The Titanosaurus (in?)validity project Part 1 - Saltasaurus

The whole Alamosaurus vs. Utetitan discussion on the Dinosaur Mailing Group made me curious about Titanosaurus' validity, since it's not been tested since Wilson and Upchurch (2003) declared it a nomen dubium in 2003. Their justification was "In summary, all six features forwarded by Lydekker in his diagnosis of T. indicus are now broadly distributed within Titanosauria. ... Because no diagnostic characters could be identified, T. indicus must be regarded as a nomen dubium." But you can't just take an original diagnosis and say "well, those characters are all shared with other taxa, time to sink it" and "we don't see any differences", you have to put in the work yourself to evaluate the existing differences (since no two bones are identical) and show they're at least plausibly individual/positional/ontogenetic/etc. variation. Not only has that never been done for Titanosaurus, there's never even been any suggested taxa it is identical to, of which we would need two to make it indeterminate (because only one identical taxon would make them synonyms). 

So as a little project I decided I'll start from the tip of my sauropodomorph cladogram and go stem-ward, comparing all procoelous titanosaurs with known caudals to see which, if any, could make Titanosaurus a nomen dubium. I'm not a sauropod worker, but at least titanosaur mid caudals have a pretty simple morphology with e.g. no complicated laminae, so the anatomy should be within my grasp. Here's a picture we'll be seeing a lot- the lectotype of Titanosaurus indicus.  Note it is complete except for the fragmented anterior centrum cotyle edge and incomplete neural spine, which would have included the postzygapophyses. Thought to be lost for decades until Mohabey et al. (2013), nobody has figured a posterior or dorsal view.

Titanosaurus indicus lectotype GSI 2191 in ventral (top), anterior (left), left lateral (center) (all after Lydekker, 1879) and left lateral (right) views (after Mohabey et al., 2013; scale = 50 mm).

And our first contender is Saltasaurus loricatus, from the Lecho Formation of Argentina. Described by Bonaparte and Powell (1980) and then in more detail by Powell (1992; 2003), the holotype is PVL 4017-92, a synsacrum fused to both ilia, but mid caudals are represented by up to ten paratype elements including PVL 4017-28, 4017-32 and 4017-33, figured by Powell (1992- Figs. 22-23; 2003- Plate 34, Plate 52 Fig. 3, Plate 53 Fig. 1, 5-6). 

Saltasaurus loricatus paratype PVL 4017-33 in ventral (top), anterior (left) and left lateral (right) views (after Powell, 1992; scale = 100 mm).

Immediately noticeable is despite PVL 4017-33 having a more elongate centrum (length vs. posterior height 2.16 vs. 1.73; a proxy for proximodistal position), Saltasaurus still has prominent transverse processes. Titanosaurus' vertebra is much narrower transversely, but may be taphonomically compressed. Ventrally, Titanosaurus has a median groove anteriorly and posteriorly, but the middle fourth of the ventral midline is so convex it actually protrudes ventrally as a broad keel in both drawing and photo. Both figured Saltasaurus mid caudals instead have a continuous median groove that contains a narrow keel in its middle half. The anterior chevron facets are longer than Saltasaurus', and the posterior chevron facets are distinct circles but not obvious in Saltasaurus. These are apparent in lateral view as a posteroventrally facing surface anterior to the condyle, while Saltasaurus' condyle extends to the posteroventral corner of the centrum. The prezygapophyses are more elongate (especially noticeable in ventral view), and in anterior view form tall triangular processes while Saltasaurus' are very low. The neural peduncle is placed further posteriorly in Titanosaurus (16% of centrum length past the anterior edge vs. 4% in Saltasaurus), which is consistent in all figured Saltasaurus caudals so is not positional variation. The neural peduncle also ends much more anteriorly than Saltasaurus, with 53% of the centrum behind the peduncle compared to 38-39% in Saltasaurus, which combined with the previous ratio makes a narrower peduncle base for Titanosaurus. This posterior peduncular concavity is angled in Titanosaurus (note Lydekker's drawing gets this wrong), but rounded in Saltasaurus caudals until you get to much more distal elements like PVL 4017-38 that have a ridge for a neural spine and an elongation index of 2.62. The centrum also has a deeply concave dorsal border posteriorly unlike Saltasaurus (slightly concave in distal caudals).

Note while PVL 4017-33 has a more basally positioned postzygapophysis than Titanosaurus, PVL 2017-28 does not, so this could be positional variation. Also, I'm not sure if Titanosaurus' neural spine lacks an obvious prespinal lamina distally (like Saltasaurus) and is displaced laterally, or has a thick prespinal lamina distally (unlike Saltasaurus) and is missing its right (left in the figure) side, but someone could potentially check the specimen to find out. The fragmented cotyle edge is why I think we can see the prezygapophyses meet in ventral view, so I'm not counting that as a difference.

So in conclusion, it's pretty obvious Saltasaurus is not Titanosaurus, which is why Powell erected Saltasaurinae in the first place to contrast with Titanosaurinae. His (1992) Saltasaurinae diagnosis includes "The centra of the caudal vertebrae are broader than tall ... The spines of the caudal vertebrae are posteriorly inclined", though the former could be taphonomic for Titanosaurus and the latter is not different between the genera. 

Next up, Neuquensaurus australis.

References- Lydekker, 1879. Indian pre-Tertiary Vertebrata. Part 3. Fossil Reptilia and Batrachia. Palaeontologica Indica (series 4). 1, 20-33.

Bonaparte and Powell, 1980. A continental assemblage of tetrapods from the upper Cretaceous beds of El Brete, northwestern Argentina (Sauropoda-Coelurosauria-Carnosauria-Aves).  Mémoires de la Société Géologique de France. Nouvelle Série. 19, 19-28.

Powell, 1992. Osteologia de Saltasaurus loricatus (Sauropoda - Titanosauridae) del Cretácico Superior del noroeste Argentino. In Sanz and Buscalioni (eds.). Los Dinosaurios y Su Entorno Biotico: Actas del Segundo Curso de Paleontologia in Cuenca. Institutio "Juan de Valdes", Cuenca, Argentina. 165-230.

Powell, 2003. Revision of South American titanosaurid dinosaurs: Palaeobiological, palaeobiogeographical and phylogenetic aspects. Records of the Queen Victoria Museum. 111, 173 pp.

Wilson and Upchurch, 2003. A revision of Titanosaurus Lydekker (Dinosauria-Sauropoda), the first dinosaur genus with a Gondwanan distribution. Journal of Systematic Paleontology. 1(3), 125-160.

Mohabey, Sen and Wilson, 2013. India’s first dinosaur, rediscovered. Current Science. 104(1), 34-37.