Abstract Historically, humeri and femora have been treated as diagnostic elements for fossil phocid (Carnivora, Pinnipedia, Phocidae) identification. This resulted in the naming of a plethora of extinct phocid taxa with isolated humeri and femora as type specimens. However, the documented evidence of sexual dimorphism in pinnipeds, including qualitative and preliminary quantitative studies of long bone shapes, indicates the existence of both intra‐ and interspecific variation; their respective strengths will help determine whether long bones should or should not be used for taxonomy. The present study quantitatively assesses the morphological variability of phocid long bones, using 3D geometric morphometrics on a large sample ( n = 145) covering 16 extant taxa. Morphospaces resulting from principal component analyses sort both extant phocid subfamilies (Phocinae and Monachinae) well, for both humeri and femora. Thus, this study suggests that humeri and femora might be diagnostic enough only to differentiate between subfamilies and maybe some genera, but usually lack clear genus‐ or species‐specific shapes when variability is considered. As such, isolated long bones are preferably not to be used as type specimens. A taxonomic reassessment of holotypes of fossil species based on isolated long bones on a taxon‐by‐taxon basis is therefore warranted.
Dewaele et al. (Thu,) studied this question.