The Structure of Dmichthys 



193 



case . . . firmly united with the cranial roof The thickenings on the front part of PrO 

 (Text-figure 13 FP) ''may be regarded as the lateral ethmoid" (or ectethmoid). Finally he 

 called the deep paired impressions on the hind part of the head roof (Text-figure 13 PL) 

 "the branchial chambers." 



We have learned that no "ossified cartilage" or "ossified walls of the brain'case" are 

 to be found in Dmichthys. All these parts named by Woodward are, in reality, only 



Text-figure 82. 

 A reconstruction of the neurocranium of Dmichthys : 



A, according to Stetson; x-x, plates bounding posterior wall of anterior division of occipital region; 



y-y, branchial chambers: 

 B, according to Stensio (the contour of the head is added by the present writer). 



thickenings on the plates composing the head roof and not parts of the neurocranium. 

 The impressions for branchial chambers, as we have seen before, are each really occupied 

 by a musculus depressor capitis. They are, in all cases placed too high up to be occupied 

 by gill apparatus. 



The next to describe the neurocranium of Dinichthys was Stensio (1925). On the 

 basis of his detailed investigation of the neurocranium in Macropetalichthys, Stensio tried 

 to reconstruct the neurocranium in the Arthrodira (Text-figure 82B). It must here be 

 pointed out that in the present writer's opinion Macropetalichthys is not related to the 

 Arthrodira, and, therefore, the structure of the brain-case in this form can not give us any 

 evidence about the brain-case in the Arthrodira. (A discussion on the relationships of 

 the Arthrodira is to be found in the last section of this paper.) The sketch of the neuro- 

 cranium given by Stensio, is based on the figures of the head roof of Dinichthys given by 

 Newberry and by Woodward. From drawings it is very difficult to get the correct pic- 

 ture of the inner surface of the head of Dinichthys. Therefore Stensio's reconstruction in 

 some directions does not correspond with the real facts we find in Dinic/itfi)'5. The pre- 

 orbital processes shown by him could hardly fit into the impressions on the under side of 

 the head. The same is true of the large side processes occupying Woodward's "branchial 

 chambers." In fact, these chambers are divided from the central impressed part of the 



