POSSIL FISHES. 21 



the postero-lateral piece with the so-called ' radius,' the more especially 

 as the antero-lateral piece corresponds with that part of the thoracic shield 

 of Pterichthys which supports the plated appendage representing the pec- 

 toral fin in that genus. 



" On the other hand, it must be confessed that the closer connection of 

 the antero-median piece with the thoracic plates than with the hj'oidean 

 cornua, and the very backward position of the postero-lateral plates, ap- 

 parently out of reach of any connection with the iins, militate against this 

 view ; which, in addition, leaves the median rhomboidal plate unac- 

 counted for. 



"In the Siluroids to which I have referred (Clarias, Bagrus, Arms, 

 etc.) and in Loricaria, a vast latero-ventral shield is produced by the pro- 

 digious expansion and coalescence of the bony elements which are homo- 

 logous with those termed 'coracoid' and ' radius ' in other fishes. 



" Viewed from the ventral surface, these bones form four great plates, 

 those of each side being closely united, or even amalgamated together, 

 while the opposite pairs are joined in the middle line by a strongly 

 serrated suture. 



" VV hen the pectoral fin is provided with an anterior spine, this is articu- 

 lated by a curiously complicated joint with the so-called coracoid. The 

 cornua of the hyoid are large stout bones, and the urohyal, also a large 

 and strong bone, which is particularly broad in Lorica/ria, connects the 

 hyoidean with the pectoral apparatus." 



Mr. W. Kitchen Parker, in his " Monograph on the Shoulder-Grirdle " 

 (p. 23, et seq.), discusses, with, much learning and ability, the homologies 

 of the pectoral shields of the Siluroids and that of Goccosteus, but does not 

 fully adopt the suggestion of Prof. Huxley in regard to these. By him 

 itlie bones of Clarias and Lorioaria, marked o and d in the above wood- 

 cut, are considered to be, respectively, clavicles and interclavicles, while 

 the anterior pair of plates of the ventral shield of Cocoosteus (marked c' in 

 the figures), he regards as interclavicles, and considers them as the coun- 

 terparts of the lowest bone in the shoulder-girdle of the sturgeon. Of 

 the posterior pair of plates in Ooocosteus, he says : " The bones lettered 

 d' may either belong to the post-clavicular cincture, or they may answer, 

 in their sub-cutaneous portion, to the second pair of interclavicular bones 

 of the LophobranoMi {Hippocampus, Syngnathus, etc.)." He says, fur- 

 ther : " The keystone pieces (&' e of figures) are the exact counterparts of 

 the first two abdominal line plates of the Lophdbranchii." 



From the almost complete identity of structure in the ventral shields of 

 Goccosteus and Dinichthys, the questions here raised are of special inter- 

 est in the study of the latter genus ; and, whatever conclusion shall be 

 reached in regard to one, will equally afEect the other. 



