. asagrayana (Mont) S. Hatt. F. riparia Hampe Herbertus aduncus subsp. tenuis (A. Evans) A.A. Mill. & E.B. Bohrer Lophozia bicrenata (Schmidel) Dumort. Porella pinnata L Radula obconica Sull. Scapania scandica (Arnell & H. Buch.) Macvicar b) Species that were found only at the ACAD sites in Maine: Mosses Aulacomnium androgynum (Hedw.) Schwägr. Hylocomium splendens (Hedw.) Schimp. Hypnum plicatulum (Lindb.) A. Jaeger Isothecium stoloniferum Brid. Mnium hornum Hedw. Racomitrium microcarpum (Hedw.) Brid. Sphagnum compactum Lam. & DC. Liverworts Cephalozia lunulifolia (Dumort.) Dumort. Gymnocolea inflata (Huds.) Dumort. Lophozia longidens (Lindb.) Macoun Nowellia curvifolia (Dicks.) Mitt. Scapania undulata (L.) Dumort.