The elusive arms of the ammonite

Paleontology is littered with unsolved questions. Some can be answered, but just not now; maybe additional specimens are needed, or new techniques. Others are bound to remain unanswered because they require evidence that can't be obtained (for example attributes that can't be fossilized, or organisms that lived in times and places that neglected to leave a depositional record). One longstanding question is what did the business end of an ammonite look like? Presumably they had some kind of appendages appropriate to cephalopods; it's been assumed that they had ten arms, based on their closest living relatives, but they've been coy about leaving direct physical evidence.

A paper just published by Smith et al. (2021) aims to shed light on the brachial apparatuses of certain Late Cretaceous ammonites in the Scaphitidae. If you're mildly conversant in Upper Cretaceous marine rocks but not an ammonite fanatic, "Scaphitidae" will probably put you in mind of the famous Scaphites and its lazy outer whorl, but other members of the family were more traditional in appearance. Over the years, little hook-like bits have been found with specimens of a couple of scaphitid genera, specifically Hoploscaphites and Rhaeboceras. There was initially some thought that these were radular "teeth" (the radula being the molluscan's scraping food processor), but they seemed to be too large, and then genuine radular teeth turned up.

This still left the hook-like bits without an identity. Smith et al. used CT-scanning on several ammonites to show the numbers and locations of the objects within the shells. They differ significantly between Hoploscaphites and Rhaeboceras. In Hoploscaphites they are all a few mm long and have two paired tips, a bit like someone making air quotes. In Rhaeboceras, they range in length from 1 mm to more than 1 cm and come in nine different shapes. Six of them have two cusps, but they are generally strongly unequal in shape, whereas other are tricuspid, pointed, or rounded. They were found in clusters in the shells, with no pattern to distribution except that they were not found inside the jaws in those shells with jaws preserved. The different morphologies of the Rhaeboceras bits are arranged in straight or curved longitudinal rows with the cusps pointing in one direction (straight row) or outward (arc), and certain morphotypes tend to be found in association. The number of elements in a shell can be as many as 168.

Figure 2 from Smith et al. (2021). A shows a specimen of Rhaeboceras halli with hooks, B is a detail image of the hooks in situ, C shows examples of radular teeth, and D and E show eight of the nine hook morphologies. CC BY 4.0.

Smith et al. interpreted the elements as belonging to a pair of tentacular clubs, and provided a reconstruction of Rhaeboceras in life with the clubs extended. Hooks are known from belemnoids and squids, but there is evidence that belemnoid hooks and squid hooks derive from different tissues. The presence of hooks in scaphitid ammonites would show that a third cephalopod group acquired them. They also provide some resolution on the arrangement and nature of ammonite arms.

Rhaeboceras in action; interestingly, this ammonite would have had more diverse hooks than belemnoids and squids. A wider range of prey? Figure 6 in Smith et al. (2021). CC BY 4.0.

References

Smith, C. P. A., N. H. Landman, J. Bardin, and I. Kruta. 2021. New evidence from exceptionally "well-preserved" specimens sheds light on the structure of the ammonite brachial crown. Scientific Reports 11:article 11862. doi:10.1038/s41598-021-89998-4.