Scales and tooth whorls of ancient fishes challenge distinction between external and oral 'teeth'.

The debate about the origin of the vertebrate dentition has been given fresh fuel by new fossil discoveries and developmental studies of extant animals. Odontodes (teeth or tooth-like structures) can be found in two distinct regions, the 'internal' oropharyngeal cavity and the 'external' skin. A recent hypothesis argues that regularly patterned odontodes is a specific oropharyngeal feature, whereas odontodes in the external skeleton lack this organization. However, this argument relies on the skeletal system of modern chondrichthyans (sharks and their relatives), which differ from other gnathostome (jawed vertebrate) groups in not having dermal bones associated with the odontodes. Their external skeleton is also composed of monoodontode 'placoid scales', whereas the scales of most early fossil gnathostomes are polyodontode, i.e. constructed from several odontodes on a shared bony base. Propagation phase contrast X-ray Synchrotron microtomography (PPC-SRµCT) is used to study the polyodontode scales of the early bony fish Andreolepis hedei. The odontodes constructing a single scale are reconstructed in 3D, and a linear and regular growth mechanism similar to that in a gnathostome dentition is confirmed, together with a second, gap-filling growth mechanism. Acanthodian tooth whorls are described, which show that ossification of the whorl base preceded and probably patterned the development of the dental lamina, in contrast to the condition in sharks where the dental lamina develops early and patterns the dentition.The new findings reveal, for the first time, how polyodontode scales grow in 3D in an extinct bony fish. They show that dentition-like odontode patterning occurs on scales and that the primary patterning unit of a tooth whorl may be the bony base rather than the odontodes it carries. These results contradict the hypothesis that oropharyngeal and external odontode skeletons are fundamentally separate and suggest that the importance of dermal bone interactions to odontode patterning has been underestimated.