Sunday, February 11, 2018

Titanosaurs all the way down

There are a lot of titanosaurs. Over at The Compact Thescelosaurus, there are currently 101 species within Titanosauria, and another 30 non-titanosaurian somphospondyls, which probably include a few things that will be eventually be classified within Titanosauria. (If you're unfamiliar with the term, "somphospondyls" will take some explanation, which I'll get to in a minute; also, "somphospondyl" is a truly unappealing word.) Together they make up a little less than nine percent of the dinosaur chart. Also, as of this weekend, I've removed all of the internal divisions in Titanosauria; it's just titanosaurs all the way down. This is not an admission that all titanosaurs were alike, but rather a recognition that we are still a long way from knowing how they were related to each other.

It's not as though nobody is looking at titanosaurs; with several new species a year lately, we've been getting numerous studies. Titanosaurs are a growth industry. The problem is that different authors have a tendency to come up with significantly different phylogenies. To illustrate this, I transcribed examples from a group of publications into ASCII cladograms and turned those into figures, so they should look the same on everyone's devices. Obviously, you're going to need to expand the figures to read them. I also generally only selected one phylogeny from each paper, but there might be multiple trees in a given paper for different methodologies. The publications I chose all came from at least 2016:

Bandeira et al. (2016)
González Riga et al. (2016)
Poropat et al. (2016)
Carballido et al. (2017)
Carvalho et al. (2017)
Gorscak et al. (2017)
Tykoski and Fiorillo (2017)
González Riga et al. (2018)
Sallam et al. (2018)

Part 1, with González Riga et al. 2016, Bandeira et al. 2016, Poropat et al. 2016, Carballido et al. 2017, and Tykoski and Fiorillo 2017.

Part 2, with Gorscak et al. 2017, González Riga et al. 2018, Sallam et al. 2018 (-ulna and +ulna), and Carvalho et al. 2017.

Basically, I took everything after the split with Brachiosauridae, and did not add any higher-level classifications except to lump euhelopodids if they showed up. If you'd like to play "Pin the Name on the Clade", there are definitions below from D'Emic (2012), Carballido et al. (2017), and Tykoski and Fiorillo (2017). I'm using mathematical symbols to simplify things a bit. "+", as in "Brachiosaurus altithorax + Saltasaurus loricatus", means the clade made up of the most recent common ancestor of those two species plus all of its descendants (a node, by the way). "<" or ">", as in "Brachiosaurus altithorax < Saltasaurus loricatus", means the clade of all forms more closely related to Saltasaurus than to Brachiosaurus (a stem or branch). Although the definitions stay the same, in practice the membership of many of these clades can vary greatly from publication to publication.

Titanosauriformes: Brachiosaurus altithorax + Saltasaurus loricatus
Somphospondyli: Brachiosaurus altithorax < Saltasaurus loricatus; the opposite clade is Brachiosauridae. By definition, all titanosauriforms are either one or the other, but we aren't sure which side some of them are on.
Euhelopodidae: Euhelopus zdanskyi > Neuquensaurus australis
Titanosauria: Andesaurus delgadoi + Saltasaurus loricatus
Lognkosauria: Futalognkosaurus dukei + Mendozasaurus neguyelap
Rinconsauria: Muyelensaurus pecheni + Rinconsaurus caudamirus
Eutitanosauria: Epachthosaurus sciuttoi < Saltasaurus loricatus
Lithostrotia: Malawisaurus dixeyi + Saltasaurus loricatus
Saltasauridae: Opisthocoelicaudia skarzynskii + Saltasaurus loricatus
Saltasaurinae: Opisthocoelicaudia skarzynskii < Saltasaurus loricatus
Opisthocoelicaudiinae: Opisthocoelicaudia skarzynskii > Saltasaurus loricatus
Aeolosaurini: Aeolosaurus rionegrinus + Gondwanatitan faustoi

After doing this for a while, my mind began to wander into philosophical questions like "Why is Titanosauria placed where it is?" and "Brachiosauridae really is just a historical accident, isn't it? We have this split between the brachiosaurs and the titanosaur line, and there's like 10 brachiosaurid species on a good day versus at least 130 of the other guys. Big whoop." Also, I realized I couldn't think of a single useful thing to go with about half of the genera in the cladograms, except maybe the general geography.

A few punchlines:
1) Euhelopodidae doesn't show up as often as I would have thought. When it doesn't, Euhelopus is basal to Titanosauriformes (so it doesn't show up in the sections I transcribed). Interestingly, Carballido et al. (2017) came up with essentially Euhelopodidae minus Euhelopus;

2) Somehow, despite the chaos raging around it, Andesaurus always manages to show up in about the same place with respect to everyone else. This means that the membership of Titanosauria is relatively stable, despite whatever antics the other titanosaurians are getting up to. It's good to know who you can count on;

3) On the other hand, Epachthosaurus and Malawisaurus, key taxa for Eutitanosauria and Lithostrotia, do whatever they feel like. In some phylogenies, Eutitanosauria is bigger than Lithostrotia because Epachthosaurus is more basal than Malawisaurus (González Riga et al. 2016, Tykoski and Fiorillo 2017, Carballido et al. 2017). In some phylogenies, Lithostrotia is bigger than Eutitanosauria because Malawisaurus is more basal (Bandeira et al. 2016, Poropat et al. 2016, Gorscak et al. 2017, González Riga et al. 2018, both Sallam et al. 2018 results). In one phylogeny, Eutitanosauria and Lithostrotia are the same because the two genera are in a polytomy (Carvalho et al. 2016). This suggests to me that a) it may be helpful to only use one or the other of Eutitanosauria and Lithostrotia; and b) Epachthosaurus and Malawisaurus may not be ideal anchor taxa. Unfortunately, it seems like there is only one Andesaurus when it comes to reliability;

4) The smaller clades tend to be useful for their two anchor taxa, e.g., Muyelensaurus and Rinconsaurus are usually near each other for Rinconsauria, as well as Futalognkosaurus and Mendozasaurus for Lognkosauria and Saltasaurus and Opisthocoelicaudia for Saltasauridae (except for Sallam et al. 2018, in which Saltasauridae is larger than Eutitanosauria). Supporting casts are more variable. Neuquensaurus usually hangs out with Saltasaurus. Interestingly, Isisaurus and Rapetosaurus tend to be found in the vicinity of each other;

5) Finally, taxon selection for titanosaur phylogenies varies quite a bit.

So, what is the deal with titanosaurians? Three factors come to mind:

1) Many taxa are only known from very incomplete remains
Stop me if you've heard this one before: the holotype of [insert your favorite dummy name here] is three dorsal vertebrae and a rib. Titanosaurs are getting better for skulls and partial skeletons, but a lot of species are still only known from one specimen, consisting of a handful of bones. We've only glimpsed at titanosaurian anatomical diversity.

2) There are so many, and most of them are recent names
It's already an uphill battle when you have as many as a hundred species to consider within Titanosauria. It just gets harder when every year a few more are described, so that in the time between writing a paper, submitting it, addressing reviews, and finally getting it published, there are already a few more species. 27 of the 101 species I currently have listed as titanosaurians have been published from 2011 to the present. To put that in perspective, the total number of stegosaurian species in the sheet is 32, and the total number of choristodere species is 37. Aside from the volume of new descriptions, there is also the simple fact that when a name is new, there hasn't yet been time for further consideration. You end up with a bunch of species with one paper under their belts.

3) They're everywhere
No well-established clade of non-avian dinosaurs can match the titanosaurs for their combination of numbers and distribution. Many dinosaur groups are conveniently localized to a few regions, especially the Cretaceous groups that get the most love (tyrannosaurs, ceratopsians, hadrosaurids, etc.). If you're going to do a comprehensive study of titanosaurs that includes studying original specimens, you'd better have a heck of a travel budget, or the ability to organize and coordinate a widely dispersed multinational team. Titanosaurs have been named from 20 countries on all continents except Antarctica (which you can add if you count unnamed fossils). Having large numbers of specimens shipped to you is not logistically feasible, either; titanosaurs do not fit in envelopes!

Maybe a solution is some sort of vast shadowy titanosaur collaboration, wherein a group of researchers decides on a bunch of characters to measure and standards for measuring them, and pools their data.


Bandeira, K. L. N., F. Medeiros Simbras, E. Batista Machado, D. de Almeida Campos, G. R. Oliveira, and A. W. A. Kellner. 2016. A new giant Titanosauria (Dinosauria: Sauropoda) from the Late Cretaceous Bauru Group, Brazil. PLoS ONE 11(10):e0163373. doi:10.1371/journal.pone.0163373.

Carballido, J. L., D. Pol, A. Otero, I. A. Cerda, L. Salgado, A. C. Garrido, J. Ramezani, N. R. Cúneo, and J. M. Krause. 2017. A new giant titanosaur sheds light on body mass evolution among sauropod dinosaurs. Proceedings of the Royal Society B: Biological Sciences 284(1860):20171219.

Carvalho, I. S., L. Salgado, R. M. Lindoso, H. I. de Araújo-Júnior, F. C. Costa Nogueir, and J. A. Soares. 2017. A new basal titanosaur (Dinosauria, Sauropoda) from the Lower Cretaceous of Brazil. Journal of South American Earth Sciences 75:74–84.

D'Emic, M. D. 2012. The early evolution of titanosauriform sauropod dinosaurs. Zoological Journal of the Linnean Society 166(3):624–671.

González Riga, B. J., M. C. Lamanna, L. D. Ortiz David, J. O. Calvo, and J. P. Coria. 2016. A gigantic new dinosaur from Argentina and the evolution of the sauropod hind foot. Scientific Reports 6:article number 19165. doi:10.1038/srep19165.

Gonzàlez Riga, B. J., P. D. Mannion, S. F. Poropat, L. D. Ortiz David, and J. P. Coria. 2018. Osteology of the Late Cretaceous Argentinean sauropod dinosaur Mendozasaurus neguyelap: implications for basal titanosaur relationships. Zoological Journal of the Linnean Society in press.

Gorscak, E., P. M. O'Connor, E. M. Roberts, and N. J. Stevens. 2017. The second titanosaurian (Dinosauria: Sauropoda) from the middle Cretaceous Galula Formation, southwestern Tanzania, with remarks on African titanosaurian diversity. Journal of Vertebrate Paleontology 37(4):e1343250.

Poropat, S. F., P. D. Mannion, P. Upchurch, S. A. Hocknull, B. P. Kear, M. Kundrát, T. R. Tischler, T. Sloan, G. H. K. Sinapius, J. A. Elliott, and D. A. Elliott. 2016. New Australian sauropods shed light on Cretaceous dinosaur palaeobiogeography. Scientific Reports. 6:article number 34467. doi:10.1038/srep34467.

Sallam, H. M., E. Gorscak, P. M. O’Connor, I. A. El-Dawoudi, S. El-Sayed, S. Saber, M. A. Kora, J. J. W. Sertich, E. R. Seiffert, and M. C. Lamanna. 2018. New Egyptian sauropod reveals Late Cretaceous dinosaur dispersal between Europe and Africa. Nature Ecology & Evolution in press.

Tykoski, R. S., and A. R. Fiorillo. 2017. An articulated cervical series of Alamosaurus sanjuanensis Gilmore, 1922 (Dinosauria, Sauropoda) from Texas: new perspective on the relationships of North America's last giant sauropod. Journal of Systematic Palaeontology 15(5):339–364.


  1. Great post, I've been looking at this problem from the outside for years, wishing I had the time and a mythical benefactor who could fly me around the world several times over to see all the material.
    re. Malawisaurus. I'm fairly sure there is a big problem of association with this taxon. We generally associate the rather basal looking jaws and skull pieces with the postcranial bits more or less by historical accident. However I think the idea that the more derived 'Karongasaurus' jaw and teeth (found after the traditional Malawisaurus association had been established) better fits with the bulk of the rather advanced looking postcranium.

    1. Thank you very much!
      Re: Karongasaurus and Malawisaurus: Then, of course, the one publication to include Karongasaurus (Sallam et al. 2018) finds it more basal than Malawisaurus, at least in the Bayesian trees they focus on. The titanosaurs are not making it easy to figure them out!

  2. Re: Malawi dinosaur revisions, I have a paper in revision, stay tuned.

  3. Nice article - puts the sheer 'speciosity' of titanosaurs in perspective.
    As a general rule, sauropod workers have been very reluctant to erect (or resurrect) family-level taxa for these titanosaur clades. Available names like Antarctosauridae, Argyrosauridae, Andesauridae, and Aeolosauridae have been avoided. Instead, we have names like Lognkosauria and Rinconsauria (which you mention). This is undoubtedly due to the instability of titanosaur phylogeny.

    1. We do have Saltasauridae - but it's a very narrow clade.
      With Nemegtosaurus and Opisthocoelicaudia now likely synonyms, it'll be interesting to see what impact this has on titanosaur phylogeny.