Abstract
The first instance of sympatric megacephalic map turtle species (Graptemys barbouri and Graptemys ernsti) along the Gulf Coast of the United States, and below the fall line, is in the Choctawhatchee River system in Alabama and Florida. This unique population was discovered in a river system that was believed to be devoid of Graptemys species. Barbour's map turtle (G. barbouri) was discovered and documented in 1997 and the Escambia map turtle (G. ernsti) in 2002 in the Pea River, a western tributary of the Choctawhatchee River by James C. Godwin. From 2007-2008, I collected data from basking surveys and capture efforts along the Choctawhatchee and Pea rivers in Alabama and Florida. I found data from basking surveys reliable in the Choctawhatchee River for identifying Graptemys down to species, but not in the Pea River, because of the overlap in range and the presence of hybrids. This project began as a mark-recapture study at five locations but was converted to a survey and mark endeavor up the southern Choctawhatchee and Pea rivers in Alabama, as well as the northern Choctawhatchee River in Florida. The discovery of putative hybrids of these species by James Godwin added an additional layer of complexity to this new range extension. Godwin included a general range map based on a low number of captured specimens in his unpublished report on the range of G. barbouri in this drainage in Alabama, presenting the occurrence of both species in the river system. I provide an updated range map of both species in the river system, as of 2008, including delineation of a hybrid zone where specimens with shared morphological traits have been observed. I captured 115 Graptemys specimens and measured carapace length, plastron length, carapace width, height, weight and head width on each in the field. All turtles were marked, and released back at the point of capture. I also measured 8 specimens from the Auburn University Museum. Of these 115 turtles I captured, I classified 72 as G. barbouri and 38 as G.ernsti. The remaining 5 turtles with morphological traits of both species, at a ratio of nearly 1:1, were labeled hybrids. All five of these apparent hybrid turtles were either juvenile or older females. No male hybrids were found. Morphological characteristics of G. barbouri and G. ernsti from the Choctawhatchee River system, such as head and carapace markings, were compared with specimens from neighboring river systems. Turtles that were considered hybrids shared near equal characteristics of both species or had jumbled/indistinguishable patterns. The pigment widths on the upper and lower 5th marginal scutes in G. barbouri from the Choctawhatchee river were compared with specimens from the parent drainages and were found to be significantly different. Comparisons of relative carapace height and relative carapace width, using the Mann-Whitney U Test, of G. barbouri and G.ernsti from the Choctawhatchee River system and the parent drainages were not significant among adult and subadult males and females, but were significant in unsexable juveniles. Comparisons between G. barbouri and G. ernsti from the hybrid zone and outside the hybrid zone in the Pea and Choctawhatchee rivers were also tested, using the same non-parametric test, and were not significant at any size (juvenile, male, or female). Juvenile hybrid relative carapace lengths and relative carapace widths were also compared with juvenile G. barbouri and G. ernsti in the drainage and hybrid carapace measurements were significantly closer to G. ernsti than G. barbouri. Three hypotheses as to the method of establishment of G .barbouri and G. ernsti in the Choctawhatchee and Pea River were considered: (1) that both species were accidentally or purposefully introduced by humans; (2) that both species occur in the Pea River system as a result of stream capture (the connection of a tributary of one river system to another as a result of environmental changes); and (3) that both species entered the river system by traveling over land during extreme flood events. I concluded that stream capture is the most plausible hypothesis for a sympatric distribution of G.barbouri and G.ernsti in the Pea River. This explanation assumes that G. ernsti was present in the Pea River (previously a tributary of the Yellow River) when it was captured by the Choctawhatchee River. This connection allowed G. barbouri to enter the Pea River and expand its range upriver. G.barbouri may have entered the Choctawhatchee River from the Chattahoochee River in Alabama or from the Chipola River in Florida by stream capture or by brief connections of close neighboring tributaries. Basking and capture data from the Choctawhatchee River, south of the Alabama border, shows a much higher abundance of G. barbouri, as opposed to upriver of the confluence of the Pea River, where stream capture was originally suspected.