GEOLOGICAL SURVEY IN SILURIAN ROCKS OF SOUTH OF SCOTLAND. 855 



distribution of fishes in geological time, however, which is given on the next page of 

 the same work, though still placing the Pteraspids, Cephalaspids, and Pterichthyids under 

 the Marsipobranchii, he qualifies the position by the judicious use of points of inter- 

 rogation. A few pages further on (p. 66), speaking of the Ostracodermi, including 

 the Pteraspidae, he indicates his opinion that they " are in no way closely connected 

 with the ancient shark types." 



For this doctrine of an affinity between the Marsipobranchii and any of the groups 

 reckoned to the " Ostracodermi," I myself never could perceive any real justification, 

 and I consequently found myself quite in agreement with the opinions expressed by 

 Prof. Lankester in the short paper (xvi.) which he wrote on the opposite side of 

 the question. Prof. Lankester lays principal stress on the entire absence of 

 any proof that the Ostracodermi were monorhinal like the Lampreys and Hags. I 

 would go further and ask, Where is the evidence that they were really "agnathous" ? 

 These fossils never show any trace of endo-skeleton at all — it must have been entirely 

 cartilaginous — so that it would be just as reasonable to affirm that they had no chon- 

 drocranium. Again, it is by no means so certain that the term " Agnatha," if taken to 

 indicate a primitive deficiency of the mandible or mandibular arch, can properly be 

 applied even to the Cyclostomes, for though the distinguished embryologist F. M. 

 Balfour believed that their ancestors never had lower jaws, and that they themselves 

 are the " remnants of a primitive and praegnathostomatous group," * the opinions 

 of those who consider these creatures to be degenerates from originally gnath- 

 ostomatous forms cannot be entirely overlooked (see Howes, x.). Huxley (xi.), and 

 following him, Howes (op. cit.) have even maintained that, in the cartilaginous frame- 

 work of the Marsipobranch head, elements are present which represent parts at least 

 of the mandibular arch in the true Gnathostomata,f though the latter author, from 

 other developmental reasons, emphasises the enormity of the gap which lies between 

 the Marsipobranchii and the other and higher Vertebrata. 



It is not, however, necessary to enter further into the agnathous question, if the 

 palaeontological facts described in this report indicate thai the Pteraspidae, Drepanas- 

 pidaa, and Psammosteidae, are derivable from the Ccelolepidae, and that the latter are 

 of Elasmobranch origin. 



For the same reason we need not enter into any very detailed discussion of the 

 second theory to which I have referred — namely, that of Prof. W. Patten, who seems 

 to be endeavouring to revive, on a scientific basis, what we have long considered to be 



* Comparative Embryology, London, 1881, vol. ii. p. 69. 



t Huxley says regarding Petromyzon (op. cit., p. 427) : — " The posterior lateral cartilages are directly connected 

 with that end of the suborbital arch, which answers to the articular end of the suspensoriuin in the frog, and in their 

 position exaggerate the peculiar arrangement of the tadpole's meckelian cartilage. That,they are parts of the mandib- 

 ular arch I believe to be certain, but in the absence of any knowledge of their mode of development, I leave the ques- 

 tion as to their exact homology open." This view is supported by Howes, who also finds a representative of the man- 

 dible in the prepalatine cartilage of Myxine, remarking, — " In its relationships to the superficial branches of the 

 trigeminal nerve, this ' prepalatine ' closely corresponds with that which Huxley claimed as Meckel's cartilage in the 

 Lamprey, and with that I hold it to be homologous notwithstanding Parker's view to the contrary" (op. cit., p. 133). 



