DEVONIAN FISHES OF IOWA 173 



is the only plate traversed 'by all three of these canals. The 

 pre- and' post-orbital systems are sometimes confluent in other 

 forms, but the occipital (or 'more properly, the occipito-central) 

 does not unite with these other systems save only in Macropetal- 

 ichthys. 



Evidence of the primitive nature of this genus, as compared 

 with other Arthrodires, is furnished by the following characters, 

 which are strongly indicative of embryonic or ancestral condi- 

 tions: (1) continuity of the sensory canal systems; (2) discrete- 

 ness of the central elements and their separation on either side 

 of the middle line by the elongated median occipital; (3) re- 

 duced number of median series of plates ; (4) complete enclosure 

 of the orbits within the headshield; and (5), absence of any 

 evidence of articulation, overlap, or other connection between 

 the headshield and a system of dorsal body plates. 



That the headshield was produced posteriorly over the nuchal 

 region, similarly, and in fact to about the same extent as in 

 Homosteus, is apparent from the configuration of the under 

 surface and structures seen within the interior of the cranial 

 buckler. An interesting feature first noted by Cope is that a 

 passageway for the notochord is provided by an ossified tubular 

 sheath of small diameter, which is supported by the narrow 

 posterior extension of the parasphenoid, and perforates by 

 means of a triangular orifice the thin, vaulted and backwardly 

 sweeping partition or septum depending from the under side 

 of the occipital plates, and interpreted by the present writer as 

 the ossified posterior wall of the chondrocranium. That this 

 septum actually closed the chondrocranium behind seems ex- 

 tremely probable both from its form and position, which are 

 highly suggestive of the conditions observed by Traquair * in 

 Dipterus ; and by its suspension from the cranial roof in a man- 

 ner recalling that in Neoceratodus. The septum is, however, 

 very thin-walled, and there is no evidence of separately ossi- 

 fied exoccipital plates adjacent to the triangular foramen 

 magnum. Another thin transverse septum depends vertically 

 from the posterior border of the headshield, and like the first, 

 is rigidly united with the narrow extension of the parasphenoid 



* 



Ann. Mag. Nat. Hist. 1878, ser. 5, 2, p. 5, pi. 3, fis 



