Flora and Soils of Wells Savannah, an Example of a Unique Savanna Type

Flora and Soils of Wells Savannah, an Example of a Unique Savanna Type

Shelingoski, Susan


Wells Savannah is a unique wet pine savanna located in the Lower Coastal Plain of North Carolina. It consists of two utility rights-of-way and a fire-suppressed pond pine (Pinus serotina) woodland. The objectives of this study of Wells Savannah were to characterize its soils, inventory its vascular flora, and compare its soils and flora with those of reference sites. Soil drainage, texture, and redoximorphic features were described at all locations. Wetland indices were calculated to represent the degree to which the plant communities were adapted to wet substrates. We identified 209 species in 107 genera and 48 families at Wells Savannah. Eight community associations were recognized. Wetland indices for Wells Savannah were significantly lower than those for the reference savannas. The unique combination of fine textured, very wet soil and plant species composition present at Wells Savannah has not been identified in previous studies of savannas in the region.


Savannas are communities with widely scattered trees or shrubs and a continuous graminoid layer that includes shade intolerant, poorly competitive species that begin to disappear if fire is excluded for only a few years (Frost et al. 1986, McPherson 1997). Prior to European settlement, species-rich savannas were maintained by naturally occurring, relatively frequent (recurring every 1-6 years), low intensity fires throughout millions of hectares in the Southeast from Virginia to Florida and west to Texas (Taggart 1990). Proactive efforts to suppress fires began in the early 1920’s and nearly eliminated natural fires in much of the United States by the 1950’s (Taggart 1990). As a result of this fire suppression and landscape development, extant savanna is now a mere fraction of its original range. Remaining savannas are islands within a highly managed and manipulated landscape (Johnson and Tothill 1984).

The small scale (i.e., less than 1,000 m^sup 2^) species richness of North Carolina savannas is greater than that of most other community types in North America (Walker and Peet 1984), so it is not surprising that scientists have been drawn to this remarkably diverse ecosystem for decades. Wells and Shunk’s (1928b) floristic inventory and community descriptions for the 1,500-acre Big Savannah in Fender County constituted the first reported study of a North Carolina savanna. They indicated that Big Savannah was completely treeless and supported an unusually dense and rich herbaceous flora. It has been the belief that this unique community existed because of annual burning and fine-textured soil that remained saturated to the surface for 10 months or more in most years. Strenuous efforts were expended to conserve Big Savannah but, unfortunately, this botanical treasure had completely succumbed to the plow and bulldozer by the 1950’s.

Wells Savannah (WS), a site discovered in 1997 in Fender County, supports a plant community similar to that of the former Big Savannah. This similarity is not surprising given that the two savannah sites are only 8 km apart and both reside within a continuous band of the rare Liddell soil series (Barnhill 1990), a wet Inceptisol. WS may represent the only remnant of this uncommon community type in North Carolina.

The three specific objectives of this study were to (1) inventory the flora, (2) describe the plant communities, and (3) describe soil drainage class, texture, and rcdoximorphic indicators at WS. We also used the flora and soil information to evaluate the uniqueness of WS relative to other savannas on the Lower Coastal Plain of North Carolina.


Wells Savannah is located approximately 140 km southeast of Raleigh in Fender County, North Carolina (Figure 1). Throughout the 19th Century, naval stores were the primary source of income for Fender County residents (Barnhill 1990). By 1940, much of the land had been converted to pine plantations and now the county relies largely on pulpwood and sawtimber production. Currently, 82% (185,866 ha) of the county is forested. The temperature regime is thermic and precipitation averages 1,350 mm per year.

WS is located on the north side of Pelham Road (SR 1319) at a point 2.7 km northwest of the intersection of Pelham Road and Watha Road (SR 1313; Wallace East USGS 7.5 minute topographic quadrangle), only 8 km northwest of the historic Big Savannah. WS and the Big Savannah site occur within a continuous band of Liddell soil, which is rare on the Lower Coastal Plain (Barnhill 1990). This occurrence of Liddell soil covers 9,762 hectares of northwest Pender County, with only 464 additional hectares of Liddell soil scattered in the remainder of the county, none of which occurs in any of the surrounding counties. Liddell silt loam is classified as a coarse-silty, siliceous, subactive, acid, thermic typic endoaquept (Barnhill 1990). Its high content of silt and very fine sand is unusual in this predominately sandy-textured region of the state. The fine soil texture and flat topography severely retard drainage at WS.

Aerial photos taken in 1938 (United States Department of Agriculture 1938) show that, prior to habitat fragmentation, the Liddell soil band connecting the two savanna sites supported a continuous corridor of open vegetation with widely spaced trees, likely pines, and minimal shrub cover. Such an open community could have developed in response to frequent fire. Although the historic fire frequency for this corridor is unknown, the circumstantial evidence from the 1938 aerial photos, in conjunction with charcoal found in the WS soil at a depth of about 20 cm, suggest that fire was frequent. Aerial photos taken during 1944, 1955, 1965, and 1972 show that the open plant community at WS gradually became a woodland with denser tree and shrub growth, most likely as a response to fire suppression.

WS currently consists of two mowed rights-of-way extending northeastward from Pelham Road and a 43 ha fire-suppressed woodland located between the two rights-of-way. Approximately 4 hectares of open savanna exist within the two rights-of-way and support a diverse assemblage of native plants that typically occur in wet pine savanna communities. The smaller and northwestern-most right-of-way, owned by Four County Electric Membership Corporation, was established in 1952. The larger and southeastern right-of-way was established in 1972 by Progress Energy. The corridors have been maintained by mowing since their installations and are currently on a 3-year mowing rotation. The Four County Electric Membership Corporation right-of-way was last mowed in January of 2003; the Progress Energy right-of-way was last mowed in April of 2002. Herbicides have never been used to manage the Progress Energy right-of-way, but were in use until 1999 within the Four County Electric Membership Corporation right-of-way. The North Carolina Natural Heritage Program (NCNHP) and the two utilities recently signed a Memorandum of Understanding to not use herbicides and to continue the current mowing regime (NCNHP, pers. comm.).

Although WS has been logged since the early 1900’s, its soils have not been tilled or drained so they still exhibit characteristics typical of the Liddell series. Since the site has not been influenced by disturbance that can cause unnatural soil and plant community conditions for years to come, it is an excellent candidate for restoration. Accordingly, the 47 ha tract was purchased by the North Carolina Coastal Land Trust in April of 2002. The North Carolina Coastal Land Trust is currently working with the utilities to develop and implement a controlled burn plan. It is believed that the woodland habitat will be restored to savanna with scattered pond, loblolly, and longleaf pines under a 3-year prescribed fire regime.

Savannas located in Holly Shelter Game Land (HSGL) in southeastern Fender County were used for comparison to that of WS. These savannas were selected because HSGL is close to WS, has soils that are similar to those of WS, and is maintained with frequent fire. LeGrand and Sorrie (1997) and soil profile samples were used to identify savanna communities within HSGL that most closely matched the soils and vegetation at WS.


Floristic Inventory

The flora was inventoried during the growing seasons of 2001-2003. Plants were collected throughout Wells Savannah from areas exhibiting different topographies, surface hydrology, and indicators of different disturbance histories. In addition, 10 permanent plots were established and extensively inventoried within the rights-of-way at WS, and 6 within savannas at HSGL, following Carolina Vegetation Survey (CVS) guidelines (Peet et al. 1998). Voucher specimens were deposited in the North Carolina State University Herbarium (NCSC). Nomenclature followed Kartesz (1999). Endangered, Threatened, and Rare species status was determined through the NCNHP County Status List for Fender County (NCNHP 2003).

Community Association Descriptions

While two general plant communities, savanna in the rights-of-ways and pine woodland in adjacent habitat, were recognized at WS, topographic, hydrologic, and plant species variation throughout each general community facilitated description of more specific associations. The associations occurred throughout WS, each with restricted distribution. Descriptions of each specific association include indications of dominant species, areal extent, location, and topographic or hydrologic features.

Soil Description

We described 10 soil profiles from the rights-of-way and 4 from the woodland at WS. Visual observation of soil color using the Munsell Soil Color Chart (Munsell Color 2000) and texture were used to describe the soil profiles. The 14 profile locations collectively represented the range of topography and disturbance indicators at the site. We also evaluated 6 profiles from 4 of the wettest savannas identified by LeGrand and Sorrie (1997). Two of these savannas were classified as ‘wet Ultisol’ variants, the other two as ‘wet Spodosol’ variants. We described the soil redoximorphic features (Vepraskas 1992) of each profile and determined which soils were hydric based on these descriptions (Hurt et al. 1998). Visual observation was used to determine the redoximorphic features present within the soils. Such features include redox concentrations (accumulations of iron and manganese oxides producing rust-colored masses), redox depletions (grey masses where iron and manganese oxides have been stripped), oxidized rhizospheres (rust colored areas surrounding roots where oxygen leaving the roots has oxidized iron in the soil), and reduced matrices (low chroma soils resulting from the presence of reduced iron). Hurt et al. (1998) used redoximorphic features in conjunction with color and texture to classify hydric soils by their geographic location. Soil samples were taken from each horizon of each soil profile. A particle size analysis (Gee and Bauder 1986) was performed on each sample at the North Carolina State University Soil Laboratory in August of 2002.

Wetland Indices

Ten permanent 10 × 10 m plots were established in the rights-of-way at WS and 6 in the HSGL reference savannas. Areal cover class (trace, 0-1%, 1-2%, 2-5%, 5-10%, 10-25%, 25-50%, 50-75%, 75-95%, 95-100%) (Peet et al. 1998) of each species in each of these plots was determined during June of 2002. A wetland index that expresses the degree to which the plant community in each plot was adapted to wet soil conditions was calculated by applying weighted averaging methods (Wentworth et al. 1988) to the species’ cover classes and the Region 2 wetland fidelity indicators for each species (United States Fish and Wildlife Service 1988). The calculation of weighted averages uses ecological indices and importance values for all species divided by the sum of all importance values.


Flora and Associations

There were 209 species in 107 genera and 48 families found at WS. Families including the most species were Poaceae (41), Cyperaceae (32), and Asteraceae (28), while Rhynchospora (Cyperaceae) was the largest genus represented with 20 species. Thirty-two percent of the species were Obligate Wetland (OBL) and 39% were Facultative Wetland (FACW). Seven plant species listed as Endangered, Threatened, or Rare in the state of North Carolina were identified within the rights-of-way at WS (Table 1). The State Rare broad-winged sedge grasshopper [Stethophyma celatum (Otte)], previously reported from Big Savannah, was also observed at WS. Twelve NCNHP Watch List species, including southern winged loosestrife (Lythrum alatum var. lanceolatum) previously known from only three other locations, also were found at WS (LeBlond 2001).

The amount or kind of landscape disturbance appears to be a major factor determining the distribution of three woodland plant associations. The most extensive association is a dense woodland with a dominant shrub layer. Areas supporting this association lack evidence of soil disturbance. Overstory species include Pinus serotina, P. taeda, Symplocos tinctoria, and Nyssa sylvatica. Long-term (over 60 years) fire suppression has allowed a dense shrub understory to develop that is composed primarily of Lyonia lucida, Magnolia virginiana, Persea palustris, and Cyrilla racemiflora. Although the herbaceous stratum is sparse, Zigadenus densus, Ctenium aromaticum, Pteridium aquilinum var. pseudocaudatum, and Sarracenia flava are present, evidence that this community had formerly been a savanna.

The second association is scattered throughout the woodland, occurring in areas where disturbance resulting from timber harvesting has opened the canopy and compacted the soils so much that shrub growth is limited. Old timber harvest landings, drag trails, and harvested areas support an association with a species composition similar to that of the dense woodland but with the addition of Acer rubrum var. trilobum in the overstory and Arundinaria gigantea ssp. tecta, Morella caroliniensis, Morella cerifera, and Gaylussacia frondosa in the shrub stratum. Areas closer to the rights-of-way have a richer herbaceous stratum than those embedded in the woodland at greater distances from the rights-of-way. Species typically found in all areas include Carex glaucescens, Rhynchospora cephalantha var. cephalantha, R. chalarocephala, Rhexia lutea, and Polygala lutea.

The third woodland association is a ‘red maple drain’ that appears to have been left undisturbed by timber activity due to its wet soils and hardwood-dominated overstory. This drain is located in the southwest corner of the woodland, approximately 30 m north of Pelham Road. Overstory species include Acer rubrum var. trilobum, Liquidambar styraciflua, Quercus nigra, and Nyssa sylvatica. The shrub stratum is sparse, composed mainly of saplings of overstory species, while the herbaceous stratum is mostly Pteridium aquilinum var. pseudocaudatum and Osmunda cinnamomea. Other species include Ilex coriacea, I. opaca, and Smilax glauca.

In the WS rights-of-way, plant species composition varies along microtopographic and moisture gradients. All five of the recognized associations are predominantly herbaceous and devoid of trees. A sparse shrub layer (cover less than 20%) is present in the higher areas. Although a number of species are found in all five associations, a given species is dominant in only one or two.

The most extensive right-of-way association occurs on high sites and is dominated by Arundinaria gigantea ssp. tecta, Muhlenbergia capillaris var. trichopodes, Ilex glabra, Andropogon virginicus var. uirginicus, and Schizachyrium scoparium. Vegetation is extremely dense and varies only slightly within the association.

The second most extensive association is dominated by Dichanthelium scabriusculum, D. scoparium, Aristida purpurascens var. uirgata, and Scleria minor. This association occurs along the woodland edge of the southeastern boundary of the larger right-of-way and in small pockets along the northwestern edge of the woodland in the western right-of-way. The association is found on the grade between high areas and low areas within the rights-of-way.

Low-lying areas that remain saturated to the surface throughout most of the growing season but rarely pond constitute the third association, dominated by Sarracenia flava, S. purpurea, Ctenium aromaticum, Calopogon pallidus, Lachnanthes caroliniana, and Rhexia alifanus.

A fourth association is common in areas disturbed frequently enough to prevent the occurrence of most graminoid species. Equipment associated with the maintenance of the rights-of-way and private vehicles utilizing the utility companies’ access roads have compacted soils and created large ruts that pond water. These areas are sparsely vegetated and support species, such as Drosera capillaris and D. intermedia, which are poor competitors and are frequent in open habitats. Other species common in these areas are Lycopodiella alopecuroides and Aletris farinosa.

The last right-of-way association occurs in a large swale within the Four County Electric Membership Corporation right-of-way. The swale formed in an area that had been a landing for a timber harvest. It remains ponded longer than any other area within either right-of-way, and supports the largest number of sedge and rush species of all the associations. Dominant species include Fimbristylis miliacea, Juncus acuminatus, Eleocharis microcarpa, Rhynchospora macrostachya, Saccharum giganteum, Carex striata, and Xyris ambigua.

Compositional Comparison to Big Savannah and Holly Shelter Game Land

A comparison of the dominant species listed above to the dominant species listed by Wells and Shunk (1928b) in their inventory of the Big Savannah supports the idea that WS shares a very similar plant community to that reported for Big Savannah. Nearly 80% of the dominants listed for Big Savannah are dominants at WS. Dominant species present at both savannas include Arundinaria gigantea ssp. tecta, Andropogon virginicus var. uirginicus, Ctenium aromaticum, Dichanthelium scabriusculum, Rhynchospora chapmanii, Schizachyrium scoparium, and Scleria minor. While a number of the dominant species of Big Savannah and WS are present within HSGL savannas, they are not dominant at HSGL. Only one of the dominant species at WS and Big Savannah, Ctenium aromaticum, was also a dominant at HSGL in three of the six plots. Aristida stricta Michx. is present and dominant in all inventoried HSGL plots, but absent from both WS and Big Savannah. Additionally, 20 species of Rhynchospora occurred at WS, and double the number of species of Rhynchospora were observed during the permanent plot inventories at WS than at HSGL.


Visual evaluations of the soils at WS and HSGL suggest that the soils at WS remain wetter for longer periods of time than those at HSGL. Soils from all 14 WS sampling locations had signs of aquic conditions (Vepraskas 1992). Their redoximorphic features, indicators of aquic conditions, included redox concentrations (iron accumulation), oxidized rhizospheres, redox depletions, and reduced matrices. Redoximorphic features were found closer to the surface in WS profiles than in all but two profiles from HSGL. All profiles from WS were poorly or somewhat poorly drained (Buol et al. 2003, Table 2), while three of the six locations at HSGL were moderately well drained. The “F13-Umbric Surface” and “F3-Depleted Matrix” hydric indicators were noted in most profiles from both WS and HSGL.

Particle size clearly distinguished the soils sampled at WS and HSGL (Table 3). WS soils had lower percentages of sand and higher percentages of silt and clay than did HSGL soils. Although these textural differences were present throughout the soil profile, they were more pronounced within the upper 60 cm of the soil.

Wetland Indices

The plant communities in the WS plots demonstrated a greater community-level adaptation to wet soils than did the communities of the HSGL plots. The mean wetland indices for plant communities in the WS and HSGL plots were 2.0 and 2.3, respectively, and were significantly different (t = 2.84, df = 15; P = 0.013).


The greater adaptation of the Wells Savannah plant community to wet soils, relative to those of the HSGL communities, showed that saturation nearly to the soil surface has prevailed at WS for the long-term. The continually wet acidic soil is one of the unique features of WS.

A companion study to that reported here showed that the species composition of the WS plant community was unique when compared with a wide sample of savanna communities in the southeastern NC Coastal Plain. The sites with communities most similar to that of WS were from HSGL.

Based on our comparison of shared dominant plant species, the community at WS bore a stronger similarity to that previously reported for Big Savannah (Wells and Shunk 1928a, 1928b) than it did to HSGL. The strong affinity of vegetation of WS and Big Savannah is not surprising since (1) Big Savannah was a very wet savanna on Liddell soil, (2) WS is, perhaps, the wettest savanna on acidic soil currently known in NC and occurs on Liddell soil, and (3) the Big Savannah and WS sites are only 8 km apart. The rich flora and wet soils of WS do not reside comfortably in either of the previously described ‘wet Ultisol’ or ‘wet Spodosol’ savanna types, such as those studied at HSGL. The most striking difference in the floras of WS and HSGL was the complete lack of wiregrass (Aristida stricto) at WS and its relative abundance (25% or more cover) in every permanent plot at HSGL. WS is a species-rich, moist savanna, one of the most threatened of all southeastern grasslands (Frost et al. 1986), and its soils and flora are similar to those of Big Savannah. WS, the “Ghost of Big Savannah,” warrants conservation and restoration.

Although at least three savanna inventories and classifications have followed the pioneering study on Big Savannah (Taggart 1990, Schafale 1994, LeGrand and Sorrie 1997), no identified savanna community matches that described by Wells and Shunk (1928b). Taggart (1990) found that soil moisture and texture were the most highly correlated factors with species composition in savannas. Accordingly, he classified the wet savanna communities as Ultisol and Spodosol types. Schafale (1994) recognized ‘very wet clay’ and ‘wet Pleea flat’ variants, as well as Ultisol and Spodosol savanna community types. Although ‘very wet clay’ and ‘wet Pleea flat’ variants might seem suitable names for the communities of WS, further investigation into the specific soil features of WS revealed that WS does not fit into those of either variant. The ‘very wet clay’ and ‘wet Pleea flat’ variants typically occur on basic soils, oftentimes over limestone. Although soil moisture and texture of such variants resemble those of WS, their high pH and high nutrient levels do not match the acidic, nutrient poor soils of WS. Finally, LeGrand and Sorrie (1997) concluded that plant communities in wet, fire-maintained savannas and flatwoods of HSGL were best classified as Spodosol and Ultisol types. The wet Inceptisol soil and rich plant community at WS are not accommodated by any of the three previous savanna classification systems.

All other Lower Coastal Plain savannas that have recently been inventoried (Taggart 1990, Schafale 1994, LeGrand and Sorrie 1997) fit into the more general classifications for wet pine savannas described by Schafale and Weakley (1990), and NatureServe (2003). The former classification system identifies a ‘Pine Savanna’ category. By definition, this community type occurs in “wet, generally flat areas, and occasionally low ‘islands’ in peatlands or swamps, saturated part of the year, with frequent fire.” Wet mineral soils are noted as the substrate, but there is mention only of the Ultisol, Alfisol, and Spodosol soil orders. WS occurs on a soil series and order not previously identified as supporting a savanna association in southeastern North Carolina. WS could represent a new variant whose primary characteristics are soil texture, moisture, the lack of Aristida stricta, and the presence of high numbers of Rhynchospora species.

NatureServe (2003) lists an herbaceous, perennial graminoid, saturated temperate or subpolar grassland formation. WS resembles the Rhynchospora oligantha, Sarracenia spp. – (Aristida beyrichiana, Ctenium aromaticum) – Osmunda cinnamomea/Sphagnum spp. saturated herbaceous alliance (V.A.5.N.m.17). Certain aspects of WS could place the site in this alliance typical of ‘herbaceous bogs, hillside seepage bogs, wet prairies, muck bogs and poor fens, most often occurring on peaty soils with significant organic accumulation’. The alliance supports up to 10 different species of Rhynchospora. Despite the preceding general affinities, WS, with a silty surface layer and 20 species of Rhynchospora, does not entirely support its placement in this alliance. There are a number of other alliances within the formation dominated by species found at WS, yet all are associated with sandhill communities or wet, sandy soils with an impermeable subsurface layer that perches the water table. The woodland taxa (Class II) listed in the NatureServe classification also contrast with WS. WS best fits into the Pinus palustris, Pinus (elliottii, serotina), saturated woodland alliance (II.A.4.N.f.6). The associations within the alliance that most resemble WS are all dominated by Pinus palustris, have very minimal herbaceous layers, and occur on wet Spodosol soils. The woodland association having the closest resemblance to WS, the Pinus palustris – Pinus serotina/Sporobolus pinetorum – Ctenium aromaticum – Eriocaulon decangulare var. decangulare woodland (CEGL004502), is referenced as an Atlantic Coastal Plain Very Wet Clay Longleaf Pine Savanna. Both NatureServe (2003) and Schafale and Weakley (1990) fail to identify predominantly herbaceous communities occurring on saturated, fine textured soils. The ‘Pine Savanna’ community is listed by the North Carolina Natural Heritage Program (2003) as a natural community in need of protection. Pine Savannas have an S2 state rank, making the type imperiled in North Carolina because of rarity or otherwise vulnerable to extirpation in the state. Expanding this community type to include sites such as WS could help in the protection of the community type.

There are currently no other representatives of this kind of savanna ecosystem under protection. At present, HSGL is the largest protected area of fire-maintained savannas in the state. Even so, none of the savannas found there closely compare in species composition or soil characteristics to those of WS. The species composition and associations identified at WS most closely resemble those of the extirpated Big Savannah.


The 209 plant species observed at Wells Savannah during this study are listed in the following annotated checklist of taxa. Families, genera, and species are arranged alphabetically within major taxonomic divisions. Each species is followed by the location where it was observed (R = rights-of-way, W = woodland), the wetland fidelity indicator, protection status (where applicable), and collection number.


The authors thank the managers of Holly Shelter Game Land for permission to sample their savannas in order to conduct this study. We thank all of the volunteers who participated in the plot inventories of Wells Savannah and Holly Shelter Game Land. They were crucial in gathering a tremendous amount of data in a short period of time. We also thank the Soil Science Department of North Carolina State University for use of their facilities, and Barrett Richards, Justin Ewing and Dr. Michael Vepraskas for their help with soil profiling and particle size analysis. We thank the Carolina Vegetation Survey (CVS) for use of their data. All data from the CVS database were extracted by Dr. Robert Peet, who also lent advice and expertise in the data management. We thank Donna Wright, Janice Allen, and numerous others who contributed both time and ideas to the development of this study. The study was partially funded by The North Carolina Agricultural Research Service, The North Carolina Wildflower Preservation Society, and The North Carolina Academy of Science.


BARNHILL, W.L. 1990. Soil survey of Pender County, North Carolina. United States Department of Agriculture, Soil Conservation Service, Washington, D.C.

BUOL, S.W., R.J. SOUTHARD, R.C. GRAHAM, and P.A. MCDANIEL. 2003. Soil genesis and classification, 5th ed. Iowa State Press, Blackwell Publishing Company, Ames, Iowa.

FROST, C.C., J. WALKER, and R.K. PEET. 1986. Fire-dependent savannas and prairies of the Southeast: original extent, preservation and status. In: Kulhavy, D.L. and R.N. Connor (eds.). Wilderness and natural areas in the eastern United States: a management challenge. Center for Applied Studies, School of Forestry, Stephen F. Austin St. University, Nacogdoches, Texas.

GEE, G.W. and J.W. BAUDER. 1986. Particle size analysis, p. 383-409. In: Methods of soil analysis. Part 1. Physical and mineralogical methods. Agronomy Monograph no. 9.

HURT, G.W., P.M. WHITED, and R.F. PRINGLE (eds.). 1998. Field indicators of hydric soils in the United States, version 4.0. U. S. Department of Agriculture, Natural Resource Conservation Service, Fort Worth, Texas.

JOHNSON, R.W. and J.C. TOTHILL. 1984. Definition and broad geographic outline of savanna lands. p. 1-14. In: Proceedings of the symposium: International savanna symposium. Australian Academy of Science, Canberra, Australia.

KARTESZ, J.T. 1999. Synthesis of the North American flora. North Carolina Botanical Garden, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina.

LEBLOND, R.J. 2001. Site survey report: Pellam Road Powerline Savanna Significant Natural Heritage Area. Unpublished report to North Carolina Natural Heritage Program, Raleigh, North Carolina.

LEGRAND, H.E., Jr. and B. SORRIE. 1997. Biological inventory of Holly Shelter Game Land, North Carolina. Department of Environmental and Natural Resources, Division of Parks and Recreation, North Carolina Natural Heritage Program, Raleigh, North Carolina.

MCPHERSON, G.R. 1997. Ecology and management of North American savannas. The University of Arizona Press, Tuscon, Arizona.

MUNSELL COLOR. 2000. Munsell Soil Color Charts. Gretag Macbeth, New Windsor, New York.

NATURESERVE. 2003. NatureServe Explorer. An online encyclopedia of life [web application]. Version 1.8. NatureServe, Arlington, Virginia. http://www.natureserve.org/explorer

NORTH CAROLINA NATURAL HERITAGE PROGRAM. Division of Environmental and Natural Resources, Division of Parks and Recreation. 2003. County status list: Pender County. http://www.ncsparks.net/nhp/elements2.fm

PEET, R.K., T.R. WENTWORTH, and P.S. WHITE. 1998. A flexible, multipurpose method for recording vegetation composition and structure. Castanea 63:262-274.

SCHAFALE, M.P. 1994. Inventory of longleaf pine natural communities in North Carolina. North Carolina Natural Heritage Program, DPR, DEHNR, Raleigh, North Carolina.

SCHAFALE, M.P. and A.S. WEAKLEY. 1990. Classification of the natural communities of North Carolina, 3rd approx. North Carolina Natural Heritage Program, Raleigh, North Carolina.

TAGGART, J.B. 1990. Inventory, classification and preservation of coastal plain savannas in the Carolinas. Ph.D. dissertation, University of North Carolina, Chapel Hill, North Carolina.

UNITED STATES DEPARTMENT OF AGRICULTURE. 1938. Aerial photographs, Pender County. Provided by Pender County Natural Resource Conservation Service (NRCS), Agricultural Stabilization and Conservation Service, Agricultural Adjustment Administration, College Park, Maryland.

UNITED STATES FISH AND WILDLIFE SERVICE. 1988. National list of plant species that occur in wetlands: Southeast (Region 2). Biological Report 88(26.2). United States Department of the Interior, United States Fish and Wildlife Service, Washington, D.C.

VEPRASKAS, M.J. 1992. Redoximorphic features for identifying aquic conditions. North Carolina Agricultural Research Service, North Carolina State University Technical Bulletin no. 301.

WALKER, J.W. and R.K. PEET. 1984. Composition and species diversity of pine-wiregrass savannas of the Green Swamp, North Carolina. Vegetatio 55:163-179.

WELLS, B.W. and I.V. SHUNK. 1928a. Plant communities of the coastal plain of North Carolina and their successional relations. Ecology 9:230-242.

WELLS, B.W. and I.V. SHUNK. 1928b. A southern upland grass-sedge bog: an ecological study. North Carolina Agricultural Experiment Station Technical Bulletin no. 32.

WENTWORTH, T.R., G.P. JOHNSON, and R.L. KOLOGISKI. 1988. Designation of wetlands by weighted averages of vegetation data: a preliminary investigation. Water Res. Bull. 24:389-396.

Received January 17, 2004; Accepted September 1, 2004.


1 Department of Botany, Campus Box 7612, North Carolina State University, Raleigh, North Carolina 27695-7612;

2 North Carolina Natural Heritage Program, 302 South Onslow Street, Richlands, North Carolina 28574

* Present address: 279 Hogan Farm Road, Apex, North Carolina 27523. email address: susanshel@hotmail.com

Copyright Southern Appalachian Botanical Society Jun 2005

Provided by ProQuest Information and Learning Company. All rights Reserved