GREGORY, PRESENT STATUS OF ORIGIN OF TETRAPODA 319 



show the least indication of an internal skeleton, nor is there ever any 

 trace of branchial arches and internal jaws. The pouch-like markings 

 on the inner side of the carapace in Pteraspis, the large sack-like mark- 

 ings in Thelodus pagei, the small round gill openings of several genera 

 indicate a branchial apparatus somewhat similar to tliar of the larval 

 Petromi/zon. The median opening in front of the paired eyes in Cepha- 

 laspis and its allies suggests also the median nostril of Petromyzoii. Due 

 caution is necessary in accepting any of these resemblances as indicative 

 of real affinity with cyclostomes, yet they apparently give evidence of the 

 mode of functioning of the mouth and phar}Tix. 



In brief, the ostracoderms, instead of being free swimming predatory 

 types, with powerful locomotive organs, strong internal jaws and true 

 teeth^ as in the later fishes and ancestral Tetrapoda, were, with the ex- 

 ception of the Birkeniiclae, sluggish and bottom-living forms accustomed 

 to feed on small organisms, which they obtained in the mud. or by draw- 

 ing water into their capacious pharm-x. 



In these animals the process of cephalogenesis, or synthesis of elements 

 into a complex head, was probably in a low stage of development, most 

 of the ^'cephalic buckler"' in forms like Drepanaspis and Cephalaspis 

 ]epresenting an expanded pharyngeal region rather than a large brain, 

 which may have been as small as it is in recent cyclostomes. 



The ostracoderms represent a stage in chordate phylogeny immediately 

 preceding the acquisition of an endoskeleton impregnated with mineral 

 salts ; their first visceral arches, if present, had not yet been transformed 

 into primary or cartilaginous jaws; the process of cephalogenesis was in 

 them in a low stage, and the elements of their shelly exoskeleton were 

 potentially homologous with the cosmine, vasodentine and isopedine of 

 ])rimitive ganoids. To that extent they stand in a pre-gnathostome stage 

 of evolution and probably represent the remote forerunners of the ganoids 

 and Tetrapoda, while possibly having remote relationships also in another 

 direction with the ancestral elasmobranchs. 



The Antiarchi have advanced beyond the typical ostracoderms in hav- 

 ing the head sharply differentiated from the thorax and the mouth armed 

 with functional ja^AS, which are fashioned from the dermoskeleton. But 

 not even the excellently preserved specimens of Bothriolepis described by 

 Patten (1912) show any traces of car.ilaginous jaws, branchial arches or 

 cartilaginous axial skeleton. 



The systematic position and relationships of the Arthrodira constitute 

 so controversial a subject that it would be the part of discretion to avoid 

 it if -possible. Unfortunately this question as well as certain still larger 

 cnes cannot be altogether kept out of the discussion as to the origin of 



