A bit more about dissorophids and temnospondyls

This all started
a few days ago with some interest by readers in the nesting of dissorophoids (Cacops and kin; Fig. 1) apart from temnospondyls. The large reptile tree (LRT) nested dissorophoids at the base of the lepospondyls, contra traditional studies. I tested this heretical nesting several times over and the nesting is robust. Today we’ll put that nesting to yet another test.

Here’s the problem
Cacops looks like a temnospondyl. It’s big. It has a big head, short torso and tiny tail. It was probably terrestrial, judging by the robust limbs. Even the palate looks like that of a temnospondyl. The question is: can all this be by convergence?

In this case, as in many others…
it’s better not to eyeball it, or play favorites, or follow tradition, but to let the computer decide.

Over the last few days
I’ve been combing the Internet for traditional dissorophid outgroups in the literature. Iberospondylus was one candidate, but it nested only with temnospondyls in the LRT, far from dissorophids.

Figure 1. Cacops and its sisters.

Figure 1. Cacops and some of its sisters.


Another, perhaps better candidate  
is Parioxys fericolus (Cope 1878, Carroll 1964; Early Permian). It shares several traits with Cacops, like a big curved squamosal. Cope (1882) later suggested his specimens were actually young Eryops (Fig. 2), but subsequent workers considered Parioxys a separate genus. Moustafa (1955) allied Parioxys with the Dissoroophidae in the super-family Dissorophoidea. Carroll (196) described an earlier and more primitive species (Parioxys bolli, Fig. 3).

Figure 2. Eryops, a temonspondyl, shares many traits by convergence with Cacops (fig. 1). Even the palate is a close match. This is where phylogenetic analysis really shines, separating convergent taxa from close kin.

Figure 2. Eryops, a temonspondyl, shares many traits by convergence with Cacops (fig. 1). Even the palate is a close match. This is where phylogenetic analysis really shines, separating convergent taxa from close kin.

Carroll reports,
“It is primarily on the basis of the configuration of the pelvis and the possession of two pairs of sacral ribs, as well as the lack of a fourth trochanter on the femur, that Moustafa allied Parioxys with the dissorophids.”

Among basal tetrapods, Cacops is atypical in having two sacral ribs, although Eryops has one “true sacral” and another vertebra very much like it. Carroll further notes,

Carroll continues:
“Since the features that Moustafa used to ally the dissorophids with Parioxys have developed separately within the two groups, these characters cannot be cited to indicate close relationship.
 The possession of a posterior proximal ramus of the adductor ridge in P. bolli, and the presence of a fourth trochanter, further separate the genus from dissorophids, which do not show these features even in the later Middle Pennsylvanian genera.”

Figure 3. Parioxys is a temnospondyl sister to Eryops and, despite sharing several traits, is not close to Cacops.

Figure 3. Parioxys is a temnospondyl sister to Eryops and, despite sharing several traits, is not close to Cacops in the LRT. Note the large fourth trochanter below the femur and the long ilium connecting to two sacrals, but covering three. Note the deeply curved squamosal. No complete skeleton is known yet for this genus, so this is a chimaera.  Images compiled from Carroll 1964

After phylogenetic analysis
the dissorophids remain nested at the base of the lepospondyls. Parioxys nested with Eryops. Only with the removal of ALL intervening taxa do dissorophids nest with temnospondyls, and then there is loss of resolution.

With the removal of Parioxys from the dissorophids, the former clade, Dissorophoidea,
now appears to be paraphyletic

Yet another heresy.
I know the basal tetrapod workers don’t like this new insight into temnospondyl and dissorophid relations, or rather the lack thereof. Maybe this will solve some of the problems they’ve been having on their own in phylogenetic analyses.

And add this discovery to the pile
of pterosaur origins, turtle origins, whale origins, snake origins, dinosaur origins, multituberculate origins, bat origins, diadectid origins, reptile origins and many more that the large reptile tree brings insight to. I never thought it would go this far.

As always,
if anyone can produce a taxon or a set of taxa that can attract Cacops and the dissorophids to the temnospondyls, please send them over. I am more than willing to test any serious candidates.

Carroll RL 1964. The relationships of the Rhachitomous amphibian Parioxys. American Museum Novitates 2167:1-11.
Cope ED 1878. Descriptions of extinct Batrachia and Reptilia from the Permian formation of Texas. Proc. Amer. Phil. Soc., vol. 17, pp. 505-530.
Cope ED 1882. Third contribution to the history of the Vertebrata of the Permian formation of Texas. Ibid., vol. 20, pp. 447-461.
Moustafa YS 1955. The skeletal structure of Parioxys ferricolus, Cope. Bull. Inst. d’Egypte 36: 41-76.



19 thoughts on “A bit more about dissorophids and temnospondyls

  1. So this is a case where consulting the specimens would be helpful. Parioxys is a mess, and basing a phylogeny on the published descriptions (rather than personal observations of the specimens) is going to be misleading. There are three specimens of note: the type of P. ferricolus (at AMNH), the type of P. bolli (also AMNH), and the Moustafa specimen (at MCZ). P. ferricolus is a cacopine dissorophid, probably the same thing as “Cacops woerhi.” P. bolli is a trematopid postcranium, probably attributable to Acheloma pricei from the same locality. The MCZ specimen is a crushed skull of Acheloma, again probably A. pricei. You wouldn’t know this from looking at the drawings, because the drawings do not represent what is actually preserved in the specimens.

      • As I said before, you need to sample the range of characters associated with early tetrapd evolution much more completely than you currently are.

        And again, cherry-picking taxa and/or characters to “support” a topology is methodologically unsound.

      • Adding taxa without adding characters soon hits the threshold where the tree just becomes chaotic. I’m speaking from experience. Check out stereospondylomorph phylogeny in my preprint – it’s all just random and snout length, with no phylogenetic signal in most of it.

        You are several hundred taxa beyond that threshold. That’s why your “discoveries” are always so “astounding”.

        Today I started adding characters to my matrix for the next publication.

      • No cherry-picking here. Why would you think that? I have no preconception what the data will recover. I’m just adding taxa. If you can’t deliver taxa or traits that will, in your words, “support’ a topology that you prefer, then what you’re accusing me of, you are accusing yourself of. Seriously. I’m just following established procedures and the data documents a gradual accumulation of traits, which, in the end, is what we’re all looking for.

      • David, evidently your paradigm about a ‘threshold” is wrong, as the topology documents. It is the opposite of chaotic at present. The tree is fully resolved with strong Bootstrap scores. In Science sometimes you have to give up cherished hypotheses when the data does not support them. I’m not crazy about being the pariah and the bearer of bad news, but I have to report findings, just as you would do, too.

      • Perhaps I wasn’t clear.

        If I cherry pick taxa or characters that I think might support temnospondyl monophyly and tell you to include them, that is methodologically unsound. It biases the results. I realize that you want me to hand you changes on a silver platter, but I literally cannot because to do so would be methodologically wrong.

      • Thanks, Jason, that wasn’t clear. If the tables were turned, however, I would suggest taxa that might help clarify your tree topology — if I had them already.

      • It is the opposite of chaotic at present. The tree is fully resolved

        That does not mean it’s not chaotic.

        with strong Bootstrap scores.

        How many replicates did you run?

        And please stop with the psychology, it’s beside the point and just part of talking past one another.

      • Include all the characters you can find that are parsimony-informative for your taxon sample, then merge the redundant ones. For your current taxon sample I expect there are easily 1000 useful characters, likely 2 or 3 times that.

      • For the part of the tree of interest, both my character set and taxon sample are substantially more extensive. I do in fact find some important heterodox results, and am not at all defending orthodoxy qua orthodoxy. However, temnospondyls are not polyphyletic. Thanks for your concern.

  2. Maybe this will solve some of the problems they’ve been having on their own in phylogenetic analyses.

    Uh, how would that work?

    Let me add that the Parioxys specimens are really, really difficult. The specimen Moustafa described is encrusted in the usual redbeds ironstone crust; the sutures he drew are basically clairvoyance – there’s not actually evidence for anything as bizarre as a temnospondyl with parietals that reach all the way to the front end of the eye sockets…

    • I took my data from Carroll 1964. If you have a beef about sutures, take it up with him. At this stage, that’s the literature. When and if the Ruta study sheds more light on the taxon, I will score it accordingly.

      • Taking anything up with Carroll is difficult: for practical purposes he’s dead – his dementia caught up with him a few years ago. :-(

        Is there anything about Parioxys in his ’64 paper, though? I can’t remember it being there.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s