An Oviraptorosaur Worth the Name?

The newest oviraptorosaur on the blog is, in fact, one of the oldest. Nick Longrich and colleagues (Ken Barnes, along with Scott Clark and Larry Millar from Paleo Field Excursions, who collect in the Big Bend area in which the Aguja and Javelina Formations expose) have addressed the question that was first approached by William Parks in 1933, the nature of a tiny tarsometatarsus (ROM 781).

Metatarsi referred to Ornithomimus elegans (Parks, 1933). A & C, ROM 781 (holotype) in cranial or extensor view (A) and proximal view (C); B & D, RTMP 82.39.4 in extensor (B) and proximal (D) views. Despite their similar size, they indicate disparate levels of preservation, but preserve the same diagnostic “tongue-like” process (to the upper left in C and D). E, MOR 752, partial pes in extensor view. Scale bar for all specimens, 3 cm. Hatching indicates broken surfaces exposed in the plane of the image, while dotting indicates completed margins.

Parks, lacking any other possible option at the time, referred the partial foot to Ornithomimus as a new species, elegans (Latin, delicate, in comparison to the more robust form in other Ornithomimus recognized at the time). It would only be later, with Sternberg’s description of Macrophalangia canadensis and Currie and Russell’s 1988 referral of that species as a synonym of Chirostenotes pergracilis, that the fused form of the foot would be found to represent an unusual trend in an otherwise unfused-footed clade. Phil Currie would return to the subject in 1989 by referring Parks’ elegans to Osmólska’s Elmisaurus (1981), given that the morphology of the proximal tarsus were nearly identical. He would specifically point out (1) the third metatarsal (MTIII) is “pinched” between MT’s II and IV, only the proximal end of MTIII is excluded from the flexor surface of the metatarsus, so that the medial end of MTIV and lateral end of MTII converge toward one another and lie atop MTII, with (2) small flanges of the distal MTII and IV that overlap over the flexor surface of the distal MTIII, (3) the metatarsus coossifies, with (4) a sharply emarginated caudolateral margin, whereas in other oviraptorosaurs and indeed theropods this margin is rounded, and which (5) may form an elongated flange-like process, and (6) at the proximal and mid-shaft positions, MTII is extremely narrow flexor-plantarly, or from front to back, making up significantly less than 50% of the available depth of the other metatarsals, and forming a deep sulcus on the metatarsus caudal/plantar face. Comparatively, Parks’ elegans has a thinner aspect, but is relatively longer, than Osmólska’s rarus, but elegans was overall shorter and narrower than Chirostenotes pergracilis, leading one to project a series, with pergracilis at one end, rarus at the other, and elegans in the middle. Elmisaurus rarus also possesses a very “sharp” caudal margin to MTIV, Indeed, Varricchio (2001) would specify precisely this, while at the same time referring new foot bones to elegans; rather than deriving from the Campanian Dinosaur Park Formation, these derived from the Maastrichtian Hell Creek Formation, which implicitly extends the range of elegans through the Maastrichtian.

Metatarsus in cross-section, demonstrating the relative position and “thickness” of MTIII in Caenagnathidae. A, RTMP 79.20.1, referred to Chirostenotes pergracilis, reconstructed following Currie & Russell, 1988; B, ZPAL MgD-I/172, holotype of Elmisaurus rarus (Osmólska, 1981); C, ROM 781, holotype of Ornithomimus  elegans (Parks, 1933). B and C follow Currie (1989), and A is heavily reconstructed from a distorted specimen, which obscures its usefulness. Note that in A and C, the caudal margin of MTIII appears to bear a sulcus, however the degree of this depression in B cannot be determined so readily, nor in A where the specimen is compressed.

In 1994, Phil Currie and colleagues described several partial mandibles of caenagnathids which differed in shape and size from that of CMN 8776, holotype of Caenagnathus collinsi (Sternberg, 1940). These jaws were fused, but shorter overall, with a shorter symphysis compared to width and height, and with an “upturned” aspect to the tip of the jaw. The dorsal ornamentation, comprising longitudinal and radial ridges of the dorsal mandibular symphysis, also differed; moreover, Currie et al. implied there were more than one jaw moprhology represented by these specimens. They suggested that some of these jaws probably belonged to Caenagnathus sternbergi, whilst others would be a different, probably new species. They also suggested that elegans and sternbergi were synonyms; this was further supported by Sues who, in 1997, described for the first time postcranial material consistent with Chirostenotes pergracilis which also included edentulous jaw fragments (although not of the mandible) — but note that this material is now the holotype of Epichirostenotes curriei (Sullivan et al., 2011), the best known early Maastrichtian oviraptorosaur. Sues more firmly argued  that elegans and sternbergi, despite being based on incomparable holotypes, were probable synonyms, due mostly to “gracility” and size.

Skeletons in silhouettes of Elmisaurus, showing Elmisaurus rarus based on ZPAL MgD-I/172 and 98; B, Elmisaurus rarus, projected size of ZPAL MgD-I/20; C, Ornithomimus elegans, based on ROM 781. Scale bar equals 50cm.

It is this philosophy, that size can allow referral, that Longrich et al. take up when addressing elegans. Rather that following Currie then Sues, Longrich et al. imply that elegans is a unique taxon, and refer several of the RTMP mandibles to the species. They then formally named a new “genus,” Leptorhynchos, to emphasize the intermediacy implied by elegans, between rarus and pergracilis. Longrich et al. further argue that the relative fusion of the mandibles, technically class IV symphyses (Scapino, 1981) indicate maturity, and thus the morphology represented by these jaws would be “final.” Problematically, despite mentioning it briefly, the authors did not regard the odd case of Caenagnathasia martinsoni (Currie et al., 1994), in which the holotype jaw is fused, but the similarly-sized paratype right dentary has an open, rugose symphyseal plate (a class III symphysis). The two specimens are nearly of identical size, yet show disparity in fusion, implying one can transition from an open, “immature” morphology to a closed, “mature” one without much change in size. They further distinguish these jaws amongst their new formulation of elegans and collinsi, leaving sternbergi without a dentary; this is handled by referring sternbergi as a synonym of collinsi. In a fairly convoluted manner, Longrich et al. presume that, as the Dinosaur Park Formation seems to preserve two skeletal morphs of “Chirostenotes,” where Caenagnathus collinsi represents the larger form and contains foot material referred to Macrophalangia canadensis, and Chirostenotes pergracilis represents the smaller form and jaws referred to Caenagnathus sternbergi. Bizarrely, Longrich et al. also confuse the quality of preservation of one specimen of either Elmisaurus rarus or elegans, in which the tarsals fuse together and to the proximal metatarsals (ZPAL MgD-I/172 and RTMP 82.39.4, respectively), but ignore the other, near-identically-sized specimens in which tarsometatarsal fusion does not occur (ZPAL MgD-I/20 and ROM 781, respectively). ZPAL MgD-I/20 is larger than 172, the holotype, yet shows less complete fusion. This disparity is probably not taxonomically useful, but represents individual variation. To assume size and state of fusion may be useful for diagnostic reference to taxonomy misplaces the unstable development of fusion, but also the problem of interpreting broad patterns in paleontology with such an incomplete record as caenagnathid pes material. Ultimately, it calls into question Longrich et al.s framework for dividing species.

Metatarsals of Elmisaurus rarus (Osmólska, 1981). A, ZPAL MgD-I/172, the holotype, showing complete tarsal fusion and a tarsometatarsus; B, ZPAL MgD-I/20, the larger specimen, showing separate tarsals. Scale bar equals 3 cm.

Longrich et al. play loose with the reasoning for referral of specimens to species, especially as they fail to establish, though implying, that there are merely a few skeletons involved. Jaw morphology amongst Dinosaur Park caenagnathids differ in markedly unique ways:

1. There are not two, but three possibly distinct jaw types, a shallow, long-symphysis type (collinsi), a deep, short-symphysis type, and an additional deep, short-symphysis type with more upturned jaw tip;
2. There are two types of pes in the formation, but these are represented by elegans and pergracilis, whereas the holotype of canadensis is merely large in total size but not much different in proportions;
3. There are two types of articular morphology, the long, oval form (collinsi) and the short, circular form (sternbergi), among which relative proportion of medial and lateral cotylar facets are distinct.

These imply that there are a minimum of two distinct species, but they would correspond to pergracilis/canadensis/collinsi and elegans/sternbergi; if anything, this implies (as other authors have suggested) that elegans and sternbergi are synonyms. Currie (2005) further referred a new mandible to sternbergi, RTMP 2001.12.12, which involves a similar articular morphology as the holotype (CMN 2690), but had a very collinsi-like dentary morphology, but with a shorter, upturned beak tip. It is further puzzling why Longrich et al. do not distinguish the established differences between sternbergi and collinsi, namely the shapes of the articular. Currie (2005) further emphasized these differences, which included a relatively shorter, higher and more circular profile of the articular, and implied that many of the mandibles referred by Currie et al. (1994) belonged to sternbergi. As such, it becomes problematic for Longrich et al. to support elegans distinctly from sternbergi; by all logic, if one is keen on referring species based on incomparable holotypes, going with the flow would suggest elegans/sternbergi instead; they could still find ways to support unqiue nomenclature. Instead, Longrich refers Macrophalangia canadensis to Caenagnathus collinsi, refers Caenagnathus sternbergi to Chirostenotes pergracilis, and separates Ornithomimus elegans from Elmisaurus as a new taxon, Leptorhynchos[n1]. And in a bizarre twist, despite referring CMN 2690 to Chirostenotes pergracilis, Longrich et al. refer the nearly identical MOR 1107 (Two Medicine Formation, Varricchio, 2001; a partial articular and conjoined surangular) to Leptorhynchos as ?Leptorhynchos. The only reasoning given is size.

(Of course, we assume elegans is the type species of Leptorhynchos; Longrich et al. do not actually designate a type species. The ICZN, however, indicates in Art. 70.1 that in the absence of a designated type species, if you refer a previously established species to a “genus,” even if you name new species, that previous species is the type. In this case, elegans is the type species of Leptorhynchos, even though Longrich et al. do not explicitly state so, even if they should have.)

Finally, there’s Leptorhynchos gaddisi[n2], named from several skeletal elements from the late Campanian Aguja Formation, including a partial fused mandibular symphysis.

Mandible of TMM 45920-1, holotype of Leptonychos gaddisi.

The holotype (TMM 45920-1) is joined with partial postcranial remains suggestive of a small animal around 1 meter in length. The mandible is strongly reminiscent of Caenagnathasia martinsoni in having a nearly vertical rostral margin of the dentary, but with a dorsal symphyseal morphology similar to mandibles referred to Leptorhynchos elegans. Indeed, the similarity amongst mandibles not referred to Caenagnathus collinsi are all collectively distributed to Leptorhynchos, with elegans containing those mandibles not “clearly” collinsi, including RTMP 2001.12.12. Diagnostic attributes among the species implied by their symphyseal morphology or differences in “projection” of the jaw tip or its angle are numerous, but as noted by Currie et al. (1994) these features do not readily allow one to distinguish species: while numerous, they differ from one another in small details, and so placed on a spectrum one might get that several seeming diagnostic features were, in fact, suggestive of individual variation. Longrich et al. specifically diagnose gaddisi by the following:

(1) “a more anteriorly projecting tip of the beak, a strongly rounded anteroventral margin of the symphysis,” (2) “lateral occlusal margins of the dentaries not as strongly divergent in dorsal view,” (3) “with the tip of the beak being narrower and more spoon-shaped in dorsal view.”

Each of these characters is found in Caenagnathasia martinsoni, and a close comparison of the two might find them quite a bit more similar than Longrich et al. seem to have discovered. There are distinctions, however: gaddisi differs from martinsoni in the presence of additional marginal symphyseal ridges, definding circular sulci on the lingual surface of the tomia, while there appears to be more knobbly ornamentation in gaddisi, but in martinsoni there are none.

I am hesitant to accept their arguments about distribution of these jaws to taxon as this was not done systematically; but also due to the problem of lack of correspondence of holotypic material and referred material: One should not refer specimens to taxa known so poorly, when the holotype and reference specimen is a partial jaw, or a toe bone, or other such fragment — regardless of how diagnostic that fragment is. One can make claims (as Longrich et al. do) that there is a probability of correspondence, and thus follow through; that there are distributions of small and large specimens in a formation, thus two taxa. But provisional referral by these authors to Leptorhynchos gaddisi of postcranial remains, including a proximal MTIV, puts the argument to the test, and the authors do not follow their own advice; they are inconsistent.

Longrich et al. justify their separation of Ornithomimus elegans from Elmisaurus rarus sensu Currie (1989) merely through the time-honored tradition of saying that “’Ornithomimus‘ elegans, although similar in size and proportions to Elmisaurus rarus, seems to be more closely related to Chirostenotes and Caenagnathus,” thus “requiring that it be placed in a separate genus.” (Longrich et al., 2013:pg.27, emphasis added.) Now, in their defense, they did produce a cladistic analysis, a modification of the Osmólska et al (2004) analysis which Longrich et al. (2010) used to support diagnosis of another oviraptorosaur, Machairasaurus leptonychus; and that analysis does suggest that Elmisaurus elegans is not quite the same animal as Elmisaurus rarus, but not all is well with it. Into this analysis the authors inserted several new characters, and coded their OTUs as complexes of cranial and postcranial material. Despite coding new characters for the analysis, however, the analysis became so poorly resolved that they excluded two taxa (Epichirostenotes curriei and Nomingia gobiensis, both lacking any manual, pedal, or mandibular material) to resolve it, which became their figure 14:

Phylogenetic analysis, excluding Nomingia gobiensis and Epichirostenotes curriei.

Lack of resolution within Caenagnathidae is suggested by tests of this analysis excluding particular, incomplete taxa (Hagryphus giganteus, Machairasaurus leptonychus, Caenagnathasia martinsoni), but their exclusion does not effect the total topology, merely the consistency index and bootstrap values for surrounding nodes. It is my hypothesis that Caenagnathidae will remain poorly resolved in the future barring extensive skeletons of known taxa including cranial and epipodial remains, remains which permit resolution of the problems that plague phylogenetic analysis in Oviraptorosauria. Despite this, there is also the concern that much ado is made of the particular morphology of parts of these taxa, but little to their variation: Longrich et al. make it when dealing with mandibular structure, and again with tarsometatarsal fusion. We must expect that we know little about relative fusion patterns among oviraptorosaurian taxa, much as we know so little lacking ontogenetic series in most nonavian theropod dinosaurs. We know more about these trends in ornithischians due to their ubiquity (they are, after all, theropod-food, and must be readily available to feed their betters) and because in several taxa (mostly hadrosaurs and ceratopsians) extensive bonebeds capture juvenile to adult growth patterns. But theropods are different, and despite some bonebeds (Sinornithomimus, Coelophysis), fusion patterns tend to be speculative, rather than certain. And they almost certainly varied with size, as they do with crocodilians and to a degree with birds today.

Thus, on nearly the sole criterion of relative size, Longrich et al. have produced a thesis of caenagnathid taxonomy. It is clear that some of the species referenced by Longrich et al. are unique, but in many case, substantiation is made by referring to species specimens which cannot be compared through the correspondence of holotypes or referred specimens. Only one described specimen, and two further undescribed ones, include enough postcrania and crania together, and these represent distinct species themselves, Epichirostenotes curriei (Sullivan et al., 2011) and the “Hell Creek caenagnathid” (Lamanna et al., 2011), respectively.

Were I to offer a commentary on the value of this new taxonomy, and I would do so hesitantly, I would claim that the name Leptorhynchos is unnecessary: No strong case is made, and barely any at all, that Ornithomimus elegans as different from rarus as it is from pergracilis, nor that elegans is not best off left as a species of Chirostenotes. Their own cladistic analysis cannot form a monophyletic clade for Leptorhynchos, and their characters supporting this taxon are merely suggestive of jaw morphologies intermediate between Caenagnathasia martinsoni and Caenagnathus collinsi, which seems quite a lot of space (western Asia to North America) and time to cover, yet lumped down into a new “genus.” It is also disconcerting to see with so ready a hand the dismissal of value of clearly distinct morphologies (e.g., Caenagnathus sternbergi) while at the same time over-emphasizing morphology. As per Currie (2005), there does seem to be two very distinct mandibular morphologies in the Dinosaur Park Formation, but one of these appears correlated to the articular morphology presented by Caenagnathus sternbergi. At the same time, there are two distinct metatarsal morphologies, one of which is represented by Ornithomimus elegans, and having a close (potentially closer) similarity to Elmisaurus rarus from Asia. The later age of the latter (early Maastrichtian for the Nemegt Formation) implies that it is the Chirostenotes pergracilis morphology that grades into the Ornithomimus elegans morphology, then into the Elmisaurus rarus one, grading into an arctometatarsus, and this seems at odds with Longrich et al.’s result shown above, which places rarus more basal to elegans and pergracilis (though there is a lack of resolution in several sections of the tree).

In the end, I can except provisionally that each of these species is distinct:

Chirostenotes pergracilis Gilmore, 1924, containing Macrophalangia canadensis Sternberg, 1932
Caenagnathus collinsi Sternberg, 1940
Caenagnathus sternbergi Cracraft, 1971
Leptorhynchos elegans (Parks, 1933) [Longrich et al., 2013]

However, when it comes to Leptorhynchos gaddisi, there’s some difficulty, not the least which is lack of affirmation of relative ontogeny, lack of concern for potential growth while exhibiting symphyseal fusion — as they are certainly incorrect about whether a fused mandibular symphysis is an indicator of maturity, as if that meant cessation of skeletal growth — and the remarkable actual similarity to Caenagnathasia martinsoni. Perhaps gaddisi is a unique species, and Longrich et al. are right to be concerned that their analysis did not recover a unique clade; and perhaps that means they had a cause to name another new “genus” for gaddisi. I suspect that had they I might still have been critical: I follow Phil Currie in being curious about these jaw specimens, but at the same time cautious about making assumptions of distinction when we have little postcranial association by which to then compare them further. And a lack of postcranial/cranial association in general is a point of major queasiness for me. Keeping taxa separate barring firm cause for synonymy (morphological being at the top of the list of hurdles to jump, not merely geographic and stratigraphic sameness) is best than assuming that some esoteric quality (size, elegance of synonymy, some predilection for lumping, etc.) is more appropriate.

In the end, while the paper is useful and describes at the least a new species, I wonder if it would have been wiser for Longrich et al. to leave elegans alone and simply coin a new taxon Leptonychos gaddisi and leave it at that. It’s not like elegans being in Elmisaurus (where I prefer it) was hurting anyone’s sense of aesthetic, and elegans being in Chirostenotes would have cleansed that sense of ill-feeling more easily than new taxonomic nomenclature. As it is, as Longrich et al. have technically created (albeit unnecessary) new name separating elegans from rarus, I am inclined to support its creation, and will refer to the taxon as Leptorhynchos elegans from here on. It might take some time to get around to a more formal reaction as I am currently attempting to gather more data on Caenagnathasia martinsoni, but base reaction suggests “Leptorhynchos” gaddisi may be more basal than described.

Edit: Commenters Cay and Vahe indicate that, contrary to my statements above, that Art. 70.1 does not allow automatic type fixation; elegans is not the type species of Leptorhynchos automatically. This means that, currently, “Leptorhynchos” is not an available name. Because of this, it should be possible to produce errata to the publication in which a type species is fixed, in which case I hope to prevail upon Nick Longrich that gaddisi may be more useful a type species.

[n1] Leptorhynchus derives from the Greek leptos (λέπτος), “small,” and rhynchos (ρύνχος), “jaw or beak.” The name is a reference to the smaller size of the purported beak between referred jaws and in gaddisi[n2] than to Caenagnathus collinsi and referred jaws to sternbergi.
[n2] “gaddisi” is coined to honor the Gaddis family. Typically, this would be written “gaddisorum,” as the suffix -orum is used to designate a group consisting of at least one male (and in cases where gender is not specified amongst the group’s members). However, the ICZN does not mandate that improper use of -i or -ae to -orum or -arum must be corrected: ICZN Art. 32.5 describes the appropriate necessary emmendations. It is assumed that all others are unnecessary, and thus -i does not need to be emmended to -orum, as it should.

Cracraft, J. 1971. Caenagnathiformes: Cretaceous birds convergent in jaw mechanism to dicynodont reptiles. Journal of Paleontology 45: 805-809.
Currie, P. J. 1989. The first records of Elmisaurus (Saurischia, Theropoda) from North America. Canadian Journal of Earth Sciences — Revue de Canadienne des Sciences de la Terre 26 (6): 1319-1324.
Currie, P. J. 2005. Theropods, including birds. pp.367-397 in Currie & Koppelhus (eds.) Dinosaur Provincial Park: A Spectacular Ancient Ecosystem Revealed. Indiana University Press (Bloomington, Indiana).
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. [Published in 1994, dated 1993.]
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.
Gilmore, C. W. 1924. A new coelurid dinosaur from the Belly River Cretaceous of Alberta. Canada Department of Mines Geological Survey Bulletin (Geological Series) 38 (43): 1-12.
Lamanna, M., Sues, H.-D., Schachner, E. & Lyson, T. 2011. A new caenagnathid oviraptorosaur (Theropoda: Maniraptora) from the Upper Cretaceous (Maastricthian) Hell Creek Formation of the western United States. Society of Vertebrate Paleontology, 2011 Meeting Abstracts: 140A.
Longrich, N. R., Currie, P. J., Dong Z.-m. 2010. A new oviraptorid (Dinosauria: Theropoda) from the Upper Cretaceous of Bayan Mandahu, Inner Mongolia. Palaeontology 53 (5): 945-960.
Longrich, N. R., Barnes, K., Clark, S. & Millar, L. 2013. Caenagnathidae from the Upper Campanian Aguja Formation of west Texas, and a revision of the Caenagnathinae. Bulletin of the Peabody Museum of Natural History 54 (1): 23-49.
Osmólska, H. 1981. Coossified tarsometatarsi in theropod dinosaurs and their bearing on the problem of bird origins. Palaeontologica Polonica 42: 79-95.
Osmólska, H., Currrie, P.  J. & Barsbold R. 2004. Oviraptorosauria. pp.165-183 in Weishampel, Dodson & Osmólska (eds.) The Dinosauria, 2nd Edition. University of California Press, Berkeley.
Parks, W. A. 1933. New species of dinosaurs and turtles from the Upper Cretaceous formations of Alberta. University of Toronto Studies, Geological Series 34: 1-33.
Scapino, R. 1981. Morphological investigation into functions of the jaw symphysis in carnivorans. Journal of Morphology 167: 339-375.
Sternberg, C. H. 1932. Two new theropod dinosaurs from the Belly River Formation of Alberta. The Canadian Field-Naturalist 46: 99-105.
Sternberg, R. M. 1940. A toothless bird from the Cretaceous of Alberta. Journal of Paleontology 14 (1): 81-85.
Sues, H.-D. 1997. On Chirostenotes, a Late Cretaceous oviraptorosaur (Dinosauria: Theropoda) from western North America. Journal of Vertebrate Paleontology 17 (4): 698-716.
Sullivan, R. M., Jasinski, S. E. & Van Tomme, M. P. A. 2011. A new caenagnathid Ojoraptorsaurus boerei, n. gen., n. sp. (Dinosauria, Oviraptorosauria), from the Upper Ojo Alamo Formation (Naashoibito Member), San Juan Basin, New Mexico. Fossil Record 3. New Mexico Museum of Natural History and Science Bulletin 53: 418-428.