Although the bog turtle, Clemmys muhlenbergii, was described at the beginning of the 19th Century (Schoepff 1801), basic ecological requirements were only recognized in the past several decades and its distribution is still being defined. Bog turtles were unknown from Maryland until 1941 (McCauley and Mansueti 1944) and the majority of known sites of occurrence in the state were not located until the 1980's.

The lack of successful field work on this species prior to the 1980's is not unique to Maryland. In Virginia, bog turtles were reported as early as 1924 (Brady 1924), but that record was based on an incorrectly identified juvenile wood turtle, Clemmys insculpta. Barton (1960) reassessed the situation in Virginia and demonstrated that there was no valid evidence of its occurrence there. Hutchison (1963) provided the first authentic evidence for Virginia, and presently it is known from numerous sites in four counties (Herman and Tryon 1993). The species was not recorded from Georgia until 1980, South Carolina until 1982, and Tennessee until 1986 (Hale and Harris 1980, Hemman and Putnam 1982, Herman and Putnam 1983, Herman and Wamer 1986). New colonies are still being discovered in Pennsylvania (Hulse and Hulse 1992), New Jersey (Amdt 1986), Delaware (Amdt 1977), North Carolina, Virginia, and Georgia (Dennis Herman, pers. comm.). While the species has been known to exist in the mountains of North Carolina since the 1880's (Yarrow 1882, Dunn 1917), its actual distribution in the mountain and piedmont regions was slow to be documented. Herman (1994) reported 99 records in 18 North Carolina counties, including a 1991 record in the North Carolina piedmont, 100 miles outside of its previously known range. Herman et al. (1993) showed that the species was more broadly distributed in North Carolina than had been previously recognized.

Figure 1

Now that the distribution of this species is generally understood, it is apparent that there are major gaps in its overall range (Figure 1), suggesting that at one time the bog turtle was much more common and widespread than today. A hiatus of 250 miles occurs between the closest known Maryland and Virginia localities. The core of the range is an area encompassing eastern Pennsylvania, the upper piedmont of Maryland and Delaware, and running north through New Jersey, eastern New York and western Connecticut. There is also a record from Rhode Island (Babcock 1917), and there are distinct populations in northwestern New York, northwestern Pennsylvania, the southern Appalachians, and the North Carolina piedmont.

This fragmented broad scale geographic distribution is difficult to explain with current zoogeographical information, but it is clearly a result of Pleistocene events. The disjunct micro-distribution (by county or drainage system) is apparently the result of limited habitat availability coupled with the species' need for certain early successional wetland communities. This species' disjunct (macro-distribution) and patchy occurrence (micro-distribution) indicate that its overall range has been contracting. As will be shown, this decline is partly natural, and partly a result of human disturbance, while the macro-distribution is a result of Pleistocene and post-Pleistocene climatic changes.

For biological information on the bog turtle see Barton and Price (1955), Ernst and Barbour (1972), Ernst and Bury (1977), Holub and Bloomer (1977), Bury (1979), and Ernst et al. (1994). Tryon and Herman (1990) and Hemlan and Tryon (1993) provided information on the conservation status of the species in the southeastern United States from Virginia to Georgia, and Amdt (1977) studied the natural history of bog turtles in Delaware. Landry (1979) provided a bibliography of literature on the bog turtle containing 158 references. Conservation status has been addressed by many authors, and a balanced summary was presented by Arndt (1978).

In 1974 Lee prepared a status report on this species in the southeastern United States for a publication then being proposed by the Tall Timbers Research Station. The focus of his account was the dependence of bog turtles on sedge meadows, the narrow successional window in which populations prospered, and the fragile nature of this subclimax seral stage as it related to local hydrology and land alterations. Although the published version of the Tall Timbers symposium never materialized, much of the habitat information from that study is presented herein.

The genus Clemmys is restricted to North America. It is the only member of the subfamily Emydinae (Ernst et al. 1994). Clemmys is comprised of four extant species, three which occur in the eastern United States and are aquatic or semi-aquatic to terrestrial, and one western species, C. marmorata, which is the most aquatic in the genus. Additional species of Clemmys are known as fossils dating from the Paleocene -(Romer 1966), Eocene (Hay 1908), and Pliocene (Brattstrom and Stum 1959).

Figure 2. Pattern of orange pigment on head and neck of bog turtles from Baltimore County, Maryland (A), and the mountains of North Carolina (B).

Clemmys mahlenbergii is the smallest species of Emydidae, and one of the smallest of the world's turtles. The maximum verified shell length is 108 mm (Ryan 1981), and the average length of adults is less than 100 mm. Since its original description by Schoepff (1801) as Testudo muhlenbergii, it has been placed in at least nine other genera by various authors. The first use of the combination Clemmys muhlenbergii was by Fitzinger in 1835. Dunn (1917) described bog turtles from the southern Appalachians as a distinct species, Clemmys nuchalis. While this taxon is not currently recognized, we are unaware of detailed systematic work evaluating differences in any of the disjunct populations. Minimally, there appear to be differences in the pigment patterns on the necks and heads of individuals from northern and southern populations (Figure 2), but the extent of this variation alone is insufficient to warrant separating these as distinct taxa (Ernst et al. 1994). Carl Ernst and Jeffrey Lovich do have some data indicating that these populations are different (C. Ernst, pers. comm.). Bern Tryon and Dennis Hemman (pers. comm.) are collecting blood samples from bog turtles at various localities for use in genetic analysis. Amato et al. (1993), reporting on preliminary results of the DNA analysis of these collections, found that the ten samples they tested exhibited unusually low levels of mitochondrial DNA variability. In contrast, Parker and Whiteman (1993), using DNA finger printing, showed greater genetic diversity in fragmented populations of spotted turtles. Studies on other freshwater turtles have shown genetic divergence to be greatest when intervening habitat is terrestrial (Scribner et. al. 1986).

The relationship of Clemmys mahlenbergii to other species in the genus is problematic. Electrophoretic studies by Merkle (1975) indicated that it is most closely related to C. guttata. Because of the number of protein systems sharing bands with C. marmorata, he concluded that C. mahlenbergii was closer to a common ancestral form than the other extant species of Clemmys. This same conclusion had been reached by Hay (1908), based on comparison of living and fossil Clemmys. However, Lovich et al. (1991) examined plastral morphology and found that C. muhlenbergii shared fewer similarities than the other three species in the genus, and were surprised at what they considered the relative dissimilarity of C. muhlenbergii and C. guttata. This is in contrast to Merkle's data, and the fact that C. muhlenbergii and C. guttata are known to hybridize (Ernst 1983).

The type locality of Clemmys mahlenbergii was listed simply as "Pennsylvaniae" and later restricted to Lancaster County by Stejoeger and Barbour (1917). The karyotype was described by Bickham (1975).

Descriptions of this turtle are available in numerous publications. The most recent and probably most readily available descriptions and illustrations are provided by Conant and Collins (1991), and Ernst et al. (1994). This species is quite distinct and problems posed by identification of adult individuals are minimal. Hatchlings and young juveniles have been confused with the young of other Clemmys species, but this was before illustrations of the young of all three eastern species were readily available.

While the first bog turtle from Maryland was reported by McCauley and Mansueti in 1943, Cope (1873) had reported it in the "northeastern portion of the State" during the previous century. Cope did not refer to any particular Maryland specimen or locality, and his statement may have been based on the proximity of that area to Lancaster County, Pennsylvania, where bog turtles were then well known.

The first Maryland specimen was collected on 8 August 1941 along Poplar Run near Grave Run Road, about one mile north of Grave Run Mills and two miles WSW of Rockdale in Baltimore County (Figure 3a). The circumstances under which it was found were atypical, which is unfortunate since the account was widely cited and believed typical of the biology and ecology of this then little known turtle. The animal was found embedded in hard clay beneath a board 15 or 20 feet from the stream (McCauley and Mansueti 1943, 1944). Robert McCauley (1945) discussed the turtle from Grave Run Mills in detail. That specimen now resides at the Natura1 History Society of Maryland (NHSM 450).

John Cooper (1947, 1949) reported additional specimens he and others obtained along the Susquehanna River floodplain in Cecil County, Maryland. The habitat was described as small, intermittent, vegetation-choked ponds. Cooper did not otherwise characterize the habitat but did provide a list of other reptiles and amphibians known from the immediate area.

Howard Campbell (1960) reported on a previously unknown colony he monitored in northern Baltimore County, one mile SW of Whitehall. In that same paper Campbell mentioned two other sites, one near Eko (=Eklo) in Baltimore County and one from Broad Creek in Harford County. The areas occupied were referred to as "swamps", but the descriptions of the habitat suggest that they were marshes bordered by wet meadows. Reed (1956) and Prince et al. (1957) had previously reported the species from the Broad Creek area.

MAPS WILL GO HERE
Figure 3. Previously published maps depicting known distribution of bog turtles in Maryland. A, McCauley (1945); B, Harris (1969); C, Cooper et al. (1973); D, Harris (1975). Maps reproduced with permission from the Natural History Society of Maryland.

By 1960, when John Cooper published his survey of the distribution of Maryland amphibians and reptiles, bog turtles were known from only the above mentioned sites in three counties. By 1969, when Herbert Harris updated and expanded the format of the state herpetological survey, the number of sites had risen to nine, in the same three counties (Figure 3b). By the time Harris revised his survey (1975) the total number of colonies had risen to 30, and the species was also known from northeastern Carroll County (Figure 3d). These early state surveys did much to promote herpetological record-keeping in Maryland, and resulted in a great increase in the number of known records for various species. Efforts by several local naturalists, specifically focusing on bog turtles, were responsible for the three-fold increase in the number of localities discovered between 1969 and 1975.

In the mid-1960's Nemuras (1966, 1967) reported additional sites for bog turtles he located in northeastern Maryland. These were important finds and provided the first hints that the earlier specimens from Maryland represented more than isolated, peripheral records. His 1967 study also gave valuable insight into the habitat requirements of this species. Previous published information was anecdotal, and much came from specimens found in marginal habitats.

In March of 1969 Lee headed up a committee of the Maryland Herpetological Society (Natural History Society of Maryland) established to survey for new bog turtle sites, and to evaluate those already known. In three months, three people working on weekends added 20 new sites to the list of known localities. The project was terminated because it was felt that the additional attention was encouraging collecting that was draining local populations, and one member of a field party antagonized land owners to the extent that one site was deliberately filled-in for no other reason than the property owner did not want to deal with amateur herpetologists. In a subsequent letter to the president of the Maryland Herpetological Society, Lee remarked that the species was much more widespread in the Baltimore, Harford, and Cecil County area than previously suspected, and that the real issue appeared to be management of the early successional bog/meadow communities that appeared to be optimum habitat for Clemmys mahlenbergii. In the course of that survey the term "sedge meadow" was first used to describe good bog turtle habitat. The number of new colonies located is reflected in the distribution maps presented by Cooper et al. (1973) and Harris (1975) (Figures 3c and 3d). Table 1 shows how the number of known Maryland sites grew between 1873 and l984.

Table 1. Total number of known bog turtle localities in Maryland

Year  Number of known sites              Source
1873   no specific record                Cope (1873)
1937   one unverified record             Brady (1973)
1944           1                         McCauley and Mansueti (1944)
1945	       1                         McCauley (1945)
1949	       2                         Cooper (1949)
1960	       5                         Campbell (1960)
1969	       9                         Harris (1969)
1973	      30                         Cooper et al. (1973)
1975	      30                         Harris (1975)
1984	     177                         Taylor et al. (1984), Dawson (1948)

Because of the listing of the species as endangered by the State of Maryland in 1972 (Annotated Code of Maryland, Article 66C, Section 125) and the recommendations made by Cooper et al. (1973), the Maryland Department of Natural Resources initiated a study in 1976 to more accurately determine the distribution and abundance of the bog turtle in Maryland. The primary purpose of that study was to locate and catalogue sites, but it also sought information on habitat requirements (Dawson 1984). Six hundred and eighty-nine "potential" sites (identified by soil type) were examined in Carroll, Baltimore, Harford, and Cecil Counties, and 173 sites supporting bog turtles were located. Twenty-three of the 30 "historic" sites were also resurveyed at that time, but turtles were found at only four. Because of the unexpected number of new sites located, the species was removed from the state's protected species list by the Maryland Department of Natural Resources in 1982, despite evidence that a large percentage of the known historic sites no longer supported turtles.

Chase et al. (1989) studied habitat characteristics, population size and home range of Maryland bog turtles. They found considerable variation in population size at the 20 sites studied in detail, and attributed the variation to habitat characteristics and differences in carrying capacity. They found home ranges to be larger in adult males ~ = 0.176 ha) than in adult females (~ = 0.066 ha), and noted considerable overlap between areas occupied by individuals of both sexes. This important study was an essential step in understanding the ecology of bog turtles in the central portion of their range.

Charles Stine (unpublished) studied the seasonal movements of bog turtles at a site in northern Carroll County. With radio-telemetry he tracked four turtles for a two year period to document migration to and from hibernation sites, and daily activity patterns at different seasons. The turtles studied were active from late March through the first week in November. The period spent hibernating and the nature of the sites selected were consistent with data presented by Ernst et. al. (1989). He also found that turtles were active throughout the day and did not estivate during the summer.

Between 1953 and 1956 Lee repeatedly attempted to capture an adult bog turtle living in a small pond at the edge of an open woodland well south of the other sites where this species was then known to occur in Maryland. That individual was clearly observed on numerous occasions and, although never captured, its identity is certain. The site was in Baltimore City in a large, then relatively undisturbed wood lot at the northwest corner of the intersection of Northern Parkway and Loch Raven Boulevard. Loch Raven Boulevard had recently been expanded and the construction corridor bisected the pond. Clumps of sedges at the upper end suggested that the site had been relatively open previously, and that the pond had been created by road construction. The locality was reported to Herbert Harris for his 1963 distribution survey, but Harris considered the animal to be a released captive. At that time the Loch Raven Boulevard/Joppa Road area (a few miles to the north) contained extensive areas of spring-fed sedge meadows and small meandering streams, but that habitat was not surveyed for bog turtles prior to its destruction.

In the latest version of his distribution survey of the amphibians and reptiles of Maryland, Harris (1975) showed two other records of bog turtles located southeast of their known local range. One of these, plotted as just east of Baltimore City (Harris 1975), was based on a specimen collected from "Hall Spring" in Herring Run Park by Glen Fullano and Dave Sieminski on 13 May 1973. That specimen now resides in the collection of the Natural History Society of Maryland (NHSM 1816). Hall Spring is a small resurgence generating about eight or nine gallons of water per minute (Otton and Hilleary 1985). It is located within a city park, adjacent to a major urban roadway. The spring itself has a concrete wall and floor, with discharge pipes and a drain. No habitat suitable for bog turtles is present today, but we have no record of the condition of this site when Fullano and Sieminski collected their turtle.

The second specimen was collected in a marsh along the north side of Maryland Route 198, east of Laurel in Prince George's County. William Grogan (pers. comm.), who lived in the area, recalls that the locality was a shallow wetland with tussock sedges, and it struck him as likely habitat for this species. When we visited the site in June 1992, it was an open cattail marsh with relatively deep standing water. At that time the only turtles in evidence were painted turtles (Chrysemys picta) and snapping turtles (Chelydra serpenana). The dramatic change in this site was apparently due to hydrologic changes resulting from the widening of Route 198, and there is no reason to believe that it was not acceptable habitat two decades previously. We were unable to locate this specimen among the series of bog turtles at the Natural History Society of Maryland.

Both of these records were mapped by Harris (1975) (see Figure 3d). However, their occurrence beyond the expected local range persuaded him that they had been released and did not represent remnants of natwal populations. Looking back, the probability that two bog turtles would be released (remember that this was a time when few local naturalists had even seen this species), then recaptured by someone who knew what they were and brought them to the Natural History Society of Maryland, seems very small. We consider it more probable that these specimens came from colonies that were dying as a result of environmental changes generated by development. It seems likely to us that other now extirpated, disjunct populations of bog turtles in the greater Baltimore Metropolitan Area suffered the same fate.

Other interesting accounts of bog turtles in Maryland and adjacent Virginia are those reported by Brady from the Plummers Island-Stubblefield Falls area just west of Washington D.C. (Brady 1937, McCauley 1945, Manville 1968). These sightings were made between 1924 and 1955, and have generally been disregarded by regional herpetologists. Brady's 1924 record from Stubblefield Falls, Virginia was shown to be a young wood turtle by Barton (1960). While this certainly casts doubt on the Plummer's Island records, it does not prove them erroneous. We do not wish to suggest that the bog turtle actually occurred in the D.C. area in historic time. In fact, a zoogeographic analysis conducted by Lee and Dennis Smith (1957) presented evidence for a "prairie peninsula corridor" which allowed many prairie species of plants and animals to extend their distributions eastward into the area now occupied by middle latitude deciduous forest. Stuckey and Reese (1981) compiled a considerable amount of information on the prairie peninsula concept. Most of the information they presented, however, is from Ohio and other states west of the Appalachians. Today, there are many relict populations of prairie species in the central Appalachians and middle Atlantic states, and there are relict and highly distinct populations of other prairie species in western New York and western Pennsylvania. Smith (1957) discussed the biogeography of the prairie peninsula in detail and showed how glacial influence could have affected the distribution of bog turtles and other species of reptiles and amphibians. We agree with Smith (1957) that the bog turtle may be of midwestern prairie origin. Glacial retreat opened up unoccupied habitats to the north, and glacial gouging created basins for bogs, including some which are extant today. We suggest that the gradual reforestation of glaciated areas fragmented the bog turtle's northern most populations after the last glacial advance. The major portion of the bog turtle's current range is south of the glacial maximum.

We suggest that this species originally evolved in wet prairie habitats west of the Appalachians and expanded eastward during inter-glacial periods. During subsequent glacial periods eastern bog turtles were confined primarily to wet prairies of the lower Susquehanna basin, and in the Southern Appalachians, more or less where they occur today. During these glacial periods, most regions of North America were drier than they were during the interglacials, since vast quantities of freshwater were stored as ice (as thick as 10,000 feet) and snow. Thus, wet prairie regions south of the glacial advance became fragmented and were eventually lost as suitable habitat for bog turtles through drying, and those to the north were covered by ice. During the more recent interglacial periods, and at the end of the last glacial advance, the southern Appalachian population remained isolated, as it is today. The turtles in the Susquehanna refugium expanded northward into habitats in basins left by glacial scouring (many to later disappear because of natural succession). Others expanded north and east into adjacent river systems, primarily the Delaware and Hudson Rivers.

Based on evaluation of a fossil herpetofauna recovered from a cave in eastern Quebec and another from Highland County, Virginia, Fay (1984) concluded that in nonglacial Quaternary environments the local reptile and amphibian fauna of 30,000 years ago was nearly identical to that of present times. Climatic conditions may have been fairly similar to those now found in the mountains of southwestern Virginia, suggesting that a bog turtle refugium could have been present in the Southern Appalachians for a very long period.

The native Susquehanna fish fauna includes northern Mississippi basin species that used post-glacial outlets. Hocutt et al. (1986) list 12 fish taxa that are clearly in this category. By contrast, the Delaware River has a less rich fauna, and was not directly linked to Pleistocene glacial outlets. Thus, if bog turtles do represent part of a fauna of midwestern prairie origin, the northern population had direct access at least to the Susquehanna and Hudson Rivers, and, through subsequent lateral captures (or at least shared areas occupied by tributaries), access to adjacent systems in Maryland, Delaware, New Jersey, and New England. The current shared fauna suggests that prior to formation of the Chesapeake Bay (starting about 15,000 years ago) the Susquehanna River had as tributaries all piedmont streams and rivers west to at least the Patapsco River. There is also good evidence that the Susquehanna previously turned east and crossed the Delmarva in the vicinity of Salisbury (Lee 1976, Coleman et al. 1990, Kerhin et al. 1996). Thus, the Susquehanna may have been connected to the Delaware basin as well. Bog turtles using the Susquehanna and Delaware river flood plains as corridors for dispersal would have had access to a network of river systems that drained over 39,500 square miles, until about 15,000 years ago.

If one looks at the distribution of bog turtles in Maryland from an aquatic zoogeographical perspective, the documented range in the Chesapeake Bay drainage falls almost entirely in the eastern division of the piedmont. Nearly all sites are confined to the river systems between the Susquehanna River on the east and the Patapsco River on the west. This corresponds to area "C" as defined by Lee et al. (1981) and, excluding the unlikely Plummer's Island reports mentioned above, includes almost all of the previous historical (non-fossil) records from the state. The only exception is an occurrence in the Monocacy River drainage of Carroll County. This portion of Maryland's drainage system supports 56 native species of fishes (Lee et al. 1981) several of which are of midwestern origin (i.e., stoneroller, Campostoma anomalum; silver jaw minnow, Ericymba buccata; walleye, Stizostedion vitreum; and logperch, Percina caprodes). These are mostly distinct populations of Mississippi basin species which invaded the Atlantic slope either through a headwater stream capture by the Susquehanna, or through Pleistocene glacial outlets. This aspect of distribution patterns has been reviewed on a regional basis by Jenkins et al. (1972) and locally by Lee (1976). Additionally, the hellbender (Cryptobranchus alleganiensis) and the map turtle (Graptemys geographica) are both Mississippi basin species with disjunct populations in the Susquehanna River (the map turtle also occurs in the Delaware River). Their recent isolation in this area is evidenced by a lack of detectable differences between lower Susquehanna and Mississippi basin populations. The 444 mile length of the Susquehanna River, a system with a drainage area of 27,500 square miles, would be a likely corridor of dispersal for bog turtles and other prairie elements that invaded the eastern United States. The headwaters of the Susquehanna River are in Otsego County, New York and, therefore, would provide a connection with the area of glacial scouring in the proximity of upstate New York's relict populations of bog turtles. It is clear that the Susquehanna basin is a key element in the current distribution of Clemmys muhlenbergli. Nearly all major colony sites are in this or neighboring basins, and over 70% of the known sites for the species occur there. The proximity of the Susquehanna to the Delaware River basin, and the Delaware basin to colony sites in New Jersey and New England is obvious. The semi-aquatic habits of the bog turtle and its small size would tend to limit dispersal, and it seems likely that its current distribution is influenced by proximity of available habitat and natural dispersal routes. The one other area in Maryland where bog turtles are now known to occur is just west of Parrs Ridge in Carroll County. There, these turtles are known from a few sites along Big Pipe Creek, a tributary of the Monocacy River (Potomac River drainage). These localities, the only recent ones documented anywhere in the extensive Potomac River basin, undoubtedly represent an isolated invasion of this drainage from the adjacent Susquehanna or Gunpowder Rivers. Since bog turtles are limited to this small portion of the Potomac River basin, we suggest that this crossover occurred quite recently.

American Elk

The bog turtle's characteristic open meadow habitat (see Ecological Distribution) would have been difficult to maintain once the prairie peninsula receded. We propose that in pre-colonial times bog turtle populations centered around successional communities tied to beaver ponds, and perhaps to open areas maintained by periodic burning by native Americans or natural fires. Burning by native Americans was a common practice and may have played an important role in the survival of bog turtles in some areas during the past 10,000 years. Because of the frequency with which we have found bog turtles in grazed areas we also suggest that, prior to the colonial period, the grazing activities of large herbivores like the American elk (Cervus canadensis) and the American bison (Bison bison) may also have been important in the maintenance of bog turtle habitat.

John Donelson, in his 1780 journal (Roe 1951), noted that in Pennsylvania "The open space around and near the sulphur or salt springs, instead of being an 'old field', as had been supposed by Mansker, at his visit here in 1769, was thus free from trees and underbrush by the innumerable herds of buffalo and deer and elk that came to these waters." Shoemaker (1915) reported many such buffalo clearings and cited place names such as Clearfield, Pennsylvania, as evidence that such circumstances were common.

Portions of the known historic range of these large, grazing mammals rather closely parallel the current range of the bog turtle. We know from Seton (1909) that in Snyder County, Pennsylvania large numbers of bison and a few elk were present until the mid-1700's. The last bison in Pennsylvania were killed between 1783 and 1800 (Garretson 1938). Historically, salt licks along the Susquehanna River were known to be heavily used by elk and currently the lower Susquehanna basin is one of the major areas of concentration for bog turtle colonies. Elk were once so numerous in this area that "leading to the licks the paths of these animals were as large as many of the great public roads of our country" (Rhoads 1903). The last elk were killed in Pennsylvania in 1877 (Shoemaker 1915).

In North Carolina, elk were known from the mountains and northwestern piedmont but were eliminated by hunting by the late 1700's. In this same state bison were still common between 1720 and 1750, but had totally disappeared by 1765. Like bog turtles, they were known only from the western portion of the state (Lee and Funderburg 1977). The distribution of elk in Maryland is not well documented (Mansueti 1950, Lee 1984), but there is good evidence that they ranged east at least to the eastern edge of the piedmont. We know that they did occur around the District of Columbia, and elk are known from archaeological sites as far east as Frederick, Montgomery and Prince George's Counties. From place names (Elkridge, Elk Neck, Elk River, etc.) in Cecil, Harford, and Howard County, we assume they occurred in these counties as well (see Mansueti 1950). In Maryland, bison were present throughout the mountains and piedmont into the historical period (Lee 1984). Thus, these large grazing animals could have been key species in the maintenance of open, early successional meadow communities in all areas of the bog turtle's range from central Pennsylvania southward.

If one considers the landscape from an earlier perspective, the local occurrence of open meadows maintained by grazing may have been more pronounced than it was during the pre-colonial period. In the late Pleistocene large grazers and browsers were an important component of the fauna of eastern North America. Research conducted by David Stedman (pers. comm.) on various fossil sites within the Susquehanna drainage provides a good overview of the paleo-history of that region. He documented the local presence of such prominent herbivores as the ground sloth (Megalonyx jeffersonii), giant beaver (Castoroides ohioensis), mammoths (Mammuthus columbi and M. primigenius), mastodon (Mammut americanum), horses (Equns complicatus and E. fraternus), tapir (Tapirus veroensis), and peccaries (Mylohyus nasutus and Platygonus compressus). All died out or were hunted to extinction when the first humans expanded into eastern North America (about 11,000 ybp). The early hunters concentrated on the larger (2 44 kg) species of mammals.

As an aside, it is interesting to note that we have observed bog turtles in or adjacent to cattle droppings, where they appeared to be foraging on insects associated with the droppings. While bog turtles are dietary generalists, this type of feeding activity could have been important when open habitats were maintained by native late Pleistocene to pre-colonial grazers. We have also found bog turtles in the depressions made by cattle walking around and through wet meadows.

The beaver, Castor canadensis, undoubtedly created much habitat for bog turtles, and it is clear that this rodent was a major component of the fauna in the pre-colonial period. For instance, populations of beaver in Maryland must have been phenomenal. Guthein (1949) provided some quantification of the numbers removed from the area of the Potomac River in the early 1600's. In 1632 a single merchant exported beaver skins at a total value of 40,000 gold crowns. Another trader had "gotten 1500 weight of beaver." At one point there were 110 Indians "loaded with 4,000 weight of beaver skins." Another excerpt tells of seven thousand Indians, each with forty beaver skins. Even allowing for exaggeration, Maryland beaver populations must have been quite healthy in the early 1600's.

By the end of the 1800's fur trapping had all but eliminated the beaver from the southeast, and it was uncommon everywhere east of the Mississippi River. Bonwill and Owens (1939), and Mansueti (1950) document the extirpation and eventual return of beaver populations to Maryland. Evidence for edge use of beaver ponds by bog turtles in Maryland is mostly anecdotal. Robert Simmons told us that he found bog turtles in the upper ends of beaver ponds at the Conowingo site in Cecil County. His turtles were found by lifting the overhanging thatch of tussock sedges. This was in the same area mentioned by Cooper (1949), but it is not clear if the small intermittent ponds Cooper referred to were originally beaver ponds. While beaver ponds certainly provide at least marginal habitat under certain conditions, stream impoundment could also create sedge meadows upstream of the ponds that would be ideal for sustaining small bog turtle populations.

The habitat needs of beavers have been studied in northern forests where quaking aspen (Populus tremuloides) makes up an important component of their diet. Typically, a beaver colony will occupy a given site for a number of years but when all the food plants within a reasonable distance are eaten the beaver colony vacates the site. The untended dam eventually gives way and the former beaver pond becomes a meadow. Aspens again fill the meadow and once the seedlings mature beavers will reoccupy the site. While this sequence of beaver colonization and recolonization has not been worked out for the central Atlantic states, there is no reason to assume that the pattern was not similar. Thus, along any given post-Pleistocene stream course bog turtles would have had a spectrum of open meadows, early successional woodlands and beaver ponds. In such a system, dispersal and colonization of reopened sedge meadows by bog turtles is easy to visualize. The constantly changing habitat mosaic at any given site might even stimulate dispersal of bog turtles. Kiviat (1978) also felt that beavers were instrumental in maintaining and creating bog turtle habitat.

Throughout their range bog turtles occupy a variety of open wetland habitats. Most literature focuses on sphagnum-bog communities, resulting in the species' current common name. Netting (1927) noted that records from the northern edge of the species' range are from sphagnum-bogs, and considered tamarack-bogs their original habitat. He felt that their occurrence in clear streams further southward could be explained by the fact that boreal plant communities disappeared as post-glacial plant communities closed in upon such areas. At the time of Netting's study the habitat of bog turtles in the southeast had not been characterized and his assumption about habitat in the south was incorrect. Other early authors had described bog turtle habitat in terms which would not indicate bogs in the traditional sense, i.e. meadow ditches (Abbott 1882) and meadow streams (Street 1914).

Since bog turtles occur in a variety of open wetlands lacking a canopy layer, and over a relatively broad latitudinal span (34°N to 44°N), it is difficult to explain their fragmented geographic distribution based simply on habitat availability. Their spotty micro-distribution can be explained by local habitat availability, succession, limited dispersal ability, and current and past land use practices. In Figure 4 we illustrate the narrow geographic and geological availability of land area for bog turtle habitat.

Variation in terminology used to describe sites occupied by bog turtles, as well as the animal's common name, has caused confusion in characterizing typical habitat. Researchers from different areas and time periods have understandably been inconsistent in their use of terminology. Characterizations of habitat for this species are a mixture of geological terminology (fens, bogs, river floodplains, ponds), habitat descriptions (Appalachian bogs, marshes, swamps, beaver ponds), plant community characterizations (open canopy, grassy meadows) and combinations of these terms (tamarack-bogs). We feel that it is important to define both the physical characteristics and the plant community that a site supports when it is inhabited by these turtles.

Figure 4. Generalized 32 mile NE-SW cross section through the Maryland piedmont (Baltimore and Carroll Counties) showing geologic and topographic units, and a profile of landforms (modified from Cleaves et al. 1974, and Brush 1975). Note limited areas of allovial soil and the patchy occurrence of potential bog turtle habitat.

Bogs are typically composed of mats of vegetation which grow over ponds, small lakes and other natural depressions. The aquatic system is fed primarily and directly by rain. The Sphagnum and rooted vegetation of the shore grows out into the water, gradually advancing while slowly decomposing plant material fills the depression. Younger bogs often have open water in the center, while mature bogs become filled with encroaching shrubs and eventually shift completely to a terrestrial community. The amount of time required for this process varies, depending on the original depth of the basin (in the north many deep basins are the result of glacial scouring), drainage volume, frequency of fire, and annual temperature. Sphagnum is a major bog building plant. It decomposes slowly, retains a large volume of water even in advanced stages of decomposition, and has high acidity.

Shallow water communities with emergent vegetation are best referred to as marshes, and when the emergent vegetation is dominantly woody shrubs and trees, the communities are properly called swamps. Unlike bogs and sedge meadows, neither marshes or swamps provide optimal habitat for bog turtles.

Tryon and Herman (1990) consider fens to be spring fed habitats, and describe bogs as areas where water accumulation is from rain and run-off. Fens are traditionally flooded clay or peat areas (e.g., the Fenland District of eastern England). In the southeast (Virginia to Georgia) fens occur at high elevations, have a high pH, and have floristic affinities with calcareous fens of midwestern and northeastern North America (Weakley 1991). At this time bog turtles are documented from only one true fen in the Southern Appalachians, indicating that fens generally do not provide habitat suitable for bog turtles. An overview of southeastern mountain bogs and fens is provided by Richardson and Gibbons (1993). According to the U.S. Fish and Wildlife Service (1997), a significant number of newly discovered bog turtle sites in Sussex County, New Jersey occupy calcareous fen habitats.

A recent publication discussing management of forested wetlands (Welsch et al. 1995) illustrates the degree of confusion that currently exists over the habitat requirements of these turtles, even among scientists. It recommends "Within calcareous fens where chalky, crumbly deposits are evident in surface pools, preserve and encourage scrub and/or shrub habitats as important overwintering habitat for rare bog turtles." While mention of bog turtle habitat in a publication for wetlands ecologists is commendable, since very few fens actually support bog turtles, persons venturing into the field with this search image are likely to overlook most bog turtle habitat. More importantly, in most areas woody shrubs should be discouraged (see below).

Clemmys mahlenbergii was named by Loannis Schoepff in 1801 for its discoverer, the Reverend Gotthilf Henry Emestus Muhlenberg (1753-1815). Muhlenberg was a well known botanist responsible for the discovery and description of numerous plants, including many rushes and sedges. We suspect that Mublenberg's attention to rushes and sedges led him to the habitat where this secretive turtle was discovered. It is likely that the type specimen came from a sedge meadow community. Beck (1928) noted that Muhlenberg studied rare wetland plants at a swamp in eastern Providence Township, Lancaster County, Pennsylvania, where bog turtles have been found, and speculated that the type may have come from that site.

Bog turtles inhabit true mid-to-late stage bogs in some portions of their range. They are normally not found in swamps or marshes. In Maryland they typically inhabit wet meadows formed by subclimax communities of grasses and sedges (Figure 5). These sites are usually located in level and low-gradient (3-8% slopes) spring fed areas that have modest amounts of permanent running water. Characteristically, they do not have standing water of significant depth. High gradient areas do not support bog turtles, possibly because high runoff velocity prevents accumulation of the soils necessary for sedge meadow development. Most stable sites are on flood plain terraces well above areas subject to regular flooding.

Tussock sedges, Carex stricta, are important to these turtles in Maryland. Bog turtles hide under the matted vegetation of the tussocks and use runways among them, and the tussocks are important egg laying and incubation sites. The matted dead vegetation of tussock and other sedges, rushes and grasses also forms a thick protective mat which may discourage predators. The mats of vegetation may be particularly important in the spring when they provide shelter prior to the appearance of new growth meadow vegetation. We have frequently noted the droppings of raccoons and foxes on top of these mats, demonstrating that these communities are regularly patrolled by predators. Campbell (1960), in discussing a site that apparently lacked tussock sedges, felt that raccoons were primary predators of bog turtles. The matted tussocks also appear to limit the germination of woody vegetation, and ice and snow mats the vegetation into a thick thatch that may inhibit invasion of other plant species.

Additionally, tussock sedges provide thermal protection and maintain high relative humidity Zovickian (1971) noted that high relative humidity may be important for successful growth of hatchling bog turtles. Measurements of temperature and humidity taken beneath and adjacent to sedge mats at a Maryland bog turtle locality on 19 July 1992, between 10:00 and 11:00 AM, showed extreme differences of 10.3° C and 29.6% relatively humidity, with the runways under the tussocks being cooler and more humid than measurements made above the tussocks.

Platania and Sanders (1977) characterized the habitat of the bog turtle in the southern portion of its range. They noted that the species is found in sedge communities on silt loam soils (Baile, Codorus, Glenville, Hatboro, and similar soils). Herman (1991) noted that in the Blue Ridge between Virginia and Georgia, bog turtles prefer spring-fed wetlands that contain highly organic mucky soils (Codorus, Hatboro,

Toxaway and other allovial silt loams). The Maryland Department of Natural Resources has used U.S Department of Agriculture soils maps to identify potential bog turtle habitat for survey purposes (Pitts 1978).

Figure 5. Typical sedge meadow habitat for bog turtles in Carroll County, Maryland. Note springfed flow of shallow water meandering between clumps of sedges. Woody vegetation is encroaching from the upper right.

Chase et al. (1989) characterized the Maryland habitat of bog turtles using 36 habitat variables to relate indices of population density to habitat characteristics. They found that the greater the index of population density, the more likely the colony site was to be located in a circular basin with spring-fed pockets of shallow water, a bottom substrate of soft mud and rock with interspersed wet and dry pockets, and dominant vegetation of low grasses and sedges. Sixty-one percent of the turtles they found were not in the water, but most were less than 20 cm from water. Seventy-eight percent of their captures were next to low grasses and sedges. We agree with their habitat assessment, and add that meadows should have sufficient elevation not to be flooded when stream levels rise and, while the shrub layer should not be dense, some scattered shrubs are desirable since adult turtles frequently hibernate among woody root systems and may also use these root masses as a refuge during other seasons.

In the early summer of 1974 Lee and Elmer Worthley set up three transects in a sedge meadow along a floodplain near Bel Air, Harford County, Maryland. The 1.2 ha site was chosen because it had a small colony of bog turtles (ca 30 adults) and was obviously undergoing fairly rapid succession. Permanent standing water had caused the site to be invaded by common cattail (Typha latifolia). The upper meadow section of the floodplain was being invaded by smooth alder (Alnus serrulata), green ash (Fraxinus pennsylvanica), and blackberry (Rubus spp.). The core of the habitat (about one acre) was still open and dominated by sedges, rushes and grasses. When the locality was revisited in July 1992, the cattail marsh had expanded and filled much of the site. The higher areas were on damp soils dominated by reed canary grass (Phalaris arundinacea), green ash, alder, buttonbush (Cephalanthus occidentalis), and halberd-leaved tearthumb (Polygonum arifolium). Despite the near total replacement of the original community, some remnant species of the original meadow, wool-grass sedge (Scirpus cyperinus) and skunk cabbage (Symplocarpus foetidus), were still present in modest numbers. At that time, no bog turtles could be found, even though turtles could always be located with reasonable effort in the 1970's. We cite the history of this locality because it documents the gradual succession of a habitat that had declined steadily since it was discovered in 1969. Despite the lack of change in land use at or near the site, we saw considerable change between 1969 and 1974, and a nearly complete shift in the plant community from wet meadow to cattail marsh between 1974 and 1992. These observations show that smaller sedge meadow sites can all but disappear in less than 20 years.

Large areas of one of our study sites in northern Carroll County, and several other Maryland sites examined by us, exhibited dense accumulations of ferrihydrate precipitated by 'iron bacteria'. These bacteria occur where iron-bearing groundwater discharges at the surface. SeveIa1 genera of these microorganisms (Sphaerotilus, Siderocystis, Leptothrix, Siderocapsa and Gallinella) precipitate red flocculants of ferric iron that cover the substrate (Robbins et. al. 1992, Robbins and Norden 1994). Characteristically associated with them is Leptothrix discophora, which forms an oil-like film on the surface of the water. These precipitates are typical of groundwater-fed habitats in this area and have been observed at many of the bog turtle sites that we have visited. In fact, disturbance of these flocs and surface films can be a useful indication of recent bog turtle activity. On several occasions we were able to locate turtles by tracing their "trail" through the surface film to places where they had burrowed into the substrate. It is interesting that one of the earliest descriptions of a Maryland bog turtle site (Campbell 1960) described the water as being covered by a ''red scum" and having an occasional "light oil slick."

The following is a listing of plants identified at one northern Carroll County site, with dominant species marked by an asterisk (*). Mosses: Sphagnum* (at least five species are common in Maryland sedge meadows). Ferns: New York fem (Thelypteris noveboracensis), marsh fem (Thelypteris palustris), cinnamon fem (Osmunda cinnamomea), sensitive fem (Onoclea sensibilis*), hay-scented fem (Dennstoedtia punctilobula). Grasses and Sedges: lurid sedge (Carex lurida), tussock sedge (C. stricta*), other sedges (Carex spp*), spikerushes (Eleocharis spp.), straw-colored sedge (Cyperus strigosus), rice cut-grass (Leersia oryzoides*), fowl manna grass (Glyceria striata), velvet grass (Holcus lanatus), tickle grass (Agrostis hyemalis), and rushes* (Juncus effusis, acuminatus, J. marginatus). Other Herbs: skunk cabbage (Symplocarpus foetidus*), many-flowered agrimony (Agrimonia parviflora), sweet white violet (Viola blanda), blue marsh violet (V. cacullata), small bedstraw (Galium trifidum), arrowhead (Sagittaria latifolia), jewel-weed (Impatiens pallida*), joe-pye-weed (Eupatoridelphusfistulosus), New York iron weed (Vernonia noveboracensis), narrow-leaved mountain mint (Pycnanthemum flexnosum), boneset (Eupatorium perfoliatum), halberd-leaved tearthumb (Polygonum arifolium), purple-stemmed medaster (Aster puniceus), rough goldenrod (Solidago patula), swamp milkweed (Asclepias incarnata), nodding ladies'-tresses (Spiranthes cernua), skullcap (Scutellaria sp.), water horehound (Lycopus americanus), purple dead-nettle (Lamium purpureum), deer-tongue (Dichanthelium clandesffnum), bugleweed (Lycopus virginicus), and square-stemmed monkey-flower (Mimulus ringens). Woody Shrubs, Vines, and Saplings: Japanese honeysuckle (Lonicera japonica), smooth alder (Alnus serrulata*), American meadow-sweet (Spirea latifolia), steeple-bush (Spiraea tomentosa), red maple (Acer rubrum), black willow (Salix nigra), northern arrow wood (Viburnum recognitum), swamp rose (Rosa palustris*), multiflora rose (Rosa multiflora*), and maleberry (Lyonia legusmna). Prior to this survey Galium trifidum was believed extirpated from the state.

At the same site Stine (unpublished data) found that in the winter the middle and upper portion of the sedge meadow experienced freezing and thawing of the surface layer, causing uplifting of the soil. Spring runs could be heard flowing under the uplifted areas. Chase et al. (1989) in their habitat characterization of turtle colony sites noted that the most densely occupied areas exhibited interspersed wet and dry pockets. We feel that this freezing and uplifting of the surface layer is important in the annual maintenance of wet and dry pockets, and causes adult bog turtles to migrate to specific wintering sites where the substrate is more stable. The young of the year certainly do not undertake these migrations, and this may give them several weeks in the fall and early spring to forage in areas free of adults. In captivity adults occasionally prey on young (Amdt 1972) and there is no reason to assume that this does not occur in the wild. As far as surface upheavals exposing hibernating young to freezing temperatures, we expect that this is not a problem. Studies on young painted and box turtles (Terrapene carolina) show that they can survive freezing. How these very young turtles survive freezing temperatures is fairly well known (Pack and Packard 1995, Storey et al. 1988), and we suspect that young bog turtles also adjust physiologically to low temperatures.

Compared to other aquatic temperate species, a fair amount of information is available on the hibernation of bog turtles. Amdt (1977) reported that bog turtles in northern Delaware apparently hibernated for six months (mid-autumn to early spring) and Ernst (1977) observed no active turtles after October. Chase et al. (1989) found that the eight turtles they followed moved in and out of winter retreats from November to March. The turtles followed by Stine (unpublished data) moved in and out of hibernation in November and early March but always remained in close proximity to hibernation sites. During late December, January and February, they were essentially inactive and were found in the same position on each visit. Some inhabited communal winter dens that were used for more than one season. In New Jersey Holub and Bloomer (1977) found that bog turtles entered winter retreats in late October and emerged in early to mid-May. Ernst et al. (1989) studied over-wintering of C. muhlenbergii in Pennsylvania and New Jersey, and found the turtles to use a variety of sites.

A wide variety of other amphibians and reptiles occur in sedge meadows with bog turtles. Species found by us using a bog turtle site in northeastern Carroll County include dusky salamanders (Desmognathus fuscus), four-toed salamanders (Hemidactylium scutatum), long-tailed salamanders (Eurycea longicauda), American toads (Bufo americanus), pickerel frogs (Rana palusms), milk snakes (Lampropeltis triangulum), queen snakes (Regina septemvittata), ribbon snakes (Thamnophis sauritus), snapping turtles (Chelydra serpentina), and spotted turtles (Clemmys gutfata). Spotted turtles do not regularly occur syntopically with bog turtles, and we have a clear impression that these two small Clemmys prefer different micro-habitats.

In the southeast the bog turtle's distribution may be correlated with the distribution of two native mammals, the star-nosed mole (Condylura cristata) and the meadow vole (Microtus pennsylvanicus). The mole constructs subterranean runways which are often water filled, and the vole makes surface runs just beneath the vegetation. Bog turtles make regular use of, and enlarge, both types of runs and are seldom encountered on the surface except when basking. Various aspects of the ecological and geographical distribution of the bog turtle in the southeast are closely mirrored by the star-nosed mole (Lee 1987) and meadow vole (Had1 1981). This turtle's regular use of vole runs was discussed by Lee (1970). The width of adult bog turtle shells (65-72 mm), compared to the normal width of Condylura (about 55 mm) and Microtus (about 60 mm) runs, seems large, but the runs are ideal for young turtles and are quickly widened by larger individuals. The frequency with which we have found bog turtles in these runs demonstrates a clear association.

Bog turtles are opportunistic feeders, capturing a wide variety of terrestrial and aquatic organisms, primarily invertebrates. They also consume vegetation and berries, and occasionally take carrion. These turtles are able to fed on land or in and under water.

In suitable habitats bog turtles exist in high densities. Chase et al. (1989) estimated a high of 213 turtles per hectare based on amark-recapture study. Holub and Bloomer (1977) reported average densities of up to 40 adults per acre (0.4 ha). At the site studied by Stine (unpublished data) the density of bog turtles was estimated to be 100 turtles in a 1.6 ha area. Stine also found that the summer home range of both sexes overlapped, and it was not uncommon to find several adults within a one meter area. Chase et al. (1989) found that bog turtle home ranges were less that 0.2 ha, and were slightly larger for males (x= 0.176 ha) than females (0.066 ha). This is quite a bit smaller than known home ranges for other eastern Clemmys. Spotted turtles have home ranges of approximately 0.5 ha (Netting 1936). Wood turtles have larger home ranges, estimated in various studies to be from 0.8 to 4.9 ha (Harding and Bloomer 1979), or from an average of 3.3 ha for females to 4.99 ha for males (KauLmann 1995), to as much as 24.3 ha (Quinn and Tate 1991).

Although aggression between bog turtles has been noted, particularly toward smaller conspecifics (Holub and Bloomer 1977), they seem less aggressive that spotted turtles or wood turtles. In spring, male spotted turtles are quite aggressive and exhibit combat by biting the front legs of opponents. We have observed bouts of combat lasting ten minutes or more, and the scales on the front legs of many adult males show significant scarring. Combat in male wood turtles has been reported by Harding (1990), and both sexes are solitary except during mating. The high population density exhibited by colonies of bog turtles is unusual for Clemmys, and is possible because of the animal's small size, relatively passive nature, and, presumably, the high productivity of its wet meadow habitat.

There is some evidence of bog turtle populations surviving in marginal habitats. In South Carolina the one known locality is an area of small interconnected pools separated by former beaver dams (Herman and Putnam 1982). The pools are surrounded by deep mud and large stands of rushes. The Georgia localities are in areas with dense brush and a canopy, along a small mountain stream in a dense forest with little ground level vegetation. This apparently represents a remnant habitat largely succeeded to mature forest. Tryon and Herman (1990) noted a lack of potential nesting and incubation sites at this location and found that the few individuals observed were old adults. While such populations are of questionable viability, it is significant that adult turtles can survive in such habitats. Bog turtles are long lived, with a recorded life of 35+ years (Herman 1990). Because of this, animals holding on in suboptimal situations would serve as "seed stock" for colonies which could resurge after fire, impoundment of water courses and opening of the canopy layer by beavers, or a variety of anthropogenic changes might result in a return to suitable habitat.

Local land use from the colonial period until the present would appear to have maintained, if not enhanced, bog turtle habitat. Clearing for agriculture or grazing of domestic stock suppressed succession of open meadows, as grazing by bison and elk did previously. By the mid-1900's, mowing and regular canopy removal for utility line rights-of-way and other permanent disturbance may have created more "permanent" habitat than had existed in the pre-colonial period. Several existing Maryland bog turtle sites, for example, are in mowed areas under power lines.

The site monitored by Stine (unpublished data) is a 1.6 ha area that, until the late 1970's, had been used as pasture for daiIy farming, and had been grazed for nearly a century. When the local dairy industry became less profitable the cows were removed but the land owner continued to periodically burn the sedge meadow to keep it open. In the late 1980's new county ordinances discouraged burning and, since that time, while the site still supports an active bog turtle colony, it has undergone uninterrupted succession and is now rapidly being invaded by woody shrubs and saplings.

Where suitable patches of habitat occur in close proximity, within a common drainage, it is likely that bog turtles wander from patch to patch. However, the significance of this exchange of turtles on population structure and the genetics of the colony is unknown. Such wandering turtles probably once played a significant role in repopulating newly opened sites. However, now that so many sites have been destroyed, and many intervening stretches of streamside habitat have been cleared or developed, the movement of bog turtles to new sites is being made more difficult, if not precluded. Bog turtles are less likely than spotted or wood turtles to wander outside defined colony areas, and turtles such as painted and snapping turtles, which are more aquatic, are much more frequently found on roads and in areas well removed from suitable habitat than are bog turtles. Research examining the significance of inbreeding and reduced genetic diversity of isolated groups of bog turtle would be very useful.

Some recent population surveys have indicated an increase in the number of known bog turtle "colonies," suggesting that the bog turtle's apparent rarity reflects inadequate survey rather than a scarcity of turtles. This results from adding the number of known historic sites to those recently discovered. The totals are impressive only to people not directly involved in field work with the species, and we fear that most published tallies do not reflect reality. First and foremost, sites are being lost at a rapid pace as a result of changing land use and natural succession. The actual number of currently inhabited sites is seldom known for any region. This is due to the widely scattered nature of the sites, the fact that most are on private property and seldom monitored, and that the few people with access to site locations are often unable to spend the time necessary to conduct follow-up field studies.

Another aspect of this problem is how colonies have been defined. Many of the sites listed and totaled for various regions and states (see Table 2) are places where only a few turtles have been found, often only one animal. Records, sites, and populations are terms used interchangeably by persons writing reports and attempting to summarize known records. None of these necessarily represent viable colonies. "Occupied colonies" may represent localities with only one or two turtles. Only in rare cases (i.e., Herman and Tryon 1993) has current reproductive success been factored into efforts to define what constitutes a viable colony. Various researchers have used different levels of verification to classify extant, viable, and extirpated sites. We would be surprised if 150 of the sites totaled for the entire range (Table 2) of the species are viable. Tryon and Herman (1990) believed that only 52 to 54% of the total recorded localities in the southern Appalachians have potential for current bog turtle populations.

Another problem with using individual "colony" totals to express population status, is that an unknown but significant number of sites are in such close proximity that counting them as separate colonies is misleading. Most surveys in Maryland and elsewhere during the late 1970's and early 1980's simply documented locations of occurrence with little regard for relevant factors which would be useful in predicting population size or viability. Lest this seem like harsh criticism of colleagues who conducted these surveys, we should acknowledge that their work was timely and important to our current understanding of this species and its biological requirements.

The problem, as we see it, is that administrative decisions regarding the bog turtle have frequently been based on compiled regional reports of field studies conducted over the last two decades. A recent report by Bourg (1992) demonstrated all of the pitfalls listed above. Terms like colony and population were used repeatedly with little indication of the actual number of turtles or the viability of isolated "colonies." Nor was there any indication of the degree or significance of genetic exchange between those sites that were in close proximity.

To provide consistency of terminology, we suggest the following. A "record" is a verified report of the occurrence of one or more turtles. A "site" is a place. A "colony" represents a site or a number of interconnected sites with an interacting group of breeding turtles. A "viable colony" is one that should be able to maintain itself and occurs in a place where the plant community, land ownership and management is reasonably stable (i.e., 30-50+ adults, two or more ha, indication of recruitment of young animals into colony, etc.). A "population" is the total number of turtles living in sites within a defined and separate geographic area (i.e. Southern Appalachian population, western Pennsylvania population).

Even with the most optimistic of population estimates, there are not many bog turtles. This in itself is not necessarily a serious concern. There are many species that, because of various biological factors, are naturally uncommon. The problem in this instance is the small area of total habitat available, the vulnerability of this habitat, and this turtle's documented decline. There are many species currently listed as federally Endangered which clearly have larger populations, higher annual reproductive rates, larger geographic ranges, and wider spectra of ecological tolerance than the bog turtle.

Reviewing the situation in Maryland, Taylor's 177 (Taylor et al. 1984) sites were clusters of interconnected localities, and the number of colonies (both active and historic) is much smaller. A number of sites/colonies have been lost since the mid-1970's (at least 43% decline in historic sites surveyed) and there are no important sites on protected public land. None of the sites on public land have potential to support substantial colonies in the future, even if they were to receive effective management. Furthermore, when Lee outlined a population monitoring/site management study he wanted to initiate and asked wildlife managers in Maryland if they knew of a sizable colony (ca 50+ turtles) of bog turtles that they thought would be stable enough for a 10-year population study, they knew of none. Only several known sites were of that magnitude, and all of those were on private property owned by elderly landowners, and their future was uncertain. To us, this was alarming information.

Of the 12 states in which the bog turtle is now known to occur, four (Massachusetts, Tennessee, South Carolina and Georgia) have only peripheral records or turtles known from small, isolated sites. In most of these states the turtle is known from a single county, and if one excludes historic records and single occurrences of animals found on roads etc., each of these states has only one or two known sites. However, of these four states, Massachusetts and Tennessee are known to still have viable colonies. Of the remaining states (Connecticut, New York, New Jersey, Delaware, Pennsylvania, Maryland, Virginia and North Carolina), the turtle occurs only in a small area of each, except for New Jersey. A typical site is 2 ha or less. If we assume that an average site occupies 1.5 ha and that there are 227 known extant sites (Table 2), the total area occupied by these turtles represents about 341 ha (Tryon and Hemman [1993] estimated 130 ha for the entire southeast). Hemman and Tryon (1993) estimated that perhaps only 52% to 54% of the total recorded localities in the southeast (Virginia to Georgia) had potential for bog turtles, but Tryon (pers. comm.) now thinks the 54% estimate is probably much too high for viable long-term populations in the south, and feels that 20% is more realistic. Herman and Tryon (1993) additionally document a 46% net reduction of habitat in the southeast in just the last 15-20 years. Accordingly, the combined total area now occupied by the species is probably less than half the size it was only two decades ago.

Chase et al. (1989) calculated an average of 44.2 turtles per site in Maryland. Their sites, compared to those in the Southern Appalachians, seem to have larger populations so this figure is too high for use throughout the animal's range. While some large sites do exist in Floyd Co., Virginia, and perhaps in Alleghany Co., North Carolina and Sussex Co., New Jersey, the average colony in these states is much smaller than 44 turtles. However, if this estimate is used in combination with the approximation of 227 occupied sites (Table 2)1, it g*es a total estimate of only 9,988 individuals. If we add to this the number in captivity in zoos and other responsible institutions (84 individuals, International Species Information System 1992), the estimated total number of existing bog turtles is still less than 11,000.

In terms of density of turtles per hectare, the studies cited by Bourg (Eglis 1967, BaTton 1957) represented large healthy colonies. Chase et al. (1989) gave a figure of 7 to 213 turtles per hectare. The average of their figures shows a mean of 58.7 turtles per hectare. Again, we think this figure is high for areas outside Maryland, and probably high for Maryland, where our field experience indicates typically lower densities. Using a figure of 357 occupied hectares of habitat, and the density figures given by Chase et al. (1989), we arrive at an estimate of less than 21,000 turtles for the species' total range. Thus, we have very crude population estimates of from 10 thousand to 21 thousand individuals in the global population. We consider these figures to be maximum estimates, and suspect that the global population may be much smaller. We include these estimates only to show that, even with figures obtained from high population density estimates, the estimated total number of bog turtles is still small. By these calculations, the bog turtle's estimated global population is smaller than the number of female Atlantic loggerhead turtles (Caretta caretta) nesting in the southern United States (14,000 to 29,000), or the global population of nesting female leatherback sea turtles (Dermochelys corracea) (estimate of 120,000 at time of listing, but now down to 20,000-30,000). These sea turtles are listed respectively as Threatened and Endangered by USFWS (Federal Register; 40685, 11 August 1977: Federal Register; 8495, 2 June 1990).

Now, to look at these population figures more realistically, Dawson (1984) noted that "the majority of sites [in Maryland] were less than 1 acre (0.40 ha) in size." In North Carolina (by far the state with the largest population in the southern Appalachians) Tryon and Hemman (1990) noted that most sites are from one to five acres in size (0.40 to 2.0 ha), and the average size was later calculated to be 1.06 ha for all colonies, and 1.47 ha for viable core habitats (Hemman and Tryon 1993). Therefore, in two states where bog turtles are believed to be doing well, the average site has less than 2 ha of habitat. While the total number of currently occupied sites is unknown, the number may be smaller than the 227 figure presented in Table 2. In 1978 Chase et al. (1989) randomly selected and revisited 20 occupied sites identified in the original survey. In 25% of these sites no bog turtles were found. In a recent conversation with wildlife managers from the Maryland Department of Natural Resources we leamed that by 1993 less than one half of the Taylor/Chase sites still supported turtles.

They also informed us that the site with the most turtles studied by Chase et al. (1989) had changed ownership, was no longer being grazed and multi-flora rose was dominating the habitat. One 2 acre (0.8 ha) site that we have been monitoring in Maryland since the late 1960's had been completely lost through natural succession by 1991, and it seems likely to us that many of the Maryland sites identified in the mid-1970's have similarly deteriorated through succession. It is reasonable to assume that as few as 100 sites (and probably fewer than 25 colonies) are currently extant in Maryland.

'Calculations on this and following pages are based on the numbers taken from publications cited in the previous text. While the numbers given by the U.S. Fish and Wildlife Service (1997) for the northern populations differ slightly (see Table 2), their figures would only increase the total number of extant sites by s, and would not significantly change the results.

Population sizes estimated from mark-recapture studies are interesting, but we must keep in mind that Chase et al. (1989) captured only 229 turtles in a minimum of 240 site visits, and in Maryland's original intensive surveys of the mid-1970s only 346 turtles were found. The statistical calculations are an important contribution, but remember that these calculations represent predictions and not known numbers of turtles. Furthermore, a large percentage of the captures at many sites are old adults, raising the possibility that reproduction may not be occurring with regularity.

Bog turtles are neither randomly nor evenly distributed throughout the sites they occupy. We have studied a 1.62 ha sedge meadow in Maryland. Turtles ranged throughout this site, but were concentrated in a central core which occupied about 25% of the total area. Population densities in that 25% were high. If we made density estimates in the core area and used them to calculate densities for the entire site, we would produce a misleadingly high estimate of the number of turtles present. Experienced bog turtle researchers focus effort where the number of captures will be greatest, in places where the turtles are most likely to be found. Densities calculated for these areas should not be used to estimate density distribution throughout an entire site. Additionally, using radio telemetry Stine (unpublished) showed seasonal migrations within his study site. If these movements are not considered, mark-recapture estimates may be misleading because the original marked animals might later be in less carefully searched portions of the site, and not likely to be recaptured. We do not know how to factor these considerations to get a realistic assessment of total population, but the number is probably lower than estimated above. Considering an estimated minimum 25% range wide habitat loss since the 1970's results in an estimate of only about 269 hectares of habitat and 179 sites throughout the entire range. Since there is no way to correct for possible bias with calculations of average number of turtles per site or turtles per hectare, we will use the previous density figures. Still, one must keep in mind that they are probably unrealistically high. For example, the large number of turtles at one Chase et al. (1989) site (213 individuals), which no longer exists, weighed heavily in their density estimates. If that site is not included in our calculations, the average density per occupied site becomes 39.5 turtles, a decrease of over 11%. The revised number of turtles calculated by average density per number of sites would be 7,912 (number of existing sites x 44.2 turtles per site). The number calculated by the reported average density per hectare is 11,890. It seems reasonable to us that the global population is somewhere between these last two figures, although one could argue that the calculations, because of overestimates of density, may be as much as 20 to 50% too high.

If one considers the number of historic sites now known to no longer support bog turtles, these population estimates are even more alarming. Particularly when we remember that many of those "historic" records date only from the 1960's and 1970's. For the few areas where there is enough information to document "historic" versus current occurrence, we find the following: Massachusetts, a 25% known loss; Connecticut at least a 75% loss; New York, a 45% loss with complete extirpation of one semi-relict population and another reduced to one site; New Jersey, 65% of historic sites lost; Delaware, about a 36% loss; Maryland, a 82.7% loss among 30 historic sites in less than 30 years, followed by a 43% loss of remaining surveyed sites in 15 years; Pennsylvania, at least a 50% loss of known sites and extirpation of the western disjunct populations; Virginia southward, only 50% of the total records now represent occupied sites (various regional reviews). Although these figures are crude (but conservative) we estimate at least a 50% average loss of known sites for this species throughout its entire range in a relatively short period. Add to this the fact that the majority of records (NY 13%, NJ 20%, PA 9%, MD 29% = 71% total) are from areas that are being rapidly developed.

Population size does not necessarily relate to the area available at a particular site, and some small sites support greater numbers of turtles than do some larger ones. In 1973 Ken Nemuras (in Cooper et al. 1973) estimated that there were about 50 "colonies" in Maryland with an average of 50 turtles per

colony. That estimate of the number of turtles per site proved to be quite accurate when compared to population estimates based on mark-recapture studies that were later conducted by Chase et al. (1989).

Detailed population estimates were made by Chase et al. (1989) at twenty sites in Carroll, Baltimore and Harford counties. The number of turtles at the sites studied ranged in estimated size from one to 230 (x = 44.2 individuals per site). Population density of turtles per ha ranged from seven to 213 (x = 58.77). Based on their figures, a crude estimate would indicate that seven to eight thousand bog turtles inhabited Maryland in the late 1970's (but see previous discussion). Because of habitat loss and succession, we suggest that there are fewer colonies in existence today and, based on an 82.7% loss of historic sites in less than a 30 year period, and the accelerated rate of land development, it is likely that only several thousand individuals currently inhabit Maryland. Recall that Maryland had a large number of sites, which in the late 1970's represented about half of the total known sites for the species (Table 2).

Unfortunately, the bog turtle, because of its small size, semi-terrestrial habitat, docile nature and unusually attractive appearance makes an ideal pet. This has always been a problem. Long before there were field guides and other books for the lay public on North American amphibians and reptiles, there was The Golden Nature Guide Series. In the reptile and amphibian volume of that series (Zim and Smith 1953) bog turtles were illustrated in an attractive terrarium and the text stated that "Turtle makes an excellent pet, eating chopped meat, earthworms, fruits and greens. It does not have to be fed in water, as do Spotted Turtles." In more recent versions of this book the illustrations and text do not emphasize captivity. As early as mid-1960, articles in the Bulletin of the Philadelphia Herpetological Society (Robotham 1963) were explaining the plight of bog turtle populations as they suffered from the pet trade and land use changes. Over the past three decades the problem has grown as the reptile pet trade has become a major business, and sky-rocketing prices have made international marketing very profitable. We are aware of considerable anecdotal information indicating that significant numbers of bog turtles have been, and still are being, shipped out of the United States, and the US Fish and Wildlife Service (1997) provided anecdotal and documented information on other instances of illegal collecting and export of this species for the pet trade, amounting to several thousand turtles.

The pet trade value of specimens has increased from about $100 an animal in 1973 to over $500 today. One 1992 add offered a pair of bog turtles for $1,500, and four adults were listed for sale at $750 per pair by a Florida animal dealer in that same year. With such high monetary value, it is difficult to protect colonies from poaching. Even acquisition of sites by bonafide conservation organizations does not guarantee protection from the pet trade. For instance, The Nature Conservancy acquired land in southeastern Pennsylvania because it contained a sizable colony of bog turtles. The site was the subject of population monitoring efforts and radio telemetry studies which showed a decline in population. Later, a researcher found approximately 60 turtle traps in the bog turtle habitat (L. Master, Memo Eastern Regional Office Nature Conservancy, 28 May 1992). This was clearly an illegal effort to obtain bog turtles by commercial collectors. Another regrettable incident happened during the summer of 1995 when the entire captive colony of bog turtles maintained at the Atlanta Zoo was stolen. Unknown persons entered the zoo grounds after hours and removed the turtles from their enclosure. Not only was this a very productive breeding colony, but it included some animals that had been in captivity for decades and had established longevity records. These turtles were part of a notable research effort, and the theft represents a loss of incalculable value. We are certain that these turtles were taken for sale to private collectors.

As noted by Dawson (1984) the majority of Maryland's occupied sites are on private land. This, along with their disjunct micro-distribution pattems, makes individual bog turtle sites difficult to protect

from collecting and habitat alternation, and problematic to manage. State and federal agencies are reluctant or unable to spend funds for habitat management on private land, and the landowners themselves have little incentive to maintain bog turtle habitat unless it is being done through activities already adopted, such as grazing. We feel that effective management of private bog turtle habitat will require financial incentives to encourage land owners to protect colony sites. As will be mentioned later, this could be accomplished through tax advantages associated with conservation easements, particularly if funds were made available to purchase such easements from landowners.

Another problem may be landowner concerns over local, state and federal regulation of private land. We have spoken with several landowners who are concerned about the future of their bog turtle habitat, but are unwilling to work with state or federal agencies. In these instances, greater cooperation would result from contact initiated by a private or quasi-private conservation group working toward bog turtle conservation. Such a group could be a national or local conservation organization, or a local land trust.

Most bog turtle colonies are small and many are less than an acre in size. Natural succession proceeds at a rapid pace in such small habitat islands, and it is difficult and time-consuming to monitor scattered, small sites. Nevertheless, this species' long term well-being depends less on the protection of individual turtles than it does on protecting habitat to maintain viable populations. Small, isolated sites on private land will be impossible to maintain without the active participation of land owners. The scattered nature of bog turtle sites in Maryland, coupled with the fact that very few are on public land, makes it clear that the cooperation of private landholders must be a major element in any effective program to protect bog turtles.

As discussed previously, land use changes resulting from the "Europenization" of North America have resulted in several cycles of change that have affected the distribution and abundance of this small turtle. For instance, after the pre-colonial period and the local elimination of elk, bison and beaver (species that we believe were once key agents of bog turtle habitat maintenance), land clearing by colonists and grazing by farm stock probably kept many sites open. Prior to the widespread use of tractors early in this century, wet meadows were often maintained as grazing areas with horse drawn plows (Isaac McGee, owner of one of our study sites, pers. comm.). However, when heavy tractors and other modern farm equipment were introduced, they easily became stuck in the saturated soils of sedge meadows. Consequently, these areas were drained, or were avoided and no longer kept open by occasional plowing.

Another factor that may have been significant was the promotion of farm pond construction in the mid-l900's. Because of the types of habitat occupied by bog turtles, we suspect that many sites were converted to piedmont farm ponds. The springs and seepage areas that fed the sedge meadows became the water source for ponds as land owners converted this "unproductive" land to warm water impoundments. Circular basin type habitats fed by hillside springs would have been particularly easy to convert. These habitat types were considered by Chase et al. (1989) to be among the best sites for bog turtle colonies. Unfortunately, the steep edges of farm ponds limit the amoumt of emergent vegetation, and transitional areas between land and water tend to be abrupt. Because of this, farm ponds do not provide suitable habitat for bog turtles.

The area north of the Fall Line between Philadelphia, Wilmington and Baltimore is the region now most heavily populated with bog turtles. In this area there are more colonies than there are in the combined remainder of the species' range. Not only is the area highly urbanized and bisected by numerous interstate highways, but even the rural remnants of this region support a rapidly expanding maze of suburban environments. In the 1950's, for example, there were vast areas in Dulaney Valley and Potts Springs in central Baltimore County which provided almost continuous stretches of potential bog turtle

habitat. This habitat was lost to development long before the ecological requirements of bog turtles were recognized, but despite the lack of records we suspect that the area supported these turtles. There is not likely to be any significant decrease in the rate of expansion of suburban communities in the Philadelphial Baltimore area.

Another major change in land use throughout this region is exhibited by the dairy industry. From the colonial period through the late 1800's dairy animals were commonplace and were allowed to graze in small numbers on private farms throughout the piedmont of Maryland. In 1884 the first Maryland creamery was established. The industry was at first concerned primarily with butter production, but in the Baltimore-Washington area milk was shipped to cities for direct consumption. This trade, because of the limitations of then available refrigeration, was originally restricted to areas within a twenty-five mile radius of large cities, and Baltimore, Harford and Montgomery were the first major milk-producing counties in Maryland (Vokes 1968). The development of better roads, faster transportation and refrigerated trucks gradually widened the area from which milk could be shipped.

According to information provided by the Maryland Department of Agriculture (Maryland Agricultural Statistics, County Data, various years) the four Maryland counties known to support bog turtles (Baltimore, Carroll, Cecil and Harford) had 123,489 dairy and beef cattle in 1964. However, by 1980 that number had decreased to 37,450. That trend continued, and by 1991 those four counties had only 30,500 cattle. If the reduction in numbers of dairy and beef cattle in these counties continues at this rate, the role played by these large herbivores in maintaining open meadows will be eliminated in the near future.

We feel that the dairy industry was important because the grazing activity of dairy herds kept sedge meadow areas open. Many of the sites known to us in the 1970's were grazed by dairy animals and it was probably this activity that allowed bog turtles to continue to be rather common locally throughout that decade. The dairy industry was the last in a series of events (presence of bison/elk, land clearing during the colonial period, private subsistence farms, and dairy farms) that maintained the open, wet meadow habitats inhabited by bog turtles in this area.

One other major factor in the recent loss of local bog turtle habitat was an increase in property value that has made it difficult for property owners to maintain family farms. As agricultural land, these properties are taxed at a relatively low rate that has not risen significantly in recent years. However, increased housing needs coupled with the desire of persons in the Baltimore Metropolitan Area to move into the surrounding countryside has greatly increased the demand for single family homes and commercial centers in areas where bog turtles occur. Projections developed by the Maryland Office of Planning (1992) predict that the three Maryland counties supporting the greatest number of bog turtles (Baltimore, Cecil and Harford) will experience 91,305 acres of new development between 1990 and 2020, and that 46,294 of the acres converted to commercial and residential uses will be agricultural land. So, while the monetary value of this land as rural open space and agricultural land has remained relatively stationary, its value under subdivision has increased significantly. Several older individuals who presently own agricultural land with bog turtle sites have indicated to us their desire to retain the property in a rural state. However, they are concerned about the fate of this family farmland when it passes into the hands of their children. It would seem that this change in emphasis from agricultural to residential uses is coming at a time when the attachment of the younger generation to historic family land uses is weakening.

The Maryland Non-tidal Wetlands Protection Act, and implementing regulations, allow for historic agricultural practices to continue, including grazing and periodic burning to maintain pasturage. However, there seems to be some misunderstanding on the part of some local jurisdictions, and the owner of one bog turtle site told us that he had been informed that he could not continue to pasture his farm animals

in the wet meadow on his property, or burn it to discourage the growth of woody vegetation (specifically multi-flora rose). This confusion emphasizes the need to merge ongoing efforts to manage bog turtle habitat with parallel efforts to protect wetlands. In addition, existing Maryland regulations allow for special consideration of wetlands having "exceptional ecological or educational value of Statewide significance" by designating them Non-tidal Wetlands of Special State Concern. We feel that the ecologically diverse sedge meadow habitats currently occupied by bog turtles, and other rare species such as Galium trifidum, deserve such designation.

Although existing regulations seem to adequately protect such small nontidal wetlands within areas undergoing development, they would not ensure that these sites remain habitat suitable for bog turtles. For instance, a wetland within a property to be subdivided could be retained with a surrounding buffer. However, it is virtually certain that the surface and subsurface hydrologic changes resulting from adjacent grading and construction of housing units, roadways, utility lines and other appurtenances would result in a wetland unsuitable for bog turtles. We feel that preservation of bog turtle sites requires that they either be granted expanded protection under existing federal, state and local nontidal wetland regulations, or that other mechanisms be developed to effectively protect these isolated habitats and the aquifers that feed them. These mechanisms could include fee simple acquisition, protective zoning, and conservation or land preservation easements. Although current emphasis seems to be on fee simple acquisition, we suggest that this option should not be the primary tool used to protect this species. There are several reasons for this. In this time of fiscal austerity, state and federal acquisition money will become more scarce than it is today. In addition, large bureaucracies may be less sensitive to the needs of single species, and necessary action occurs more slowly. For instance, due to the nature of its established land acquisition process, the purchase of a typical property by the State of Maryland, once it is identified as a priority, takes from twelve to twenty four months (Diane Ramsey, pers. comm). Clearly, private conservation groups or other non-public organizations could act more quickly and efficiently where acquisition is indicated.

As of January 1, 1995, over 50 land conservation organizations operated in Maryland, including 41 local land trusts and two major state land preservation agencies. All of these exist for the sole purpose of preserving land in its natural state, or for agricultural or historic purposes, either by acquiring land or enacting deeds of easement or other legally binding restrictions. Including preservation of bog turtle habitat into these goals seems to us to be a logical extension of their ongoing activities. The Maryland Agricultural Land Preservation Foundation (Maryland Department of Agriculture) actively pursues acquisition of land preservation easements for agricultural lands. As of June 30, 1994, they had protected 109,909 acres throughout Maryland, including 336 properties comprising 42,065 acres in Baltimore, Carroll, Cecil and Harford Counties. The Maryland Environmental Trust (Maryland Department of Natural Resources) generally does not purchase easements, but does accept conservation easements donated by landowners. As of January 1, 1955, they had placed restrictive easements on a total of 41,250 acres throughout Maryland, including 106 properties comprising 11,117 acres in these same four counties. Both of these state agencies, and all pertinent local land trusts, could be enlisted in the effort to protect bog turtles. Establishing a group to identify appropriate habitat, submit those recommendations to a land preservation agency, and work to ensure that an appropriate easement is written and appended to the deed would be a very useful element of any local bog turtle preservation effort.

Preservation of bog turtles in Maryland will require protection of the sedge meadows that they inhabit, and the groundwater systems that feed them. In addition, these preservation areas should be strung together along drainage corridors to allow for natural movement of turtles between sites of occurrence. Where possible, historic habitat that has been modified should be included within these corridors and appropriate management should be initiated to return damaged areas to suitable habitat. This concept is fully consistent with the "Greenways" strategy now being developed by the Maryland Greenways Commission, and .associated activities being conducted by the Maryland Department of Natural Resources, Maryland Department of Transportation, and other state agencies.

Although acquisition of habitat can be an effective preservation tool, this entire effort need not involve purchase of private property. In fact, considering that traditional private land management has been instrumental in maintaining bog turtles in this area, we suspect that most of the presently inhabited sites could, and perhaps should, remain in private hands. At the present time, only a few of the known sites are publicly owned. Since management of small, isolated patches of habitat would be difficult for any public agency with limited staff and resources, we suggest that acquisition by the state only be used where habitat is located adjacent to existing state land. Other areas could be protected most easily and quickly through an integrated program of regulatory control, protective zoning, conservation and land preservation easements. All of these are presently available, although no effort seems to have been made to use them to protect bog turtle habitat in this area.

Habitat management for animals like bog turtles is relatively simple, but protecting wildlife and assessing the need for various levels of protection has become a complex business which involves private landowners, local, state and federal agencies. The exact location and size of extant populations cannot be published for fear that private and commercial collectors would raid colonies. To some extent this practice also hampers conservation efforts. Some of the information we present is vague and detailed maps showing known Maryland sites are not included because of our desire to discourage visitation. However, that effort is probably pointless. Using readily available published information, USGS quad maps and the latest Maryland distribution maps (Harris 1975), we feel that we would have little difficulty determining the location of many local sites supporting bog turtles. The same result could undoubtedly be obtained by anyone else who was equally motivated.

Several authors have pointed out the need for habitat management in order to maintain viable bog turtle habitat (Lee 1973, Tryon and Hemman 1990). The most detailed treatment, presented by Tryon and Hemman (1990), is for habitat in the southern Appalachians, and includes concurrent distributional surveys, density studies, landowner education, habitat protection, and captive breeding and head start programs. Habitat management is essential for several reasons. Most obvious is natural succession. Most sites occupied by this turtle are small, 0.40 to 2.0 ha, and the characteristic open sedge meadows are easily swamped by invading woody vegetation. This process accelerates if the volume of groundwater seepage decreases, or if ponding or overland flow increases.

Tryon and Hemman (1990) thought that in the southern bog turtle habitats studied by them, ecological succession was progressing at an increased rate. This may be a result of nutrient runoff from farmland, but this has not been demonstrated. Of primary concern is shading caused by alders, willows and red maples. These plants may also lower the water table through transpiration. Tryon and Herman prescribed pruning, burning, and grazing by small-to-moderate sized herds of cattle or horses to maintain suitable habitat, and noted that well over 50% of the existing bog turtle sites in the southern Appalachians are located in pastures actively grazed by these animals. In instances where succession of woody plants has occurred, unwanted vegetation could be removed by selective cutting (which is practical because of the small size of many colony sites), burning or grazing (Tryon and Hemman 1990). Many small colony sites would also benefit from selective cutting of adjacent trees. This would not only improve the habitat, but cutting back the adjacent canopy could enlarge the area acceptable to bog turtles.

In addition to the natural succession of native species, the introduction of non-native species may also be a serious problem. For instance, the invasion of the introduced purple loosestrife (Lythrum salicaria) has altered the character of natural wetland communities in New York (Smith 1964). This exotic species could become a problem in other parts of the bog turtle's range, just as multiflora rose has done in Maryland. The Maryland Department of Natural Resources (Cooley 1994) has identified 56 invasive, exotic plants that occur on floodplains and in wetlands in Maryland. It is not known how many of these

regularly invade local bog turtle habitat, but the Carroll County site that we studied does support several plants on that list (halberd-leaved tearthumb, purple dead-nettle, Japanese honeysuckle, and multiflora rose).2

Management of areas where water tables have been altered, where springs were covered or capped, and where natural flow of water has been controlled would be difficult. In such places bog turtle colonies have disappeared or, at best, survive as remnants. Correcting local hydrology could be costly, difficult and probably not an efficient or effective tool for broad scale management. On the other hand, plant community management, where hydrology has not been altered, could be most effective. Systematic long range studies of the impact of habitat management on bog turtle population density would provide extremely useful information.

To ensure that site preservation and habitat management efforts are conducted effectively and efficiently, we recommend that active sites be monitored closely, and their status reviewed on a regular basis. All active sites should be photographed to allow periodic evaluation of successional change and permanent transects and botanical studies should be established in the more important localities. Property ownership and long range plans of property owners should be determined and cataloged, and property owners should be provided significant incentives to maintain habitat.

Bog turtles are known from 12 states in eastern North American. The species was included by Groombridge (1982) in the IUCN Amphibian and Reptile Red Data Book (Pt 1:31-33) and listed as intermediate, but it is not yet listed as Endangered or Threatened by the U.S. Fish and Wildlife Service. However, Fish and Wildlife Service published a proposed rule to list the bog turtle as Threatened on January 29, 1997 (Federal Register 62[19]: 4229-4239). In that rule, the northern population (from New York and Massachusetts south to Maryland) would be considered Threatened, and the southern population (Virginia to northern Georgia) would also be listed as Threatened due to the similarity of its appearance to the northern population. The comment period extends to April 29, 1997. We assume that the currently available information on this species, combined with comments that will be submitted by interested persons, will lead to fommal listing of the bog turtle sometime after that date.

Most states provide legal protection for this turtle under their existing nongame wildlife regulations, and it is listed in many of these states as Endangered or Threatened (Allen 1986). In Maryland the bog turtle was relisted as State Threatened by the Department of Natural Resources on October 24, 1994. As of that date possession of bog turtles without a special permit was prohibited in Maryland.

Even though the species is protected virtually throughout its range by state laws and regulations, there continues to be trafficking in wild caught turtles, particularly for international trade. Although the Lacey Act makes it illegal to transport illegally taken animals across state and international boarders, stronger legislative action is needed. Under CITES an export permit is required before a listed species can be sent out of the United States. The bog turtle was added to CITES Appendix II in 1975 (Honegger 1985). Because of the volume of trade in bog turtles, CITES upgraded the species from Appendix II to Appendix I on June 11, 1992. Federal protection under the Endangered Species Act would further protect the species from collectors. However, most collecting within their range is now illegal and it is becoming difficult for collectors to claim that turtles were obtained legally. Federal protection may not have the

2The magnitude of the problem posed by the invasion of introduced plant species is demonstrated by information provided by John Kartez from the Biota of North America Program housed at the North Carolina Botanical Garden. That database lists 784 introduced plant taxa for Maryland, fully 26.2% of the total flora.

advantage for bog turtles that it does for some other species because the majority of known colonies are on private lands. Because of this, "critical habitat" designations would not be practical. Modest enforcement by state agencies and the US Fish and Wildlife Service would effectively control legal commercial trade. However, illegal trade in reptiles is difficult to monitor and it is difficult to assess what impact commercial trade may continue to have if there is lack of vigorous enforcement. We feel that the high monetary value placed on bog turtles guarantees that a black market trade in illegal animals will flourish.

While we believe that bog turtles deserve protection under the Endangered Species Act because of habitat loss and the impact of the pet trade, there is another issue which deserves discussion. Because the species is not yet federally listed, funding for land acquisition, population monitoring, land management and scientific study is less readily available. Limits on state revenues force most states to focus conservation resources on federally listed species. Unlike many species of plants and animals which are federally protected, a relatively small amount of effort is needed for protection and management of bog turtles. This is because of the small size and the nearly self-contained nature of the sites that they occupy, the inexpensive and effective means of controlling succession (pruning, fire, grazing), and the small size of the turtles themselves. We recognize that protection of only the site without protection of the larger watershed, is a potential problem for long-range stewardship. However, current water quality and sediment and erosion control laws and regulations are probably adequate, and many sites are largely self-contained on hillsides or in small basins, or have a very limited upstream drainage area.

Elaborate trapping, tracking and housing facilities like those needed for eagles or wolves would not be necessary for the bog turtle. The species is easily maintained and bred in captivity. While we recognize that being federally listed is a dubious form of notoriety, it is a valuable asset in terms of rallying the interest of individuals and conservation groups in long term protection . The additional legal penalties that are provided under the Endangered Species Act would also help reduce illegal pressure from the pet trade, and would mandate protection of bog turtle colonies found on federal lands.

Since captive propagation of bog turtles is presently being conducted successfully at several locations, we feel that the potential impact of this practice on commercial trade should also be considered. While we believe that protection of colonies and successional management of occupied sites is the best long range strategy for the preservation of this species, we also advocate captive breeding programs. Youmg produced by turtles from known localities could be used for restocking purposes, and for the preservation of genetic stock until protection, land acquisition and habitat management options can be put into place. Restocking programs could be used to start new colonies on protected state or federal lands, to re-establish extirpated colonies, or to add new "blood" to depleted colonies where inbreeding of small populations could lead to abnormalities. It is important for captive breeding programs not to mix stock if the release of captive bred turtles into the wild is planned. At the present time we know very little about the level of genetic difference between bog turtle populations. In fact, we do not even know if more than one taxon is present. To prevent a drain on native stocks, captive bred animals could be used for research purposes and for exhibit in zoos and nature centers.

Making it legal to sell or trade captive bred bog turtles does have some potential to adversely impact conservation efforts. If trade in bog turtles is totally prohibited, identification of illegally obtained turtles will be easier than if the sale of captive bred individuals is permitted. While most breeders and hobbyists are highly reputable, there will always be some individuals who will attempt to illegally market any animal having a high commercial value. Proposals to identify and track captive-bred or otherwise legally permitted animals are laudable, but probably not foolproof, at least with existing technology. Current emphasis is typically on microchip marking systems using passive integrated transponders (PIT's). These

devices are effective and accurate (Camper and Dixon 1988), but with the value of bog turtles pushing into the $500 to $1000 range, we suggest that the appearance of counterfeit or black market PIT tags is predictable. On the other hand, captive breeding projects could produce enough turtles to significantly reduce demand and market value, and if set up properly could generate funds to be used for habitat protection or acquisition.

Captive husbandry techniques for bog turtles are well established (Tryon and Hulsey 1972). Bog turdes were first bred by professionals in a zoo in 1973, but private individuals were successfully breeding this species as early as 1964 (Arford 1967, Eglis 1967, Galli 1978). Currently, there are at least eight institutions housing captive bog turtles, but only half of these are engaged in breeding programs (International Species Information System 1992). In the spring of 1992 Lee, in a cooperative agreement with the Maryland Department of Natural Resources, stalted a captive breeding program for this species using stocks from northern Carroll County, Maryland. Offspring from that effort are currently being loaned to zoos to expand captive breeding of Maryland bog turtles. Another successful captive breeding program involving Maryland bog turtles is presently underway at the Baltimore Zoo.

We thank Mary K. Clark (North Carolina State Museum of Natural Sciences), Carl Ernst (George Mason University), William Grogan (Salisbury State University), Dennis Herman (Zoo Atlanta), Donnell Redman (Natural History Society of Maryland), and Ann Sommers (University of North Carolina, Greensboro) for reviewing earlier drafts of this paper. Karl Krombein (National Museum of Natural History, Smithsonian Institution) provided access to the Plummers Island logbook maintained by the Washington Biologists' Field Club. Elmer Worthley, Jean Worthley, Tom Benassi and Donnell Redman assisted with the plant community analysis. John Kartez provided information on introduced plant species from the Biota of North America Program. David Stedman (Florida State Museum) and Charles Stine (Johns Hopkins University) granted permission for us to use their unpublished data. Tim Almager provided information on turtles in the collection of the Natural History Society of Maryland and William Grogan, Herbert Harris and Ken Nemuras discussed their many years of field experience with this turtle. Herbert Harris also granted permission to reprint the maps in Figure 3. A number of persons from the Maryland Department of Natural Resources provided useful information, including Glenn Therres and Scott Smith who shared information on this turtle's current status in Maryland, George Beston who discussed the Maryland Nontidal Wetland Protection Act, and Diane Ramsey who explained the State's land acquisition process. Robert Miller provided several distribution records and loaned us the Maryland specimen from the Towson State University collection on which Figure 2A is based. Renaldo Kuhler (North Carolina State Museum of Natural Sciences) prepared the illustrations. Cathy Wood typed the manuscript.

Literature Cited Abbott, C. C. 1882. Notes on the habits of the "Savannah Cricket Frog." American Naturalist 16:707.

Alford, D. 1967. Clemmys muhlenbergi. Bulletin Chicago Herp Society 2(4): 1.

Allen, W. B., Jr. 1986. State lists of endangered and threatened species of reptiles and amphibians including laws and regulations of each state. Privately Printed. Pittsburgh Zoological Park.

Amato, G., J. L. Behler, B. W. Tryon, and D. W. Herman. 1993. Molecular variation in the bog turtle, Clemmys muhlenbergii. Proceedings: Conservation, Restoration and Management of Tortoises and Turtles- an International Conference. SUNY, Purchase, New York (July 1993).

Amdt, R. G. 1972. Additional records of Clemmys muhlenbergi in Delaware, with notes on reproduction. Bulletin Maryland Herpetological Society 8(1): 1-5.

Arndt, R. G. 1977. Notes on the natural history of the bog turtle, Clemmys mahlenbergi (Schoepff), in Delaware. Chesapeake Science 18: 67-76.

Amdt, R. G. 1978. The bog turtle...an endangered species? Delaware Conservationist 22(2): 18-21, 25.

Amdt, R. G. 1986. Notes on the bog turtle, Clemmys mahlenbergi, in Warren County, New Jersey. Bulletin Maryland Herpetological Society 22(2): 56-61.

Babcock, H. L. 1917. An extension of the range of Clemmys muhlenbergii. Copeia 42: 32.

Barton, A. J. 1957. Our knowledge of the bog turtle, Clemmys mahlenbergi, further augmented. Unpublished Masters thesis. University of Pittsburgh, Pittsburgh, Pennsylvania. 74 pp.

Barton, A. J. 1960. Deletion of Virginia from the known range of Clemmys mahlenbergi. Herpetologica 16(2): 120.

Barton, A. J., and J. W. Price, Sr. 1955. Our knowledge of the bog turtle, Clemmys muhlenbergi, surveyed and augmented. Copeia 1955: 159-165.

Beck, H. 1928. Henry E. Muhlenberg, botanist. Papers Read Before the Lancaster County Historical Society (6 January 1928) 32(1): 99-107.

Bickham, J. W. 1975. A cytosystematic study of turtles in the genera Clemmys, Mauremys and Sacalia. Herpetologica 31(2): 198-204.

Bonwill, A., and H. Owens. 1939. The retum of a native. Bulletin Natural History Society of Maryland 10: 35-45.

Bourg, N. A. 1992. Status of the bog turtle (Clemmys muhlenbergii) in North America. Report submitted to U.S. Fish and Wildlife Service. Newton Comer, MA 02158.

Brady, M. K. 1924. Muhlenberg's turtle near Washington. Copeia 1924: 92.

Brady, M. K. 1937. Natural history of Plummers Island, Maryland. VI. Reptiles and Amphibians. Proceedings of the Biological Society of Washington 50: 137-140.

Brattstrom, B., and A. Stum. 1959. A new species of fossil turtle from the Pliocene of Oregon, with notes on other fossil Clemmys from western North America. Bulletin Southern Califomia Academy of Sciences 58(2): 65-71.

Brush, G. S. 1975. Forest Ecology of the Piedmont Region, Maryland. A Progress Report. The Johns Hopkins University, Baltimore, Maryland. 50 pp.

Bury, R. B. 1979. Review of the ecology and conservation of the bog turtle, Clemmys mahlenbergu U.S. Fish and Wildlife Service, Special Scientific Report - Wildlife No. 219.

Campbell, H. W. 1960. The bog turtle in Maryland. Maryland Naturalist 30(1-4): 15-16

Camper, J. D., and J. R. Dixon. 1988. Evaluation of a microchip marking system for amphibians and reptiles. Texas Parks and Wildlife Dept., Res. Publ., 7100-159:1-22.

Chase, J. D., K. R. Dixon, J. E. Gates, D. Jacobs, and G. J. Taylor. 1989. Habitat characteristics, population size and home range of the bog turtle, Clemmys mahlenbergi, in Maryland. Joumal of Herpetology 23(4): 356-362.

Cleaves, E. T., D. W. Fisher, and O. P. Bricker. 1974. Chemical weathering of serpentenite in the eastern Piedmont of Maryland. Geological Society America Bulletin 85: 437-444.

Coleman, S., J. Halka, C. Hobbs III, R. Mixon, and D. Foster. 1990. Ancient channels of the Susquehanna River beneath Chesapeake Bay and the Delmarva Peninsula. Geological Society of America Bulletin 102: 1268-1279.

Conant, R., and J. Collins. 1991. A Field Guide to Reptiles and Amphibians. Third Edition. Houghton Mifflin Company, Boston. 450 pp.

Cooley, G. 1994. Invasive exotic plants of wetlands and floodplains of Maryland. Maryland Department of Natural Resources (22 July 1994). 3 pp.

Cooper, J. E. 1947. Some observations on winter herpetology with reference to early spring collecting. Natural History Society of Maryland, Junior Society News 3(14): 4.

Cooper, J. E. 1949. Additional records for Clemmys mahlenbergii from Maryland. Herpetologica 5: 75-76.

Cooper, J. E. 1960. Distributional Survey of Maryland and the District of Columbia. Bulletin Philadelphia Herpetological Society, May-June, pp. 18-24.

Cooper, J. E., R. Franz, F. Groves, J. Hardy, H. Harris, D. Lee, P. Wemple, and R. Tuck. 1973. Endangered amphibians and reptiles of Maryland: A Special Report. Bulletin Maryland Herpetological Society 9(3): 43-100.

Cope, E. D. 1873. Sketch of the zoology of Maryland. Pp. 16-18. In Walling and Gray's New Topographical Atlas of Maryland. O. W. Gray, Philadelphia.

Dawson, S. A. 1984. The status of the bog turtle (Clemmys mahlenbergi) in Maryland. Pp. 360-362. In A. W. Norden, D. C. Forester, and G. T. Fenwick (eds). Threatened and Endangered Plants and Animals of Maryland. Maryland Department of Natural Resources, Natural Heritage Program Special Publication 84-I.

Dunn, E. R. 1917. Reptile and arnphibian collections from the North Carolina mountains, with special references to salamanders. Bulletin American Museum of Natural History 37(23): 593-634.

Eglis, A. 1967. Clemmys muhlenbergi, rarest of North American turdes. Animal Kingdom (New York Zoological Society) 70: 58-61.

Emst, C. H. 1977. Biological notes on the bog turtle, Clemmys mahlenbergii. Herpetologica 33(2): 241-246.

Emst, C. H. 1983. Clemmys guttata (spotted turtle) and Clemmys mahlenbergii (bog turtle). Natural hybrid. Herpetological Review 14: 75.

Emst, C. H. 1985. Endangered bog turtle. Pp. 270-273. In H. H. Genoways and F. Brenner (eds). Species of Special Concern in Pennsylvania. Camegie Museum of Natum1 History, Special Publication No. 11. Pittsburgh, Pa.

Emst, C. H., and R. W. Barbour. 1972. Turtles of the United States. University of Kentucky Press, Lexington, Kentucky. 347 pp.

Emst, C. H., and R. B. Bury. 1977. Clemmys muhlenbergli (Schoepff). Cat. Amer. Amphib. Reptiles 204: 1-2.

Emst, C. H., R. W. Barbour, and J. E. Lovich. 1994. Turtles of the United States and Canada. Smithsonian Institution Press, Washington, D.C. 578 pp.

Emst, C., R. Zappalorti, and J. Lovich. 1989. Overwintering sites and themmal relations of hibemating bog turtles, Clemmys muhlenbergii. Copeia 1989(3): 761-764.

Erwin, T. L. 1981. Natural history of Plurnmers Island, Maryland. XXVI. The ground beetles of a temperate forest site (Coleoptera: Carabidae): an analysis of fauna in relation to size, habitat selection, vagility, seasonality, and extinction. Bulletin Biological SocietyWashington (5): 104-224.

Fay, L. P. 1984. Mid-Wisconsin and mid-holocene herpetofaunas of Eastern North America: a study in minimal contrast. Pp. 14-19. In Genoways, H. H., and M. R. Dawson (eds). Contributions in Quatemary Vertebrate Paleontology: a Volurne in Memory to John E. Guilday. Camegie Museum of Natural History, Special Publication No. 8.

Fitzinger, L. J. 1835. Entwarf einer systematischenAnordnung der Schildkroten nach den Grundsatzen der Naturlichen Methode. Ann. Mus. Wien 1:103-128.

Fowler, J. A. 1945. The arnphibians and reptiles of the National Capital Parks and the District of Columbia Region. U.S. Department of the Interior, National Park Service, National Capital Parks, Washington 115498. 4 pp.

Galli, J. 1978. The bog turtle an endangered species. New Jersey Outdoors 5(2): 8-10.

Garretson, M. S. 1938. The Arnerican Bison: The Story of its Extemmination as a Wild Species and its Restoration Under Federal Protection. New York Zoological Society. 254 pp.

Gidley, J. E. 1971. The Pleistocene history of the Appalachian marnma1 fauna. Pp. 233-262. In P. C. Holt (ed). The Distributional History of the Biota of the Southern Appalachians. Palt III. Vertebrates. Virginia Polytechnic Institute and State Un*ersity, Blacksburg, Virginia. Research Division Monograph 4.

Gidley, J. W., and C. L. Gazin. 1938. The Pleistocene vertebrate fauna from Cumberland Cave, Maryland. U.S. National Museum, Bulletin 171. 99 pp.

Groombridge, B. 1982. The IUCN Amphibian-Reptilia Red Data Book, Testudines, Crocodylia, Rhynchocephalia. International Union for Conservation of Nature and Natural Resources. Gland, Switzerland. Part 1. 426 p.

Hale, P. E., and M. J. Harris. 1980. Geographic distribution. Clemmys mahlenbergi. Herpetological Review 11(1): 13.

Hall, E. R. 1981. The Mammals of North America. John Wiley and Sons, New York. Vols. 1-2.

Harding, J. H., and T. J. Bloomer. 1979. The wood turtle, Clemmys insculpta...anatural history. Bulletin New York Herpetological Society 15(1): 9-26.

Harris, H. S. 1969. Distributional Survey: Maryland and the District of Columbia. Bulletin Maryland Herpetological Society 5(4): 97-161.

Harris, H. S. 1975. Distributional Survey (Amphibia/Reptilia): Maryland and the District of Columbia. Bulletin Maryland Herpetological Society 11(3): 73-167.

Hay, O. P. 1908. The fossil turtles of North America. Camegie Institution of Washington, Publication. No. 75. 568 pp.

Hemman, D. W. 1990. Captive husbandry of the eastern Clemmys group at the Zoo Atlanta. Pp. 54-62.

In K. R. Beaman, F. Caporaso, S. McKeown and M. D.Graff (eds). Proceedings of the First International Symposium on Turtles and Tortoises: Conservation and Captive Husbandry. Chapman Un*ersity, Orange, Califomia.

Hemman, D. W. 1991. Natural history of Clemmys muhlenbergi (bog turtle) inthe Southern Appalachians. Association Southeastern Biologists Bulletin 38(2): 110. (abstract).

Hemman, D. W. 1992. Status and distribution of the bog turtle, Clemmys mahlenbergii, in North Carolina, 1992. Progress report submitted to North Carolina Wildlife Resources Commission, 31 December 1992. 67 pp.

Herman, D. W. 1994. The bog turtle, Clemmys mahlenbergli, in North Carolina. An action plan for its conservation and management. Report to North Carolina Wildlife Resources Commission. 144 pp.

Hemman, D. W., and C. E. Putnam. 1982. Geographic distribution. Clemmys muhlenbergi. Herpetological Review 13(2): 52.

Hemman, D. W., and C. E. Putnam. 1983. Two new records of the bog turtle, Clemmys muhlenbergi Schoepff, in Georgia. Herpetological Review 14(2): 55.

Hemman, D. W., and B. W. Tryon. 1993. Land use, development and natural succession and their effects on bog turtle habitat in the southeastern United States. Proceedings: Conservation, Restoration and Management of Tortoises and Turtles - an International Conference. SUNY, Purchase, New York (July 1993).

Hemman, D. W., B. W. Tryon, and A. C. Boynton. 1993. Clemmys mahlenbergii. Herpetological Review 24(4): 154.

Hemman, D. W., and J. L. Warner. 1986. Geographic Distribution. Clemmys muhlenbergi. Herpetological Review 17(4): 92.

Hocutt, C. H., R. E. Jenkins, and J. R. Stauffer, Jr. 1986. Zoogeography of the fishes of the central Appalachians and central Atlantic Coastal Plain. Pp. 161-211. In C. H. Hocutt and E. O. Wiley (eds). The Zoogeography of North American Freshwater Fishes. John Wiley and Sons. New York.

Guthein, F. 1949. The Potomac. The Johns Hopkins University Press. Baltimore, Maryland. 414 pp.

Holman, J. A. 1977. The Pleistocene (Kansan) herpetofauna of Cumberland Cave, Maryland. Annals Camegie Museum 66: 157-172.

Holub, R., and T. Bloomer. 1977. The bog turtle, Clemmys muhlenbergi... a natural history. Bulletin New York Herpetological Society 13(2): 9-23.

Honegger, R. E. 1985. Clemmys mahlenbergi (Schoepff, 1801). Code A-301.007.007.004:1-2. In P. Dollinger (ed). Convention on International Trade in Endangered Species of Wild Fauna and Flora Identification Manual. Volume 3: Reptilia, Amphibia, Pisces. Secretariat for the Convention, Lausanne, Switzerland.

Hulse, A. C., and K. L. Hulse. 1992. New county records for amphibians and reptiles from Pennsylvania. Herpetological Review 23(2): 62-64.

Hutchison, V. H. 1963. Record of the bog turtle, Clemmys muhlenbergi, in southwestern Virginia. Copeia 1963: 156-157.

International Species Information System. 1992. Reptile Abstract. Apple Valley, MN.

Jenkins, R. E., E. A. Lachner, and F. J. Schwartz. 1972. Fishes of the central Appalachian drainages: their distribution and dispersal. Pp 43-117. In P. C. Holt (ed.). The Distributional History of the Biota of the Southern Appalachians. Part III. Vertebrates. Virginia Polytechnic Institute and State Un*ersity, Blacksburg. Research Division Monograph 4.

KauLman, J. 1995. Home ranges and movements of wood turdes, Clemmys insculpta, in central Pennsylvania. Copeia 1995(1): 22-27.

Kerhin, R., J. Halka, J. Hill, D. Wells, R. Conkwright, L. Hennessee, W. Panageotou, and M. June Parlc. 1996. The Maryland Geological Survey's contributions to the geology of Chesapeake bay. Pp. 5769. In D. Brezinski and J. Reger (eds). Studies in Maryland Geology- in Commemoration of the Centennial of the Maryland Geological Survey. Maryland Geological Survey Special Publication No. 3. Baltimore, Maryland.

Kivat, E. 1978. Bog turtle habitat ecology. Bulletin Chicago Herpetological Society 13(2): 29-42.

Landry, J. L. 1979. A bibliography of the bog turtle, Clemmys muhlenbergi, biology, ecology and distribution. Smithsonian Herpetological Information Service, No. 44. 21 pp.

Lee, D. S. 1970. Ecological associations of amphibians and reptiles with certain Maryland mammals. Bulletin Maryland Herpetological Society 6(2): 19-21.

Lee, D. S. 1973. Creatures in distress. Maryland Conservationist 49(6): 14-19.

Lee, D. S. 1976. Aquatic zoogeography of Maryland. Atlantic Naturalist 31: 147-157.

Lee, D. S. 1984. Maryland's vanishedbirds and mammals: reflections of ethics past. Pp. 454-471. In A. Norden, D. Forester, and G. Fenwick (eds). Threatened and Endangered Plants and Animals of Maryland. Maryland DepartmentofNatural Resources,Natural Heritage Prograrn Special Publication 84-1.

Lee, D. S. 1987. The star-nosed mole on the Delmarva Peninsula: zoogeographic and systematic problems of a boreal species in the south. hIaryland Naturalist 31(2): 44-57.

Lee, D. S., and J. B. Funderburg. 1977. Mammals. Introduction and species accounts. Pp. 385-389, 392-407. In J. E. Cooper, S. S. Robinson, and J. B. Funderburg (eds). Endangered and Threatened Plants and Animals of North Carolina. N.C. State Museum of Natural History, Raleigh.

Lee, D. S., and D. W. Hemman. 1995. Zoogeographic history of the bog turtle Clemmys mahlenbergi. 1995 SSAR/