The uppermost Hell Creek and related beds on the Laramidian side of the regressing interior seaway of North America during the Maastrichtian has played host to a number of titanic ceratopsians. Given the names Triceratops, Torosaurus, Nedoceratops, debate has plagued these taxa as to whether they are synonyms. Outliers have been named, included apparently basal relatives such as Eotriceratops, Ojoceratops, and even Tatankaceratops.
Triceratops used to be very speciose, with 10 and more species encapsulated under the umbrella, all displaying reasonable variation that, in the early days of dinosaur paleontology, we can potentially excuse the describers to thinking reflected systematic variation. But advances mean changes. These species have been whittled down, and now only two remain: horridus (the type species) and prorsus. On the other hand, the taxon Torosaurus has only had three species named for it, latus (the type), gladius (subsumed into the former), and utahensis (which some suggest should be split off). Additional possible species have since been named, but phylogenetic analysis renders them as outliers, with the caveat that Torosaurus and Triceratops are generally considered distinct clades, which Nedoceratops hatcheri often finds itself in the middle of.
Speaking from the finding of samples of skulls across a broad swath of ontogenetic stages from across the Hell Creek of Montana, Johns Scannella and Horner have presented an argument that renders all Hell Creek ceratopsians (generally) into one taxon, Triceratops, although the question of which species fall where has been left open (and I would say problematically so, as this renders the debate of synonymy questionable). In response, however, Andrew Farke and Nicholas Longrich have produced a series of responses, to which in turn Scannella and Horner have responded, which have both either supported the validity of Torosaurus with respect to Triceratops (it is an outlier to the Triceratops clade comprising the two species), or Nedoceratops hatcheri (as a distinct taxon intermediate between the two). This has become known as “Toroceratops,” a spirited debate that spans the literature for the last few years and is allowing researchers to bring tool after tool into the arsenal of answering that most important of taxonomic questions:
“Is this taxon valid?”
Andrew (Andy) Farke has a new paper out in PLoS in which he and his colleagues (Maiorino, Farke, Kotsakis & Piras) use 2D lateral view geometric morphometrics to assess total and partial skull shape among a broad sampling of Triceratops, Torosaurus, and Nedoceratops skulls to answer whether the shapes of these skulls tend to cluster together regardless of or in conjunction with ontogeny. Principle to Scannella and Horner’s theory is the argument that the Torosaurus skull morphology reflects a full maturity of the cranial morphology that, despite size, no Triceratops skull has ever reached. This would suggest that Torosaurus is a peramorphic adult of the Triceratops skull, which retains a decidedly “adult” aspect for most of its life, but then drastically changes. Unfortunately, as Farke and his coautors note, the absence of coordinate postcranial long bone (femur, tibia, rib) histology associated with skulls and placed in stratigraphic sequence confounds work that has, effectivelly, focused on a highly variable set of bones, the skull. Morphometrics is a good way to sample this work, but it seems that it is merely one step out of many. Next would be to sample geometrically different views of the skull (dorsal, ventral) as well as to potentially use full 3D morphometrics to capture conclusively the full range of variation possible. This latter option preserves the ability to create landmark point clouds, but refines them in that it can do this regardless of perspective and with higher difelity: 3D point clouds are generated by meshes created from scanning in three dimensions the skull directly, rather than a map created from a side view that, barring some artsy skills, is plagued by the distortion of size relative to the camera.
But this is not to say that the study is bad; it is not. Farke et al. fiund that Torosaurus skulls do not group with Triceratops skulls, though some odd specimens that have been considered integral to the debate seem to be on the margin between then. This isn’t the final answer, and the “Toroceratops” debate is far from over.
Maiorino, L., Farke, A. A., Kotsakis, T. & Piras, P. 2013. Is Torosaurus Triceratops? Geometric morphometric evidence of Late Maastrictian ceratopsid dinosaurs. PLoS ONE 8 (11): e81608.