Azhdarchoid topology is surprisingly stable (Cerqueira et al.)

I’ve been helping a friend on a project involving pterosaur phylogeny. One thing I suggested doing was analyzing which characters are informative or uninformative for azhdarchoids. I decided to run the recent analysis of Cerqueira et al.and see if removing successively more distant sister taxa to Azhdarchoids affected the internal topology of Azhdarchoidea. Here’s what I found. The internal topology of Azhdarchoidea is surprisingly consistent regardless of whether the following combinations of taxa are removed from the matrix:

  1. Pteranodontoids save Pteranodon, Muizquiopteryx, and Nyctosaurus;
  2. The three aforementioned pteranodontoids;
  3. All outgroups except just Rhamphorhynchus and Darwinopterus spp. as well as removing these and only using Pterodactylus as the outgroup;

The topology we get looks like this:

This made me wonder, especially within tapejaromorphs, which taxa were responsible for the tree instability. So I tested this further. I experimented with deleting the tapejaromorphs that were in a polytomy with tapejarids and thalassodromids as well as trying to resolve the relationships inside sinopterines.

Deleting Caupedactylus moves Aymberedactylus to be the sister to sinopterines+tapejarids, but leaves a giant polytomy for that group with no resolved internal nodes. Deleting Aymberedactylus still leaves Caupedactylus in a polytomy with thalassodromines and tapejarids, but restores the tapejarine node. Deleting both taxa reduces tapejarines and sinopterines back to a giant polytomy, just like when Caupedactylus was deleted. Clearly Aymberedactylus is important for the internal resolution of tapejarines+sinopterines.

If Eopteranodon lii is deleted, sinopterines all nest together, whereas its inclusion but trying to delete Sinopterus or Huaxiapterus corollatus does not help. However, Huaxiapterus benxiensis’s deletion results in Huaxiapterus corollatus (Eopteranodon + Sinopterus) suggesting there are character conflicts between Huaxiapterus benxiensis and Eopteranodon lii as far as which taxon is closer to Sinopterus. This led me back to testing the removal of Caupedactylus and Aymberedactylus again. Clearly, even though while Aymberedactylus is apparently likely close to tapejarids, Caupedactylus is important for resolution within tapejarids.

Obviously, this points to a need to explore which character scorings of these taxa are responsible for these conflicts as well as explore and seek out additional characters which might help resolve the conflict between these tapejaromorphs. None of these taxa really should be deleted a priori or excluded from consideration when creating a super matrix of azhdarchoids.

It’s worth noting that once the ornithocheiromorphs (sensu Andres et al.) are deleted, computational times on this matrix drastically shrink, so most of the time to analyze the tree is spent trying to resolve the position of ornithocheiromorphs (and this is true, even if more fragmentary ones are deleted, but others are retained).

Just my two cents…


Andres, Brian, et al. “The earliest pterodactyloid and the origin of the group.” Current Biology, vol. 24, no. 9, 2014.

Cerqueira, Gabriela M., et al. “A new azhdarchoid pterosaur from the Lower Cretaceous of Brazil and the paleobiogeography of the Tapejaridae.” Acta Palaeontologica Polonica, vol. 66, 2021, in press. (WARNING: PDF LINK)

I write about the evolution and ecology of animals from the past (and sometimes the present). I also sometimes write about library and information science.