122 Bashford Dean Memorial Volume 



A paper on Dinkhthys was published by Woodward in 1922. He gave here the 

 first more detailed drawing (unhappily on a very small scale) of a Dimchthys head shield 

 from the inside. In the same paper we also find a drawing showing the relative position 

 of the jaw elements of Dinichthys. 



The most recent paper (Stetson 1930) describes a new reconstruction of Dimchthys 

 terrelli Nwb. exhibited in the Museum of Comparative Zoology, Harvard College. The 

 mounted specimen is nearly complete (only the ventral shield is absent), and the whole 

 reconstruction must be accepted as the best one. Nevertheless it shows some mistakes. 

 The front part of the head is concave instead of being convex. The pre-maxilla is too 

 small in relation to the maxilla and to the whole head. It certainly belongs to another and 

 smaller individual. Both the combinations proposed by Stetson of placing pre-maxilla and 

 maxilla in relation to the sub'orbital are in the present author's opinion incorrect. Espe- 

 cially that combination shown on his PL II must be regarded as wrong, as we shall see 

 later when describing the gnathal elements in Dinichthys. The mandible is too strongly 

 curved and comes too near the antero-lateral plate. Possibly it belongs to another speci- 

 men. The body carapace lacks the postero-lateral plate. 



Since all attempts to rebuild Dinichthys have been more or less incorrect or incom- 

 plete, the present writer will try to give a new reconstruction of this animal, based on 

 his latest investigations, leaving its acceptance or rejection to the judgment of others. 



NOMENCLATURE OF HEAD AND BODY PLATES 

 AND SENSORY CANALS IN THE ARTHRODIRA 



If we look at the carapace structure of different Arthrodires, we find that in spite 

 of the great difference in form, adaptation and size, the number and relative arrangement of 

 the single plates are unusually constant. Such a constant arrangement of skeletal elements 

 is hard to find in other primitive fishes even when they belong to one order. This shows 

 that the Arthrodira are a group of very closely related forms and are also a very conserva- 

 tive group, since their structure changed very little during the enormously long time 

 from Lower to Upper Devonian. 



For this reason, it is possible to make a schematic drawing of an "ideal arthrodire," 

 showing all the plates typical of the Arthrodira in their characteristic positions. Such a 

 drawing might be a great help in the study of new and unsatisfactorily known forms 

 or fragments. 



In Text-figures 9, 10, and 11, I have tried to make such schematic drawings of the 

 plates of the head and of the ventral and dorsal body carapace of such an ''ideal arthrodire.'" 

 These represent not a definite form, but a composite — in which the most characteristic 

 points of all known forms have been assembled. The single plates are separated to show 

 the overlapping margins. (In reality the head plates of the Arthrodira were seldom found 

 isolated). All the sensory canals found in the various forms are included. The ossification 

 centers and ossification rays are also indicated on the plates at the right side of 

 the drawings. 



