A Transitional Turtle

Turtles: how did they evolve?

Turtles (a group which includes within it tortoises) are most famous for their shells, which are made of their fused ribs. No other animal has a shell constructed in this manner, and it is a feature that all turtles share. The turtle evolution question, therefore, is somewhat synonymous with how turtle shells evolved. It’s also a case where creationists might reasonably ask “where are the transitional forms” – how do you get to shell from no shell? Unfortunately for them, fossils that fit the bill do exist. In 2009, for example, fossils of Odontochelys were discovered in China. This turtle had a complete plastron, the bottom of the shell, but instead of the upper part (the carapace) it merely had broadened rib bones. That sounds like a transitional form to me.

Backtracking for a moment, however, there have been two competing models of shell evolution: the palaeontologists tended to favour the composite model, which is the idea that parts of the skin were integrated into the shell when it evolved, while the developmental biologists inferred from their studies of turtle embryo development that this was not the case (the “de novo” model). The discovery of Odontochelys has apparently shown this latter view to be true, “liberat[ing] the paleontological search for the even deeper history of the turtle stem from its previously self-imposed constraint of osteoderm [hard skin structures]-bearing forms.”

The authors of a new paper (pdf), from which the above quote is taken, use this freedom to investigate an older and previously-known fossil, Eunotosaurus, determining that it is closely related to turtles and using it to investigate “deeper” into the evolution of the shell. Like Odontochelys, Eunotosaurus had only broadened ribs instead of a carapace, but also lacked the plastron – filling, in the process, another gap. With this information they can construct a basic timeline of turtle evolution, as shown in this nice video:

As I said, this isn’t a great position for creationists. On the 12th of June Brian Thomas tried desperately to defend against the evidence in an article titled Follow the Eunotosaurus: An Evolutionary Shell Game? He asked:

A real evolutionary transition—morphing between kinds—should show an in-between critical feature, like muscle attachment points half way between lizard and turtle, or ribs that were still flexible enough to permit reptile-like breathing, yet part-way stiffened like the turtle’s. What were Eunotosaurus’ telling transitional features?

Tellingly, he never actually answers that question, not even telling his audience that this fossil does not have a complete shell. Instead, he jumps straight to talking about stem and crown animals and similar jargon:

The study authors wrote that it “only differs from undisputed stem turtles…in sharing fewer derived characters with crown turtles.” “Stem turtles” denote evolutionary ancestors, “derived characters” describe body features supposedly inherited from those hypothetical stem turtles, and “crown turtles” refer to modern turtles, which have all their proper turtle parts in place.

The quote-mine given is interesting in that it’s not the ellipses that remove the necessary context, but simply what was said immediately afterwards. Here’s the full paragraph, with the most important sentence fragment bolded:

The morphology of Eunotosaurus is consistent with the explicit prediction of Kuratani et al. that the early stages of the turtle shell, prior to the emergence of Odontochelys, was marked by a vertical scapula positioned rostral to the ribcage. This condition is expressed in Eunotosaurus, in contrast to some other putative turtle sister groups, which have a scapula dorsal to the ribcage. Furthermore, the gross morphology of Eunotosaurus only differs from undisputed stem turtles, such as Odontochelys, in sharing fewer derived characters with crown turtles, as would be expected for an earlier member of the turtle stem. For example, Eunotosaurus lacks the derived conditions of neural plates, a hypoischium, and a co-ossified plastron (though all of the bones that form the plastron are present in Eunotosaurus).

So we have a transitional form which is very similar to the more turtle-like Odontochelys but lacks certain features as you would expect from an earlier animal. Thomas tries to defend by saying:

Possessing fewer features which some evolutionists might consider “derived” is not at all the same as having at least one transitional core feature.

Perhaps so, but this fossil does have “at least one transitional core feature” – those broadened ribs, which are part-way between normal ribs and turtle shell. All we need is for Thomas to mention that elephant in the room, but he’s not going to do it.

Also, all three quoted phrases are entirely subjective! One researcher could reassign a “stem turtle” to “evolutionary dead end” status, or swap a “derived character” for a “coevolved” character at any time and for any number of reasons. Indeed, secular scientific literature is filled with just such assertions and counter-assertions. It appears that Eunotosaurus’ transitional status rests on little more than a few conjectures.

In other words, Brian thinks all cladistics is bunk. But cladograms – “evolutionary trees” – are not build arbitrarily, but instead represent the best tree that can be built from the available evidence. New evidence can, in theory, lead to a tree being significantly revised, with the status of some features being changed in the process, but this doesn’t happen “at any time and for any number of reasons.” But the fact that Eunotosaurus has some, but not all, of the features of a modern turtle seems pretty solid, and does not “rest on little more than a few conjectures.”

So if Eunotosaurus was not “a critical transitional form,” then what was it? The study compared a range of fossil turtle anatomy details to provide the answer: “There is strong support for a turtle + Eunotosaurus clade [group].” Is that very different than just saying, “Eunotosaurus was a turtle?”

Yes, yes it is. The claim that “there is strong support for a turtle + Eunotosaurus clade” means that the relationship between turtles and Eunotosaurus looks like this:

From left we have first “turtles” and “more turtles,” collectively making up the “turtle” clade and symbolising the fact there is more than one variety of turtle in the world. One level outwards we now include Eunotosaurus – it branches from the “stem” of the turtle group. Then we have the group I’ve called “not turtles,” standing in for other reptiles and the like that aren’t turtles but are still nearer to them than to some other animals, which are represented by “bumblebees” because why not?

In this diagram, Eunotosaurus is closely related to turtles, but isn’t one – just like how we’re closely related to chimpanzees, but aren’t chimpanzees ourselves. If Eunotosaurus were a turtle the diagram would look more like this:

Here, Eunotosaurus comes within the turtle group, which cannot be defined in a way as to exclude it. In the real world, tortoises fit in the place where Eunotosaurus has been placed above, hence tortoises are turtles.

But this wasn’t actually the alternative that the authors had in mind. The real other arrangement, which used to be accepted but which this study argues strongly against, looks more like this:

In other words, Eunotosaurus is not closely related to turtles at all. Frankly, this is the position that Thomas should be arguing, failed though it may be. In insisting that an animal that doesn’t even have a shell is within the same “kind” as fully-shelled turtles makes a mockery of the already pseudoscientific field of baraminology.

Before we get the conclusion, here’s an interesting footnote:

In what could be translated as, “The evidence shows that this fossil turtle’s shell originated suddenly, by direct creation,” the Current Biology authors wrote, “The lack of osteoderms [bony plates embedded in skin, unconnected to the skeleton] in the recently discovered stem turtle Odontochelys semitestacea strongly supports the de novo model of shell origination…”

This gets back to the “de novo” vs composite models I mentioned earlier, and it’s clear that Thomas has no idea what they are about: those sentenses are in no way equivalent, when the context is known. Thomas also says:

Eunotosaurus already had the organization of leg and breathing muscles unique to turtles, inferred by muscle attachment points seen on its fossilized bones.

That explains the mention of muscle attachments earlier. He adds in another footnote:

No evolutionist has yet proposed a detailed hypothetical series of transitions between lizard-breathing and turtle-breathing anatomy, let alone constructing and testing models or finding appropriate transitions among fossils. Instead, hand-waving magic words like “reorganization” by “evolution” substitute for actual evidence.

It’s true that we don’t yet know how turtle breathing evolved, but as usual that doesn’t mean we never will and now that we know more about the shell we can start on that problem. As the press release says:

[Tyler] Lyson says he and his colleagues now plan to investigate various other aspects of turtles’ respiratory systems, which allow them to manage with their ribs locked up into a protective outer shell. “It is clear that this novel lung ventilation mechanism evolved in tandem with the origin of the turtle shell,” he says.

After all that, Thomas concludes:

By failing to showcase even one transitional core feature, this particular fossil best fits the creation model, which proposes that God created turtles to reproduce after their kind—not between different kinds.

On the contrary: by failing to mention the most important detail here Thomas has gone over and above his usual failings at science journalism.

…well done? It is certainly an achievement.