Invagination and erosion of the turtle cranium

Turtles have no temporal fenestra,
but some of them have enlarged their jaw muscles by greatly enlarging the cranium, or by invagination of the cranium from the occiput, or both (Fig. 1). Skull temporal fenestra are important traits to categorize most reptiles, but turtles do not follow other clade morphologies. That has made turtles difficult to categorize and nest in traditional studies.

Figure 1. Macrochelys (Macroclemys) skull colorized. Most workers label the bone above the curled quadrate as a squamosal, but here it is considered a supratemporal, which has horns in basal turtles. This skull shows a minimum of occiput invagination, but note the great height of the cranium.

Some paleontologists
think turtles are diapsids related to placodonts, but that is not supported by the large reptile tree.

Other paleontologists
think turtles are anapsids related to Eunotosaurus, but that is not supported by the large reptile tree.

Still other paleontologists
USED to think turtles are anapsids related to pareiasaurs, and that IS supported by the large reptile tree. Basal turtles, like pareiasaurs and all basal tetrapods, have both an external (dermal) skull  surrounding and protecting the smaller internal (braincase) skull.

Basal turtles have a solid cranium – with horns!
Elginia and Meiolania are basalmost turtles they have horns and at least we know that Meiolania had a solid carapace and plastron. The outgroup, Sclerosaurus, has horns, but no shell and no broad ribs. In Meiolania the large, horned supratemporal sends a ventral process to contact the quadratojugal leaving a hole for the quadrate and stapes (ear bone). The supratemporal is a large bone in basal turtles that does not go away in derived turtles. Rather, the squamosal continues to shrink.

Figure 2. Elginia and Meiolania, two basal horned turtles without skull invagination. In Meiolania the supratemporal sends a ventral process to contact  the quadratojugal leaving a hole for the quadrate and stapes (ear bone). The supratemporal is a large bone in basal turtles that does not go away in derived turtles. Rather, the squamosal continues to shrink.

Proganochelys (Fig. 3) has long been recognized as a basal turtle. It has no horns or skull invagination, so, in this context, it is not a basal turtle, but a transitional turtle, between horned and invaginated-skull turtles.

Figure 3. The skull of Proganochelys, a basal turtle without skull invagation and without horns. Note the identification of the supratemporal on the right matching that of basal turtles like Elginia and Meiloania in figure 2.

Softshell turtles
have an invaginated cranium, no horns and sometimes reduce their bony shells. A basal turtle with teeth, Odontochelys, nests with a soft-shell turtle, Trionyx, in the large reptile tree. The cranium of Trionyx is invaginated from the occiput, creating space for large jaw muscles. The skull of Odontochelys (Fig. 7) is difficult to study with available data, but it appears to have large round holes in the crushed cranium. At least it does not appear to have the solid cranium that was illustrated originally (Fig. 7). Rather the cranium appears to be so badly crushed, even in the low resolution image available, that it may indeed have had a more fragile, less box-like, Trionyx-like cranium. I requested high rez images, but was informed that another paper focusing on the skull of Odontochelys is in progress. Looking forward to that!

Figure 3. Trionyx, a softshell turtle with bones colorized.

Other tested turtles have a an increasingly invaginated cranium
Chelonia, the sea turtle, is a basal turtle that has a rather solid skull with a little posterior invagination.

Macrochelys, the alligator snapping turtle (Fig. 1), has a deeper invagination and a much taller cranium.

Pelomedusa and Foxemys are similar to the snapping turtle, but without the grand enlargement of the cranium.

Terrapene (Fig. 5), the box turtle, has lost most of its original cranium, revealing a braincase, like a mammal, snake, amphisbaenid or bird.

Figure 5. Terrapene, the box turtle, with skull bones colorized. Note the lack of a dermal skull and the appearance of the cranial skull, the braincase, as in birds, mammals, snakes and amphisbaenids.

The large reptile tree
nests Kayentachelys between the soft-shell turtles, Trionyx and the hard-shelled turtles, like Chelonia (Fig. 6). Kayentachelys has a complete cranium without invagination. Separate nestings of soft-shell and hard-shell turtles with skull invagination indicate this trait was convergent, not homologous. Such a tree topology has not been recovered before, but then no prior study (that I can recall) has included Sclerosaurus, Elginia and Stephanospondylus.

Figure 6. Turtle phylogeny showing extent of horns and cranial invagination. Here the skull invagination of soft-shell turtles is convergent with that of most other turtles.

Parts of the skull of Odontochelys cannot be accurately reconstructed with available data (Fig. 7). There are apparent temporal fenestrae in the in situ specimen exposed in dorsal view. These would ordinarily have the appearance of diapsid openings and would lend credence to the diapsid hypothesis of turtle origins. Instead, let’s wonder if these holes represent either: 1) geological erosion; or 2) erosion of the posterior cranium in spots transitional to the morphology seen in Trionyx. There’s nothing else I can say at present until better data comes along.

There is a large circular plate
(Fig. 7) in the palatal view of the smaller Odontochelys that was labeled a possible squamosal. I don’t think there is room on the skull for that bone at present. So that elliptical bone may be from elsewhere, perhaps on the carapace or plastron.

Most turtles have anterior nares.
The anterolateral placement of the large naris in Odontchelys is different from all other turtles and similar only to Elginia (Fig. 2).

Figure 7. Data and tentative interpretations of skull elements for Odontochelys. I can’t make more sense than this of the bones. Sorry. Gray areas appear to represent holes in the cranium. Note the difference between the original drawing and the photo with color overlay. The skull bones of Odontochelys appear to be more fragile than boxy turtle skulls are.

The skull of Stephanospondylus (Fig. 8) is a good starting point for both pareiasaur and turtle skulls. It nests (Fig. 7) at the base of both clades.

Figure 8. Stephanospondylus skull in two views. Note the rotation of the post parietals to the dorsal skull along with the transformation of the supratemporals into small horns.

Perhaps more taxa
will someday unite soft-shell turtles with hard-shell turtles, but at present, the convergence is remarkable among all turtles with an invaginated skull. With regard to Odontochelys, I think we’ll see a strong revision of the original drawing.

References
Li C, Wu X-C, Rieppel O, Wang L-T and Zhao L-J 2008. An ancestral turtle from the Late Triassic of southwestern China. Nature 456: 497-501.