Scathing Book Review – Pterosaurs (Witton 2013) Phylogeny (family tree) issues

Earlier we discussed many of the sins of omission and commission in the new book, “Pterosaurs” by Mark Witton (2013). Here we’ll take a look at the family tree of pterosaurs that Witton presents. It also includes several sins of omission and commission.

Figure 1. Pterosaur phylogenies (above) according to Witton (2013) based on Unwin and Lü 2005, Below, simplified results of the large pterosaur tree based on specimens. Note there are no outgroups in Witton's tree. In Witton's tree there are many strange bedfellows and no tiny pterosaurs (in pink), which is the main reason why the trees look so different.

Figure 1. Pterosaur phylogenies according to Witton (2013) based on suprageneric taxa employed by Unwin and Lü 2005 (above) versus simplified results of the large pterosaur tree by Peters (below) based on specimens (fig. 2), but here simplified by lumping specimens into clades. There are no outgroups in Witton’s tree. In Witton’s tree there are many strange bedfellows (see text) and no tiny pterosaurs (in pink). Countra the paradigm, the tiny pterosaurs don’t all nest together, but rather at the bases of distinct clades of larger pterosaurs according to their traits.

Witton’s (2013) book includes a family tree of the Pterosauria (Fig. 1 above). Unfortunately it has no outgroups. As Witton noted earlier, he follows traditional paleontologists in assuming, without testing, that pterosaurs nested with archosaurs. They don’t. On the other hand, the large pterosaur tree and large reptile tree (modified from Peters 2007, also see Fig. 6) includes more than a dozen outgroup taxa within and beyond the Lepidosauriformes, all demonstrating a gradual serial acquisition of pterosaurian traits.

Sins of Commission
Witton’s two basal taxa (Dimorphodontidae and Anurognathidae) do not include any Triassic taxa. So his Jurassic pterosaurs are more primitive than his Triassic pterosaurs.

Witton’s third clade up, the Campylognathoididae, includes Triassic pterosaurs with many multi-cusped teeth. This is a primitive trait inherited from Longisquama, a basal fenestrasaur. Witton’s nesting of campylognathoidids makes it an unlikely sister clade to Anurognathidae, which have only a few widely spaced simple teeth. (Fig. 2). Eudimorphodon does not preserve a tail, but other campylognathoidids have the largest tails among all pterosaurs, which is distinct from anurognathids. In the large pterosaur tree (Fig. 6), the common ancestor of these two clades is closer to the base of all pterosaurs, deep in the Triassic.

Figure 2. Click to enlarge. Basal anurognathid (Dendrorhynchoides) compared to basal campylognathoidid (Eudimorphodon). Witton considers these clades to be sister clades. The large pterosaur tree did not confirm these results.

Figure 2. Click to enlarge. Basal anurognathid (Dendrorhynchoides) compared to basal campylognathoidid (Eudimorphodon). Witton considers these two to be members of sister clades. The large pterosaur tree did not confirm these results.

Witton follows earlier workers in lumping super-sharp-snouted pteranodontids with toothy ornithocheirids (Fig. 3) rather than with other sharp-snouted germanodactylids. Witton’s nesting is based largely on the warped deltopectoral crest, which basal taxa in both clades don’t have. And this is the danger that comes from using suprageneric clades.

Figure 3. Comparing toothy ornithocheirids to toothless pteranodontids. There is convergence here, but the large pterosaur tree demonstrates these two clades are not related until you get back to Scaphognathus.

Figure 3. Click to enlarge. Comparing toothy ornithocheirids to toothless pteranodontids. There is convergence here, but the large pterosaur tree demonstrates these two clades are not related until you get back to Scaphognathus.

Witton’s tree lumps long-necked, long-legged azhdarchids and chaoyangopterids together with short-necked, short-legged tapejarids and thalassodromids largely based on the shared trait of an enlarged antorbital fenestra (Fig. 4). The large pterosaur tree found this trait to be convergent between tapejarids and azhdarchids and the tiny anurognathid, Anurognathus.

Tapejarids and thalassodromids vs azharchids and chaoyangopterids.

Figure 4. Click to enlarge. Tapejarids and thalassodromids vs azharchids and chaoyangopterids. The former find closer relations with Germanodactylus cristatus according to the large pterosaur tree. The later find closer relations with n44 and n42 among the tiny pterosaurs.

Witton’s tree nests sharp-snouted dsungaripterids with toothy ctenochasmatids based on the presence of a sometimes shared rostral crest (Fig. 5). Considering the many differences between these two strange bedfellows, there are much better matches, according to the large pterosaur tree results.

Figure 6. Ctenochasma and Germanodactylus. Hard to imagine two more different pterosaurs, yet Witton's tree nests these two as sister clade members.

Figure 5. Ctenochasma and Germanodactylus. Hard to imagine two more different pterosaurs, yet Witton’s tree nests these two as sister clade members, basal Lophocratia.

Witton’s Wukongopteridae includes Darwinopterus, his transitional taxon with a pterodactyloid head and neck and rhamphorhynchoid tail and toes. No taxa in Witton’s tree are shown to gradually evolve toward Darwinopterus or beyond it. The Peters tree (Fig. 1 below) nests the sterile clade, Wukongopteridae, with Scaphognathus (Fig. 6), which begets many descendant clades. In that sense, Darwinopterus and its kin sort of jumped the gun, gaining certain pterodactyloid traits before the Pterodactloidea actually arose.

The Sin of Omission – The Tiny Pterosaur Issue
Witton’s tree (Fig. 1) includes no tiny pterosaurs. Peters’ tree found tiny pterosaurs to be transitional taxa  at the bases of several clades (Fig. 1). Certain paleontologists have strongly suggested that tiny pterosaurs should not be included in phylogenetic analysis. Deleting them is no problem. Here (Fig. 6) are the results. In detail the tree topology changes very little. Overall, however, a monophyletic Pterodactyloidea is recovered. That’s a big difference. See notes below.

The large pterosaur tree with 40+ tiny pterosaurs deleted. The result now includes a monophyletic "Pterodactloidea." Otherwise, not too many changes.

Figure 6. The large pterosaur tree with 40+ tiny pterosaurs deleted. The result now includes a monophyletic “Pterodactloidea.” Otherwise, not too many changes from the original large pterosaur tree.

When you delete 40+ tiny pterosaurs
from the large pterosaur tree you get 1050+ most parsimonious trees (Fig. 6). There is some loss of resolution here and there, chiefly in Dorygnathus. The biggest change is the return of a monophyletic and traditional Pterodactyloidea in the abridged tree. Here (Fig. 6) Darwinopterus nests with Scaphognathus in a distinct clade, far from the base of the new “Pterodactyloidea.” The little pterosaur nicknamed “Rhamphodactylus” (Fig 7) nests as the basalmost pterodactyloid in the abridged tree.

Distinct from the Witton tree, here (Fig. 6) all the sharp-nosed taxa nest together. Tapejarids and dsungaripterids are derived from Germanodactylus. Azhdarchids and ctenochasmatids are basal clades close to Beipiaopterus, Huanhepterus and the outgroup Dorygnathus. Without tiny pterosaurs, it’s still a large morphological leap from Dorygnathus to “Rhamphodactylus.” And… there is a size reduction (Fig. 7).

Figure 7. Dorygnathus and Rhamphodactylus. By deleting tiny pterosaurs this becomes the transition point between basal pterosaurs and "Pterodactyloids."

Figure 7. Dorygnathus and Rhamphodactylus. By deleting tiny pterosaurs this becomes the transition point between basal pterosaurs and “Pterodactyloids.”

By deleting tiny pterosaurs, anurognathids are still derived from dimorphodontids. Rhamphorhynchus is still derived from Campylognathoides.

The Pterodactylus clade becomes topsy-turvy with the former larger derived taxa nesting at the base and giving rise to smaller taxa. We’ve seen that flip happen before, in Mortimer’s theropods.

The Germanodactylus clade (including Pteranodontia, Dsungaripteridae and Tapejaridae) remains essentially untouched with tiny pterosaurs gone, but then it never included many tiny pterosaurs, except at its base.

So
Tiny pterosaurs are not essential for uncovering many pterosaurian lineages, but if you want full resolution in your family tree, they are key. These transitional taxa need to be included, and no longer ignored, in future analyses. The Peters’ tree, based on more specimens, has more gradual changes and fewer strange bedfellows than the Witton tree (Fig. 1), based on suprageneric taxa and too few characters.

References
Peters D 2007. The origin and radiation of the Pterosauria. Flugsaurier. The Wellnhofer Pterosaur Meeting, Munich 27.
Witton M. 2013. Pterosaurs. Princeton University Press. 291 pages.

4 thoughts on “Scathing Book Review – Pterosaurs (Witton 2013) Phylogeny (family tree) issues

    • Actually, no. I’m talking strange bedfellows here. There are better sister and ancestral taxa in the large pterosaur tree that avoid these obvious morphological mismatches. A gradual blend is better than a jarring change.

  1. “MISCONCEPTION: Taxa that are adjacent on the tips of phylogeny are more closely related to one another than they are to taxa on more distant tips of the phylogeny.”
    That is part of your “strange bedfellows” hypothesis. You argue that they are closely related so they must look alike.

    “MISCONCEPTION: Taxa that are nearer the bottom or left-hand side of a phylogeny represent the ancestors of the other organisms on the tree.”
    You have directly stated that you think certain taxa you have tested are ancestral to other taxa in your tree. In fact you repeat this error in your response to my link (did you read it?) when you say, “There are better…ancestral taxa in the large pterosaur tree…” .

  2. If it would make you feel any better, I’ll try to say more often, “sisters to the ancestors.” By ancestral taxa, I mean “from among these taxa you’ll find ancestors.”

    Moreover as a dictum to evolution, closely related taxa DO look more alike*, and share more traits, than those who are less closely related. That’s why they nest together in analyses. That’s how we give them names and clades.

    The misconception you refer to above is actually true, more or less. But, as you know, scientists like to be as ultra-precise and foolproof as possible, so we like to say unknown “sister” to that known taxon down and to the left. That gets a little long if you do it often enough. The common ancestor is never known precisely, but you can get pretty damn close. And each of those taxa down and to the left DO give the best possible take on that unknown ancestor, as you already know. It’s no so much a miscalculation as a shortcut. I know you don’t believe everything you read, hence your note, but just because it’s on the ‘net doesn’t mean it’s gospel (referring to your “misconceptions.” You still have to pull the meaning out for every situation and circumstance and when we all speak the same language, which we do, then verbal shortcuts are understood by all concerned.

    * The large reptile tree and the large pterosaur tree are currently large enough to handle incredible convergence, such as when you have several turtle-like taxa with shells, the one most closely related, Stephanospondylus, doesn’t even have a carapace or plastron! The suite of other traits nests it. So, when I say something looks more alike, I mean in every measurable way compared to all the other 338+ taxa on the tree, that is, according to the computer software results.

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 )

Google photo

You are commenting using your Google 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 )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.