Quantification of Color Pattern Variation in the Common Garter Snake (Thamnophis Sirtalis) from Two Coastal Islands and Mainland Massachusetts, USA

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/261473107

Quantification of color pattern variation in the common garter snake (Thamnophis sirtalis) from two coastal islands and mainland Massachusetts, USA.

ARTICLE in HERPETOLOGICAL REVIEW · MARCH 2014

3 AUTHORS, INCLUDING:

Scott Smyers Andrew Mckenna-Foster Oxbow Associates, Inc. Nantucket Maria Mitchell Association

11 PUBLICATIONS 37 CITATIONS 7 PUBLICATIONS 2 CITATIONS

SEE PROFILE SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Scott Smyers letting you access and read them immediately. Retrieved on: 22 March 2016 Herpetological Review, 2014, 45(1), 8–12. © 2014 by Society for the Study of Amphibians and Reptiles Quantification of Color Pattern Variation in the Common Gartersnake (Thamnophis sirtalis) from Two Coastal Islands and Mainland, Massachusetts, USA

The Common Gartersnake (Thamnophis sirtalis) inhabits partially stretched skin and scales on two specific regions of live a wide variety of ecosystems and has been extensively studied T. sirtalis. We then quantified the area of lateral blotches (black) throughout most of North America (Gregory and Larsen 1993; relative to scales and dorsal stripe completeness of T. sirtalis Rossman et al. 1996). Conspecific variation in both morphol- from Nantucket and mainland Massachusetts. ogy (Grudzien et al. 1992; King 1988) and color patterns (Kelley Study Sites and Methods.—We used high-resolution digital et al. 1997; Mooi et al. 2011) have been documented within and macro-photography to examine patterns of pigment on scales between populations of several Thamnophis species (Brodie and skin of adult T. sirtalis from the islands of Nantucket and 1989; Ruthven 1908; Westphal and Morgan 2010). In general, Tuckernuck and the mainland of Massachusetts. Nantucket Is- color pattern variation in snakes is presumed to reduce preda- land is 118 km2 and is the largest of three islands within Nan- tion risk through interacting mechanisms normally including tucket County, Massachusetts. Tuckernuck Island is 3.6 km2 and background matching and color pattern-specific antipredator approximately 1 km W of Nantucket. These islands are approxi- behaviors (Brodie 1992), mimicry (Brodie 1993), or aposematic mately 40 km S of the closest mainland (Cape Cod) and 15 km coloration (Williams et al. 2004), but it has also been linked to E of the nearest large island, Martha’s Vineyard, Massachusetts. thermoregulatory efficiency (Bittner et al. 2002; King 1988). We analyzed scales for stripe and blotch on 32 Nantucket Although color patterns are not a definitive method to deter- snakes, 28 Tuckernuck snakes, and 26 specimens from four mine species, subspecies, or distinct genetic lineages, variable counties on the mainland of Massachusetts (Middlesex [N = 5], patterns are important to consider because differential selective Worcester [N = 16], Hampshire [N = 1], and Hampden [N = 4] pressures can exist in each population (Brodie 1992). Thamno- counties) between 2008 and 2011 (Fig. 1). All snakes had a body phis sirtalis exhibits population-specific phenotypes without mass of greater than 10 g and a snout–vent length (SVL) greater necessarily being genetically distinct (Gregory and Larsen 1996; than 250 mm and were released after being photographed. One King 2003; King and Lawson 2001; Mooi et al. 2011; Westphal person could obtain digital images for analysis in the field by us- 2007; Westphal and Morgan 2010). In spite of considerable re- ing the following procedure. Snakes were gently pressed against search on various T. sirtalis subspecies from central and western a firm surface, usually the photographer’s upper leg, while kneel- North America and the Maritime Gartersnake (T. s. pallidulus) ing on the opposite leg to provide stability for the camera and from Nova Scotia, Canada (Allen 1899; Barnes et al. 2006; Bleak- specimen. While holding the snake, the photographer gently ney 1959; Gilhen 1984, 2010; Rye 2000), there have been very few stretched the skin and scales to expose the lateral blotches on quantitative measurements of color patterns comparing popu- the left side of the snake approximately one-fourth the SVL as lations or subspecies of the Common Gartersnake in the north- measured from the snout. Instead of pressing the snake too eastern USA. firmly, holding the snake and watching the expansion and con- Our objectives were 1) to develop a standardized method traction of the skin during breathing provided a passive method for measuring color pattern variability, and 2) to compare color of obtaining the best image. When a snake inhales, the anterior patterns between T. sirtalis populations from two islands and portion of the body expands and the skin is exposed between the mainland Massachusetts. To achieve our first objective, we used scales to provide a good image of the lateral skin pattern. After a pattern analysis similar to that developed by Westphal (2007); a snake was in position, it relaxed and began breathing rhyth- however, we based our analysis on macrodigital images of mically and we timed the photograph with the breathing. The snake was then slightly rolled on to its ventral scales so that a SCOTT D. SMYERS1* second set of photographs could be taken of the dorsal stripe in ANDREW MCKENNA-FOSTER2 the same region of the snake’s body. JONATHAN D. SHUSTER2 Our analysis of the dorsal stripe focused on the amount of black pigment on the scales where we counted seven scales along 1Oxbow Associates, Inc., P.O. Box 971, Acton, Massachusetts 01720, USA the dorsal stripe, normally row 10, and estimated the amount of 2Maria Mitchell Association, 4 Vestal St. Nantucket, Massachusetts 02554, USA black pigment on each scale (Fig. 2). No black on the scale was *Corresponding author; e-mail: [email protected] scored zero; black on 1–25% of the scale was scored 0.25; 26–50% was scored as 0.5; 51–75% was 0.75; and 76–100% was scored

Herpetological Review 45(1), 2014 ARTICLES 9

as 1.0. The total score was summed to provide a stripe score for all males and females, and there was no significant difference each snake (range from 0–7). We developed a scoring system for (Blotch, P = 0.23, Stripe, P = 0.70). Therefore, we pooled males analysis of lateral blotches using the same concepts described and females in the color pattern analysis. We used ANOVA Gen- by Westphal (2007). We analyzed an area defined by five rows of eral Linear Model to compare mainland and island population scales dorsal to the lateral stripe (scale rows 4–8) and seven scales data sets. The blotch and stripe data were normally distributed anterior to the site of reference. For each of the 35 scales in this and comparisons were tested at ฀ a = 0.05. area, we estimated the percent of the scale’s border adjacent to dark interscalar skin that formed a dark, pigmented blotch (Fig. 2). If there was no dark pigment on the interscalar skin adjacent to the scale, the score was zero. Any dark pigment on the interscalar skin from approximately 1–25% of the scale’s border resulted in a score of 0.25, 26-50% of the scale bordering black pigment was 0.5, 51–75% was 0.75, and 76–100% was 1.0. The total score for each row was summed then divided by five (number of rows scored) to provide an average blotch-per-row score for each snake. In many instances more than one image had to be reviewed, often from the dorsal image, to see around the curve of the snake for rows 7–8. We used miniTab 12.1 (1999 Minitab, Inc., State College, Penn- sylvania) to test the data for normality, examine basic statistics, and compare the two response variables, lateral blotches and dorsal stripe between mainland snakes (pooled all samples) to the island specimens (pooled per island) using ANOVA General Linear Model and Tukey pairwise comparisons. To confirm that Fig. the variation in color pattern was not a sexually dimorphic trait, 1. Approximate location and number of Thamnophis sirtalis photographed for color pattern analysis. we used a t-test to compare stripe and blotch scores between

Fig. 2. Examples of scores for each scale for the dorsal stripe and black blotches from one mainland snake (A. and B, Holyoke, Hampden Co.) and one island snake (C. and D.,Nantucket, Nantucket Co.).

Herpetological Review 45(1), 2014 10 ARTICLES

Fig. 3. Representative photographs of patterns of snakes from non-coastal counties and from the islands of Nantucket and Tuckernuck, Massachusetts. The specimens in A-C are from the mainland (A. and B., Wachusett Mountain, Princeton, Worcester Co.; C. Dracut, Middlesex Co.) while those in D-G are from the islands (D. and H. Tuckernuck Island, Nantucket Co.; E-G Nantucket Island, Nantucket, Co.).

Herpetological Review 45(1), 2014 ARTICLES 11

Results and Discussion.—There was significantly less dark pallidulus in northeastern Maine and the Maritime provinces pigment within the dorsal stripes of mainland snakes (N = 26, of Canada, providing validity to the historic range map of the 0.63 ± 0.13 (mean ± 1 Standard Error [SE]) compared to snakes subspecies by Bleakney (1959). This subspecies has been most from Nantucket (N = 32, mean = 2.20, ± 0.12) and Tuckernuck thoroughly studied in Nova Scotia, Canada, where morphologi- islands (N = 28, mean = 2.09, ± 0.10, F 2, 85 = 58.9, P < 0.001). The cal variation, life-history, and general ecology have been docu- stripe scores from Tuckernuck and Nantucket were not signifi- mented (Gilhen 1984). Populations on smaller islands of Nova cantly different (P = 0.74). Lateral blotches on snakes from Nan- Scotia have also exhibited variation in color patterns (Barnes et tucket (N = 30, mean = 3.53 ± 0.09) were significantly smaller al. 2006). than snakes from the mainland (N = 25, mean = 4.23 ± 0.083, P Methods used in this study are likely to be useful range-wide

<0.001), and Tuckernuck (N = 28, mean = 4.06 ± 0.14, F2, 82 = 11.7, for T. sirtalis as well as other species. Additional studies of pat- P = 0.002). However, lateral blotches on snakes from Tuckernuck terns of T. sirtalis from other nearby islands and populations did not differ from those on mainland snakes (P = 0.532). along the coast and inland could reveal different patterns than Our results indicate that T. sirtalis from two coastal islands those reported herein. This may provide insight into how popu- have significantly more black pigment on the dorsal stripe com- lation-specific selective pressures influence isolated populations pared to snakes from the Massachusetts mainland. The snakes of T. sirtalis. Our methods could be used to compare color pat- from Nantucket and Tuckernuck had discontinuous stripes with terns of T. sirtalis and T. s. pallidulus from the maritime provinces dorsal blotches that were sometimes fused with lateral blotches of Canada or northern New England to determine if historically to form dark cross bands. There was no significant difference recognized subspecies are functionally different. between lateral blotch areas of snakes from the mainland com- Collecting data from live snakes is advantageous because pared to Tuckernuck, but snakes from Nantucket had signifi- the elasticity of the skin allows close examination of true color cantly less blotch area than mainland snakes. patterns, which is often difficult or impossible when examining There are several possible explanations for the significant preserved specimens. Although the snakes often bite and strug- differences resulting from our color pattern analysis. The spot- gle to escape, by restricting data collection to photographs of a ted pattern on island snakes may be related to antipredator be- relatively small region of the body, we reduce the specimen’s dis- haviors, thermoregulation, or genetic drift. A snake’s likelihood comfort and overall resistance allowing for a relatively consistent of fleeing potential visual predators has been linked to color image. We intend to continue studying this species’ predisposi- patterns (Brodie 1992; Jackson et al. 1976), and the contrast- tion to color pattern variability, both on the individual and pop- ing pattern exposed when a T. sirtalis flattens its body to reveal ulation level. Future research may lead to a better understanding rows of dark blotches or pigment on the skin between the scales of how this wide-ranging species not only persists and disperses, may elicit a startle response from visual predators. Indeed, oth- but how it is successful throughout variable regions of North ers have shown that within Thamnophis ordinoides, striped America including island habitats. individuals are more likely to flee predators compared to spot- ted individuals which are more likely to remain motionless and Acknowledgments.—This study was conducted under Scientific rely on crypsis to avoid predation (Brodie 1989). Unlike their Collecting Permits (#087.08, 087.09, 087.10, 087.11 SCRA) from the mainland counterparts, Thamnophis sirtalis from the islands Massachusetts Division of Fish & Wildlife. We thank The Nantucket of Nantucket and Tuckernuck are not exposed to mammals that Biodiversity Initiative and Oxbow Associates, Inc. for providing fund- potentially prey upon snakes, though feral cats (Felis domestica) ing. We are grateful for permission to conduct field work by Nantuck- et Conservation Foundation, Tuckernuck Land Trust, The Trustees of are an introduced predator that may or may not be ecologically Reservations, Linda Loring Nature Foundation, Wachusett Mountain important in some areas (Massachusetts Department of Fish State Reservation, and many other generous land owners. We are and Game, 2011). Additional studies investigating antipredatory grateful to many field assistants: S. Alger, J. Blyth, R. Carr, R. S. Egan, behaviors could assist in determining the relative importance of A. Feldman, A. Heighton, R. Kennedy, R. Kessler, D. Lang, E. LoPresti, predation in maintaining such color pattern variation in T. sirta- C. Mirbach, J. Pelletier, and T. Westlake. We are thankful to L. Perotti lis on these islands and the adjacent mainland. of Roger Williams Zoo, for providing field training for field assistants. Thermoregulation is another possible explanation for dif- A. Richmond from University of Massachusetts at Amherst Natural ferences in lateral blotches between Nantucket and mainland History Collections, as well as J. Martinez and J. Rosado of the Muse- snakes. A study on differences in thermoregulation of melanistic um of Comparative Zoology provided access to preserved specimens gartersnakes demonstrated that the larger area of black pigment and encouragement to pursue the research. J. Martinez, J. Losos, and R. Clancy provided constructive comments on this manuscript. T. is functionally beneficial to adult snakes and likely most impor- Tada and R. Strohsahl assisted with preparation of figures. tant to gravid females (Bittner et al. 2002). Finally, the varying color patterns may be due to genetic drift, and the varying color patterns are the result of relatively small founding populations, thus have no selective advantage. liTeraTure CiTed Lazell (1976) inferred that color variation of T. sirtalis from the coastal islands and Cape Cod Massachusetts may be relict allen, g. m. 1899. Notes on the reptiles and amphibians of Intervale, populations of T. sirtalis pallidulus, suggesting a postglacial dis- New Hampshire. Proc. Boston Soc. Nat. Hist. 29:63–75. barnes, s. m., C. m. dubesky, and T. b. Herman. 2006. Ecology and mor- persal-driven hypothesis: cold-tolerant snakes advanced north phology of Thamnophis sirtalis pallidulus (Maritime garter snake) with the glacial retreat and were stranded on the islands as sea on Georges Island, Nova Scotia. Northeast. Nat. 13:73–82. level increased and cut off any overland dispersal vectors. Bleak- biTTner, T. d., r. b. king, and J. m. kerFin. 2002. Effects of body size and ney (1959) provided a detailed review of T. s. pallidulus including melanism on the thermal biology of garter snakes (Thamnophis descriptions of color patterns, scale formulas, and hybrid zones sirtalis). Copeia 2002:477–482. from the northeastern USA and adjacent Canada. Rye (2000) bleakney, s. 1959. Thamnophis sirtalis sirtalis (Linnaeus) in eastern reported a genetically distinct maritime lineage of T. sirtalis Canada, redescription of T. s. pallidula Allen. Copeia 1959:52–56.

Herpetological Review 45(1), 2014 12 ARTICLES

brodie, e. d., Jr. 1989. Genetic correlations between morphology and king, r. b. 1988. Polymorphic populations of the garter snake Tham- antipredator behavior in natural populations of the garter snake nophis sirtalis near Lake Erie. Herpetologica 44:451–458. Thamnophis ordinoides. Nature 342:542–543. ———. 2003. Mendelian inheritance of melanism in the garter snake ———. 1992. Correlational selection for color pattern and antipreda- Thamnophis sirtalis. Herpetologica 59:486–491. tor behavior in the garter snake Thamnophis ordinoides. Evolu- ———, and r. lawson. 2001. Patterns of population subdivision and tion 46:1284–1298. gene flow in three sympatric natricine snakes. Copeia 2001:602– ———. 1993. Differential avoidance of coral snake banded patterns 614. by free-ranging avian predators in Costa Rica. Evolution 47:227– lazell, J. d. 1976. This Broken Archipelago. Demeter Press Quadran- 235. gle/The New York Times Book Co., New York, New York. 260 pp. gilHen, J. 1984. Amphibians and Reptiles of Nova Scotia. Nova Scotia mooi, r. d., J. P. wiens, and g. s. CasPer. 2011. Extreme color variation Museum Press, Halifax, Nova Scotia. 162 pp. within populations of the common gartersnake, Thamnophis sir- ———. 2010. Erythrism in the maritime garter snake, Thamnophis talis, in Central North America, with implications for subspecies sirtalis pallidulus, in Nova Scotia. Can. Field Nat. 124:99–103. status. Copeia. 2011:187–200. gregory, P. T., and k. w. larsen. 1993. Geographic variation in repro- rossman, d. a., n. b. Ford, and r. a. seigel. 1996. The Garter Snakes: ductive characteristics among Canadian populations of the com- Evolution and Life History. University of Oklahoma Press, Nor- mon garter snake (Thamnophis sirtalis). Copeia 1993:946–958. man, Oklahoma. 332 pp. ———, and ———. 1996. Are there any meaningful correlates of geo- ruTHVen, a. g. 1908. Variations and genetic relationships of the gar- graphic life-history variation in the garter snake, Thamnophis sir- ter-snakes. U.S. Nat. Mus. Bull. 61:1–195. talis? Copeia 1996:183–189. rye, l. a. 2000. Analysis of areas of intergradation between described grudzien, T. a., b. J. Huebner, a. CVeTkoViC, and g. Joswiak. 1992. Multi- subspecies of the common garter snake, Thamnophis sirtalis, in variate analysis of head shape in Thamnophis s. sirtalis (Serpentes: Canada. Dissertation, University of Maine, Orono. Colubridae) among island and mainland populations from north- wesTPHal, m. F. 2007. On the evolution of correlated color traits in eastern Lake Michigan. Am. Midl. Nat. 127:339–347. garter snakes. Dissertation. Oregon State University, Corvalis. JaCkson, J. F., w. ingraHam, and H. w. CamPbell. 1976. The dorsal pig- ———, and T. J. morgan. 2010. Quantitative genetics of pigmenta- mentation patterns of snakes as an antipredator strategy: a multi- tion development in 2 populations of the common garter snake, variate approach. Am. Nat. 110:1229–1053. Thamnophis sirtalis. J. Hered. 101:1–8. kelley, k. C., s. J. arnold, and J. gladsTone. 1997. The effects of sub- williams, b. l., e. d. brodie Jr., and e. d. brodie iii. 2004. A resistant strate and vertebral number on locomotion in the garter snake, predator and its toxic prey: persistence of newt toxin leads to poi- Thamnophis elegans. Funct. Ecol. 11:189–198. sonous (not venomous) snakes. J. Chem. Ecol. 30:1901–1919.

Herpetological Review 45(1), 2014