Splitting Chirostenotes

Seems audacious of me, eh? Chirostenotes (specifically, Chirostenotes pergracilis) enjoys one of the more interesting synonymy lists among dinosaurs, as it contains (or has contained) no fewer than three generic synonyms and four specific synonyms, but also that it has almost never contained all of these at once, with various authors adding or removing various names depending on newer finds. This has been aided by the difficulty of finding complete or reasonably replete skeletal material of singular individuals of small dinosaurs in the Dinosaur Park Formation of southern Alberta. Instead, partial material abounds. This partial material has led to a rich lesson in synonymy. So … here goes.

Caenagnathidae, comparison of various known skeletal material. Note the exclusion of Caenagnathus collinsi; this is intentional, although I should include it anyways. The three skeletons on the right are considered Chirostenotes (the top is Chirostenotes sp., a non-determinate designation, while the others seem comfortably Chirostenotes pergracilis), the two on the left Elmisaurus rarus.

First. The Background

Note: For the purpose of this discussion, I will be referring to all taxa by their original designation, such that Ornithomimus elegans Parks, 1933 while certainly not a species of Ornithomimus but a caenagnathid instead, will continuously be referred to by that designation until such time I consider it necessary to “refer” the material by summation of opinion, which I will include below.

Note the second: This is a very long post … you have been warned.

In the early 192os, while prospecting for fossils in southern Alberta, near the old township of Steveville (now near Brooks, between Medicine Hat and Calgary), Charles Whitney Gilmore discovered a series of fossils that would become the first of many famous Canadian dinosaurs: Styracosaurus albertensis, Albertosaurus sarcophagus, Struthiomimus altus, Corythosaurus casuarius … and Chirostenotes pergracilis [1][n1] (at that time, it was the first any paleontologist knew of the particularly gracile manus of maniraptoran theropods — Henry Fairfield Osborn [2] would describe three other maniraptorans that year {Velociraptor mongoliensis, Saurornithoides mongoliensis, and Oviraptor philoceratops [n2]}, including similar hands, but these were unknown to Gilmore). This last taxon was only one of several fossils discovered during the 1920s, including a particularly larger but less distinct fossil (more later), that would end up in a hodge-podge of oviraptorosaurian goodyness. Gilmore thought the hand would belong to a “coelurid”[n3] of sorts, then undefined and odd, restricted to animals like Ornitholestes hermannii [3] for which the term (as used) “compsognathid” was also used, often interchangeably.

Gilmore’s taxon would be followed by Chas Sternberg’s description of Macrophalangia canadensis [4]  (which he discussed as a “robust” ornithomimid with particularly short feet), but based on a complete pes including all phalanges. William Parks would follow with a description of a partial pes from a different region of southern Alberta he would also ascribe to Ornithomimidae [5]. Unfortunately, more substantive fossils would not relate of these fossils to one another for another 40 years, with the exception of a fairly large nearly complete pair of conjoined mandibles named Caenagnathus collinsi by Richard M. Sternberg [6]; this jaw would go unnoticed for some time.

Eventually, beginning in the 1970s, scientists began to suspect that some of these taxa may, in fact, belong to one another. But, lacking in comparable material to conjoin them, they would be retained separately for some time. Cracraft [7] would endorse Sternberg’s theory that Caenagnathus collinsi was a bird, supporting this with another mandible fragment (an articular fused to the posterior ramus of the jaw) named Caenagnathus sternbergi. Lacking familiarity with then-known but largely ignored Oviraptor philoceratops (which possessed a similarly convex articular morphology), Cracraft assumed the jaw was a particularly odd bird, and coined the names Caenagnathidae for the “family” and, in keeping with current ornithological tradition, it’s own “suborder” Caenagnathiformes.

The 1980s saw a shift when, first in 1981 when Halszka Osmólska described the remains of an Asian taxon, Elmisaurus rarus [8], recognizing that it was similar to the forms Chirostenotes pergracilis and Macrophalangia canadensis, and implied the two could be synonymous with one another. In 1988, Phil Currie and Dale Russell described a single specimen[9]  in which forelimb, hindlimb, and vertebral, pelvic, and pectoral material was closely enough associated to assume it was a single animal, and used this to synonymize Chirostenotes and Macrophalangia into one another, and pergracilis and canadensis, producing only Chirostenotes pergracilis. Greg Paul would further this sentiment [10] by subsuming sternbergi into pergracilis, but even further by subsuming Elmisaurus into Chirostenotes without firmly justifying this process (the same argument has continued in Paul’s work as I noted here and here due to the assumption that modern generic “lumping” is consistent with a “true” definition or usage of the “genus” concept). Paul is also one of the first to enforce the concept that Oviraptor and Chirostenotes were closely related, a position later affirmed by various systematic analyses using cladistics [n4]. Paul would also posit, followed by Currie et al. [11; see below], that Caenagnathus collinsi might represent the mandible of Chirostenotes pergracilis. Paul [10] subsumes sternbergi into pergracilis, noting it (along with collinsi) with a “?”, so that this referral is questionable even by him.

The 1990s would see further arguments presuming synonymy among the taxa, for various reasons. Currie, John Rigby and Robert Sloan’s analysis of teeth from the Dinosaur Park Formation [11]  from which virtually all of these taxa derived, they described the jaws of Richardoestesia gilmorei, and firmly asserted that the mandible of Caenagnathus collinsi was likely to belong to Chirostenotes pergracilis. This, unfortunately, left the question of Caenagnathus sternbergi open, as a smaller, differently proportioned articular shape appeared to firmly prevent its synonymy with Caenagnathus collinsi in the first place; hence, Chirostenotes sternbergi. Currie would later summarize what was considered Elmisauridae [12] and included only the newly re-minted Chirostenotes and Elmisaurus in it, supporting Parks’ Ornithomimus elegans as a second species of Elmisaurus. Further hints that Caenagnathus collinsi might belong to this group were made, but in the same volume Barsbold Rinchen included it (and its “family” Caenagnathidae, as the sole inhabitant) in his Oviraptorosauria [13]

Later, Hans Sues would describe a larger caenagnathid skeleton from the Horseshoe Canyon Formation [14] that preserved portions of the skull and repeated regions of the skeleton previously reported by Currie and Russell [9]. Sues not only sought to affirm Currie and Russell, but used this to firmly cement the idea that Caenagnathus collinsi was synonymous with Chistoenotes pergracilis, due to the shape of the edentulous maxilla, and also argued that Parks taxon (originally Ornithomimus elegans [5]) was a likely synonym of Chirostenotes pergracilis. Thus Sues also subsumed all North American caenagnathid/elmisaurid material into Chirostenotes, recognizing only pergracilis and rarus as valid species, founded under the premise that the diagnoses used to separate various species were “insufficient.”

Not shortly thereafter, Dave Varricchio, studying collections from the Two Medicine Formation of Montana [15] (roughly equivalent to and considered an extension of the Judith River Group — dominantly, the Dinosaur Park Formation) argued that the gracile nature of Ornithomimus elegans was not only distinct from the more robust Chirostenotes pergracilis, but was far more similar to that of Elmisaurus rarus from Asia, which is even further robust in comparison. This was primarily achieved because of the proximal fusion of the distal tarsals to the metatarsals, forming a tarsometatarsus (otherwise unknown in caenagnathids, or oviraptorosaurs for that matter); based on size and general gracility, Varricchio also set Chirostenotes sternbergi aside and referred it to his differentiated taxon as Elmisaurus elegans. Teresa Maryańska, Osmólska and Mieczysław Wolsan [16] divided Elmisaurus rarus from all other species (including Elmisaurus elegans, now apparently Chirostenotes elegans by implication), while asserting that it might actually be even more avian-like due to characteristics of the proximal tarsometatarsus (e.g., an insertion ridge for the m. tibialis cranialis, a slit between the third and fourth metatarsals indicating a likely lateral proximal vascular foramen, all of which the authors considered more like birds and less like nonavian theropods such as Caenagnathidae. Ironically, the extreme avian-like jaw would not be avian enough, while the very dinosaurian-like pes would be.

Finally, and bizarrely, Paul [17] reversed some earlier positions in retaining Caenagnathus collinsi apart from Chirostenotes pergracilis, provisionally retaining Chirostenotes elegans, and following Sues [14] in dividing the ROM specimen as a potential new species, adding in an additional possible species in the form of the Sandy specimens [see 18; see n5 and 19]

And this remains the last of the lumping/splitting we have seen. Various analyses have treated as split or lumped in their own way, but none have formally argued for or against. Most of these (like [16] above] have been cladistic analysis or merely along the lines of a North American versus Asian split among taxa.

Second. The Concern

The problem with some of this synonymy is that it is assailable from a fairly strong position: that of material overlap. Typically, similar morphology is used to develop a synonymy, where material literally overlaps between two different taxa and that material is literally identical. In cases like above, there has either been doubt among the authors as to which species a given specimen may belong (e.g., Sues referred his Horseshoe Canyon specimen [14] to Chirostenotes sp., not to a particular or even new species) while others have made the argument simply due to the “likelihood” of a single species in a given formation, or the amount of variation implied in a given region/time frame.

These implications leave a lot of room to argue, as I may, that a formation may contain many divergent and related taxa, as it does in Asia where the Djadokhta, Baruun Goyot, and Nemegt formations each contain multiple dissimilar taxa (from caenagnathids to oviraptorids) which are difficult to synonymize without detailed studies of ontogeny and morphology (the initial descriptions indicating the latter at least, which appears to contradict the synonymies proposed by Paul [17] for the most part when it comes to Oviraptoridae).

And finally, it may be argued that strict morphological differences at the same or very similar size can reject or affirm assumptions of synonymy due to the referral of material that does overlap. E.g., the holotypes of Chirostenotes pergracilis and Macrophalangia canadensis are virtually identical in morphology and size to that of Currie and Russell’s specimen [9], while Sues’ ROM specimen is significantly larger [14], and that of the Sandy specimens larger still; the latter rwo are also known from different stratigraphic levels, a general implication for taxonomic separation in some researcher’s viewpoint. And additional consideration is the clear morphological variance of similar elements in the same formation, such as the presence of three distinct dentary morphologies in the Dinosaur Park Formation: Following work from Currie, Stephan Godfrey and Lev Nessov [19] which identified an elongated, shallow mandible (holotype of Caenagnathus collinsi) and two shorter mandibles with upturned and angular “chins” but with distinct morphology of the symphysis, although this may be variation; and either of these jaws may or may not pertain to Caenagnathus sternbergi.

Similarly, postcranially oviraptorosaurs appear much more conservative, and much of their variation and diagnostic features appear in the skull — such that any postcranial variation should be assumed to be minor compared to cranial variation, it would be assumed that missing cranial material would be dramatically divergent in skeletally diverse taxa (this is an assumption, and certainly falsifiable). I would, however, expect skeletal similarity to be less useful in referring taxa than I would cranial material; this would also be approached differently between the two branches of Caenagnathoidea, as most oviraptorids are diagnosed on cranial features, whereas most caenagnathids are diagnosed on postcranial features.

Third. The Solution … For Now

A presumptuous argument in the title notwithstanding, I propose that following the above criteria, that all taxa not directly comparable using the above three criteria (material, stratigraphic, and morphologic overlap) be separated as unique taxa until such a time that work demonstrates at least two of the above. Three is cake.

Thus, then, is my proposal (and the premise for the title of this post):

1. Caenagnathus collinsi would be distinct from Chirostenotes pergracilis — if Sues [14] cannot demonstrate that his Maastrichtian specimen is the same taxon, that the stratigraphy and morphology do not compare (they do not), then the material similarity (edentulous maxilla with a curved tomial margin comparing to that of the mandible of the other taxon) without overlap cannot affirm that the two are actually likely to be the same (this is especially important when there are similar but non-identical jaws from the Dinosaur Park Formation that cast the “species” issue into doubt).

2. Similarly, Caenagnathus sternbergi is directly comparable to only three specimens: the holotype of Caenagnathus collinsi, from which it differs morphologically; the Maastrichtian partial jaw [19], from which it also differs and for similar reasons but including others such as shallower rami, higher profile, larger overall size, extreme convexity of the articular, etc.; and an articular from the Two Medicine Formation [15] with which it agrees morphologically by differs in location. Using the above criteria, but ignoring apparent complete fusion of the element implying ontogenetic maturity, we can presume Caenagnathus sternbergi to be synonymous with Caenagnathus collinsi. However, the doubt arising from multiple mandibular specimens and skeletal material may imply separation, and this has been the general consensus among researchers. Thus, I would treat this as Caenagnathus sternbergi.

3. Elmisaurus rarus is fairly distinct on all three grounds, and this only contradicts one author [10,17] who did not apply critical analysis to subsuming the “genus.”

4. Ornithomimus elegans differs in morphology from all North American caenagnathid metatarsals described to date, while it compares well with Elmisaurus rarus with which it shares several unique characteristics, contra [16] where it is implied there is no difference and with which it shares two of the distinct features the authors ascribe to separate Elmisaurus from Chirostenotes. This presents a stratigraphic separation from another very similar taxon, while presenting a possible morphological similarity across the continents, thus I would agree with other authors in affirming its distinctiveness as a taxon. The only issue is that there are more reasons to assume it belongs in Elmisaurus than in Chirostenotes (fusion of the distal tarsals, the proximal “spur,” and the crest for the m. tibialis cranialis), thus Elmisaurus elegans it would be.

5. Macrophalangia canadensis would thus be the only real slam-dunk synonymy involved here, as it presents a stratigraphic (same formation and horizon) overlap, while it requires an additional specimen to affirm a morphologic and material overlap (including virtually identical in size, leaving out the issue of ontogenetic change). Thus, as nearly every author before me, I would affirm the synonymy with Chirostenotes pergracilis.

Conclusion

Caenagnathidae represents a diverse range of taxa, including manus, pes, and cranial features, that are hindered by the paucity of the material to which we can ascribe to them. What material is present, barring new discoveries, hints at odd morphology. Newer, unincluded taxa such as Hagryphus giganteus and the even more-giganteus-ic Gigantoraptor erlianensis, imply that caenagnathids were diverse in many other directions; some studies implicate tiny “avian” Avimimus portentotosus and novelty-inducing Nomingia gobiensis as possible caenagnathids, yet these are based on less complete or overlapping material with the diagnostic portions known from “secure” caenagnathids. We can thus expect to see more should further studies in the Dinosaur Park Formation yield more complete material at least, as well as the formations of central Asia.

[n1] A pair of slender dentaries were also found along the Red Deer River much earlier (late 1910s), including small, high-aspect triangular teeth, but these would not be closely inspected until much later. They would eventually provide the impetus to recognize a later set of jaws as the likely mandibles for Chirostenotes, those of Caenagnathus, by recognizing Richardoestesia gilmorei.
[n2] Ironically, Osborn also linked Oviraptor to Ornithomimidae on the sole basis of the lack of teeth.
[n3] Generally, any small theropod dinosaur that did not fit into the classic “families” was a “coelurid.”
[n4] Paul thus used Caenagnathidae appropriate when using a single “family” to contain them, as it was named in 1970 as opposed to Oviraptoridae (1986) or Elmisauridae (1971). This practice has not been supported subsequently, where Oviraptoridae is considered too dissimilar to support in the same “family,” and instead the name Caenagnathoidea is used where Paul uses Caenagnathidae.
[n5] These specimens, from the Hell Creek of South Dakota, are currently under study by Matt Lamanna and Hans Sues of the Carnegie Museum in Philadelphia, and may represent unique species; the specimen BHM 2033, also from the Hell Creek, and mentioned by Currie, Godfrey and Nessov, is possibly the same taxon.

[1] Gilmore, C. W. 1924. A new coelurid dinosaur from the Belly River Cretaceous of Alberta. Bulletin of the Canadian Department of Mines, Geological Surveys 38:1-12.
[2] Osborn, H. F. 1924. Three new theropod dinosaurs from the Protoceratops zone of Mongolia. American Museum Novitates 144:1-12.
[3] Osborn, H. F. 1903. Ornitholestes hermanni, a new compsognathoid dinosaur from the Upper Jurassic. Bulletin of the American Museum of Natural History 19(12):459–464.
[4] Sternberg, C. H. 1932. Two new theropod dinosaurs from the Belly River Formation of Alberta. The Canadian Field-Naturalist 46:99-105.
[5] Parks, W. A. 1933. New species of dinosaurs and turtles from the Belly River Formation of Alberta. University of Toronto Studies (Geological Series) 34:1-33.
[6] Sternberg, R. M. 1940. A toothless bird from the Cretaceous of Alberta. Journal of Paleontology 14(1): 81-85.
[7] Cracraft, J. 1971. Caenagnathiformes: Cretaceous birds convergent in jaw mechanism to dicynodont reptiles. Journal of Paleontology 45:805-809.
[8] Osmolska, H. 1981. Coosified tarsometatarsi in theropod dinosaurs and their bearing on the problem of bird origins. Paleontologia Polonica 42:79-95.
[9] Currie, P. J. & Russell, D. A. 1988. Osteology and relationships of Chirostenotes pergracilis (Saurischia, Theropoda) from the Judith River Oldman Formation of Alberta. Canadian Journal of Earth Sciences — Revue de Canadienne des Sciences de la Terre 25:972-986.
[10] Paul, G. S. 1988. Predatory Dinosaurs of the World: A Complete Illustrated Guide. Simon & Schuster (New York City).
[11] Currie, P. J., Rigby, J. K., Jr. & Sloan, R. E. 1990. Theropod teeth from the Judith River Formation of southern Alberta, Canada. pg.107-125 in Carpenter and Currie (eds.) Dinosaur Systematics: Perspectives and Approaches. Cambridge University Press (New York) .
[12] Currie, P. J. 1990. Elmisauridae. pg.245-248 in Weishampel, Dodson, and Osmolska (eds.) The Dinosauria. University of California Press (Berkeley).
[13] Barsbold R., Maryańska, T. & Osmólska, H. 1990. Oviraptorosauria. pg.249-258 in Weishampel, Dodson, and Osmolska (eds.) The Dinosauria. University of California Press (Berkeley).
[14] Sues, H.-D. 1997. On Chirostenotes, a Late Cretaceous oviraptorosaur (Dinosauria: Theropoda) from western North America. Journal of Vertebrate Paleontology 17(4):698-716.
[15] Varricchio, D. J. 2001. Late Cretaceous oviraptorosaur (Theropoda) dinosaurs from Montana. pg.42-57 in Tanke and Carpenter (eds.) Mesozoic Vertebrate Life. Indiana University Press/Canadian Natural Resources Press (Bloomington and Drumheller).
[16] Maryańska, T., Osmólska, H. & Wolsan, M. 2002. Avialan status for Oviraptorosauria. Acta Palaeontologica Polonica 47(1):97-116.
[17] Paul, G. S. 2010. The Princeton Field Guide to Dinosaurs. Princeton University Press (Princeton and Ozford).
[18] Russell, D. A. & Triebold, M. 1995. A new small dinosaur locality in the Hell Creek Formation. Journal of Vertebrate Paleontology 15 (supplement to 3):57A.
[19] Currie, P. J., Godfrey, S. J. & Nessov, L. A. 1994. New caenagnathid (Dinosauria: Theropoda) specimens from the Upper Cretaceous of North America and Asia. Canadian Journal of Earth Sciences — Revue de Canadienne des Sciences de la Terre 30:2255-2272.