124 ANNALS NEW YORK ACADEMY OF SCIENCES 



The mandibles in the arthrodires were not stationary, but they have no 

 sign of the ordinary piscine articulation even in the best preserved fossils 

 (Fig. 4). They appear to have been attached to the skull plates some- 

 what as the scapula in mammals is attached to the body, namely, by liga- 

 ments, muscles and connective tissue. From the mechanical point of 

 view, there seems to be three possibilities for movement in these arthro- 

 diran jaws : first, they might have had the muscles so arranged that there 

 would have, been a synchronous movement of the head and jaws, and from 

 a study of Dinichthys this seems to be the best arrangement, as it is the 

 one that lends itself best to the arrangement of the muscular system; 

 secondly, the jaws might have been arranged so as to work against the 

 skull, but the movable joint at the back of the head seems to make this 

 rather disadvantageous, as the head is not a stationary structure, and 

 part of the force of the movement would be lost, as the joint in the neck 

 would give and the head would be pushed back until it touched the dorsal 

 shield. The usual joint or condyle in the neck region is not prominent 

 and must have been an uncalcified cartilaginous articulation, if any- 

 thing, for the dorso-ventral movement of the head would tend to dislocate 

 any ordinary articulation between the head and the neck. 



In brief, these agnathous forms developed their dermal plates just as 

 the gnathostomes did, but while the gnathostomes put the emphasis on 

 the cartilaginous substratum and developed from it the principal struc- 

 tures of the head and jaws, thus making the membrane bones subservient 

 to the cartilage and finally drawing them in as a covering for the car- 

 tilage, the ostracoderms and arthrodires put all the stress on the outer 

 dermal plates and developed the movable parts from these elements, while 

 apparently neglecting the development of the cartilaginous viseral arches. 



The peculiar head structure of the Arthrodira seems to imply an 

 equally peculiar musculature as follows: (1) The joint between the 

 dermal plates of the head and dorsal shield implies the existence of 

 muscles to raise and lower the head. (2) As stated above, there appears 

 to be no surface on the mandible that could articulate with a quadrate in 

 the ordinary piscine fashion, so that from the present knowledge of 

 arthrodiran anatomy the adductor mandibulaa of the Pisces could not be 

 applied to these forms. Thus it seems probable that any system of mus- 

 culature that would be effective in its mechanical action would be entirely 

 unfishlike. Accordingly, in Plate XI the musculature of Dinichthys is 

 figured according to the mechanical requirements and follows no fish 

 type. The movements of the head would require two large muscles in 

 order to move the head up and down on the ginglymoid joint. One pair 

 would be in the posterior region of the skull where the marks are plain 



