Pitch pine (Pinus rigida Mill.) communities in the Hudson Valley region of New York.
This is one of a coordinated series of studies on New York pitch pine (Pinus rigida Mill.) barrens communities which are scattered widely throughout the State [ILLUSTRATION FOR FIGURE 1 OMITTED]. The first focused on sites in western and central New York (Seischab and Bernard, 1991), another on northeastern New York (Bernard and Seischab, 1995). The purpose of this study was to describe the barrens communities in sites in SE New York as a continuation of our study to characterize New York pine barrens [ILLUSTRATION FOR FIGURE 1 OMITTED]. We also wanted to compare, where possible, the vegetation and species composition of the Hudson Valley barrens to those in central, western and northeastern New York.
The southeastern area is of particular interest because two of the best known pine barrens in New York, the Albany Pine Bush and the Schawangunk Mountains dwarf forest are located there. The Pine Bush has been the subject of a number of publications, most addressed to a general audience (Rittner, 1975, for example), but Milne (1985) made a detailed study of upland forests and their successional development. He suggested that areas dominated by scrub oak (Quercus ilicifolia) with scattered pitch pine were characteristic of very disturbed sites, especially those that burned often. In the absence of disturbance, Milne believed they would succeed to closed-canopy pitch pine forests, then to areas dominated by a mixture of pitch pine, white pine (Pinus strobus), red oak (Quercus rubra), white oak (Q. alba), red maple (Acer rubrum) and white ash (Fraxinus americana).
Schawangunk Mountain ridge vegetation was described by McIntosh (1959), who noted that bare rock surfaces of the ridges were covered with pitch pine, sometimes dwarfed, no more than 1-3 m tall. These dwarf ridge top communities graded into taller pitch pine and oak forests on the slopes of the mountain. McIntosh compared the flora of the mountains to that of the New Jersey pine barrens and concluded the resemblance was chiefly physiognomic, the apparent similarity due more to the occurrence of certain wide-ranging xeric species common to the two areas than to a high degree of similarity of all the flora.
Our other studies in upstate New York [ILLUSTRATION FOR FIGURE 1 OMITTED] found that pine barrens occurred on deep glacial outwash sands and gravels, on sandstone and granitic gneiss outcrops, on channery soils, and in sterile old fields. Their distribution was in a narrow band around the Adirondack Mountains, on hard rock sites in the Thousand Island Region of the St. Lawrence Valley, and on outwash sand and gravel, lacustrine delta, lacustrine silt, kame moraine, dunes and sterile, rocky old-field sites on the Ontario Plain. In central and western New York they occurred on exposed sandstone, shale, lacustrine sand and kame moraine.
Many of these communities were less than 0.1 ha, surrounded by oak-hickory or mesic northern hardwood forests. Further, some were succeeding to northern hardwood forest and most showed evidence of fire, timber cutting, agriculture or various building projects. For example, two of these communities, Penfield and Junius [ILLUSTRATION FOR FIGURE 1 OMITTED] had been invaded by red maple, black oak (Quercus velutina) and black cherry (Prunus serotina); the pitch pine were either dead or dying. In other cases, larger tracts of these communities had been fragmented by development [Rome, Glens Falls [ILLUSTRATION FOR FIGURE 1 OMITTED]].
Sandy sites in central and western New York had little white oak and white pine, species that were widely encountered at sites on shallow stony soils of sandstone, siltstone or shale origin (Seischab and Bernard, 1991). Community classification in this study resulted in the separation of plots with a dominant ericaceous shrub layer, those which were sparsely ericaceous and those with an abundance of mesic herbs and shrubs. In northeastern New York, pitch pine communities (Bernard and Seischab, 1995) were on rock pavement of granitic gneiss or sandstone or on deep sands. Rock site plots were dominated by oaks (Quercus spp.) and white pine (Pinus strobus) with extensive coverage by Cladina stellaris and C. rangiferina with varying amounts of ericaceous shrubs. Sandy site plots differed primarily in the extent of the their ericaceous shrub cover, dominance of oaks and white pine in their canopy and the degree of mesophytism in the herb layer.
The three most northern sites sampled in this study, Schaghticoke, Cairo and Albany, were all on sandy soils [ILLUSTRATION FOR FIGURE 1 OMITTED]. The Albany site was on a sand dune in the Pine Bush adjacent to a highway on the W and the New York State Thruway on the S. It bordered one of the most disturbed areas where fires were common (Milne, 1985). The Schaghticoke site was on part of a lacustrine delta; the stand was adjacent to a road and bisected by an electric powerline right-of-way.
The Cairo site was on outwash sand and gravel of proglacial fluvial deposition. It was bordered on two sides by paved roads. There was evidence that this site had been farmed at one time.
All the other sites were on conglomerate rock on Schunnemunk Mountain and in the Schawangunk Mountains, in areas under control of the State Park system or in private preserves that have not allowed disturbance.
Three different sites were sampled in Minnewaska State Park (Route 44, Garage, Castle Point), three in the Ice Caves area (Lake Maratanza, Ice Caves, a wet site with Sphagnum moss) and three in the Mohonk Preserve (Trapp's Bridge, Bonticou Crag, Johanson Parcel). One site was examined on Schunnemunk Mountain, S of the Shawangunk Ridge.
Seventy-three 10 x 10 m plots were established in 1994 at 13 sites containing pitch pine in the Hudson Valley region of New York. Plots were contiguous with one another within each of these sites. Diameters of all woody stems over 2.5 cm dbh were measured in each plot. The largest pitch pine in each plot was cored at 1.2 m above ground and aged to approximate the age of each of these communities. In the open canopy stands with short trees, cores at 1.2 m height may have resulted in an underestimation of tree age. In dwarfed communities cores were removed at 20 cm above the soil surface. The height of the canopy was determined with a Suunto clinometer. Percent cover of each of the shrub, herb, bryophyte and lichen species was estimated. Incremental percent cover estimates of [less than]1%, 1-5% and then 5% intervals to 100% were used.
Relative importances (IV) of the woody species with stems 2.5+ cm dbh were calculated by the formula (relative density + relative basal area)/2 (McIntosh, 1957). A species by plot data matrix of woody species IVs and shrub and herb relative cover values was established. Vegetation data from each plot were relativized to 100%.
The data matrix was subjected to the classification program TWINSPAN (Hill, 1979a) to determine the communities present. In order to determine environmental relations to the vegetation the matrix of 73 plots and 113 species was also subjected to detrended correspondence analysis (DCA) (Hill, 1979b).
Soil samples were collected in each ploth to a depth of 10 cm. On rock sites where soils were less than 10 cm deep the samples were taken to the surface of the rock. Soil pH and specific conductance were determined on slurries of 1:1 soil to distilled water. For sites other than those on flatrock, the percent sand, silt, clay and fractions of very coarse sand, coarse sand, medium 1 sand, medium 2 sand, fine sand, and very fine sand were determined according to Bouyoucos (1936). Soil texture determinations could not be made at other sites since the soils consisted of organic matter deposited on conglomerate rock. Organic matter content was determined as percent loss on ignition.
Spearman Rank correlations (Sokal and Rohlf, 1981) were determined between ordination axes, soil variables, tree height, tree age and species richness data (total species richness, overstory richness and understory richness). A stepwise multiple regression procedure (STATISTIX, Analytical Software) was used to analyze the relationship between DCA ordination scores and environmental characteristics of plots.
Vascular plant nomenclature follows Gleason and Cronquist (1991); lichen nomenclature, Dirig (1994); and bryophyte nomenclature, Crum and Anderson (1981).
All sites had acidic soils with very low specific conductance (S) values (Table 1); however, those on conglomerate rock sites in the Shawangunk Mts. had a substantially lower pH than sandy site soils. The three sites in the northern part of the study area had sandy soils but the Schaghticoke and Cairo soils contained considerably more silt and clay than the Albany site (Table 1).
Age of the oldest tree varied considerably from site to site and was a function of time since disturbance, primarily fire history. The Cairo and Schagticoke sites had taller trees and greater species richness than any of the other sites.
TWINSPAN. - The first division of the TWINSPAN dendrogram [ILLUSTRATION FOR FIGURE 2 OMITTED] separated plots with a closed forest (b) with white pine as a dominant, red maple a major species and a layer of mesic herbs from those with shorter trees, a more open canopy and a distinct ericaceous shrub cover [[ILLUSTRATION FOR FIGURE 2(A) OMITTED] and Table 2]. The length of vertical lines in the TWINSPAN dendrogram are based on eigenvalues which represent ecological distances between paired groups (Hill, 1979a). Group (a) plots, dominated by pitch pine and ericaceous shrubs (68% cover), divided into groups (c) and (d). Group (c) plots were located in a wet depression in conglomerate rock near Lake Maratanza and its vegetation included a carpet of Sphagnum sp. and typical bog shrubs such as Chamaedaphne calyculata and Kalmia angustifolia. Pitch pine in these plots averaged 5.3 m in height: twice the height of those in the adjacent Lake Maratanza plots.
Group (d) included plots on conglomerate and sandstone substrate and all of the Albany sand dune plots. Group (d) divided into group (g) plots all on conglomerate from the Shawangunk Mrs. and (h) on conglomerate in the Shawangunks, conglomerate and sandstone on Schunnemunk Mt. and sand dunes in the Albany Pine Bush. The plots in group (g) had a greater proportion of Betula populifolia, Gaylussacia baccata, scrub oak and Pteridium aquilinum than did (h) (Table 2). Group (g) had a more dense ericaceous shrub cover (111% absolute cover), due to stratification of shrub species, than did (h) (76%) absolute [TABULAR DATA FOR TABLE 1 OMITTED] cover) (Table 2). There was also considerably more scrub oak in group (g) than (h); most of the scrub oak in (h) can be attributed to the Albany plots. Group (g) plots had an average canopy height of 6.0 m and group (h) had a slightly taller average height of 6.6 m.
All plots in groups (i) and (j) were on conglomerate slab rock in the Shawangunk Mts. These plots had varying amounts of exposed rock, shallow organic soils on which stunted [TABULAR DATA FOR TABLE 2 OMITTED] pitch pine grew. Trees in group (i) averaged 4.2 m in height while those in group (j) were slightly taller (6.6 m). Vegetatively, both groups had a distinct Gaylussacia baccata shrub layer. Group (i) included Betula species in the tree and shrub layers while group (j) had Vaccinium angustifolium in the shrub layer.
Group (k) plots were mostly from the Shawangunks but included the seven Albany plots as well. Group (1) plots were all from Schunnemunk Mt. Vegetationally, these two groups were similar to groups (i) and (j) in that they had a distinct Gaylussacia baccata shrub layer. This shrub layer also included one or more of the following in group (1): Vaccinium angustifolium, V. pallidum and/or Quercus ilicifolia. The average canopy height in group (k) was 5.5 m, however, this group included 12 plots in which pitch pine were dwarfed and averaged only 2.9 m in height. Group (1) was vegetatively similar to group (k); however, sapling and tree densities were much lower [group (k) = 277 stems/ha, group (1) = 53 stems/ha]. The greatest stem density was seen in the 12 plots containing dwarfed pitch pine (277 stems/ha).
Group (e) from Schaghticoke and (f) from Cairo were on lacustrine deltaic sands and alluvium, respectively. Vegetation at these two sites had a clearly mesophytic herb layer including Parthenocissus quinquefolia, Maianthemum canadense, Lycopodium complanatum and Trientalis borealis. These two sites had, by far, the tallest canopies (14.8 m and 15.2 m) and contained Pinus strobus, red maple, white birch and trembling aspen (Populus tremuloides) as canopy dominants.
Detrended Correspondence Analysis (DCA). - Two initial ordinations were performed, one including and one excluding pitch pine. The ordination without pitch pine separated plots better and was more interpretable. With pitch pine included in the ordination (not shown) the pine created a central tendency, pulling all of the plots towards the center of the ordination. By eliminating pitch pine from the ordination, this distortion was eliminated [ILLUSTRATION FOR FIGURE 3 OMITTED]. The Schaghticoke and Cairo plots were separated to the right and the wet site plots to the left. Wet sites lacked Gaylussacia baccata, a widespread species in the Albany, Schunnemunk and Shawangunk plots which clustered in the center of the ordination. [TABULAR DATA FOR TABLE 3 OMITTED] There was a significant increase in pH, percent silt, and clay, tree height, and species richness and a decrease in percent sand (particularly medium 1 and medium 2 sand) and tree age along the first ordination axis. The second axis was significantly positively correlated with soil conductance, percent coarse and very coarse sand, percent organic matter, tree age and understory species richness and negatively correlated with pH, percent fine sand and tree stem density (Table 3).
Sandy site plots (Schaghticoke, Albany and Cairo) separated widely in this ordination. Albany plots were clustered in the center of the ordination, Schaghticoke and Cairo to the right. These three sites were vegetationally quite distinct in that the Schaghticoke and Cairo plots contained numerous mesic herbs and a greater tree species richness than did Albany plots. Of the three sites, Cairo had the greatest tree species richness including Acer pensylvanicum, Amelanchier laevis, Betula lenta, Carpinus caroliniana, Carya cordiformis, Fraxinus americana, Ostrya virginiana and Prunus serotina, none of which were found in either of the other sandy sites. Tree height was greater at the Cairo (15.2 m) and Schaghticoke (14.8 m) sites than at Albany (7.7 m).
A stepwise multiple regression of environmental characteristics of plots on the ordination of sandy sites showed that the percent of clay (CL), coarse sand (CS), very coarse sand (VCS), very fine sand (VFS), and slope accounted for most of the variation along the first axis (DCA1 = 223.6 [+ or -] 14.6 CL + 8.3 CS + 4.9 VCS + 10.5 VFS + 2.4 SLOPE, [R.sup.2] = 0.98). Much of the second axis variation was accounted for by percent coarse sand (CS), very coarse sand (VCS) and percent silt (DCA2 = 1.41 + 3.91 CS + 0.58 VCS + 1.37 SI, [R.sup.2] = 0.98).
Since the majority of the plots were clustered tightly near the center of the ordination, it proved difficult to determine ecological relationships within this group. These clustered plots were, therefore, ordinated separately without the wet site, Schaghticoke or Cairo plots. In this second ordination [ILLUSTRATION FOR FIGURE 4 OMITTED], the Albany plots (a) on a sand substrate are clustered to the left, the Bonticou Crag (b) and Trapp's Bridge (t) plots, with the most exposed rock and on steeper slopes, are to the right. Trees on these slopes (b and t) were stunted with an average height of 4-6 m. Taller trees were found in plots in the center of the ordination; at the Minnewaska-Hwy. 44 (4) site where trees averaged 8.5 m in height. In the Albany plots trees averaged 7.7 m in height (Table 1). Dwarf pitch pine (2-2.5 m) dominated the trees at the Lake Maratanza (1) and Ice Caves (c) sites. However, pitch pine densities were much greater at the Lake Maratanza site (4500 stems/ha) than at the Ice Caves (260 stems/ha). The shrub layer at Lake Maratanza was dominated by Gaylussacia baccata (52% cover) and Vaccinium pallidum (24% cover) while at the Ice Caves Gaylussacia (87% cover) shared dominance with Kalmia angustifolium (38% cover). Sites with taller trees (Schunnemunk Mt. (s), Minnewaska-Hwy 44 (4) and Johansen Parcel (j) were at the bottom of the ordination. The plots from Minnewaska-Castle Point (p) are in the center of the ordination. A physiognomic gradient is seen in this ordination proceeding clockwise from the dwarfed ([less than]2.5 m) to the stunted (4-6 m) to the taller (7-8.5 m) trees.
Species responses to this ordination are shown in Figure 5. Pitch pine has increasing importance values along the second axis. Gaylussacia baccata decreases in cover along the first axis in those plots where exposed rock is most widespread; its greatest cover is on the sandy Albany site to the left of the ordination. Vaccinium species were found in lesser amounts in sandy sites but were widely distributed on the rocky sites. The most widely found Vaccinium species were V. angustifolium and V. pallidum. Scrub oak had its greatest cover to the left on the ordination but was missing from 17 plots in the center. Neither Kalmia species was found on the sandy site but they were widely distributed on the conglomerate sites with K. angustifolium being most widely distributed and K. latifolia being found in the steeper sloped, exposed plots.
The first axis of this ordination corresponds to an increase in conductivity, percent sand and understory species richness and a decrease in pH, percent clay, fine and very fine sand and overstory species richness (Table 3). The second axis is correlated with an increase in pH, conductance, percent fine and very fine sand and both overstory and understory species richness.
A stepwise multiple regression of environmental characteristics of plots on the ordination indicated that the first axis responded to the percent bare rock, percent clay and specific conductance (DCA1 = 49.4 + 6.6 ROCK + 2.2 S - 14.9 CL). These factors accounted for 87.9% of the variance in the alignment of plots along the first axis. The second axis showed a response to tree age and percent clay and coarse sand in soils (DCA2 = 150.1 - 0.6 TRAGE + 10.4 CL - 2.9 CS). These factors accounted for 87.3% of the variance in the alignment of plots along the second axis.
Pitch pine communities in the Hudson River region occur as three physiognomic forms. The first is a closed canopy forest with tall trees on lacustrine deposits in Schaghticoke and outwash sand deposits in Cairo. Soils in these two sites contain considerably more silt and clay than the coarser soils of the Albany Pine Bush. The heavier soils plus the lack of recent disturbance probably accounts for the large trees, few shrubs and the well-developed herb layer that characterize these sites. The first physiognomic form of forest is similar to oak-pine or barrens forests described on Long Island (Olsvig et al., 1979), the dry acidic oak-conifer forests in the Berkshire Mountains in Massachusetts (Weatherbee and Crow, 1992), and the mature forests in the Schawangunk Mountains (McIntosh, 1959) and the nearby Kittinny Ridge in New Jersey (Niering, 1953). In Pennsylvania, Illick and Aughanbaugh (1930) described this type as the most common physiognomic form in that state and McCormick (1979) noted the oak-pine or pine-oak was the most common community type in the New Jersey pine barrens. He suggested that the difference between oak-pine and pine-oak was that, in the former type, oaks cover 40% or more of the canopy and have over 50% of the stems; in the latter type, pines cover over 30% and have over 50% of the stems.
Milne (1985), in his detailed study of the Albany Pine Bush, noted that closed canopy barrens or oak-pine forests were less disturbed than the more shrubby open areas and, in the absence of fire, greater tree diversity is to be expected with northern red oak, white oak, red maple, white ash and white pine gradually becoming more important in the tree layer. Cairo and Schaghticoke had all these species plus some others that indicated an even more mature, mesic forest than the mature forest expected in the Pine Bush. This is likely because of the much higher percentages of silt and clay in soils at these two sites, allowing more mesic trees and herbs to grow there.
We found sites with similar characteristics in our other upstate New York studies (Seischab and Bernard, 1991; Bernard and Seischab, 1995), among them the forests around Ithaca (Seischab and Bernard, 1991), the Penfield, and Junius sites on the Ontario Plain and Canisteo in the central and western New York area [ILLUSTRATION FOR FIGURE 1 OMITTED] and in Saratoga Springs (Interstate median), and one of the Rome sites. All these areas had well-developed canopies, relatively high species richness, a poorly developed shrub layer and a well-developed herbaceous layer of mesic forest-floor species.
The second physiognomic type had a more open canopy of shorter, less densely stocked, stunted trees, with a generally extensive shrub layer of either scrub oaks or ericaceous shrubs. This physiognomic form was quite prominent on conglomerate sites in the Shawangunk Mountains at Minnewaska State Park, Bonticou Crag, on parts of the Trapp's Bridge site and on the sand dune in the Albany Pine Bush. This community also was identified in the Berkshire Mountains (Weatherbee and Crow, 1992), on outwash plains. Only fragments of this type remain in the Berkshires and it was predicted that this forest would succeed to a more diverse forest with white oak, northern red oak and scarlet oak (Quercus coccinea), white pine, red maple and black cherry gradually becoming more important. This type of forest also occurs elsewhere; in New York in Waverly and Irondequoit Bay (Seischab and Bernard, 1991) and in Plattsburgh, Glens Falls, Rome and Ballston Spa in northeastern New York (Bernard and Seischab, 1995). The wetland sites in Plessis and in the Shawangunk Mountains probably fit best in this category, both having had a fairly open canopy and a well-developed shrub layer of Vaccinium corymbosum.
The third physiognomic form is dwarf vegetation; there seemed to be two types, separated from each other on the second axis of Figure 4. The first type, found at the Lake Maratanza site in the Shawangunks, featured dense stands of dwarf pitch pines 2-3 m tall with associated ericaceous shrubs (primarily Gaylussacia baccata and Vaccinium pallidum) and sometimes scrub oaks and gray birch as associates. Good et al. (1979) termed this type a pygmy forest and defined it as having dense stands of pines of very low stature, many with serotinous cones, few or no tree oaks but scrub oaks and some ericaceous shrubs. As stated, we found this type on the Shawangunk Mountains and it has been described on schist or quartzite rocky summits in the Berkshires (Weatherbee and Crow, 1992) where a dense stand of pines only about 1 m high was noted in an area classified as the acid rocky summit community.
A variant of this type is dominated by a dense growth of taller shrub oaks (2 m tall), a less dense lower layer ([less than]1 m tall) of ericaceous shrubs with scattered, usually short ([less than]2 m tall) pitch pines. The communities we sampled at the Ice Caves [ILLUSTRATION FOR FIGURE 4 OMITTED] in the Shawangunks Mountains was an example of this type which has also been described from New Jersey (Good et al., 1979), Long Island (Conard, 1935; Olsvig et al., 1979) and Pennsylvania (Illick and Aughenbaugh, 1930). The latter authors noted many such sites in Pennsylvania, among them one near State College, which they described as "endless stretches of scrub oak with scattered taller pitch pine." All who have studied these areas agreed that they owe their existence to disturbance, especially fire (Milne, 1985; Olsvig et al., 1979; Weatherbee and Crow, 1992; Illick and Aughanbaugh, 1930; Conard, 1935, and Good et al., 1979). Many of these areas have been created by disturbance in recent times but some are much older; Weatherbee and Crow (1992) quote a study from 1839 which described the dwarf community in the Berkshires much as it looks today.
New York sites considered in this study, as well as those in our other studies, can be classified according to their soil type in addition to their vegetation. Those communities formed on acid rock sites in the Thousand Islands region and in the Shawangunk Mountains usually have trees of varying degrees of small stature with open canopies. Shrub layers may be well-developed if there are cracks in the rock allowing them to grow or if an extensive thick organic layer has developed. However, often the bedrock is solid and either fire or erosion has reduced the organic layer with a resulting reduction of shrub cover. The communities on sand at Albany, Glens Falls, Ballston Spa, Clintonville, Waverly and Irondequoit Bay vary in the amounts of coarse sand and have had different histories of disturbance. They therefore are not all vegetationally alike, ranging physiognomically from sites with small trees, few shrubs and a ground layer of lichens and mosses on very disturbed sites (e.g., Clintonville and Ballston Spa) to a more closed canopy physiognomy on less disturbed sites such as Saratoga Springs and Plattsburgh. The third type is found on former agricultural areas. These sites were generally in upland regions in the vicinity of Ithaca, Penfield, Junius, Cairo and Schaghticoke. They generally had heavier soils, a taller, closed tree canopy, fewer shrubs and a well-developed herbaceous layer of mesic forest floor herbs.
The last type we sampled, the wetland type, occurred in low depressions in bedrock at Plessis and in the Shawangunk Mountains. These sites had a well-developed shrub layer of either Vaccinium corymbosum or Chamaedaphne calyculata and a ground layer of Sphagnum moss.
TWINSPAN [ILLUSTRATION FOR FIGURE 2 OMITTED] separated those sites with an ericaceous shrub layer from those without and, in the ericaceous types, wet, rock or sand types separated from each other. The wet site had three taxa (Chamaedaphne calyculata, Vaccinium corymbosum and Sphagnum spp.) that were unique to the site and a high percentage of Kalmia angustifolia (Table 2). The sand and rock sites separated on cover of Pteridium aquilinum which was found mostly in sandy plots. The nonericaceous communities division had the pitch pine, white pine and mesic herbs in common (Table 2).
The DECORANA [ILLUSTRATION FOR FIGURE 3 OMITTED] showed the same relationships, the peat site and high silt and clay old fields separated from all other sites. As stated, these sites had much greater species richness in both the tree and herbaceous layers. In addition, the species importance values [ILLUSTRATION FOR FIGURE 5 OMITTED] indicate that some species, among them Gaylussacia, Kalmia angustifolia, Pteridium, Vaccinium pallidum and the scrub oak occur mostly on sandy sites, while Kalmia latifolia, Vaccinium angustifolium and white pine occur mostly on rocky sites.
Acknowledgments. - We acknowledge the assistance of Robert Zaremba, Paul Huth, Robert Dirig and Edward A. Cope in locating pitch pine study sites. The staff of the L. H. Bailey Hortorium at Cornell University kindly identified some plant specimens. The staffs of Minnewaska State park, Ice Cave Mountain and the Mohonk Preserve kindly gave us permission to collect data on their properties.
BERNARD, J. M. AND F. K. SEISCHAB. 1995. Pitch pine (Pinus rigida Mill.) communities in northeastern New York State. Am. Midl. Nat., 134:294-306.
BOUYOUCOS, G.J. 1936. Directions for making mechanical analysis of soils by the hydrometer method. Soil Sci., 42:225-230.
CONARD, H. S. 1935. The plant associations of central Long Island. Am. Midl. Nat., 16:433-516.
CRUM, H. AND L. E. ANDERSON. 1981. Mosses of eastern North America. Columbia Univ. Press, New York. 1328 p.
DIRIG, R. 1994. Lichens of pine barrens, dwarf pine plains, and "ice cave" habitats in the Shawangunk Mountains, New York. Mycotaxon, 52:523-558.
GLEASON, H. A. AND A. CRONQUIST. 1991. Manual of vascular plants of northeastern United States and adjacent Canada, 2nd ed. The New York Botanical Garden, Bronx, N.Y. 910 p.
GOOD, R. E., N. A. GOOD AND J. W. ANDRESEN. 1979. The pine barren plains, p. 283-297. In: R. T. T. Forman (ed.). Pine barrens: ecosystem and landscape. Academic Press, New York.
HILL, M. O. 1979a. TWINSPAN. A FORTRAN program for arranging multivariate data in an ordered two-way table by classification of the individuals and attributes. Cornell Univ., Ithaca, N.Y. 90 p.
-----. 1979b. DECORANA. A FORTRAN program for detrended correspondence analysis and reciprocal averaging. Cornell Univ., Ithaca, N.Y. 75 p.
ILLICK, J. S. AND J. E. AUGHANBAUGH. 1930. Pitch pine in Pennsylvania. Research Bull. 2. Commonw. of Penn., Dep. of Forests and Waters. 108 p.
MCCORMICK, J. 1979. The vegetation of the New Jersey Pine Barrens, p. 229-243. In: R. T. T. Forman (ed.). Pine barrens: ecosystem and landscape. Academic Press, New York.
McIntosh, R. P. 1957. The York Woods, a case history of forest succession in southern Wisconsin. Ecology, 38:29-37.
-----. 1959. Presence and cover in pitch pine-oak stands of the Schawangunk Mountains, New York. Ecology, 40:482-485.
MILNE, B. T. 1985. Upland vegetational gradients and post-fire succession in the Albany Pine Bush, New York. Bull. Torrey Bot. Club, 112:21-34.
NIERING, W. A. 1953. The past and present vegetation of High Point State Park, New Jersey. Ecol. Monogr. 23:127-148.
OLSVIG, L. S., J. F. CRYAN AND R. H. WHITTAKER. 1979. Vegetational gradients of pine plains and barrens of Long Island, New York, p. 265-282. In: R. T. T. Forman (ed.). Pine barrens: ecosystem and landscape. Academic Press, New York.
RITTNER, D., ED. 1976. Pine bush. Albany's last frontier. The Pine Bush Historical Preservation Project. 263 p.
SEISCHAB, F. K. AND J. M. BERNARD. 1991. Pitch pine (Pinus rigida Mill.) communities in central and western New York. Bull. Torrey Bot. Club, 118:412-423.
SELENDER, M.D. 1980. Increment borings of pitch pine (Pinus rigida Mill., Pinacea) from sites on the Shawangunk Ridge and Ramapo Mountains of southeastern New York State: age and growth dynamics. Skenectada, 2:1-9.
SOKAL, R. R. AND F. J. ROHLF. 1981. Biometry. The principles and practice of statistics in biological research. W. H. Freeman & Co., New York. 859 p.
WEATHERBEE, P. B. AND G. E. CROW. 1992. Natural plant communities of Berkshire County, Massachusetts. Rhodora, 94:171-209.
|Printer friendly Cite/link Email Feedback|
|Author:||Seischab, Franz K.; Bernard, John M.|
|Publication:||The American Midland Naturalist|
|Date:||Jul 1, 1996|
|Previous Article:||Spatial and temporal patterns of nutrient concentrations in foliage of riparian species.|
|Next Article:||Effects of microtopography and disturbance on fine-root dynamics in wetland forests of lower-order stream floodplains.|