The Structure of Dinichthys 199 



Thus the pectoral spine can not be homologous with the "Ruderorgan." What this remark- 

 able structure is, and where it was attached to the carapace, only new finds can show us. 



It is interesting to point out that according to my discoveries the spinal in Dinichthys 

 has entirely lost its character as a spine and has become a bone connecting the dorsal and 

 ventral armor. There can be no question of its being homologous with the pectoral fins 

 of other fishes. 



THE GILL APPARATUS 



Definite traces of the gill arches have never been found in theArthrodira. We may 

 therefore suppose, that the gill arches in these fishes were not ossified. The question of 

 the Arthrodiran gills is therefore a very disputable point. Miller, Pander, Traquair 

 and Woodward (in "Catalogue") have said nothing about gill structures in the Arthro' 

 dira. But in calling our PSO plate in Coccosteus a ''probable operculum" Woodward 

 (1898) indirectly implied that he thought the gill openings were placed in the cleft 

 between the head and body. 



In a description of Coccosteus published in 1902, Jaekel spoke of the side plates of 

 the body carapace (ADL, PDL, AL and PL) as "Elementen des Opercul'Apparates" and 

 we are led to assume that he believed the gill openings were placed behind the body cara- 

 pace. But in 1907 he changed his mind and wrote ''Dass die Kiemen unterhalb der Plat- 

 ten der Wangenregion untergebracht waren, ha be ich durch ... die Auffindung der 

 Kiemenbogen bei einem Wildungen Verwandten der Coccosteiden bestatigen konnen. 

 Als aussere Kiemenspalte fungierte also der Schlitz; zwischen der Wangenregion des 

 Schadels und dem Vorderrand des Halspanzers." The "Verwandte der Coccosteiden" 

 with "Kiemenbogen" mentioned in this paper was never described or figured in later 

 papers by Jaekel. Therefore, we believe that he probably changed his opinion again. 



Kemna (1904) also thought that the gills were placed on the posterior part of the 

 head and that the lower opening between the head and body carapace served as a gill 

 opening. On the contrary, in Patten's opinion (1912), the whole body carapace of the 

 Arthrodire, corresponding to that of Bothriolepis, must be regarded as a "branchial shield" 

 which "served solely for the protection of the gills and heart." The openings for the 

 "peribranchial chamber" were behind the body carapace. 



But certainly most extraordinary is the opinion Jaekel expressed in his last papers 

 (1925, 1926, 1927). In correlation with his theory about the descent of fishes from land- 

 living animals, Jaekel tried to demonstrate that the oldest known fishes had no gills at all. 

 He writes (1927, p. 284) that : "Auch bei Coccosteus und ihren Verwandten ist nichts von 

 Kiemenspalten oder Oeffnungen bekannt. . . . Wir werden also auch hier mit der Mdglich- 

 keit rechnen, dass diese Fische . . . ohne aussere Oeffnungen atmeten." 



Lastly, we may mention that Woodward (1922) and Stetson (1930) supposed the deep 

 paired impressions (Text-figure 13, PL) on the hind part of the head of Dinichthys to be 

 the "branchial chambers." 



The only remains of a gill apparatus on record were described by Broili (1929) in a 

 German Acanthaspid (Lunaspis). These are three small paired arch impressions placed 



