2007 NORTHEASTERN NATURALIST 14(2):279–292 Ecology of Thamnophis sauritus (Eastern Ribbon Snake) at the Northern Limit of its Range Sarah L.M. Bell1, Tom B. Herman2, and Richard J. Wassersug3,* Abstract - In Canada, Thamnophis sauritus (Eastern Ribbon Snake) is found only in southern Ontario and a small area of southwestern Nova Scotia. Although the Nova Scotia population is nationally designated as threatened, its distribution, seasonal activity, movement patterns, and over-wintering sites remain undescribed. We used radio-telemetry, capture-mark recapture, and direct observation to: 1) assess abundance, summer activity, and movement; and 2) to locate and characterize a hibernaculum for Eastern Ribbon Snakes in Kejimkujik National Park, NS. A total of 105 individuals were marked; among these, 13 free-ranging adults were surgically implanted with radio-transmitters and tracked from June until mid-November 2001. From late May to September, snakes were always found within 5 m of water, with summer ranges on land that rarely exceeded 5 x 10 m. From September to mid-October, snakes moved up to 173 m away from the shoreline. Eleven observations of snakes feeding on anurans (Ranidae) and fish (Cyprinidae) were made at temporary pools, marginal to the lake. Despite the use of radio-telemetry, only one hibernaculum was found. Our observations indicate that the Eastern Ribbon Snake is relatively sedentary; its low activity rate and small activity range may make it vulnerable to local extinction. Introduction In Nova Scotia, Thamnophis sauritus septentrionalis Rossman, northern subsp. (Eastern Ribbon Snake) is apparently restricted to a narrow range of habitats in two watersheds in the southwest interior (Gilhen 1984). This highly disjunct, post-glacial relict population was recently designated as threatened (COSEWIC 2002), primarily due to its small size and range, its susceptibility to environmental and demographic stochasticity, and its potential for divergence. Elsewhere, ribbon snakes are diurnal, semi-aquatic, and normally found along the water’s edge, amongst dense vegetation. Amphibians constitute their main prey (Rowe et al. 2000) although different diet profiles have been reported for different localities (Carpenter 1952, Tinkle 1957). It is not clear whether this is due to regional differences in preference by snakes or differences in prey availability at different sites. The Eastern Ribbon Snake is viviparous; mating occurs following emergence in spring (Clark 1974, Tinkle 1957), although fall mating has also been suspected in Nova Scotia (J. Gilhen, Nova Scotia Museum,J. Gilhen, Nova Scotia Museum, Halifax, NS, Canada, pers. comm.). 1Department of Biology, Dalhousie University, Halifax, NS B3H 4J1, Canada. 2Department of Biology, Acadia University, Wolfville, NS B4P 2R6, Canada. 3Department of Anatomy and Neurobiology, Dalhousie University, Halifax, NS B3H 1X5, Canada. *Corresponding author - tadpole@dal.ca. 280 Northeastern Naturalist Vol. 14, No. 2 Despite the species’ threatened status in Nova Scotia, its distribution, seasonal activity and movement patterns, and over-wintering sites remain largely undescribed. At the northeast edge of the species' range, this population is subject to long, harsh, and variable winters and short, cool summers, which may constrain reproduction and survival (Gregory and Larsen 1993, Larsen et al. 1993). To better gauge the risks that this population faces, we used radio-telemetry, capture-mark-recapture, and direct observation to assess abundance, summer activity, and movement, and to locate hibernacula for ribbon snakes in Kejimkujik National Park, NS. Materials and Methods Study site Fieldwork was conducted from mid-May to mid-November 2001, at Grafton Lake, Kejimkujik National Park (KNP; 38,100 ha), in southwestern Nova Scotia (44°23'N, 65°11'W). Grafton Lake was dammed from 1938 until 1996; in 1996, the dam was breached in an effort to restore the lake and surrounding habitats to pre-dam conditions, including natural water levels and fluctuations. The sampling area encompassed approximately 33.5 ha of wetland adjacent to the outlet of the lake. The site was chosen because of a concentration of previous sighting reports and ease of access. The terrestrial portion of the sampling area chiefly comprised emergent lakebed with slate outcrops, granite erratics, tree stumps, and rapidly revegetating organic sediment dominated by Spartina pectinata Bosc ex Link (prairie cordgrass), Scirpus spp. (bulrushes), Typha spp. (cattails), and Anaphalis margaritacea (L.) Benth. (common pearleverlasting). Pontederia cordata L. (pickerel weed), Nymphoides cordata (Ell.) Fern. (little floating heart), Nuphar variegata Dur. (variegated yellow pond-lily), and Nymphaea odorata Ait. (American waterlily) were the dominant aquatic species. Snake sampling Snakes were first located by systematic visual sweeps from the edge of the forest margin to Grafton creek, just upstream of the breached dam. The water’s edge and adjoining flood plain were intensely searched. Snakes were hand-captured, sexed, photographed, and measured. Snout-vent length (SVL) was obtained using a measuring tape while the animal was extended. Body mass (BM) was obtained by placing the animal in a mesh bag and suspending it from a hanging scale. Eastern Ribbon Snakes were marked individually by clipping ventral scales (Brown and Parker 1976). Young-of-the-year (YOY) were marked with a non-toxic permanent marker, instead of scale clipping, due to the small size of their ventral scales. Individuals were placed in one of three categories (adult, juvenile, or YOY) based on SVL and BM. A subset of animals was implanted with radio-transmitters (see below). 2007 S.L.M. Bell, T.B. Herman, and R.J. Wassersug 281 A Garmin GPS III Plus handheld global positioning systems unit (Romsey, UK and Taipei, Taiwan) was used to reference initial capture or sighting localities, and also served as a reference when snakes were released following radio-transmitter implantation. GPS locations were also collected for all recaptures and relocations; however, the GPS sensitivity was too low to represent the limited movements of these snakes. Sampling began on 22 May 2001 and concluded on 10 November 2001, with 52 days of extensive sampling distributed relatively evenly from late May to early September, as follows: May (3 d), June (12 d), July (17 d), August (15 d), and September (5 d). Additional, less extensive sampling occurred from May to November 2001. We typically surveyed for snakes between 0900–1700 h ADT in all habitat types within the study area. Sampling effort was not uniform, since some non-telemetered snakes were coincidentally located while tracking animals with transmitters. Thus, survey procedures were probably biased toward microhabitats favoured by snakes. All feeding events were directly observed; no snakes were captured and forced to regurgitate food items. Dietary determination was also done by observation of natural feeding events. During these feeding events, the observer was never closer than 2 m from the subject, which is why prey were not identified to the species level. Thermal data were also collected throughout the season, but are not reported here except in relation to hibernacula use. Radio-telemetry Temperature-sensitive radio-transmitters (model BD-2G; Holohil Systems Limited, ON, Canada) were surgically implanted in eleven females and two males. Transmitter mass was < 5% of a snake’s BM and had a diameter significantly smaller than that of the individual, following the guidelines for transmitter use outlined in Reinert and Cundall (1982). The whip antennas were shortened to 14 cm and sealed with 100% silicone rubber aquarium sealant 2–3 days prior to surgery. Surgeries were performed under lidocaine anesthesia. The telemetry study included two stages: summer—assessment of summer activity and movement; and fall—location of hibernacula. For summer, transmitters were implanted in six gravid adult females between 7 June and 28 June 2001, and snakes were tracked from June to the end of August 2001. The transmitters for this stage weighed 1.95 g, with a battery life of 16 weeks at 30 ºC. Snakes were recaptured 10–14 days after surgery in order to remove the non-dissolvable nylon sutures. For fall, transmitters were implanted in five adult females and two adult males between 5 September and 14 September 2001, and snakes were tracked from September to mid-November 2001. These transmitters weighed 1.06 g, with a battery life of 7 weeks at 35 ºC. Dissolvable sutures were used 282 Northeastern Naturalist Vol. 14, No. 2 with these snakes to ensure rapid and efficient postoperative healing in preparation for hibernation. Following surgery, snakes were monitored in the lab for 3–5 days to allow for postoperative healing. During this period, snakes were provided water ad libitum and fed 4–5 Rana pipiens Schreber (Northern Leopard Frog) tadpoles to ensure that they had normal feeding behaviour. Subsequently, the snakes were released at their original capture localities. No evidence was found in captivity, nor in the field, to suggest that the transmitters hampered the snakes’ movement in any way. All radio-transmitters were in the 172-MHz frequency band with an anticipated range of 1.0 km. Tracking was done manually using a Wildlife Materials International Inc. (Carbondale, IL) model TRX 1000S receiver with a handheld, 3-element Yagi antenna. Results Capture patterns and abundance The first sighting of a snake occurred on 30 May 2001, and the last transmitter signal was obtained on 8 November 2001. A total of 105 snakes (53 males, 52 females; 13 telemetered, 92 non-telemetered), including nine YOY, were individually marked at the study site. Snakes were active from late May through late September. From extensive sampling on 52 of 110 days between 22 May and 9 September, 40 of the 96 (42%) individually marked snakes (non-YOY) were recaptured at least once. Recapture frequencies, excluding those from radio tracking, ranged from 0–6 per individual. Capture frequency was unrelated to body size (SVL; r2 = 0.027, F = 2.598, P > 0.11), but males were recaptured more frequently than females (male: mean = 2.42, median = 2; female: mean = 1.36, median = 1 total captures/individual; Mann- Whitney U = 686.5, P < 0.002). Monthly sex ratios (May–September) did not vary from unity (􀁲2: df = 4, P = 0.903). Intervals between consecutive captures of non-telemetered, marked snakes during that period, with all recaptured individuals pooled, ranged from 1–56 d (mean = 10.3, median = 6). When mean capture intervals were calculated per individual and a frequency distribution among individuals was generated, the pattern was similar to the pooled sample (mean = 11.0, median = 7.6). Among recaptured individuals, the time between the first and last sighting ranged from 2–103 d (mean = 30, median = 20). Capture frequencies (non-YOY, male and female combined) during this sampling period fit a theoretical geometric distribution (G-test goodness of fit: Gadj = 8.29, P 􀂧 0.217). Based on this distribution, the population within the sampling area was estimated from frequency of capture (Caughley 1977) to be 􀂧 197. Chao’s Method (Chao 1988) yielded a similar estimate, i.e., 188 (95% CI = 141–284). Converting to population density, this translates to 5.6–5.9 snakes/ha. 2007 S.L.M. Bell, T.B. Herman, and R.J. Wassersug 283 The range in time between relocations (i.e., the interval between locating a marked or transmitter-implanted snake once and then again on a subsequent day) was two to 103 days. During summer, 93% of relocations for telemetered snakes were made at the same locality where they had been previously located. In contrast, during fall, only 48% of relocations were made at the same localities. Morphology and reproduction Snout–vent length varied from 140 mm to 490 mm in females and from 150 mm to 500 mm in males (Table 1), and overall, did not differ significantly between sexes (F = 0.42, P = 0.516). In contrast, body mass (BM) of males was significantly lower than that of females (F = 5.43, P = 0.022), independent of pregnancy status (Table 1). Gravid females were observed from June through September, and YOY were first observed in August. Seventeen snakes (16%) that were captured had scars from previous injuries, or were missing small portions of their tails, suggesting predatory encounters or frost damage. Three YOY (2 male and 1 female) were found dead, and one gravid adult female in captivity produced a stillborn YOY on 28 August 2001. Distribution and movement Ribbon snakes were linearly distributed along the shoreline of the study site (Fig. 1). During late May and June, snakes were found only on the north side of the lake. From July through September, adults were found frequently along both the north and south sides of the lake; however, males were more widely distributed than females. For example, juvenile females were only observed on the north side of the lake, whereas juvenile males were found on both sides. Young-of-the-year were also found only on the north side. All telemetered snakes remained on the north side of the lake. Table 1. Summary of snout–vent length and body-mass data for the different sex and size classes of Thamnophis sauritus (Eastern Ribbon Snake). Data were collected from May– November 2001. Values given are means and standard errors with ranges in parentheses. Young-of-the-year Juvenile Adult Females (n = 52) Sample size (n) 8 13 31 Snout–vent length (mm) 168.1 ± 8.9 328.1 ± 15.7 419.5 ± 6.5 (140.0–210.0) (230.0–415.0) (340.0–490.0) Body mass (g) 4.6 ± 0.3 14.6 ± 1.6 31.7 ± 1.6 (3.5–6.0) (7.0–26.5) (19.5–55.5) Males (n = 53) Sample size (n) 1 31 21 Snout–vent length (mm) 150.0 ± 0.0 310.2 ± 9.3 410.7 ± 7.5 (150) (180.0–370.0) (355.0–500.0) Body mass (g) 4.5 ± 0.0 14.0 ± 0.8 25.6 ± 1.1 (4.5) (6.0–22.0) (16.5–37.5) 284 Northeastern Naturalist Vol. 14, No. 2 From June to September, all snakes were found within 5 m of the water. For both telemetered and non-telemetered individuals, summer activity ranges on land rarely exceeded 5 x 10 m. Recapture rates for nontelemetered snakes were understandably lower; however data from those snakes indicate slightly greater movement amongst non-gravid females and males. On 14 occasions, snakes were observed swimming in open water. None of these was disturbed and all appeared to swim slowly-close to the surface, with their heads raised slightly above the water, taking minutes to cross the lake; (a distance of 38 to 71 m in 2001). On one occasion, a disturbed individual that was basking on the shoreline fled into the water and swam quickly across the lake. Of the seven radio-telemetered snakes that were tracked during fall, four crossed the open water while the other three were found only on one side. From September to mid-October, the distance traveled by the snakes between sightings increased as the snakes moved away from the shoreline. The Figure 1. Distribution of Thamnophis sauritus (Eastern Ribbon Snake) at the Grafton Lake study area, in Kejimkujik National Park, NS, Canada. The black lines represent roads. The shaded grey area represents water. The small rectangle is a building on the site. The circles indicate the first capture of each snake (n = 105). Eighty percent of the study area is shown to the left of the roads. The map shows the tight distribution of the snakes along the water’s edge. 2007 S.L.M. Bell, T.B. Herman, and R.J. Wassersug 285 maximum distance from the shoreline any snake was sighted was 173 m; the second greatest was 145 m. Despite the distances from the shore, both sightings were on the flood plain of the lake and only a meter or so above the lake level. One telemetered individual crossed the open water at least twice in the fall, and was last located on 16 September 2001 under a gravel walkway > 1 m from the water’s edge. The snake’s body temperature had dropped from 30.5 ºC to 16.4 ºC, and then remained relatively constant until 8 November 2001, when the last transmitter signal was acquired. The gravel walkway was 1.93 m wide and elevated 0.30 m above the surrounding ground. Slate rocks were embedded in various sections of the walkway, with several cracks and crevices evident. The snake’s location was only 0.6 m from the water on one side of the walkway and 2 m from the water on the other side. The transmitter signal from this snake in November and the drop in body temperature suggest that the snake was hibernating, but its depth below ground (or water) was not determined. Basking and related behaviours From June through August, ribbon snakes were often found basking in the mud along the shoreline of Grafton Lake, or floating in the water among aquatic vegetation. On four occasions, they were observed in the water diving and remaining submerged for periods > 3 min, with no evidence that they had been startled. On land, away from the water’s edge, snakes typically hid under cover with only their heads exposed. On seven occasions, we observed snakes that were largely exposed, with their heads elevated and swaying laterally with the approximate frequency that the wind was oscillating overhead vegetation, suggesting behavioral mimicry to enhance crypsis. Only four ribbon snakes were ever observed during precipitation. Telemetered individuals were usually located by radio signal, but actual sightings rarely occurred. Generally, snakes were more conspicuous at higher temperatures (􀂧 25–30 ºC) and immediately following periods of precipitation. Habitat and diet Ribbon snakes concentrated activity at three locations within the study area, all within 􀂧 30 m of each other. One site contained two small pools (􀂧 7 m x 5 m, 􀂧 5 m x 6 m) surrounded by tall, dense cordgrass. Another site (􀂧 16 m x 17 m) was in a shallow area of the cove with scattered, large, slate outcrops. The third site contained a small inflow (􀂧 23 m long) surrounded by abundant slate rubble. Fluctuations in water levels seemed to promote movement of snakes among the three sites. When one site dried up, snakes moved to one of the other two sites. When all three sites flooded, snakes moved to different ones. 286 Northeastern Naturalist Vol. 14, No. 2 Feeding was observed on 11 occasions, all at temporary pools on the margins of the lake. All observations were made in the afternoon between 1300–1600 h ADT, except for one observation which was made at 􀂧 1900 h ADT. All feeding observations were in July and August. Four snakes ate anurans (Ranidae), while seven snakes ate fish (Cyprinidae). Snakes did not move until prey were within striking distance. All snakes foraged in the water, but returned to land to ingest their prey. For 8 of the 11 (73%) feeding observations, snakes were initially on the edge of the shoreline, and for the other three (27%) feeding observations, snakes were in the water. Occasionally, Eastern Ribbon Snakes were found within close proximity (< 1 m) to one another; however, no interactions were observed. Discussion Capture patterns and abundance During summer, telemetered snakes were relocated at the same locality more frequently than in the fall; i.e., they remained in smaller areas and moved shorter distances during the summer months. Similarly, recaptures of non-telemetered snakes showed a high degree of site fidelity during the summer. However, some individuals still demonstrated erratic movements. The two population estimates suggested that 􀂧 50% of the snakes (non- YOY) in the sampling area were never seen, despite frequent sampling and apparent site fidelity. Clearly, low detectability contributes to the low recapture frequency in this population. Despite possible bias from unequal detectability, estimated summer density in the local population (5.6–5.9 snakes/ha) is surprisingly high, relative to other anecdotal observations from the Nova Scotia population. Reproduction and mortality Mating of Eastern Ribbon Snakes has not been described. Many species of Thamnophis are able to mate in both spring and fall, although there is usually one primary mating period (Gregory 1974).Gravid Eastern Ribbon Snakes were observed from June through September. However, since fieldwork did not commence until late May and few individuals were initially observed, it is unclear whether females were gravid before late May. Young-of-the-year were first observed on 7 August 2001. This is consistent with Carpenter’s (1952) findings in Michigan, where most Eastern Ribbon Snake births occurred in July and August, with a few broods produced in September and October. One gravid female was recaptured on 27 August 2001 with a failed transmitter. Prior to transmitter removal, the snake died in captivity on 28 August 2001 of unknown causes. The snake was emaciated and had lost 2007 S.L.M. Bell, T.B. Herman, and R.J. Wassersug 287 approximately 13.5 g since the time of first capture. Consistent with this finding, Gregory and Skebo (1998) and Gregory et al. (1999) found that gravid Thamnophis elegans (Baird and Girard) (Western Terrestrial Garter Snakes) in British Columbia fed less than nongravid snakes and appeared emaciated in late gestation and postpartum. Activity Ribbon snakes were rarely observed during precipitation, although activity usually increased immediately following it. This may reflect an increase in activity and abundance of prey, particularly amphibians, following precipitation. Because telemetered snakes were also not active during precipitation, it is likely that decreased visibility was not responsible for this finding. Distribution and movement During the summer, our snakes showed a strongly linear distribution precisely along the shoreline of Grafton Lake, regardless of size or sex, further suggesting the importance of proximity to water. Scribner and Weatherhead (1995) reported that 93% of Eastern Ribbon Snakes found in eastern Ontario were associated with aquatic habitats. Other studies have found that Thamnophis species preferred sites with high levels of cover, in close proximity to water (Carpenter 1952, Charland and Gregory 1995, Tinkle 1957). Eastern Ribbon Snakes, particularly gravid females, in this study were also sometimes associated with slate rocks and old tree stumps. Similarly, Charland and Gregory (1995) found that gravid Western Terrestrial Garter Snakes prefer rocky sites. All telemetered snakes in summer were gravid females, which may have contributed to their limited movements. Although recapture rates of nontelemetered snakes were low, both non-gravid females and males showed greater movement than gravid females. Charland and Gregory (1995) and Seigel et al. (1987) also documented reduced movement in gravid Thamnophis marcianus (Baird and Girard) (Checkered Garter Snakes) compared to non-gravid females. In contrast, Tinkle (1957) found that young Thamnophis proximus (Say) (Western Ribbon Snake) females (i.e., nongravid) moved less than other sex and age classes. The increased movement that we observed from September to November may in part reflect increased post-partum activity, but it also suggests that individuals were leaving summer activity sites. From September to mid- October, three of the snakes moved away from the shoreline, but remained on the flood plain of Grafton Lake in low-lying areas that would likely be submerged after snow melt the following April. The Eastern Ribbon Snake appears to be sedentary for most of the year, which is in contrast to northern populations of T. sirtalis L. (Common Garter Snake), which may migrate several kilometers annually 288 Northeastern Naturalist Vol. 14, No. 2 (Larsen 1987). However, since the majority of telemetered snakes in fall were not relocated, long seasonal migrations in this population cannot be ruled out. Emergence and over-wintering Historical records for KNP document Eastern Ribbon Snake sightings as early as 8 April and as late as 19 October. We observed the first snake on 30 May 2001, approximately one week after extensive fieldwork began. It is likely that some snakes emerged from hibernacula in May, since some individuals were muddy at the time of first capture. Populations of Common Garter Snake in northern Alberta—the northern limit for the genus—hibernate until mid-April (Larsen et al. 1993). Our latest sighting of an unmarked individual, an adult male, was 29 September 2001. No sightings were made during late October and November, suggesting that by this time the snakes had moved away from summer activity sites. Gregory (1974) documented hibernation periods of six months in populations of Common Garter Snakes in Manitoba, Canada. Based on observations from our study, hibernation for Eastern Ribbon Snakes appears to last six to seven months. Although most snakes that were located late in the season were moving away from the water, they did not move uphill and thus may be using submerged hibernation sites. Costanzo (1989) suggested that, for garter snakes, semi-submerged hibernation sites would allow snakes to maintain water balance through the winter and avoid dehydration. Carpenter (1953), in fact, showed that Common Garter Snakes in Michigan can hibernate completely submerged. This ability may also be true for Eastern Ribbon Snakes. Since only one hibernaculum was identified, and only one snake was located at the site, it is unclear whether communal hibernacula are used by T. sauritus as observed with other species of Thamnophis (Gregory 1974, Larsen and Gregory 1989, Macmillan 1995). Diet Eleven snakes were observed feeding on anurans and fish. Only amphibians were found in the diet of a Michigan population of Eastern Ribbon Snakes (Rowe et al. 2000); however, both amphibians and fish have been found in the diets of other species of Thamnophis (Arnold 1992, Carpenter 1952, Matthews et al. 2002, Tinkle 1957). Rowe et al. (2000) suggested that this difference may be due to differences in prey availability at different localities. However, there may be differences as well in both dietary preference and microhabitat use for this species across its range. For example, Rowe et al. (2001) found some ribbon snakes in adjacent forested areas, whereas none of our snakes were ever found in woodland areas, where basking opportunities would not exist. De Queiroz et al. (2001) noted an ontogenetic shift in the diet of Thamnophis validus 2007 S.L.M. Bell, T.B. Herman, and R.J. Wassersug 289 Kennicott (Mexican Pacific Lowlands Garter Snakes) from primarily anurans to primarily fish. Despite its sleek, thin body, this species appears to be relatively sedentary like more heavily bodied snakes, which rarely move quickly unless startled. Our feeding observations suggest that Eastern Ribbon Snakes in Nova Scotia is primarily an ambush predator rather than a “widely foraging” species such as slim-bodied snakes in warmer climates (Greene 1997). The Eastern Ribbon Snakes in Nova Scotia were slightly smaller than those from more southern populations. This may reflect a shorter growing season and more difficulty in finding adequate and abundant prey. We know no studies of growth rate of ribbon snakes at the temperatures that they experience in Nova Scotia when food is optimally available. Fluctuations in Eastern Ribbon Snake abundance are likely influenced by several factors, including prey availability, temperature, and rainfall. Kephart and Arnold (1982) found that Western Terrestrial Garter Snakes and Common Garter Snakes were opportunistic feeders, whose diets changed with fluctuating lake levels and abundance of prey. In our study, all feeding observations were at small, temporary pools, marginal to the main lake; this suggests that Eastern Ribbon Snakes may not actually feed within larger bodies of water, but instead depends on smaller, temporary pools. These small pools may concentrate prey as they dry up, but they are also at risk of drying completely. The feeding ecology of this snake suggests that small fluctuations in water levels could seriously affect the life span of the pools and thus the availability of prey. Conservation issues Historical records from KNP (38,100 ha) document only 118 verified sightings of Eastern Ribbon Snakes between 1970 and 2000. However, in a single year at our study site, we marked 105 snakes in 33.5 ha. This may reflect increased sampling effort, increased population due to habitat change associated with breaching of the dam, or some combination of the two. Clearly the species is locally abundant at this site, which raises several questions: 1) What is responsible for the local abundance? 2) How ephemeral is it? and 3) What risks does the population face? The adjacency of numerous shallow, ephemeral, food-rich summertime pools around the margin of the lake, plus deeply fractured slate that permits easy descent to hibernacula in the fall, may promote snake abundance at this site. Since the site is only recently emerged (due to reduced lake levels following dam breaching), it is unclear how ephemeral this high density might be. The impact of disturbance on this population is unknown. The site experiences considerable fluctuations in temperature and water levels, and is easily accessible to humans; any of these factors could potentially degrade the habitat. 290 Northeastern Naturalist Vol. 14, No. 2 Compared to other Thamnophis, the Eastern Ribbon Snake is more aquatic in both diet and movements. It also appears to hibernate closer to its feeding grounds and have limited seasonal movements. This underscores the importance of understanding water-level dynamics in wetlands and the influence of disturbance by humans and other species, especially beavers, on those dynamics. Disturbance may actually play an important role in creating Eastern Ribbon Snake habitat, but the role is poorly understood. The high population of T. sauritus at Grafton Lake appears exceptional. Limited search efforts for additional populations have failed to locate similar aggregations elsewhere in the region. Many elements of Eastern Ribbon Snakes in Nova Scotia are consistent with the generalized profile of threatened species. These include small geographic range, small population size, small home range, strong site fidelity, sit-and-wait predation, and heavy dependence on small, ephemeral pools as feeding sites. All told, the Eastern Ribbon Snake at the northern limit of its range appears to have more restricted movements and feeding limited to very specific microhabitats compared to populations in warmer climates. This more restricted habitat use makes the species more vulnerable to local extinction. Acknowledgments We thank K. Oseen for her comments on earlier versions of this manuscript. J. Gilhen shared his extensive knowledge on both the species and the localities. We are grateful to the staff of Kejimkujik National Park for providing research permits and lodging. We thank Dr. C. Harvey-Clark for his veterinary services. Drs. A. Savitzky, C. Peterson, and H. Lilywhite provided valuable information on radio-telemetry. H. McCracken kindly shared detailed topographic information. This work was conducted under a Nova Scotia Department of Natural Resources Wildlife Division Scientific Permit, a Mainland Nova Scotia Field Unit Research Permit (KEJ-2001-02), and a Dalhousie University Committee on Laboratory Animals Animal Utilization Protocol (01-028). This research was funded by the National Science and Engineering Research Council of Canada (NSERC) and the Nova Scotia Museum of Natural History. Additional funding was provided by the Environment Canada Habitat Stewardship Program for Species at Risk (2001), Human Resources Development Canada: Summer Career Placements Program (2001), and from N. and D. Bell. Literature Cited Arnold, S.J. 1992. Behavioural variation in natural populations. VI: Prey responses by two species of garter snakes in three regions of sympatry. Animal Behaviour 44:705–719. Brown, W.S., and W.S. Parker. 1976. A ventral-scale clipping system for permanently marking snakes (Reptilia, Serpentes). Journal of Herpetology 10:247–249. 2007 S.L.M. Bell, T.B. Herman, and R.J. Wassersug 291 Carpenter, C.C. 1952. Comparative ecology of the Common Garter Snake (Thamnophis s. sirtalis), the Ribbon Snake (Thamnophis s. sauritus), and Butler’s Garter Snake (Thamnophis butleri) in mixed populations. Ecological Monographs 22:235–258. Carpenter, C.C. 1953. A study of hibernacula and hibernating associations of snakes and amphibians in Michigan. Ecology 34:74–80. Caughley, G. 1977. Analysis of Vertebrate Populations. Wiley-Interscience Publication, John Wiley and Sons, London,UK. 243 pp. Chao, A. 1988. Estimating animal abundance with capture-frequency data. Journal of Wildlife Management 52:295–300. Charland, M.B., and P.T. Gregory. 1995. Movements and habitat use in gravid and nongravid female garter snakes. (Colubridae: Thamnophis). Journal of Zoology, London 236:543–561. Clark, D.R. 1974. The Western Ribbon Snake (Thamnophis proximus): Ecology of a Texas population. Herpetologica 30:372–379. Committee On the Status of Endangered Wildlife In Canada (COSEWIC). 2002. Available online at http://www.cosewic.gc.ca. Accessed March 30, 2002. Costanzo, J.P. 1989. Effects of humidity, temperature, and submergence behaviour on survivorship and energy use in hibernating Garter Snakes, Thamnophis sirtalis. Canadian Journal of Zoology 67:2486–2492. De Queiroz, A., C. Henke, and H.M. Smith. 2001. Geographic variation and ontogenetic change in the diet of the Mexican Pacific Lowlands Garter Snake, Thamnophis validus. Copeia 2001:1034–1042. Gilhen, J. 1984. Amphibians and Reptiles of Nova Scotia. Nova Scotia Museum, Halifax, NS, Canada. 162 pp. Greene, H.W. 1997. Snakes: The Evolution of Mystery in Nature. University of California Press, Berkeley, CA.. 351 pp. Gregory, P.T. 1974. Patterns of spring emergence of the Red-sided Garter Snake (Thamnophis sirtalis parietalis) in the Interlake region of Manitoba. Canadian Journal of Zoology 52:1063–1069. Gregory, P.T., and K.W. Larsen. 1993. Geographic variation in reproductive characteristics among Canadian populations of the Common Garter Snake (Thamnophis sirtalis). Copeia 1993:946–958. Gregory, P.T., and K.M. Skebo. 1998. Trade-offs between reproductive traits and the influence of food intake during pregnancy in the garter snake, Thamnophis elegans. American Naturalist 151:477–486. Gregory, P.T., L.H. Crampton, and K.M. Skebo. 1999. Conflicts and interactions among reproduction, thermoregulation, and feeding in viviparous reptiles: Are gravid snakes anorexic? Journal of Zoology, London 248:231–241. Kephart, D.G., and S.J. Arnold. 1982. Garter snakes in a fluctuating environment: A seven-year study. Ecology 63:1232–1236. Larsen, K.W. 1987. Movements and behaviour of migratory Garter Snakes, Thamnophis sirtalis. Canadian Journal of Zoology 65:2241–2247. Larsen, K.W., and P.T. Gregory. 1989. Population size and survivorship of the Common Garter Snake, Thamnophis sirtalis, near the northern limit of its distribution. Holarctic Ecology 12:81–86. 292 Northeastern Naturalist Vol. 14, No. 2 Larsen, K.W., P.T. Gregory, and R. Antoniak. 1993. Reproductive ecology of the Common Garter Snake Thamnophis sirtalis at the northern limit of its range. American Midland Naturalist 129:336–345. Macmillan, S. 1995. Restoration of an extirpated Red-sided Garter Snake Thamnophis sirtalis parietalis population in the Interlake region of Manitoba, Canada. Biological Conservation 72:13–16. Matthews, K.R., R.A. Knapp, and K.L. Pope. 2002. Garter snake distributions in high-elevation aquatic ecosystems: Is there a link with declining amphibian populations and nonnative trout introductions. Journal of Herpetology 36:16–22. Reinert, H.K., and D. Cundall. 1982. An improved surgical implantation method for radio-tracking snakes. Copeia 1982:702–705. Rowe, J.W., K.C. Campbell, and J.C. Gillingham. 2000. Diet of the ribbon snake on Beaver Island, Michigan: Temporal variation and the relationship of prey size to predator size. Herpetological Natural History 7:145–152. Scribner, S.J., and P.J. Weatherhead. 1995. Locomotion and antipredator behaviour in three species of semi-aquatic snakes. Canadian Journal of Zoology 73:321–329. Seigel, R.A., M.M. Huggins, and N.B. Ford. 1987. Reduction in locomotor ability as a cost of reproduction in gravid snakes. Oecologica, Berlin 73:481–485. Tinkle, D.W. 1957. Ecology, maturation, and reproduction of Thamnophis sauritus proximus. Ecology 38:69–76.