A new Dr. Tyler Lyson lecture video
on YouTube discusses the origin of the turtle shell. The lecture is ‘Round 6’ in a series of “Ten Rapid-Fire Presentations” sponsored by the Senate of Scientists at the Smithsonian National Museum of Natural History (NMNH) and given in 2013.
As readers know,
the origin of the turtle shell, like the origin of snakes and pterosaurs, is one of the hottest topics in paleontology today.
As part of the backstory,
the large reptile tree (LRT 1040 taxa) recovered two more or less parallel origins of turtle shells from related small pareiasurs. In this heretical hypothesis soft-shell turtles did not evolve from hard-shell turtles or vice versa.
Lyson begins his lecture with
a declaration of his interests in amniotes and the origin of their body plans. So, he and I are interested in exactly the same topics.
he was (in 2013) using an antiquated ladder cladogram in which synapsids split off first from the amniote tree topology leaving lizards the sisters of archosaurs. That’s not true when you add more taxa to your tree as we discovered 6 years ago (in 2011) here.
On that invalid tree Lyson notes
that turtles have been postulated to nest on every branch of that simplified four taxon tree. …even the croc branch and dino/bird branch (but not the synapsid branch). That adds to the humor, but comes as news to me. Lyson studied the soft tissue, bones, CT scans and developmental (embryo) patterns then integrated them into a dataset that we examined earlier here.
Lyson discusses two competing hypotheses
regarding the origin of the carapace (top shell).
- Composite hypothesis: more and more osteoderms (ossified scales) supported by fossil data
- De Novo hypothesis: broadening of the ribs and vertebrae alone (supported by developmental data)
With its broad ribs and lack of osteoderms, Lyson reports
Odontochelys falsifies the composite hypothesis. In the LRT Odontochelys is the earliest known soft-shell turtle. Without a prior understanding of the dual origin of turtles based on a large gamut of taxa, Lyson made the classic mistake of assuming a single origin of turtles with hardshell osteoderms. This is why you let the data tell you what is going on, rather than the other way around.
a 30–50 million-year gap between Odontochelys and the rest of the amniotes. He reports his interest in those transitional taxa and produces Eunotosaurus africanus, which the LRT nests closer to caseids and Acleistorhinus. Lyson produced the first detailed anatomical study of Eunotosaurus, which we talked about here, here and here.
In turtle embryos Lyson notes
the first thing to develop are broadened ribs. Of course we’re only shown one species of turtle here and it has (or will have) a hard shell. The next thing Lyson reports seeing are broadened vertebrae. Then in late stages the shell appears.
In a wonderful transforming animation
Lyson presents a speeded up evolution of the turtle shell beginning with:
- Milleretta, with slightly broadened ribs and a long series of 18 short dorsal vertebrae.
- Eunotosaurus, with 9 elongated dorsal vertebrae and broader overlapping ribs.
- Gastralia (belly ribs) then lengthen and broaden to form the plastron. Note: gastralia do not appear in Milleretta, Eunotosaurus or small pareiasaurs. So the plastron genesis remains unknown as it is already fully developed in Odontochelys and Proganochelys. Falsifying the ‘osteoderm’ hypothesis, the plastron of the basalmost known turtle with a hard domed carapace, Meiolania, has a large hole (aperture, fenestra) in the middle, as if the original development tied the pelvis to the lateral carapace and the pectoral region to the lateral carapace with only soft tissue in between. In counterpoint, the plastron of Odontochelys does radiate out from the midline as five very broad and interlocking gastralia, which is a very low number. The lateral edges of each of the anterior 4 plastron bones appears to be roughly subdivided into three or more perhaps fused slender gastralia. The ancestral taxon to Odontochelys, Sclerosaurus has no gastralia, but it does nat supratemporal horns, similar to those found in Elginia and Meiolania. Thus Sclerosaurus may turn out to be the last common ancestor of all turtles.
In the question and answer finale, Lyson notes:
the original broadening of the ribs was an adaption for burrowing.
As a reminder
Eunotosaurus is convergent with turtles in many regards, It has only 9 elongate dorsal vertebrae, similar to turtles. The ribs are broad and curve laterally before descending. Even the rib histology is similar. So why does the LRT nest Eunotosaurus apart from turtles? Because other taxa share more traits with Eunotosaurus, plain and simple. Eunotosaurus is one of several turtle mimics. This is convergence at it best, good enough to confuse a brilliant PhD from the best institutions with every resource at his command, except an large gamut taxon list. Readers, you must start with a large gamut taxon list before proceeding. It’s your Google map, your GPS to tell you where you are in the reptile tree topology. And it’s your best guide to nesting and avoiding convergent taxa. With so many taxa in the LRT, every taxon has 1039 candidates it COULD nest with, but each one finds the one with which it shares the most traits.
Tyler Lyson (lee-son) got his PhD from Yale U and is currently Curator of Vertebrate Paleontology at the Denver Museum of Nature & Science. He spoke in 2013 (see above) while working at the Smithsonian in the Vertebrate Zoology Department in the second year of his post-doctoral fellowship.
From his current DMNH online bio:
“Dr. Tyler Lyson studies fossil vertebrates, particularly dinosaurs and turtles. He is especially interested in the evolution of body plans and extinction patterns of different groups across major extinction events. In his research, Dr. Lyson integrates molecular, developmental, and morphological data from living and fossil organisms all within an evolutionary tree-based context to address complex paleobiological and paleoecological problems, including the evolution of body plans, niche partitioning in dinosaurs, and extinction patterns through time.”
Other turtle origin videos on YouTube can be found here by:
- Reiley Jacobson – haven’t seen this. May review it later.
- Benjamin Burger – same invalid traditions
- David Peters – sorry. this was made before I started adding soft-shell turtles to the cladogram. But it is otherwise correct. See, taxon exclusion can be a problem for anyone! You have to test the gamut.
Addendum and Finally, if you’re wondering…
Yes. Tyler Lyson is pulling a Larry Martin here. (identifying relatives and ‘homologous’ traits because they SEEM right, BEFORE completing a large gamut phylogenetic analysis that removes all other possibilities and the taxa nest themselves.)