The Structure of Dinichthys 



205 



the two groups. The homology proposed by Stensio is only uncertainly probable, espe- 

 cially when we notice that his '"pre-opercular canal" has only been found in a very badly 

 preserved specimen of Macropetalichthys. The absolutely constant arrangement of the 

 canals in the Arthrodira is not like that in Macropetalichthys (Text-figure 91). 



No traces of the body carapace have ever been found in Macropetalichthys. We must 

 suppose that it was not developed. If developed at all, it was certainly not movably 



Text-figure 91. 

 The head roofs of an Arthrodire (left) and of 

 Macropetalichthys (right), to show the homolo- 

 gies between the single plates according to Sten- 

 sio. The positions of the sensory canals are 



shown by dotted lines. 

 P, =PtO, P.,=M: S=EB, L, =iPrO. M,, =MB + i 

 right C + i left C+i right PrO+h left PrO; L,=hC: 



^3 "aCf. 



R, P, PM, SO, PSO, and PH in an Arthrodire are without 



homologies in MacropetaUchthys. 



M| in Macropethalichthys has no corresponding plate in an 



Arthrodire. 



connected with the head roof since no traces of joint sockets on the occiput are known. 

 Thus, Macropetalichthys had no neck joint, — a trait most characteristic for the Arthrodira. 

 This excludes also the possibility that the Macropetalichthyids had the same jaw mechan- 

 ism as the Arthrodirans. Furthermore, the whole complicated visceral skeleton of an 

 Arthrodiran (with its 14 plates) is unknown in Macropetalichthys . 



I might remark here that in Synosteus, one of the Wildungen Arthrodira briefly 

 described by Jaekel (1906.2), the head is secondarily fused with the body carapace. 

 Therefore the head could not be moved in relation to the body, and the whole jaw 

 mechanism must have been different. We may note here that in the first place this form 

 is so incompletely figured and described that it is difficult to draw any conclusions. And 

 in the second place, the whole arrangement of the plates, as far as it is possible to see it 

 from JaekePs drawing, is precisely Arthrodiran-like. This form, must have represented 

 one of the most highly developed types of the Arthrodira, which, overspeciali2;ed, lost 

 the most characteristic feature of the whole class. Nevertheless, it can be an Arthrodire, 

 just as, for instance, the Struthioideae, though wingless, are still birds. 



In Stensio's opinion, Phlyctaenaspis and Homostius ''seem to be most nearly related 

 to Macropetalichthys.'" Both these forms are typical Arthrodires, and, in my opinion, 

 are no more closely related to Macropetalichthys than are any other Arthrodira. The 

 mere fact that the eye openings in Homostius are moved nearer to the median line and 

 surrounded by PrO, PtO and C, instead of by PrO, PtO and SO as in other Arthrodires, 

 does not make it any more like Macropetalichthys. 



We can say nothing about the likeness or unlikeness between the structure ot the 

 primordial neurocranium in Macropetalichthys and the Arthrodira, so long as it is ab- 

 solutely unknown in the latter forms. We can only point out in regard to the sensory 



