204 Bashford Dean Memorial Volume 



much more difficult to find contact between these forms and the Arthrodira. The denti- 

 tion of Rhynchodontus is quite different from that of the Arthrodira, reminding one much 

 more of that of the Holocephali. Their jaw mechanism was certainly not Arthrodiran- 

 like. The condyle on the plate corresponding with ADL is only developed as a spine. 

 The shoulder girdle is far different from all plates usual in the Arthrodira. The plates 

 corresponding with PDL, PL, PVL, MV and IL are absent. The Acipenser-like (Jaekel) 

 hpad roof could not be Arthrodire-like. 



The difference between the Chondrostei and the Arthrodira is in reality very 

 obvious. It is impossible to homologi?e the structure of their head roofs. The course of 

 the sensory canals is also unlike. The position and mechanism of the jaws are of absolutely 

 different characters. All these facts make any close relationship between the Arthrodira 

 and the Chondrostei improbable. 



It only remains to dispute the relationship between the Arthrodires and the Elasmo- 

 branchs. This was sharply advocated by Stensio in his brilliant paper on the head of 

 7s/lacropetalichthys. According to his classical investigation, it is certain that Macropc 

 talichthys was related to the Elasmobranchs. Regarding Macropetalichthys as closely 

 related to the Arthrodira, Stensio then draws the conclusion that the Arthrodires also 

 were related to the Elasmobranchs. The present writer, however, cannot accept the 

 opinion that Macropetalichthys was closely related to the Arthrodira. In fact, if we 

 examine the structure of Macropetalichthys, it is hard to find much in common with the 

 Arthrodira. 



Beginning with the arrangement of the bony plates in the head roof of Macropetalich- 

 thys, we notice that all the attempts made by various authors (Eastman, Dean, Jaekel, 

 Stensio) to find an exact homology between these and the head roof plates of the Arthro- 

 dira, must be regarded as unsatisfactory. Arthrodirans, as we know, show an unusually 

 constant plan in the arrangement of the head roof. From the oldest known Acanthaspida 

 to the most recent Dinichthys we always find the same plates in the same relative position. 

 The arrangement in Macropetalichthys is of an absolutely different type. To establish a 

 homology between these two groups, Stensio must assume that one plate in Macrope- 

 talichthys corresponds to a part of, or to many plates in the Arthrodira. For instance. Mo 

 in Macropetalichthys, would correspond to L + 'j of B and % of A in the Arthrodira. Text- 

 figure 91 is a schematic drawing, showing the homology proposed by Stensio. In my 

 opinion, this homology is not satisfactory and can in no way clearly establish the relation- 

 ship between these two groups. The homology Dean proposed, wherein the head of 

 Macropetalichthys corresponded with the fused head and body armor in the Arthrodira 

 gives a better result. But Stensici's investigation has made this theory absolutely impos- 

 sible. As we know, the primordial neurocranium in Macropetalichthys goes to the hind 

 part of the head. If the head of Macropetalichthys corresponds with the head and body 

 carapace of the Arthrodira, we must expect to find that in them the neurocranium con- 

 tinues under MD. Surely nobody can accept this arrangement. 



If we turn to the sensory canals, we find that they are also differently developed in 



