Friday, January 31, 2014

Vitakridrinda the croc and croc phylogeny

Hi everyone.  With Lori "done", let's try to get more blog posts up this year.  First is something Molnar told me about last year and gave me permission to discuss since he wasn't going to write it up.  So I wrote about it in the Database in September, then forgot to mention it once I uploaded the revised pages on New Years.  Without further ado...

Figure 1. Comparison between Vitakridrinda syntype snout MSM-155-19 (top; from Malkani, 2010) and Pabwehshi holotype snout GSP-UM 2000 (bottom; from Wilson et al., 2001).  Both in sagittal section.

Vitakridrinda's syntype snout (originally described as abelisaurian by Malkani) is a really good match for the baurusuchid crocodyliform PabwehshiPabwehshi is from the same formation and about 50% larger, though the latter measurement will vary based on how far down the snout each cross section is. The nasal diverticulum seen in Vitakridrinda is listed by Wilson et al. (2001) as an autapomorphy of Pabwehshi, but the internal anatomy of other baurusuchid snouts is poorly known, though Baurusuchus, Stratiotosuchus and Wargosuchus at least lack evidence of it in the external naris. Under Molnar's interpretation, Malkani's dorsal palatal process (dpp above) is the secondary palate, the ventral palatal process (vpp) is the mandible, the cavity between these is the oral cavity, and the cavity in the ventral palatal process is the Meckelian canal. This leaves almost all of Malkani's supposed lateral surface as matrix which needs to be prepared to expose bone. Thus his supposed bite punctures are merely grooves excavated in the matrix, while his 'combat teeth' are tooth cross sections ventral to the dentary. Malkani's scenario of intraspecific combat causing these features is thus exposed as being fanciful. While most features of Vitakridrinda's snout as described by Malkani are based on misinterpretation, it should be noted Pabwehshi has labiolingually compressed tooth crowns with mesial and distal serrations and a more convex mesial edge as in abelisaurids.

I think this is a really good observation, and have since learned that Molnar heard it from Wilson, one of Pabwehshi's coauthors.  In November, Malkani himself independently contacted me for advice with manuscripts, one of which also uses Wilson's suggestion.  He came to different taxonomic conclusions than I did, but agrees on the basic point of the snout being from a croc, which further strengthens the idea.

This of course leads to the question of what the rest of the Vitakridrinda type material belongs to.  The limb bone (femur?) segments (MSM-59-19, 60-19) are possibly theropod due to their supposed inturned head, distinct neck, and hollowness, though Malkani (2009) states a section "has fibrous bone network in the hollow", so perhaps they are not actually hollow. The stated lack of an anterior trochanter may support a crocodiliform identity, or may be due to preservation. The supposed braincase (MSM-61-19) matches Baurusuchus as poorly and ambiguously as it does abelisaurids, so remains an unidentified object. The supposed tooth cross section (MSM-62-19) works as well for Pabwehshi as it does for abelisaurids.  So the basic answer is that the published data is too crummy to tell.  This is unfortunate because exactly what specimen the name Vitakridrinda should be connected to is uncertain.  Malkani refers to all of the type material as the holotype, as he thought they all belonged to one individual.  Since the ?braincase and ?femora were found 100 meters away from each other, this seems unlikely, and one (or the snout) should really be selected as a lectotype.  Malkani's manuscript will clarify some of this if it is published, but I fear without further preparation and better photography, Vitakridrinda's possible synonymy with Pabwehshi and syntypic identification will remain uncertain.

What is Pabwehshi?

Finally, because it's me, some phylogenetic tests!  I thought it would be simple to just add Vitakridrinda/Pabwehshi to the baurusuchid part of my croc tree, but then saw a more recent paper (Montefeltro et al., 2011) examining baurusuchid phylogeny excluded it from the family.  Those authors referenced another paper (Larsson and Sues, 2007; top tree in figure 2) that found it sister to a clade of sebecids and peirosaurids instead.  They named this clade Sebecia, and noted similarity between Pabwehshi and the peirosaurid Hamadasuchus in particular, which had not been included in prior analyses that found Pabwehshi to be a baurusuchid.  And so I had to enter the confusing and conflicting world of croc phylogeny.

What I found was interesting.  First, Larsson and Sues excluded Pabwehshi a priori because "it has a sagittal torus on its maxillary palatal shelves, absent in the putative baurusuchids".  So they didn't test it themselves, and indeed if added to their matrix, Pabwehshi comes out sister to other baurusuchids.  Second, adding Hamadasuchus to those earlier analyses that placed Pabwehshi in Baurusuchidae does make Sebecia more likely than before (by 1-4 steps), though it is still unparsimonious.  Sebecia takes 8 extra steps to enforce in these early analyses.  But in none of those cases does Pabwehshi move to Sebecia, needing 1-4 additional steps to go there once Sebecia is enforced.  As analyses get larger and more recent, it takes increasing numbers of extra steps to move it, and it's always closer to sebecids than to Hamadasuchus.

Figure 2. Comparison of Larsson and Sues' (2007) topology at top to Pol et al.'s (2012) below.  Note the top one has orange sebecids sister to green peirosaurids to make Sebecia, while the bottom one based on a much larger analysis has sebecids sister to red baurusuchids instead.  They also differ in that the top one has blue neosuchians sister to Sebecia, whereas sebecians are interspersed throughout the other taxa in the bottom one to form a huge Notosuchia to leave neosuchians basal instead.  Red Pahwehshi is a sebecian on top, but a baurusuchid on bottom.

The most recent analysis (Pol et al., 2012; bottom tree in figure 2) is much larger (347 characters, 88 taxa) than Larsson and Sues' (158 characters, 33 taxa) and includes a ton more postcranial characters, and postcranial codings for taxa like Baurusuchus, Sebecus, Notosuchus and Araripesuchus that were not coded postcranially by Larsson and Sues due to a lack of material/description.  It takes 4 extra steps to move Pabwehshi to Sebecia if the latter is forced to exist, and again is close to sebecids in that case.  Forcing Sebecia to exist in the first place though only takes 5 extra steps, which isn't too bad.  However, it ends up in that uruguaysuchid+pierosaurid clade which is sister to other notosuchians.  To get it to be sister to the blue neosuchians like in Larsson and Sues' tree takes 16 more steps, so seems unlikely.  As Pol et al. describe in their paper, this is due to all of the new characters and codings for Sebecus et al.'s postcrania, so while Sebecia itself isn't dead, Larsson and Sues' placement for it may be.

Importantly, even when Sebecia is enforced, Pabwehshi still clades with baurusuchids and most analyses (including the largest ones) take 4 extra steps to move it into Sebecia. This suggests there is no sebecian signal in Pabwehshi, as does the lack of a particular position within Sebecia when it is forced to be there (sister to either Bretesuchus, Iberosuchus, or a wildcard basal sebecid). Notably it is always attracted to sebecids though, which are sister to baurusuchids in non-sebecian phylogenies, as opposed to being isolated or close to Hamadasuchus as in Larsson and Sues' paper. While Hamadasuchus is important for making Sebecia itself more likely (by 1-6 extra steps), it only made Pabwehshi more likely to be a sebecian in one analysis, and then only by one step.

And there ends our foray into croc phylogeny.  See more detail on the material and analyses at the Database entry.  I'd love to add more taxa to Pol et al.'s analysis, like the large number of notosuchians that were left out, and the newly redescribed Shartegosuchidae that seems to be basal to Neosuchia+Notosuchia, so are probably important for settling the basal relationships of those clades.  But there's only so much time, and based on the rampant miscodings and poorly defined characters in theropod analyses, I'm not confident croc analyses are any better.  Which means checking matrices, which means acquiring literature past my current 372 croc pdfs, which means lots of work, so next time it's back to theropods.

References- Wilson, Malkani and Gingerich, 2001. New crocodyliform (Reptilia, Mesoeucrocodylia) from the Upper Cretaceous Pab Formation of Vitakri, Balochistan (Pakistan). Contributions from the Museum of Paleontology. The University of Michigan. 30(12), 321-336.

Larsson and Sues, 2007. Cranial osteology and phylogenetic relationships of Hamadasuchus rebouli (Crocodyliformes: Mesoeucrocodylia) from the Cretaceous of Morocco. Zoological Journal of the Linnean Society. 149, 533-567.

Malkani, 2009. New Balochisaurus (Balochisauridae, Titanosauria, Sauropoda) and Vitakridrinda (Theropoda) remains from Pakistan. Sindh University Research Journal (Science Series). 41(2), 65-92. 

Montefeltro, Larsson and Langer, 2011. A new baurusuchid (Crocodyliformes, Mesoeucrocodylia) from the Late Cretaceous of Brazil and the phylogeny of Baurusuchidae. PLoS ONE. 6, e21916.

Pol, Leardi, Lecuone and Krause, 2012. Postcranial anatomy of Sebecus icaeorhinus (Crocodyliformes, Sebecidae) from the Eocene of Patagonia. Journal of Vertebrate Paleontology. 32, 328-354. 


  1. This is very interesting! And very reasonable.

  2. No surprise to me. For Aegisuchus, I played with some of these analyses and found that constraining alternative positions (in Neosuchia, Notosuchia, or outside of both clades) for all of these taxa (and peirosaurs and araripesuchids) were all very possible arrangements (or even equally parsimonious). More characters, more taxa, and more careful coding are needed. That's why I threw up my hands and chose to focus only on where do aegyptosuchids fall in the 'relative frameworks' rather than trying to fix the problem.

    There's an outside issue as well where rearranging thalattosuchians basal to mesosuchians may also have some influence on the stability of 'definitive' neosuchians and what other taxa can float in/out around them (-I think- mostly positive, but don't recall what I found).

  3. Hi Mickey,
    I was interested in your view of the phylogeny of Neovenatoridae and Megaraptora.I'm trying to update my
    Dinosaur catalogs and these two are a problem every year.(We all know that paleontologist agree to disagree.)
    I respect your judgement.Please help

    1. This depends on whether you mean the phylogeny of Megaraptora, or their position in Theropoda.

      For the former, the main disagreement seems to be whether Australovenator is closer to Megaraptor+Aerosteon (Novas et al., 2013) or to Fukuiraptor (Benson's analyses). The former seems to have more character evidence (7 vs. 2), but I haven't checked any of them in detail, so have no strong opinion.

      For the latter, I've been a proponent of coelurosaurian megaraptorans for years. Benson's analyses always have a lack of coelurosaurs (3 taxa), so I've questioned their finding megaraptorans in Carcharodontosauridae. Novas et al. (2013) recently proposed megaraptorans were tyrannosauroids, but fall for the same problem. Besides tyrannosauroids, the non-maniraptoriform coelurosaurs consist only of Santanaraptor, Sinosauropteryx, Compsognathus and Tanycolagreus. This is better than Benson, but there are also such taxa as Bicentenaria, Zuolong, Tugulusaurus, Coelurus, Sinocalliopteryx, Juravenator, Ornitholestes and Scipionyx that would have important influence on their cladogram. For instance, their second tyrannosauroid character (shared with Megaraptor) is metacarpal III <75% length of metacarpal II. Yet this is also present in Huaxiagnathus and Sinocalliopteryx, but not in Guanlong or seemingly Dilong. So while I haven't examined most of their proposed characters, I'm skeptical.

      As for Neovenator itself, I'm happy having it as a basal caqrcharodontosaurid.