Referred specimens: NMV P16198, upper left or lower right canine with complete crown (Fig. 4), collected by J. M. Wilson.
Description: The tooth NMV P16198 is nearly complete, with only the distal tip of the root missing (Fig. 4). NMV P16198 was originally identified as a squalodontid cetacean tooth (Gill, 1957), but was subsequently identified as a pinniped, possibly representing a phocid incisor (Fordyce, 1982; Fordyce & Flannery, 1983; Long et al., 2002). While the narrowing at the base of the crown is reminiscent of incisors, NMV P16198 is interpreted here as a canine based on a relatively thick root, as opposed to the tapering, mesiodistally flattened nature of incisor roots. This thickening of the root is only present in the derived caniform incisors of Leopard seals, Hydrurga leptonyx (Blainville, 1820), which have a more exaggerated narrowing at the base of the crown, and is not present on the incisors of any other pinniped. The enamelocementum boundary allows for orientation, with a straight profile characteristic of the labial aspect of a canine and an angular enamelocementum boundary profile characteristic of the lingual aspect (Fig. 4). The cingulum is prominent on the lingual side, being slightly smaller on the labial side. Apicobasal ridges are concentrated on the labial side of the canine, being more prominent distally. This specimen is interpreted as either a lower right or an upper left canine, on the basis that the root is recurved towards the labial side, a ridge is present on the lingual side and there is a strong curvature along the tooth as a whole. While the ridge and the strong curvature are more common in mandibular canines, they are also variable between canine positions. Specimens of Hydrurga leptonyx (NMV C23614 and NMV C31561) demonstrate that the lingual ridge can be present on the lingual surface of upper canines, as well as the lowers. Upper canines of Ross seal, Ommatophoca rossii (Gray, 1844) (NMV C7354.1), and Weddell seal, Leptonychotes weddellii (Lesson, 1826) (NMV C7297, NHMUK 1951.5.1.4), also possess strong curvature, demonstrating that it is not exclusively a mandibular character. Hence these characters cannot be used to ascertain between the upper or lower canine positions.
The relatively short, curved crown and the apiciobasal ridges (an aquatic adaptation; McCurry et al., 2019) rule out possible carnivorous marsupials or terrestrial placentals. The presence of strong apiciobasal ridges and the robust nature of the crown (as opposed to the slender nature of otariids) suggest NMV P16198 may be a phocid. While some phocids do have strong apiciobasal ridges (such as Hydrurga leptonyx; NMV C23614 and NMV C31561), such ridges are highly variable in pinnipeds, and so this feature cannot be used to confidently assign NMV P16198 to Phocidae. The relatively small size is only slightly larger than the size of canines in the phocine seals Phoca Linnaeus, 1758 and Pusa Scopoli, 1771 (Supporting Information, Appendix S6). However, it should be noted that canines can be sexually dimorphic. The only monachine that approaches the size of these small seals is Australophoca changorum Valenzuela- Toro et al., 2016, for which no craniodental fossils are known (Valenzuela-Toro et al., 2016).