Re: Further Thoughts on the Origin of Chelonia

["David Marjanovic" <david.marjanovic@gmx.at>]

Elena G. Kordikova: Comparative morphology of the palate dentition in
*Proganochelys quenstedti* BAUR 1887 from the Upper Triassic of Germany and
Chelonian [sic] Ancestry [sic], Neues Jahrbuch für Geologie und
Paläontologie: Abhandlungen 2552(2), 195 -- 249 (August 2002)

"_Abstract:_ A reanalysis of skull roof and palate of *Proganochelys*
confirms the parareptilian status of ancient turtles. A paedomorphic origin
of the ancient turtle palate pattern from that of a basal pareiasauroid
lineage is suggested.
    The main types of palate dentition (acrodont, hyperacrodont,
prothecodont, pseudoacrodont, ridge thecodont and true thecodont) as well as
the origin of tooth attachment in tetrapods are discussed. Skull characters,
[sic] that are considered most important in phylogenetic systematics of the
lower tetrapods, [sic] are evaluated."

Conclusion: Turtles are paedomorphic dwarf pareiasaurs.

Walter Joyce: The presence of cleithra in the primitive turtle
*Kayentachelys aprix*, this year's SVP meeting abstracts (supplement to JVP
23(3), 66A, 12 September 2003)
"[...] was found to have clear traces of cleithra, a dermal component of the
pectoral girdle. Although these structures were previously noticed in this
taxon, they were diagnosed as simple processes of the epiplastra. This
primary homology assessment was probably influenced by the morphology of the
most primitive known turtle, *Proganochelys quenstedti*. In this taxon,
cleithra are also present, but so tightly fused to the plastron that they
appear to be outgrowths of the plastron rather than separate structures. In
*Kayentachelys aprix* the cleithra clearly demonstate their identity as
independent bones by being loosely attached to both the epiplastra (=
clavicles) and entoplastron (= interclavicle) just anterior to the scapula.
As these observations match the anatomical relationships of the cleithra of
other primitive tetrapods, diagnosing these structures as cleithra is
reasonable if topology is used as a primary homology criterion.
        The presence of cleithra in *K[...]* is intriguing as it has
implications regarding basal turtle relationships and the origin of turtles.
As all living turtles lack cleithra, its [sic] presence in *K[...]* is
suggestive that this taxon may represent the phylogenetic stem of crown
turtles. [Currently it's thought to be just inside the crown group, on the
cryptodire side. Even though it retains palate teeth.] Furthermore, cleithra
are present in most anapsid reptiles but absent in all diapsids, making it
more plausible that turtles indeed are descendants of anapsid reptiles.
Admittedly, the presence of cleithra in primitive turtles is just one
character that may be rendered homoplastic by a comprehensive parsimony
analysis. *K[...]* and other primitive turtles display numerous other
characters, however, that suggest an affinity with diapsids."

http://tolweb.org/tree?group=Diapsida&contgroup=Amniota, however, says that
the absence of cleithra is a saurian apomorphy, not a diapsid one: "Saurian
autapomorphies include: [...] Cleithrum absent. The cleithrum is a dermal
bone located on the anterior edge of the scapula, dorsal to the clavicle. It
is present in araeoscelidians, Coelurosauravus, and younginiforms." Sauria
means here Lepidosauromorpha + Archosauromorpha; it used to be used like
Lacertilia for lizards minus snakes (paraphyletic).

> Incidentally, I think that (much as I would like turtles to be surviving
> pareiasaurs) they are going to prove to be archosauromorphs.

I still think that they are pareiasaurs, and are fooling us all with their
long branch, which is in part a result of their Early and Middle Triassic
ghost lineage. :-) The molecular sampling of turtles is still rather bad,
*Alligator mississippiensis* still has quite a long branch, and
morphological analyses that find turtles as diapsids tend not to have any
pareiasaurs in the matrix (but the reverse is also true, AFAIK). Well. We'll
see. Or so I think.

BTW, unlike diapsids, but like more basal amniotes (such as ourselves),
turtles produce lots of urea and very little uric acid. This and 2 more
soft-tissue characters with the same distribution are mentioned at
http://tolweb.org/tree?group=Diapsida&contgroup=Amniota.