And I’m only going to describe this using the mandible. Let’s get started:
In the above figure, I’ve noted several features using arrows and numbers. These are omitted from the third mandible shown as it is included to compare. The mandibles are not to scale, and are size to fit the same mesiodistal length.
1. Lateral fossa of the dentary rostral to the external mandibular fossa is both distinctively offset from the remaining surface of the dentary but is large. In Gigantoraptor erlianensis, this fossa expands to most of the dentary’s lateral aspect. In Caenagnathus collinsi, this fossa bears several very large foramina, and some mandibles referred to Caenagnathus (or as Chirostenotes) bear but one large foramen, although none appear present in Gigantoraptor erlianensis. Instead, in the latter there are two distinct “fossae” (afo and pfo in the figure, respectively) that are further subdivisions of this broader depressed area. (Reason 1.)
2. The dentary and surangular have a distinctive sigmoid contact, where the posterior dorsal process of the dentary produces a tongue that sits between the anterior dorsal and ventral processes of the surangular. The posterior dentary process is lateral and ventral to the ventral surangular process, and the apex sits below the mid-height of the bar that forms the dorsal arch of the mandible. In oviraptorids, this articulation places the apex around mid-height, and the anterior dorsal process of the surangular sits slightly above the dentary and forms part of the “edge” of the dentary, unlike in caenagnathids. (Reason 2.)
3(a, b and c). The external mandibular fenestra employs a suite of features that can be attributed to reasons 3, 4 and 5, but are labeled “3” above only for simplicity. 3a is the aspect of the fenestra itself, which is elongated, ovate, and low compared to oviraptorids, in which it is at least as dorsoventrally tall as it is rostrocaudally long. 3b is the shape of the fenestra which, while ovate, is rather lenticular, or “kidney”-shaped with a small bowing into the dorsal margin from the surangular; in oviraptorids, this would be the “surangular process.” 3c is the relative size of the fenestra, which is at least 30% mandibular length and less than 60% mandibular height (erring on the side of caution, it is about 50% or less), while in oviraptorids this fenestra is less than 25% mandibular length and ridiculously tall and well over 75% of the mandibular height. These estimates are based on gross greatest lengths, rather than a comparison of length from a singular axis of length and height. Those estimates also show a clear distinction between oviraptorids and caenagnathids. This is only questionable in Oviraptor philoceratops, where the mandibular fenestra is both long relative to the mandible, and tall. (Reasons 3, 4 and 5.)
4(a and b). In all caenagnathoid oviraptorosaurs, the angular is a long, thin, splint-like bone with a U-shaped cross-section, where the U is lopsided so that the open end is oriented medially and dorsally, and the dorsal “arm” of the U is the lateral surface. This arm is relatively tall in caenagnathids, but shallow in oviraptorids (4a), overtaken by both the greater dimensions of the surangular and by the closer reach of the ventral posterior process of the dentary, which more closely contacts the surangular in oviraptorids than in caenagnathids (4b). (Reasons 6 and 7.)
5. The surangular’s ventral edge is a long sinuous margin that is often fused to the angular in oviraptorids, but never in the caenagnathids. However, in caenagnathoids the ventral edge of the surangular forms the posterior margin and extent of the external mandibular fenestra. In caenagnathids, this is a blunt terminus, almost completely vertical or sloped posteroventrally, while in oviraptorids a thin tongue extends along the angular and forms a rounded posterior extent of the fenestra while also coming much closer to sloping anteroventrally. (Reason 8.)
6(a and b). The surangular is shallow in caenagnathids while it is deep and very triangular in oviraptorids. When isolated, the surangular has a sigmoid aspect in caenagnathids, and it becomes pinched above the posterior extent of the external mandibular fenestra, while the inverse in true in oviraptorids, where this is its greatest depth. The surangular’s ventral margin anterior to the “pinch” forms a slight ventral bulge, and is analogous to the “surangular process.” In some specimens, such as FMNH PR 2081, the “surangular process” is angled into a distinct point, while in most other specimens it is very rounded. This process (6b) is extended into a long triangular process in every oviraptorid, and in no other oviraptorosaur. (Reasons 9 and 10.)
7(a andb). The articular is relatively tall in caenagnathids, thanks to a distinctly high intercotylar ridge (7a) and the placement of the articular surface of the articular bone upon a small pedicle of bone (7b) formed by the surangular. Thus, the articular sits on a small “neck” in caenagnathids, but is flush with the remainder of the surangular dorsal margin in lateral view in oviraptorids, lacking the pedicle, and possessing a shallower intercotylar ridge. (Reasons 11 and 12.)
8(a and b). The retroarticular process is a complex structure formed ventrally and laterally by the angular and surangular, with much of its medial structure by the angular, and a portion of the medial side by the articular. The surangular is restricted from the distal terminus of the process by an expansion of the angular in oviraptorids, while the opposite is true in caenagnathids (8a). However, more distinctly is that the process is very narrow and dorsoventrally tall compared to width in caenagnathids (8b), while in contrast, it is very shallow and broad in oviraptorids. (Reasons 13 and 14.)
9. The mandibular symphysis over 25% mandibular length in all caenagnathids, and less than 20% in oviraptorids. This is relatively distinctive and is one of their most distinctive traits. (Reason 15.)
10(a and b). The mandibular symphysis is fuse, with the suture completely obliterated in all caenagnathids (10a), and is shallowly inclined to the mandibular long axis (10b), although some oviraptorids approach the 20° or lower mark that separates them. This value is fairly difficult to calculate: while it might be simple to just take a line from the ventral extent of the symphysis in lateral view and draw it to the anteromesial extent in lateral view, this misses the fact that the ventral extent of the symphysis is separated from the ventral margin by a sulcus on the ventral surface, so the value in caenagnathids is lower by that additional value. And while this is an additional feature (sulcus) that distinguishes caenagnathids from oviraptorids, it cannot be determined from the published figures in Gigantoraptor erlianensis. (Reasons 16 and 17.)
11. The lingual (=dorsal) surface of the mandibular symphysis is richly ornamented in caenagnathids, partly due to the fusion of the dentaries to one another and also because of the presence of medial remnants of the sub-dental platforms that support teeth (Currie et al., 1994). In all caenagnathid mandibles, the symphysis is marked with a dorsal U-shaped sulcus in its posterior margin, although this is more or less actually two arching lines that are themselves divided in most caenagnathids by an additional, more mesial sulcus. This forms a long, thin channel that deepens posteriorly, with the posterior half being a broad “U.” This feature is absent in oviraptorids. (Reason 18.)
12. I’m kind of cheating here as while I separated these above to make two unique features noticeable in the figure, they are very strongly inter-related, and form a suite. As above, the sides of the lingual surface of the dentary preserve remnants of the sub-dental platforms in all caenagnathids, but these are formed into a pair of “ridges” or sharp profile changes in the slope of the side when viewed in section. These can be seen in the images above as forming a pair of lateral “shelves” on the symphysis, and may be part of the triturating surface. The distinction between the posterior sulcus and that of the ridges is that the ridges continue mesially while the shallow depression between them is distinct from the more posterior U-shaped sulcus. Nonetheless, the lingual surface of the caenagnathid mandible is a marvel to behold. (Reason 19.)
And this brings us to the very last, reason 20:
The profile of the mandible has its height:length ratio very, very low in caenagnathids (17% in CMN 8776) while in oviraptorids, it is very high (37% in GI 100/978). LV V0011 sits in the middle at 32%, where the measure is the greatest linear height (dorsal extent of the coronoid process to the deepest portion of the dentary, and greatest linear length to the retroarticular process. As predicted, AMNH PR6517 is close to the caenagnathid, but still high, at 32%. When removing the variable of the potentially confusing and likely diet-influenced length of the retroarticular process, and thus measuring length only to the posterior extent of the articular surface itself, these measures differ: CMN 8776 at 18% and GI 100/978 at 42% set the spectrum of size, where AMNH PR6517 is at 36% and LH V0011 is at 33%. Gigantoraptor erlianensis therefore has the second shallowest mandible among the sample, and that’s including the shallowest “true” oviraptorid, Oviraptor philoceratops.
Xu et al. (2007) note that several postcranial features are suggestive of caenagnathids, while several more point to more basal affinities. I must admit, without better photos, examination or otherwise informed argument, I cannot say anything about this. The limb anatomy implies a caenagnathid leg, with a short MCI:MCII ratio implying a more basal or oviraptorid-like manus, and the latter is similar to caenagnathids only in the relative shallow offset of the distal condyles of MCI. Large caenagnathids are known from the Kaiparowitz Formation of Utah (Hagryphus giganteus Zanno and Sampson, 2005) and Hell Creek Formation of South Dakota (Currie et al., 1994; Russell and Triebold, 2005), but these taxa imply sizes in the order of 2-3m in length, while Gigantoraptor erlianensis is at least 8.5m, and probably a bit larger. Xu et al. (2007) note that it may not have been fully grown, that despite full sutural closures and an external fundamental system formed on its limb bones, the LAGs imply less than 10 years of age (Xu et al., 2007: supplemental online material). GI 100/42, the largest oviraptorid known, was under 2m long, which implies that not only were caenagnathids larger in general, they included the largest oviraptorosaur known.
So there you have it, 20 reasons Gigantoraptor erlianensis was the biggest, baddest, nastiest thing you ever saw … and also a caenagnathid.
Clark, J. M., Norell, M. A. & Barsbold R. 2001. Two new oviraptorids (Theropoda: Oviraptorosauria), upper Cretaceous Djadokhta Formation, Ukhaa Tolgod, Mongolia. Journal of Vertebrate Paleontology 21(2):209-213.
Clark, J. M., Norell, M. A. & Rowe, T. 2002. Cranial anatomy of Citipati osmolskae (Theropoda, Oviraptorosauria), and a reinterpretation of the holotype of Oviraptor philoceratops. American Museum Novitates 3364:1-24.
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.
Osborn, H. F. 1924. Three new theropod dinosaurs from the Protoceratops zone of Mongolia. American Museum Novitates 144:1-12.
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.
Zanno, L. E. & Sampson, S. D. 2005. A new oviraptorosaur (Theropoda, Maniraptora) from the Late Cretaceous (Campanian) of Utah. Journal of Vertebrate Paleontology 25(4):897-904.
Xu X., Tan Q., Wang J., Zhao X. & Tan, L. 2007. A gigantic bird-like dinosaur from the Late Cretaceous of China. Nature 447(7146):844-847.