Turtles are an old and complex group of animals with a confusing taxonomic history. The turtles often called "freshwater turtles" of the family Emydidae have been particularly troublesome. As it happens morphology, mtDNA, and nuclear DNA produce conflicting results for their relationships. In a new paper, Uwe Fritz and colleagues review the situation and make a recommendation on what names should be applied to some of the North American pond turtles, formerly placed in the genus Clemmys.
Duméril (1806) established the genus Emys for virtually all freshwater turtles known at the time. His genus contained more than 90 species which are now known to be scattered in multiple families representing many distinct turtle lineages (Chelidae, Chelydridae, Dermatemydidae, Emydidae, Geoemydidae, Kinosternidae, Pelomedusidae, Platysternidae, Podocnemididae, and Testudinidae). Boulenger (1889) limited the genus Emys to two species, the European Pond Turtle, Emys orbicularis and the North American Blanding's Turtle, Emys blandingii. This arrangement remained until 1957 when Loveridge and Williams transferred the Blanding's Turtle to the genus Emydoidea. Blanding's Turtles have a unique skull, neck and thoracic rib morphology which more closely resembles the Chicken Turtle, Deirochelys reticularia. Some other species previously in Emys were moved to Ritgen's genus Clemmys established in 1828. From the early 19th century Clemmys contained Old and New World freshwater turtles that were considered unspecilaized species lacking distinct morphological. The exception was Louis Agassiz, in 1857 he considered each of the New World species assigned to Clemmys as a representative of a distinct genera (Actinemys marmorata, Calemys muhlenbergii, Glyptemys insculpta, Nanemys guttata). Sam McDowell's 1964 osteological study revising the ‘aquatic Testudinidae’, Restricted Clemmys to the four Nearctic species Clemmys guttata, C. insculpta, C. marmorata and C. muhlenbergii, while the remaining Old World species were transferred to the genera Mauremys and Sacalia. McDowell (1964) realized Old World and New World freshwater turtles represent highly distinct groups and placed all Old World species plus the extraterritorial Neotropical genus Rhinoclemmys in the subfamily Batagurinae and the New World species plus the Palaearctic genus Emys in the Emydinae. These two subfamilies constituted, along with land tortoises (Testudininae), the family Testudinidae in McDowell’s (1964) classification. This arrangement is the one that is retained to the present, except each of these groups is now treated as a full family and the name Geoemydidae replaced Bataguridae because of name priority. McDowell recognized the close relationship of the four Nearctic Clemmys species, with the box turtles of the genus Terrapene and the Old World Pond Turtle, Emys orbicularis. He placed all of them in the ‘Emys complex’. However he did not include the Blanding's Turtle, Emydoidea blandingii. Instead he placed it with the distinct Chicken Turtle in the ‘Deirochelys complex’ (Emydoidea blandingii + Deirochelys reticularia). In 1974 Bramble pointed out, the morphology of structures associated with the plastral hinge of Emydoidea argues rather for a close relationship of Emydoidea with Emys and Terrapene, and not with Deirochelys. The plastral hinge of Emys, Emydoidea and Terrapene consists of ligamentous tissue that allows for almost complete closure of the shell, a trait better developed in Terrapene. Based upon this, Gaffney and Meylan concluded that Emys, Emydoidea and Terrapene represent a monophyletic group within the subfamily Emydinae (as opposed to the subfamily Deirochelyinae within the family Emydidae). The three genera share not only a plastral hinge, but also a divided scapula, a unique character among living turtles. The morphological similarity of these structures of Emydoidea, Emys and Terrapene was unique enough that Bramble (1974) concluded the plastral hinge could not have evolved more than once. The four Clemmys species lacked not only the plastral hinge, but also all of the complicated morphological structures associated with this character, and were considered to have a basal phylogenetic position within Emydinae, an assumption already assumed by McDowell and Bramble. Thus, Gaffney and Meylan placed all other emydid genera (Chrysemys, Deirochelys, Graptemys, Malaclemys, Pseudemys, Trachemys) in another subfamily (Deirochelyinae) within the Emydidae. By the mid 1990's mitochondrial DNA was revolutionizing how we looked at evolutionary relationships and Bickham and colleagues presented data that Clemmys is paraphyletic with respect to all other genera of the subfamily Emydinae (Emys, Emydoidea, Terrapene), and that the Spotted Turtle (guttata) was sister to all other emydines. Thus, the Wood Turtle, C. insculpta and the Bog Turtle, C. muhlenbergii, formed the sister group to a major clade divided into a subclade with the European Pond Turtle, Emys orbicularis, Emydoidea blandingii, and C. marmorata, and another subclade with all studied Terrapene species as its sister group. Bickham and colleagues used evidence from morphology, behavior and life history, to show the hinged taxa nested within Clemmys species. This prompted Burke et al. in 1996 to expanding the genus Emys to include all emydine species except C. insculpta and C. muhlenbergii.
To add to the confusion nuclear genomic data produced conflicting results, depending on which genes were used. The Spotted Turtle, Clemmys guttata, showed up as the sister to ((Emydoidea + Emys) + Actinemys) + Terrapene or as the sister to Actinemys marmorata and these two species together are the sister group of (Emydoidea + Emys). Box turtles, Terrapene then appear to be the sister to (Actinemys marmorata + Clemmys guttata) + (Emydoidea + Emys). The contradictory branching patterns depends upon the selected loci and suggest a lineage sorting problem. Ignoring the unclear phylogenetic position of Actinemys marmorata, one recently proposed classification scheme placed Actinemys marmorata, Emydoidea blandingii, Emys orbicularis, and Emys trinacris (the in one genus (Emys), while another classification scheme treats Actinemys, Emydoidea, and Emys as distinct genera. Fritz et al. consider the inclusion of Actinemys in the same taxon as Emydoidea + Emys as unacceptable under a phylogenetic classification framework because of evidence for the non-monophyly of this clade. The genra Actinemys, Emydoidea, and Emys are morphologically distinct, and their differences exceed the differences that typically occur among species of the same genus. Thus they recommend continued usage of the distinct genera Actinemys, Emydoidea and Emys. To find the full text of this paper follow the link below.