Chelonii: Overview

The gigantic Late Cretaceous marine turtle Archelon ischyros (Testudines - Casichelydia - Cryptodira - Eucryptodira - Protostegidae - Protosteginae. This impressive animal, which inhabited the Cretaceous inland sea of central North America, had a shell 3 meters long, and snout to tail tip length of some 5 meters illustration by Nobu Tamura (Wikipedia), Creative Commons Attribution-Share Alike

Introduction

The humble turtles, or more properly Chelonii or Testudines (terminology), are among the most ancient groups of reptiles on Earth. yet these strange and highly specialised animals are often taken for granted or ignored in the history of life, eclipsed by their larger and more exotic looking but also shorter lived contemporaries such as sauropods, brontotheres, and so on. This means that few realise how truely strange these animals are. They are, for example, the only vertebrates to have their shoulder girdle inside their ribcage, an astonishing evolutionary transition achieved through a small developmental shift in their embryonic stage. They have no clear ancestors, occuring suddenly in the fossil record during the Late Triassic, a time when many important new groups, such as crocodiles, dinosaurs, pterosaurs, lepidosaurs, and mammals were also appearing, along with shorter-lived but also for their time important clades such as phytosaurs and aetosaurs. They evolved into a diverse range of forms, populating land, rivers, swamps, and the open sea, survived several mass extinctions, and continue to this day.

One of the problems with determining both the origin and the evolutionary famnily tree of turtles is that in both cases there are a number of mutually contradictory hypotheses, each persuasive to some extent, but none of which able to explain all the data.

Originally, taxonomic classifications placed the Chelonia in the Anapsida, a group that originally included all reptiles that lacked temporal skull openings. This is now considered an artificial grouping, the lack of temporal fenestra being a primitive (plesiomorphic) feature retained in many lineages. We have used Anapsida in a more restricted but phylogenetically valid sense to refer to one of the three major clades of amniotes (the alternative popular term "parareptiles" is confusing if turtles, which are reptiles, are anapsids). The other two clades are Synapsids (mammal-like reptiles and mammals) and Eureptiles (Diapsids and their immediate ancestors)). The origin of turtles in the evolutionary tree is no longer clear, if it ever was. At least seven or eight competing hypotheses have been proposed over the years, each with its own problems, and some now discarded. The three main alternatives of anapsid origin from early amniotes ("parareptiles") and lepidosaur and archosaur ancestry among the higher Diapsida, have been proposed in the light of cladistic morphology and molecular phylogeny. The contradiction between Anapsid and Diapsid origins arose because whilst several cranial characters suggest a parareptilian/anapsid origin, some appendicular (limb and limb girdles) characters suggest a lepidosauromorph affinity, and molecular analysis uniformly supports an archosaurian origin (or stem-archosaur, e.g. archosauriforms or archosauromorph (Lyson & Gilbert 2009)). The Reptiliomorph, Eunotosaur, Captorhinid, Procolophonoid, and Pareiasaur origin theories are all variants of the traditional hypothesis that Testudines are Anapsida and the sister-group to Diapsida (i.,e. to all other extant reptiles), which is why and that they retain various primitive amniote characters such as an unfenestrated skull. The Sauropterygian and Ledidosaur hypotheses argue on morphological cladistic grounds for a Diapsid origin, whilst the archosaur origin finds almost no morphological support, but is strongly supported by molecular phylogeny. Other cladistic studies support an Anapsid (Parareptilia) origin, and embryology seems to do likewise.

Moreover, even within the testudine family tree itself, it is no longer clear which groups of chelonians evolved from which. The established view is that very primitive chelonians evolved into pleurodire and cryptodire, according to how they withdraw their head and neck into their shell (Romer, 1966, Gaffney and Meylan, 1988, Carroll, 1988). This has been rejected in view of new discoveries and cladistic analyses which show a range of intermediate forms, while molecular studies sometimes give confirming, and sometimes conflicting, results as regards traditional subgroups established through morphological analysis.

The situation is made even more confusing by the fact that turtle evolution seem to exhibit frequent homoplasy (parallel evolution). Thus for example the Cretaceous Protostegid sea turtles like the great Archelon shown at the top of the page, were morphologically extremely similar to the Cenozoic leatherback turtles (the largest extant turtles) but actually evolved from much earlier and more primitive ancesors. As they lived a very similar lifestyle, natural selection resulting in their evolutionarily converging into a very similar form (the classic example is the similarity of form between the shark, tuna, ichthyosaur, and dolphin, all living a very similar ecological lifestyle of large fast-moving marine preditors). Other examples can be given as well (e.g. (Rabi & Joyce, 2012, which shows that phylogeny is anything but parsimonious.

But this is nothing next to the problem of incompatability between morphological and molecular phylogeny. If morphology is correct then both first occurance in the fossil record and molecular phylogeny are often completely unreliable and misleading. If molecular phylogeny is correct then there would be unparalleled convergent evolution between the soft-shelled turtles (Trionychoidae) and the mud turtles (Kinosternoidea); two groups unanimously united on morphological cladistic evidence, which renders cladistics, the "gold standard" of modern phylogenetic studies, more or less irrelevant. If molecular phylogeny and paleontology are correct soft-shelled and the common ancestor of all hard-shelled cryptodire turtles diverged during the early Late Jurassic (Danilov & Parham, 2006); far less ghost lineages and a more reliable fossil record, but a degree of homoplasy that is verging on the ridiculous. Stratocladistrics may throw much needed light on this problem, but as this is a methodology that lies outside the current paradigm, this is unlikely in the near or even intermediate future.

In view of all this, the "traditional" account presented in the previous (2002) version of Palaeos has been revised and replaced by a more radical presentation, including a range of alternative interpretations, as shiown by the alterative placement of several taxa in the dendrogram. The Anapsid/Parareptile, the Lepidosaur/Sauropterygian, and the Archosaur origin theories are for different reasons equally plausible, but as the first of these remains persuausive the chelonia are listed on Palaeos next to the early Anapsida, although the expanded coverage and ambiguities regarding their phylogenetic position meansd they deserve a separate unit; hence these new pages. MAK101015