Armor for Agathaumas


Charles Knight’s painting of Agathaumas sphenocerus, from Wikimedia Commons (public domain).

As one of the Old Masters of paleoart, Charles R. Knight created many iconic depictions of prehistoric animals that have influenced countless successors and imitators. Among his signature pieces is his 1897 painting of the ceratopsian Agathaumas sphenocerus. It was originally published in a magazine article as part of a series of illustrations that were supervised by Edward Drinker Cope shortly before his death (Ballou, 1897). These were Knight’s first artworks of dinosaurs and arguably the launching point for paleoart as a serious artform in America. His Agathaumas was so widely copied by other artists that it became one of the earliest paleoart memes. It even appeared as a stop-motion model in the 1925 The Lost World, the grandfather of all dinosaur films.

Despite its initial popularity, Agathaumas faded into obscurity soon after, superseded by its more famous relative Triceratops. This has often lead to the scientific background of Knight’s painting being overlooked. In reality, his Agathaumas was an amalgamation of different ceratopsians known at that time. The long nasal horn was based on Cope’s Monoclonius sphenocerus1 and the short orbital horns on M. recurvicornis (Cope, 1889), while the body was based on Othniel Charles Marsh’s skeletal reconstruction of Triceratops prorsus (Marsh, 1891b). The latter is evidenced by Cope’s concept sketch which is a direct trace of it. It is ironic that Cope gave material from his hated rival to Knight as a reference. However, the body shape was not the only influence from Marsh’s papers.

Agathaumas maquette

Knight’s Agathaumas maquette, which has similar osteoderms to the painting, from Osborn (1898).

The origin of the extensive armor that Knight portrayed on Agathaumas has long been uncertain, usually being chalked up to “artistic license”. Osteoderms are currently unknown from ceratopsians and skin impressions later discovered in species like Triceratops horridus show less elaborate scales. Even Henry F. Osborn claimed that the armor was only speculative in an article about Knight’s maquettes used for his paintings (Osborn, 1898). This explanation started almost immediately after the art was published and has persisted to this day. Yet I was not convinced by it, feeling that it underestimated Knight’s scientific knowledge.

“The frilled Dinosaur, Agathaumus [sic] sphenocerus, Cope, is based upon a prior restoration, published by Professor O.C. Marsh, of his Triceratops prorsus, this genus and species being distinguished from Triceratops by the large anterior median horn and the small posterior paired horns. […] In addition to the powerful horns the skull is protected by a great bony collar or frill, which is surrounded by heavily barbed tubercles. The tubercular character also given by Mr. Knight to the epidermis is conjectural.”

Contrary to previous belief, I have found evidence that the armor was inspired by real fossils and not mere imagination. Like the body shape, this is another idea that Knight borrowed from Marsh. Marsh (1891a), plate X, figured an assortment of isolated fossils from the Lance Formation in Wyoming. He identified them as osteoderms from Triceratops, in part because he thought it was closely related to Stegosaurus. He briefly postulated on the placement of these elements on the body, but did not describe them in detail. Although confident in his identification, they had not been found in association with a Triceratops skeleton.

“Beside the armature of the skull, the body also in the Ceratopsidae was protected. The nature and position of the defensive parts in the different forms cannot yet be determined with certainty, but various spines, bosses, and plates have been found, that clearly pertain to the dermal covering of Triceratops, or nearly allied genera. Several of these ossifications were probably placed on the back, behind the crest of the skull (Plate X), and some of the smaller ones may have defended the throat, as in Stegosaurus.”

Triceratops dermal elements

Supposed “dermal armor” from Triceratops from Marsh (1891a), actually osteoderms from Denversaurus (1-10) and a squamosal horn from Stygimoloch (11-12).

It should come as no surprise that none of these actually belong to Triceratops. Fig. 1-3 is USNM 5793, a cervical spine2 referred to the nodosaurid Edmontonia sp. by Carpenter & Breithaupt (1986). Given its provenance, it instead likely belongs to Denversaurus schlessmani, which is distinct from Edmontonia (Bakker, 1988; Burns, 2015). The two other osteoderms figured are also probably from Denversaurus, but I could not find their specimen numbers. Fig. 4-7 is a large scute that might be part of the cervical rings, while fig. 8-10 is a smaller scute of uncertain placement. Fig. 11-12 is YPM 335, a right squamosal horn referred to the pachycephalosaurid Stygimoloch spinifer by Galton & Sues (1983). Stygimoloch may represent a juvenile stage of Pachycephalosaurus wyomingensis (Horner & Goodwin, 2009) or may be a separate species based on stratigraphic separation (Fowler, 2017).

The Stygimoloch squamosal is strikingly similar to the spikes and knobs running down the back on Knight’s Agathaumas. That detail in particular is what convinced me, but the other elements line up well too. The Denversaurus scutes match those on the flank, while the spine may have been the basis for the ones adorning the frill. The arrangement of the armor on the back, frill, and throat seems to conform with Marsh’s speculations. The fact that Knight was already familiar with Marsh’s reconstruction of Triceratops makes it plausible that he would have also known about the osteoderms. The only missing part is confirmation from Knight himself, which unfortunately was never published to my knowledge.

Armored Triceratops

Illustration of Triceratops with dermal spines by Joseph Smit, from Hutchinson (1893).

Interestingly, Knight’s Agathaumas was not the first ceratopsian depicted with armor. That would be the Triceratops prorsus drawn by Joseph Smit in Hutchinson (1893). The osteoderms again look similar to the Stygimoloch squamosal, especially the cluster around the neck. Unlike Knight, the influence from Marsh is directly stated and the squamosal is even figured. Whether Knight referenced Smit’s illustration or the two artists came to their conclusions independently is unknown. It at least demonstrates that Marsh’s “Triceratops dermal armor” was known to other artists and that Knight could have easily found out about it. Overall, I think this provides a satisfactory answer to a long-standing question about one of the most famous paleoart pieces in history.


1For an overview of the taxonomic mix-up between Agathaumas sylvestris and Monoclonius sphenocerus, see this blog post by Justin Tweet.                                                                                                                                                                            2The osteoderm terminology used here follows Blows (2001).

Addendum (1/22/2021)

Destin Bogart of the E.D.G.E. YouTube channel has made a great video about Agathaumas which includes information from this post. I helped edit the script and it was fun turning my writing into a video for the first time. Go give it a watch!



Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s