Field of Science

Asilisaurus kongwe, the Oldest Avian-line Archosaur and the Early Diversification of Ornithodira

Nesbitt, S.J., Sidor, C.A., Irmis, R.B., Angielczyk, K.D., Smith, R.M.H., and L. A. Tsuji. 2010. Ecologically distinct dinosaurian sister group shows early diversification of Ornithodira. Nature. doi:10.1038/nature08718

Abstract - The early evolutionary history of Ornithodira (avian-line archosaurs) has hitherto been documented by incomplete (Lagerpeton) or unusually specialized forms (pterosaurs and Silesaurus). Recently, a variety of Silesaurus-like taxa have been reported from the Triassic Period of both Gondwana and Laurasia, but their relationships to each other and to dinosaurs remain a subject of debate. Here we report on a new avian-line archosaur from the early Middle Triassic (Anisian) of Tanzania. Phylogenetic analysis places Asilisaurus kongwe gen. et sp. nov. as an avian-line archosaur and a member of the Silesauridae, which is here considered the sister taxon to Dinosauria. Silesaurids were diverse and had a wide distribution by the Late Triassic, with a novel ornithodiran bauplan including leaf-shaped teeth, a beak-like lower jaw, long, gracile limbs, and a quadrupedal stance. Our analysis suggests that the dentition and diet of silesaurids, ornithischians and sauropodomorphs evolved independently from a plesiomorphic carnivorous form. As the oldest avian-line archosaur, Asilisaurus demonstrates the antiquity of both Ornithodira and the dinosaurian lineage. The initial diversification of Archosauria, previously documented by crocodilian-line archosaurs in the Anisian, can now be shown to include a contemporaneous avian-line radiation. The unparalleled taxonomic diversity of the Manda archosaur assemblage indicates that archosaur diversification was well underway by the Middle Triassic or earlier.

Life restoration of Asilisaurus with sail-backed poposauroid in the background. Image by M.H. Donnelly (Field Museum).

Prior to 2003 the non-dinosaurian dinosauriforms known as silesaurids were unrecognized in the fossil record. Specimens existed in collections, collected as early as the 1930s, while others were given tentative identifications (e.g., the ornithosuchid of Long and Murry, 1995 and the ornithischian Technosaurus). Dzik (2003) described the first, Silesaurus opolensis from the Carnian of Poland, with its very distinctive femoral and morphologies. Suddenly similar forms were recognized from all over the globe (e.g., Eucoelophysis, Sacisaurus; Ezcurra 2006; Nesbitt et al. 2007; Irmis et al. 2007a), whereas new specimens were being discovered from the Chinle Formation of New Mexico and Arizona (Parker et al. 2006; Irmis et al. 2007b).

Still, because the earliest pseudosuchian archosaurs were known from the Anisian (e.g., the Moenkopi Formation of Arizona), whereas the earliest ornithodirans were from the Ladinian of Argentina, there was a proposed ghost lineage for Ornithodira existing back into the Anisian.

Sterling Nesbitt looks over the Ruhuhu Valley in 2007. Photo by L. Tsuji.

One of the strengths of phylogentic analysis is the ability to make predictions as to where in time and space certain groups should and could be found. Sterling Nesbitt clearly recognized the strong possibility that the earliest representatives of ornithodira could be found in the Manda beds of Tanzania as this fauna was known for its pseudosuchian constituents. A few years back this prediction paid off, as a team lead by Christian Sidor in 2007 to explore the Permian and Triassic rocks of Tanzanian, uncovered an amazing deposit of early archosaurs including numerous specimens of a silesaurid. Named Asilisaurus kongwe (Ancient ancestor lizard) these specimens represent the earliest known member of the lineage leading to dinosaurs and strongly supports the diversification of the Archosauria by the early Middle Triassic (~243 million years ago).

The tibia of Asilisaurus, following excavation in 2007. Photo by R. Smith.

This find also strongly suggests that adaptations for an omnivorous or herbivorous diet evolved independently in silesaurids, ornithischians, and sauropodomorphs, from carnivorous ancestors. The paper also phylogenetically defines the Silesauridae and proposes that the South American archosaurs Lewisuchus and Pseudlagosuchus are members of this clade and probably also synonymous. Another early appearance of a non-dinosaurian dinosauriform in
Gondwana also provides further support for a southern origin for this group.

Skeletal reconstruction of Asilisaurus, with missing bones in gray. Image by S. Nesbitt.

One final point, silesaurs have been proposed to be basal ornithischians (e.g., Ferigolo & Langer, 2007) and represent the Late Triassic radiation of that clade. However, if this is the case, despite the lack of support in phylogentic analyses (e.g., Irmis et al., 2007b), the discovery of early Middle Triassic ornithichians would pull the split with saurischians back as early as the Early Triassic and create a sizeable ghost lineage for the base of Saurischia.
More photos and information on Ruhuhu Basin Research can be found here:


Dzik, J. 2003. A beaked herbivorous archosaur with dinosaur affinities from the early Late Triassic of Poland. Journal of Vertebrate Paleontology 23:556-574.

Ezcurra, M.D. 2006. A review of the systematic position of the dinosauriform archosaur Eucoelophysis baldwini from the Upper Triassic of New Mexico, U.S.A. Geodiversitas 28:649-684.

Ferigolo, J. and M. Langer. 2007. A late Triassic dinosauriform from south Brazil and the origin of the ornithischian predentary bone. Historical Biology 19:23-33.

Irmis, R.B., Parker, W.G., Nesbitt, S.J., and J. Liu. 2007a. Early ornithischian dinosaurs: the Triassic record. Historical Biology 19:3-22.

Irmis, R. B., Nesbitt, S. J., Padian, K., Smith, N. D., Turner, A. H., Woody, D., and A. Downs. 2007a. A Late Triassic dinosauromorph assemblage from New Mexico and the rise of dinosaurs. Science 317:358-361.

Nesbitt, S.J., Irmis, R.B., and W.G. Parker. 2007. A critical re-evaluation of the Late Triassic dinosaur taxa of North America. Journal of Systematic Palaeontology 5:209-243.

Parker, W.G., Irmis, R.B., and S.J. Nesbitt. 2006. Review of the Late Triassic dinosaur record from Petrified Forest National Park, Arizona. Museum of Northern Arizona Bulletin 62:160-161.


  1. Looks like an awesome paper. I just have one quibble. If silesaurids, ornithischians, and sauropods are all successive outgroups of the line leading to theropods, doesn't that mean the most parsimonious explanation is Dinosauriformes (minus Marasuchus - is there a name for this clade yet?) were basally herbivorous/omnivorous. and theropods secondarily re-evolved hypercarnivory?

    It seems like this is the most sensible explanation, particularly because Eoraptor seems to have an omnivorous dentition, and no one has yet found it was a basal sauropodomorph in a phylogenic study.

  2. Awesome -- great to see this early herbivorous clade being fleshed out, and really cool to see the avian total clade's earliest known occurence bumped back to the Middle Triassic.

    "The paper also phylogenetically defines the Silesauridae..."

    Haven't gotten the paper yet, but I'm interested in how this was done. I'd assume branch-based (maybe apomorphy-based)?

  3. "...bumped back to the Middle Triassic."

    Early Middle Triassic, I should say. I believe it was already known from other parts of the Middle Triassic.

    Good point raised in the first comment.

  4. @Karl - The authors actually address this in a lot of detail and with a second phylogenetic analysis (available in the supplementary data) and are not just making this assumption without support. A "carnivorous" dentition/diet is plesiomorphic for Ornithodira (and Archosauria). Thus to further test these ideas we need a well-preserved lagerpetid skull (with the teeth) to determine at the ancestral state between pterosaurs and silesaurids. Currently none have been described; however, Lewisuchus/Pseudlagosuchus is coded as "carnivorous" in the analysis, so silesaurids would actually possess both states.

    @Mike - Silesauridae is defined in the paper as follows: "The most inclusive clade for Silesauridae contains Silesaurus opolensis
    Dzik, 2003 but not Passer domesticus Linnaeus, 1758,
    Triceratops horridus (Marsh, 1889) and Alligator mississippiensis
    Daudin, 1801."

  5. Bill, is this the same taxon as 'Nyasaurus' (described by Charig in his doctoral thesis)?

  6. @Karl & Mike: We discuss quite extensively the dietary reconstruction in the supplementary information, which is freely available here:

    Given that Lewisuchus has carnivorous teeth, our analysis recover all three omnivorous/herbivorous lineages as separate aquisitions.

    @Morgan: Asilisaurus is different from "Nyasaurus".

  7. One issue, though: Langer et al. (2010) already coined and defined Silesauridae: "‘‘All archosaurs closer to Silesaurus opolensis, than to Heterodontosaurus tucki and
    Marasuchus lilloensis’’; stem-based" p. 66 in Biol. Rev. (2010), 85, pp. 55–110. doi:10.1111/j.1469-185X.2009.00094.x

  8. @Tom: That definition came out while our paper was in press, and they did not actually include a phylogenetic analysis to demonstrate their hypothesis that silesaurus-like taxa form a clade. Folks are free to use either definition because publication of Phylocode is year 0 for determining priority. Naturally I'm partial to our definition :)

  9. Or, combine the definitions! "The clade originating with the first ancestor of Silesaurus opolensis Dzik 2003 which was not also ancestral to Passer domesticus (Linnaeus 1758) Brisson 1760, Alligator mississippiensis (Daudin 1801) Daudin 1809, Triceratops horridus Marsh 1889, Heterodontosaurus tucki Crompton & Charig 1962, or Marasuchus
    (Romer 1972) Sereno & Arcucci 1994." It all works out to the same thing -- probably.

    It was also noted on the DML that the ICZN would favor Lewisuchidae Paul 1988 as a family name. However, I don't think either publication considered "Silesauridae" a family, so it's only a clade name.

  10. I wrote a long commentary on my blog-

    My basic conclusion is that I wish everyone would-
    - Google any family-level version of your new clade's genera before you give it a new name
    - Follow the ICZN's rules at least until Phylocode is enacted.
    - Don't use a Linnaean rank suffix if your intent is to name a rankless taxon that doesn't compete for priority in the ICZN.

  11. Mickey,

    I could care less about Families. They aren't real, and I don't use "family-level" names unless they've been defined as clades. Its as simple as that. Linnean taxonomy is dead. The ICZN must adapt or die.

    All of this nomenclatural discussion is kinda silly. Yes, I wish it were simpler, but you know, naming the clade is just a tiny part of our paper. Why don't people focus on the conclusions that are actually interesting! Or at least the fact that our paper is the first to actually demonstrate a clade of Silesaurus-like things!

  12. Of the names being bandied about here, only Nesbitt et al. (2010) have properly named a family-level taxon under the rules of the ICZN. Paul's (1988) Lewisuchinae (which was indeed known about by the authors before Asilisaurus was published) is a nomen nudum, because he never diagnosed his new taxon (required by Art. 13.1.1). Neither did Ezcurra et al. for their version of Silesauride. If you want to argue for the validity of the latter when the PhyloCode is enacted, that's a question for another day. But by the rules of the ICZN, these other names have no validity.

  13. "All of this nomenclatural discussion is kinda silly. Yes, I wish it were simpler, but you know, naming the clade is just a tiny part of our paper."

    I agree that trying to assign taxa to fake ranks is silly, but taxonomic nomenclature in general is an important issue for exactly the reason you provide below.

    "Why don't people focus on the conclusions that are actually interesting! Or at least the fact that our paper is the first to actually demonstrate a clade of Silesaurus-like things!"

    Therefore you should be cited in reference to this "clade of Silesaurus-like things" no matter what it is called, correct?

  14. Bill,

    Taxonomic nomenclature is important, but I'm just amazed by how many people are fixated on this one small thing in the paper. As Anonymous points out, the situation is rather cut and dry.

    We of course hope people will cite the paper for the recognition of a "clade of Silesaurus-like things" (named by both Langer and us as Silesauridae), but of course there is a history of study. Ezcurra (2006) and Nesbitt et al. (2007) were the first to explicitly recognize the similiarities and that they appeared to be basal dinosauriforms, but Ezcurra found Pseudolagosuchus, Silesaurus, and Eucoelophysis to be a paraphyletic grade. Nesbitt et al. (2007) thought they were a clade, but did not do a formal phylogenetic analysis. Ferigolo & Langer (2007) thought Sacisaurus and Silesaurus formed a clade, but hypothesized they were basal ornithischians and did not do a phylogenetic analysis. Both Irmis et al. (2007) and Nesbitt et al. (2009) recovered Silesaurus and Eucoelophysis as a clade, but did not include any of the other taxa. So I guess the best way to put it is that ours is the first to demonstrate that *all* silesaur-like taxa form a clade.

    Ezcurra, M. D. 2006. A review of the systematic position of the dinosauriform archosaur Eucoelophysis baldwini Sullivan & Lucas, 1999 from the Upper Triassic of New Mexico, USA. Geodiversitas 28:649-684.

    Ferigolo, J., and M. C. Langer. 2007. A Late Triassic dinosauriform from south Brazil and the origin of the ornithischian predentary bone. Historical Biology 19:23-33.

    Irmis, R. B., S. J. Nesbitt, K. Padian, N. D. Smith, A. H. Turner, D. Woody, and A. Downs. 2007. A Late Triassic dinosauromorph assemblage from New Mexico and the rise of dinosaurs. Science 317:358-361.

    Nesbitt, S. J., R. B. Irmis, and W. G. Parker. 2007. A critical re-evaluation of the Late Triassic dinosaur taxa of North America. Journal of Systematic Palaeontology 5:209-243.

    Nesbitt, S. J., R. B. Irmis, W. G. Parker, N. D. Smith, A. H. Turner, and T. Rowe. 2009. Hindlimb osteology and distribution of basal dinosauromorphs from the Late Triassic of North America. Journal of Vertebrate Paleontology 29:498-516.

  15. Hmm. I was unaware of Article 13.1.1. Seems Lewisuchinae is indeed a nomen nudum. I would agree with Langer that a phylogenetic definition counts for Ezcurra et al.'s proposal of Silesauridae, but the battle between Silesauridae definitions wasn't my immediate concern.

    I still say that it would have been a nice idea to use Paul's name, since he did basically get the position of Lewisuchus correct even though his taxon was monotypic. But I apologize for implying a reason for erecting Silesauridae as a clade name was to avoid priority issues.

    Yet I still view your attitude toward the ICZN as problematic. I completely agree it needs to change for modern times, especially in regard to electronic publishing. But it's still the only rulebook we have, so I stand by my statements that we should follow its guidelines at least until Phylocode comes out, and not give clades family-level suffixes unless they're intended as Linnaean family-level ranks.

    Also, besides the naming issue, the paper does seem very cool. I like that Lewisuchus was finally included in a good analysis and that its probable synonymy with Pseudolagosuchus was more explicitly discussed than before. I also like your section on fragmentary silesaurid remains. I can only say that a: paleontologists online are inordinately interested in nomenclature, and b: it's much more interesting and fun to discuss parts of a paper you disagree with than parts you like.

  16. Article 13.1.1 causes a lot of problems as it seems that not a lot of people are aware of it and actually many family group names out there are nomina dubia under the Linnaean system. I've only checked some of the Triassic ones in the past and there are obscure thing like Desmatosuchidae and Typothoracidae, which were proposed but never defined. Another is Shuvosauridae which is also a nomen dubium, whereas Chatterjeeidae is valid under the ICZN. It is almost certain that it will be Shuvosauridae that is defined cladistically in the future as "Chatterjeeia" is a junior synonym of Shuvosaurus.

    If you want a fun side project, see how many dinosaur family names are actually valid under the ICZN. I think you will be shocked. This also allows use of 'sensu' in a Linnaean rank Sys. Paleo. section as you will need to also start listing authors who actually provided a diagnosis.

    Names with Linnaean suffixes are often refined as clade names not necessarily to promote confusion, but in many cases to utilize the existing names, plus maybe we aren't all that imaginative. Even -ia suggests a Linnaean rank. What suffixes would we use, the Linnaeans took all of the good ones, or what we have been ingrained to see as good ones ;).

    @Randy - Of course it is a great and important specimen and well covered in the paper, but I have to agree with Mickey that we do get caught up many times in these tangental issues and this one is important ;).

  17. Bill--actually, Desmatosuchidae and Shuvosauridae are both fine under ICZN regulations. Desmatosuchidae was named in 1920, so it falls under Art. 12.2.4, which validates family-level names established before 1931 as long as they were based on a valid genus at the time. Shuvosauridae falls under Art. 13.5, which allows for a new family and genus to both be valid as long as one of them is diagnosed and a type species is given for the genus (both have to be newly described in the same paper for this to work, though). Ah, the joys of the Code.

  18. OK...I give up :). Victimized by Art. 13.5. So techinically almost everything is valid based on all of these subtle "However, if"'s? buried here and there in the code. What is the point of having rules if everyone gets off on these technicalities? Obviously mostly they were put in mainly to preserve pre-1930 taxonomic names.


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS