RESULTS

Snake Assemblage and Kingsnake Captures  From 1990-1996, we caught 730 snakes (including recaptures), representing 12 different species.  Ninety-two percent of all captures (98% of kingsnake captures) came from under coverboards.  Lampropeltis getula nigra (n = 400) accounted for 55% of the captures followed by Storeria dekayi (n = 151), Coluber constrictor (n = 67), and Elaphe obsoleta (n = 48), which together comprised 36% of the total captures. Less than 9% consisted of other species:  Diadophis punctatus (n = 29), Agkistrodon contortrix (n = 8), Opheodrys aestivus (n = 8) Thamnophis sirtalis (n = 8), Storeria occipitomaculata (n = 5), Carphophis amoenus (n = 4), E. guttata (n = 1) and Virginia valeriae (n = 1).

We captured 180 different L. g. nigra a total of 400 times (55% recapture rate).  The number of kingsnakes captured ranged from 28 individuals in 1996 to 41 in 1992. Neither total nor individual captures differed significantly among years (total: X² = 10.7, df = 6, P > 0.05; individual: X² = 3.7, df = 6, P > 0.5).  The minimum number alive was calculated for 1992-1996: 1992 = 43, 1993 = 37, 1994 = 33, 1995 = 37, 1996 = 38.  Because search effort was not the same in each month, captures per survey hour were compared to the percentage of total captures (Fig. 1). When all years were combined, both of these analyses showed a bimodal trend peaking in May and September.

Mean T_a at time of capture was 22.6^oC (range = 11 - 32^oC, SD = 3.7, n = 334) and mean coverboard substrate temperature was 23.3^oC (range = 11 - 36^oC, SD = 3.7, n = 325). Kingsnakes were never found under more than 4% of the coverboards on a given search day.  There was no difference in the frequency of encounters under wood versus metal (X² = 1.95, df = 1, P = 0.16, n = 213), but significantly more were captured in the afternoon (mean T_a = 23.3^oC)  than in the morning  (mean T_a = 21.2^oC; X² = 4.7, df = 1, P < 0.05, n = 249).

From 1992 - 1996, 61 males and 55 females were captured a total of 296 times with no significant difference in the frequency of recaptures between sexes (X² = 0.8, df = 1, P > 0.30).  Sex ratios did not differ significantly for any one year (see Table 1) or for individuals > 60cm that were captured in the spring (April and May; X² = 0.4, df = 1, P > 0.50, n = 42).  Thirty-two individuals (mean SVL = 66 cm, SD = 22, sex ratio = 17M:15 F) were found in more than one year, with six of these caught in three different years (mean SVL = 63 cm, SD = 11, sex ratio = 3:3) and four found in four different years (mean SVL = 55 cm, SD = 18, sex ratio = 2:2).

Size & Growth  Captured L. g. nigra ranged from 25 - 112 centimeters SVL (mean = 66 cm, SD = 24, n =171). The mean SVL of only one year (1990) differed significantly (One-way ANOVA, F_6,164 = 4.13, P < 0.001, n = 171) from other years (1994 and 1995).  Mean SVLs of males and females did not differ significantly (One-way ANOVA, F_1,117 = 0.0005, P = 0.98, n = 119), but males > 60 cm had longer tails than females with equivalent SVLs (ANCOVA with SVL as covariate, F_1,90 = 27.9, P < 0.0001, n = 93). Males and females (SVL > 60 cm ) had tail length/total length ranges of 11 - 15% and 8.7 - 13.2% respectively.

Hatchlings (n = 13) from two different females that laid eggs in captivity had a mean SVL of 22.7 cm (range = 22 - 24) and a mean mass of  6.1 g (range = 5.3 - 6.7). Young-of-the-year captured in August, September, and October had a mean SVL of 27 cm (n = 11, range = 25 - 29) and a mean mass of 10 g (n = 10, range = 8.4 - 12.2). Estimates of SVL ranges of young-of-the-year through fifth year snakes are shown in Fig. 2. 

Individuals ranged in mass from 4.6 - 521 g (mean = 156 g , SD = 142, n = 171).  A regression equation that best fit the relationship between mass (y) and SVL (x) was an exponential one of the form y = ax^b : mass (g) = 0.0005 SVL(cm)^2.95 (r = .90, df = 169, P < 0.0001, n = 171; Fig. 3). The difference in mass between males and females with equivalent SVLs was not significant (ANCOVA with SVL as covariate, F_1,116 = .46, P = 0.50, n = 119). This was also true for snakes > 60 cm SVL (F_{1, 90} = 2.76, P = 0.10, n = 93) and for males and females captured in spring and early summer (April, May, and June; F_{1, 56} = 1.49, P = 0.23, n = 59) that were > 60 cm and had equivalent SVLs.

Based on 41 recaptures of 27 individuals with a minimum of 2.5 (mean = 7.9) growing months between captures, mean growth rate was 1.4 cm per month (initial SVL and increases per month are given in Fig. 4).  Because of the small number of recaptures of first-year snakes (n = 2) as compared to other age categories, SVL differences between first-year individuals captured in the spring (n = 6) and different individuals captured in the fall (n = 3) were also calculated. The mean growing interval was 3.5 months and the mean SVL increase (2.6 cm/month) was the same as that calculated for the two recaptured individuals.  Snakes < 60 cm had a mean SVL increase of 2.4 cm per month (n = 13, SD = 0.75, range = 1.4 - 4.2). Those > 60 cm averaged 0.9 cm per month (n = 28, SD = 0.95, range = 0 - 4.6)

Diet  Six L. g. nigra disgorged food items in holding bags shortly after capture.  All snakes were captured under coverboards that had a mean substrate temperature of 25^oC (range = 21-31^oC, SD = 3).  Five of the seven prey items came from snakes with capture times between 1700 and 1900 h.  The weight ratios (prey mass in relation to snake mass) ranged from 8.4% to 31%. 

Three subadults (SVL = < 60 cm) disgorged prey items representing 3 species of snakes and one mammal.  A first year male (SVL = 30.4 cm) captured on 5 June 1995 disgorged a  S. dekayi (TL = 15.5 cm) and a C. amoenus (TL = 24 cm; Byrd and Jenkins 1996).   The second subadult (female, SVL = 57 cm) disgorged a short-tailed shrew (Blarina brevicauda).  For a third individual (male, SVL = 58.5 cm) we recorded two different regurgitation events in 1996.  On 14 June we recovered the partial remains (2 g) of a N. sipedon and an intact C. amoenus (5.8 g, TL = 28.8 cm).  On 19 September this snake (SVL = 72 cm) was recaptured under the same coverboard and disgorged a 13.8 g pine vole  (Microtus pinetorum).  Each of the adults (SVL range = 68.3 - 81 cm): two females, a male, and the recaptured male mentioned above, regurgitated a M. pinetorum (range = 10.1 - 19.6 g). 

In 1996, two snakes being radiotracked were observed feeding during mid-afternoon. Snake 1, a 112 cm (SVL) male, monitored with a temperature-sensing transmitter, was observed on 15 July (1315 to 1345 h) entering a Terrapene carolina nest and ingesting two eggs (ground surface temperatures = 26.4^oC to 28^oC, T_bs = 28^oC to 29^oC). Snake 2, a 74 cm (SVL) male (mass = 150.4 g) was observed on 16 July (1402 to 1420 h) attempting to eat a female L. g. nigra (TL = 60 cm, mass = 55.8 g).  The larger snake had 4-5 coils around the smaller snake and the snakes were writhing around on the ground.  During the observation the larger snake started to flee at which time both individuals were captured.  The T_a was 31^oC and ground surface was 28.5^oC.

Cloacal Autohemorrhaging  Captured L. g. nigra often used behaviors associated with anti-predator mechanisms, including tail vibration, tail thrashing, and cloacal discharge. Tail thrashing frequently resulted in the tail being brought over the back of the restraining hand where feces and glandular products from the cloaca were smeared. Individuals that autohemorrhaged protruded a bright red cloaca and supplemented the cloacal contents with variable amounts of blood. All snakes that demonstrated this response were more than a year old. We observed a total of 53 hemorrhaging events involving 28 individuals (8 males and 20 females) with males ranging from 65 - 96 cm (SVL) and females ranging from 44 - 91 cm (SVL). Of the individuals identified as autohemorrhagers, 5 males and 11 females were captured two or more times. Only a single male had more than one hemorrhaging event, but there were 7 females, captured a total of 52 times, for which 31 hemorrhaging events were recorded. 

Home Ranges and Movements Twelve individuals (8 M, 4 F, mean SVL = 82.3 cm, SD = 11.5, range = 67 - 106 cm) were monitored with implanted transmitters.  Snakes were in the field an average of 85 days (range = 36 - 137) and were radiotracked » 4 - 6 times per week for a total of 849 locations (see Table 2). The mean home range  was 1.97 ha (SD = 1.47, range = 0.13 - 4.79 ha) and the mean distance per move was 56 meters (SD = 54). Neither SVL nor mass correlated with home range size (n = 14, SVL:  r = 0.05, P = 0.88; mass:  r = 0.06, P = 0.83).  Male SVLs did not differ significantly from female SVLs (student t-test; t = 1.7, df = 12, P = 0.11) but male home ranges (mean = 2.48 ha, SD = 1.45) were significantly larger than those of females (mean = 0.70 ha, SD = 0.31; student t-test; t = 2.3, df = 12, P = 0.04).  Males also made significantly longer moves (mean = 60 m, SD = 57) than females (mean = 39 m, SD = 36; student t-test; t = 3.3, df = 419, P = 0.001).

On 338 of 796 (42%) tracking days, individuals made no apparent move. Snakes returned to a previous location 109 times out of 421 recorded moves.  The mean distance moved between returns was 269 m (range = 7 - 1,585 m) and time between returns ranged from 1 - 61 days (mean = 12 days).  Eleven of 12 snakes returned to a previous location at least once.  The most returns was by a male (SVL = 89 cm) that returned to the same location 8 times.  The furthest distance it was found from that location was 216 m.  Two males were tracked for two consecutive years.  Male 1 (see Table 2) had approximately the same home range both years.  The home range for Male 2 was 1.46 ha in 1995.  In 1996, this individual centered its activities in the same area, but the home range size increased to 3.40 ha. 

Cloacal Temperatures and Biotelemetry  From 16 May - 29 September 1996, T_cs were recorded for 22 different L. g. nigra (12 M:10 F). All snakes were captured under coverboards. Based on the paired t-test, T_cs (mean = 26.9, SD = 4) were significantly higher than T_as (mean = 24.2, SD = 3.3; t = 4.2, df = 21, P = 0.0002) and coverboard substrate temperatures (mean = 24.7, SD = 2.9; t = 4.3, df = 21, P = 0.0002). Substrate temperatures were also significantly higher than T_as (t = 2.3, df = 21, P = 0.016). Male and female T_cs did not differ (student t-test; t = 0.85, df = 20, P = 0.4). 

From 25 June - 20 July 1996, using temperature-sensing transmitters, we recorded 50 body temperatures (T_bs) for Snake 1 (male, 96 cm SVL, 292 g) and 37 for Snake 2 (male, 112 cm SVL, 485.5 g).  There was no difference between mean T_a (25.8°C, range = 11.5 - 36.9°C) and mean T_b (25.7°C, range = 18.5 - 35°C, paired t-test; t = 0.20, df = 86, P = 0.83), but on sunny days, the mean T_b (26.1°C) was lower than the mean T_a (28.7°C; paired t-test; t = 5.5, df = 52, P < 0.0001).  Thick vegetation made it difficult to determine whether snakes were above or below ground. 

From 3 - 4 July, temperatures were taken approximately every two hours for a 26-hour cycle (Fig. 5); during this time, both snakes remained in a fixed area. Snake 1 stayed underground, just inside a clump of small Virginia pines (P. virginiana) at the top of the landfill.  Snake 2 was located off the landfill, mostly underground, around a limestone outcrop in a field characterized by scattered groups of cedars (Juniperus virginiana) and patches of brambles (Rubus spp.). Landfill T_as ranged from 12.7 - 32.8^oC (mean = 21.3^oC, SD = 6.5, n = 13). The T_b of Snake 1 ranged from 22 - 26^oC (mean = 24.0^oC, SD = 1.3, n = 13). The T_a in Snake 2's area ranged from 11.5 - 29.5^oC (mean = 20.3^oC, SD = 6.3, n = 13). The T_b of Snake 2 ranged from 23.5 - 31^oC (mean = 26.4^oC, SD = 2.4, n = 13).

DISCUSSION

Snake Assemblage and Kingsnake Captures  Of twelve species of snakes found at our site, L. g. nigra captures made up more than half of our total captures.  The sampling biases associated with the use of coverboards in studying snake assemblage are uncertain.  However, coverboards do not appear to be a major influence on the density and high proportion of kingsnakes found in our study.  Neither total nor individual captures of L. g. nigra differed significantly between years; furthermore, calculated minimum number alive was the same for 1993 and 1995 and only differed by one in 1996.  For the past three years we have been using coverboards in old-field habitat at another site, the University of Tennessee Arboretum (UTA) in Anderson County, TN (35° 60'N, 84° 13'W), located » 6 km to the W of our study area.  Both individual and total encounters of E. guttata, E. obsoleta and C. constrictor have substantially out-numbered L. g. nigra encounters at the UTA site (unpubl. data).

Two other studies (Johnson 1964 and J. A. Klein, pers. comm. 1995) conducted within the Valley and Ridge province of Anderson County, TN, found C. constrictor (n = 85) to be almost twice as common as L. g. nigra (n = 47).  Johnson found 6 other species, including E. obsoleta and E. guttata, to be more common than L. g. nigra; furthermore,  E. guttata, A. contortrix, T. sirtalis, and D. punctatus use habitats similar to L. g. nigra, but these species composed only 6% of our total captures, while composing 38% and 26% in Johnson's and Klein's studies.  Although there were differences between capture methods (Johnson and Klein used additional collecting techniques) and the overall types of habitats surveyed in our study and those conducted by Johnson and Klein, all of the studies made use of coverboards placed in habitats known to be used by L. getula.  Some of the discrepancies could, in part, be explained by the high density of L. g. nigra at our site.  Presumably L. g. nigra has overlapping trophic niches with several of the above species (example: E. guttata and A. contortrix) and all are suitable prey for L. g. nigra. Because of the large number of coverboards available to snakes in our study and the fact that we never found more than 4% occupied by L. g. nigra  at any one time; there was ample opportunity for other species to utilize coverboards. 

Grant, et al. (1992) conducted a coverboard study from 1988 to 1990 at the Savannah River Ecology Laboratory (SREL) near Aiken, South Carolina. Coluber constrictor (n = 50), E. guttata (n = 5), and E. obsoleta (n = 1) were found under coverboards in the SREL study; interestingly, out of 227 encounters of snakes under coverboards, no L. getula were recorded. Two of the sites surveyed were near Carolina bays.  Gibbons (1977) lists L. getula as common at the SREL and often found around permanent or temporary aquatic areas.  An intriguing question is why coverboards were attractive to L. g. nigra at our study site but failed to produce any encounters for this species in the study by Grant et al. (1992).  In the future it will be important for researchers using coverboards to standardize methods and to identify physical and biological factors that could explain different results. 

The activity pattern based on encounters of L. g. nigra under coverboards was a bimodal curve peaking in May and September. Gibbons and Semlitsch (1982) reported a unimodal trend, peaking in June, for L. getula at the Savannah River Site.  The difference is probably more reflective of collecting techniques (drift fences vs. cover boards) than actual differences in activity patterns.  Captures of L. g. nigra under coverboards were normally low in July and August, but radiotracked snakes were above ground and moving in these months, suggesting that coverboards produced a seasonal sampling bias.

Radiotracking data, feeding observations, and freshly disgorged prey items by individuals captured in the late afternoon support a diurnal activity pattern for L. g. nigra.  Recent intensive tracking of an individual at the UTA site predominantly showed a midday to early afternoon activity pattern for L. g. nigra (T. J. Thomasson, unpubl. data). These data agree with what other researchers have reported for L. getula (Conant 1975, Gibbons 1977, Mitchell 1994). 

Our data indicate a relatively stable population between study years.  Capture rates and mean SVLs were similar for all years.  The overall sex ratio of L. g. nigra did not differ from 1:1 and there was no significant difference in the number of males and females captured in any one year.  Parker and Plummer (1987) cite at least 7 studies that point to males being more common in spring samples because of increased sexual activity.  We found equal numbers of males and females in the spring.  In addition, snakes recaptured across years had equivalent sex ratios, suggesting that males and females had similar survival rates.

Size and Growth  We found no published data from a single population on size differences between male and female L. g. nigra.  Mitchell (1994) reported that adult male L. g. nigra in Virginia were larger than adult females.  Fitch (1978) found L. calligaster males to be larger than females.  Male and female L. g. nigra in our study were not significantly different in size.  Furthermore, the largest male (SVL = 112cm) in our study was substantially smaller than the largest male (SVL = 145.6cm) reported by Mitchell (1994).  We also found that male kingsnakes > 60 cm SVL had longer tails than females.  The ranges that we report for tail length/total length are similar to what Ernst and Barbour (1989) report for L. getula.

The mass - SVL relationship reported for L. getula (mass = 0.0004 (SVL) ^2.94 ) by Kaufman and Gibbons (1975) near Aiken, South Carolina was similar to L. g. nigra at our site (mass = 0.0005 SVL^2.95  ).  Although subject to sampling bias and temporal variations in prey availability, these data are potentially useful for making comparisons between local populations and for comparing individuals within a population.  For example, we used mass-SVL relationship as a health indicator for snakes being radiotracked. 

Our study is the first to report growth data in nature for L. g. nigra.  Hatchling SVL and mass ranges that we report for L. g. nigra fall within the ranges reported by Ernst and Barbour (1989) and Mitchell (1994).  The pattern of size increase was similar for L. calligaster in Kansas (Fitch 1978), but these first- and second-year individuals grew considerably faster than the L. g. nigra at our site.  Mitchell (1994) reported that size at maturity for L. g. nigra is »60 cm SVL for both sexes which agrees with what Fitch (1978) reported for sexual maturity in L. calligaster in Kansas.  We found that snakes in our population attain this size during the second or third year of growth.  There is also a noticeable increase in the mass/SVL ratio at » 60 cm. 

Cloacal Autohemorrhaging  The behavior of cloacal autohemorrhaging was one of the more interesting observations in this study. We confirmed only three other reports of this behavior in snakes. Lardie (1961) reported a bloody-looking fluid released from the vent of longnose snakes (Rhinocheilus lecontei) captured near Kern County, California, H. W. Greene (pers. comm. 1996) has observed the behavior in speckled kingsnakes (L. g. holbrooki), and J. E. Fauth (pers. comm. 1998) informed us that he observed hemorrhaging in a male L. g. getula found in Francis Marion National Forest in South Carolina.

This behavior could represent a stress response or a range of possible antipredator mechanisms, from startle effect to attracting a predator to the less critical tail area (see Greene 1988, for a discussion on antipredator mechanisms). Our experience suggests that the discharge could readily be smeared about the face of certain predators. Thus, tail thrashing and discharge may be enhanced by the bright red cloaca and blood, which may be novel to a predator relative to most members of the prey community (Gans and Richmond 1957; Humphries and Driver 1967, 1970). Autohemorrhaging was not observed in L. g. nigra that were less than a year old (SVL < 44 cm). The lack of effect on predators from small amounts of blood or the energetic cost of losing blood could make hemorrhaging an inappropriate response for smaller individuals. 

Colwell (1999) evaluated parasites found in cloacal discharges of L. g. nigra found at our site.  Hemorrhagers did not differ from non-hemorrhagers in their parasite profiles (S.T. Colwell, pers. comm. 1998).  Developing hypotheses on why more than 70% of the hemorrhagers were females will require a closer look at the internal structure of the cloaca and scent glands (at least a portion of the blood came from the scent glands) of males and females.

Home Ranges and Movements  Our study is the first to report home range and movements of L. g. nigra.  Most snakes were monitored for a reasonable length of time but time frames were not standardized, thus making it difficult to detect general patterns (Macartney, et al. 1988).  Mean home range of L. g. nigra (1.97 ha) was similar to ones summarized by Macartney, et al. (1988) for C. constrictor (2.4 ha and 1.45 ha), Pituophis melanoleucus (males 1.2 ha, females 2.1 ha) and T. sirtalis (females 1.4 ha). 

There was no correlation between kingsnake size and home range size, but male home ranges were significantly (3.5 times) larger than female home ranges and males averaged longer moves.  Weatherhead and Hoysak (1989) found that dry land home ranges of male E. obsoleta (mean = 3.9 ha) were significantly (> 3 times) larger than those of females (mean = 1.22 ha) and males averaged longer moves (69.3 m) than females (43.9 m).  Mullin et al. (2000) reported a mean home range of 5.6 ha for E. o. spiloides.  Male home range (mean = 6.3 ha) was larger than that of females (mean = 3.3 ha) but the home ranges were not analyzed as a function of sex.  Many earlier studies have not found significant differences in home range sizes or movements of male and female snakes (Gregory et al. 1987).  All the above studies used radiotelemetry but the variables of sample size, duration of study, and methodology make comparisons difficult.

Monitored snakes demonstrated varying degrees of site fidelity.  Some individuals traveled considerable distances over many days only to return to a previous location.  Two kingsnakes that were radiotracked for two years remained in the same area both years.  Although tracking occurred over different time periods, there was overlap of seasons in which they were tracked.  Male 2 had a two-fold increase in home range the second year.  This snake was tracked for fewer days but more intensively during the second year.  Male 2 was nearly the same length both years but was 100 g heavier the second year.  We presently have radiotelemetry data on 20 individuals.  Future analyses will concentrate on microhabitat, nocturnal movements, and degree of home range overlap.

Cloacal Temperatures and Biotelemetry  Bothner (1973), Brattstrom (1965), and Mitchell (1994) reported mean body temperatures for L. getula ranging from 28.1 to 28.7°C.  The mean T_c (26.9°C) for L. g. nigra found under our coverboards and mean T_bs (24.5 and 27.0°C) for the two biotelemetered individuals were lower than those reported by the above researchers.  The 26-hour temperature survey provided an opportunity to record body temperatures of two individuals under extreme conditions (11.5^oC was a record low for the study area in July). Body temperatures varied only 3.5^oC and 7^oC during a period in which the ambient air temperature varied 20^oC.  Our data is suggestive of relatively precise thermoregulation (see Lillywhite 1987) but additional studies are needed to evaluate the importance of thermoregulation in L. g. nigra. 

Community Structure  How the seemingly high density of L. g. nigra at our study site is influencing the overall snake assemblage is uncertain.  Radiotracked kingsnakes were mostly found in fields and the woodland-field ecotone.  Woodlands dominate our site (» 40 of 60 ha) but monitored snakes were only briefly associated with these areas when en route to new fields.  Radiotelemetry data on E. obsoleta (n = 5) and A. contortrix (n = 3) at our site showed that both species moved into forest habitat at times when L. g. nigra remained in fields.  This was also true for monitored  E. obsoleta (n = 1) and E. guttata (n = 4) at the UTA site (A.W. Heffern, pers. comm. 2000) where kingsnake density appeared to be low. The ghosts of competition past and present may be "haunting" our fields but as Toft (1985) points out, resource partitioning patterns in reptiles are a combination of forces including competition, predation and other factors that operate independently of interspecific interactions. 

Until such time that more comparative studies are completed, we can only hypothesize about conditions that favored kingsnakes at our study site.  Two interesting differences between the UTA site and the ACWS site are the soil conditions and small mammal density.  The ACWS site has a long history of disturbance.  The soils are loosely packed and riddled with small mammal tunnels.  We often had to dig up kingsnakes in order to remove transmitters.  The challenge was trying to capture them while they moved through small mammal burrows.  The soils in the landfill were especially easy to negotiate and it was not uncommon for individuals to travel considerable distances underground.  The soils at the UTA site are clayey and contain a high percent of chert rock.  Trapping studies by F. White and F. Holtzclaw (pers. comm. 1997) showed small mammal density (especially Microtus pinetorum) to be much lower at the UTA site than at our study site.  We encourage others to document the snake assemblages in similar habitat. Tinkle (1979) stressed long-term field studies as an essential step toward testing accepted ecological theory, and critical for making practical management decisions about  particular populations or communities. We hope the baseline information provided in this paper will stimulate further investigations of this little studied species.