On the STllUCTUEE AND CLASSIFICATION OF THE TrEMATASPIDAE. 1 3 



accounted for by the openings, eithei- in the middle, or sides, of the dorsal shield. Moreover 

 a renewed examination of Fteraspis has convinced me that th(^ projection at this point, 

 shown so clearly by some specimens in the British Museum, is not due to the squeezing out 

 of the matrix from an orbital opening, as A. S. Woodward maintains. It is due on the con- 

 trary to a pronounced tubular projection of the shell itself, that is now partly worn, or 

 broken off, leaving the matrix core exposed. How far this structure protruded in its unin- 

 jured condition, or whether it was a jointed, movable appendage it is iraposible to say from 

 an examination of the known remains of this genus. However LindstrOm's most important 

 discovery of a shield of OycUhaspis, side by side with part of an oar-like appendage having 

 precisely the same sculpture, removes any doubt that may have existed as to the presence 

 of paired movable appendages in Gyathaspis and related forms. These appendages most prob- 

 ably fitted in the marginal notches mentioned above, that is in such a part of the head 

 that their identity with the similar appendages of Pterichthjs and Bothriolepis cannot be 

 questioned. We may therefore conclude that in Pteraspis, Gyathaspis, Tolypaspis, and Tre- 

 mataspis, a pair of prominent oar-like appendages, similar to those in Pteraspis and Bothri- 

 olepis, was attached to the so called eye notch on the anterior margin of the head. It is 

 obvious, however, that if appendages are attached to the first pair of openings in Tremat- 

 aspis, all the remaining eight pairs of openings must have been provided with organs of a 

 similar nature ! 



This idea, which so completely revolutionizes all previous conceptions of the structure 

 of Tremataspis receives an unexpected confirmation in my recent discoveries in Gephalas- 

 pis. For I have shown that in the head of Gephalaspis there are indications of the presence of 

 a pair of heavy crushing mandibles, so situated that they must have acted against each 

 other, at right angles to the sagittal plane, instead of parallel with it, as in true Vertebrates. 

 Moreover, there are clear indications of the presence of from twenty-five to thirty pairs of 

 small jointed and movable appendages extending along the ventral margins of the trunk 

 from the head to about the level of the cloaca. That these appendages look like segment- 

 ally arranged Arthropod appendages cannot be denied. It seems to me the least that 

 can be said is that the burden of proof lies with those who deny they are of that character. 

 The argument that it is impossible that an animal so fish-like in appearance should at the 

 same time possess many pairs of segmented appendages is not admissible until it is demon- 

 strated beyond question that we are dealing with true Vertebrates. No one has as yet fur- 

 nished any such demonstration. Aside from the crushing mandibles I have described in the 

 Edinburg specimens, nothing is known about the mouth region of Gephalaspis, and it can 

 hardly be said that reasonably adequate pains have been taken to learn more about this sig- 

 nificant region of an important animal. We are somewhat better off", as far as Pterichthys 

 and Bothriolepis are concerned, although the exact location of the mouth is still in doubt, 

 and it is not possible to harmonize Whiteaves description of the oral plates oi Bothriolepis 

 with that of Traquair. We would call attention however to the resemblances between the 



