10 William Patten. 



The second row of plates is much more clearly shown. It consists of at least three 

 distinct plates; a small triangular inner one, a larger median pentagonal one, with its long 

 axis in a transverse direction, and an irregular marginal one with its long axis nearly parallel 

 with the margin of the dorsal shield. The last two plates and the marginal one of the 

 first row have been forced outward so that their adjacent ends form a pyramidal 

 elevation. 



The two plates in the third and in the fourth rows are very clearly shown. The large 

 size of the posterior marginal one and the presence of a well marked suture on its anterior 

 border indicate that it is a compound plate. The suture extends nearly half way across and 

 then bifurcates, as though three plates were nearly fused in one. If this is really the case, 

 then we must assume that the third and fourth rows consisted originally of three plates each, 

 arranged as indicated in the figure. 



All the oral plates were apparently united by a tough but flexible membrane that 

 would allow some freedom of movement. This is indicated by the polished and rounded 

 edges of the plates and by the rounded contours of the matrix in the intervals between them. 



On the right side of the fossil, the median plates of the third and fourth rows are in a 

 nearly normal position, and they are so well preserved that they fix beyond any doubt the 

 exact location of the median plane. The first and second rows on the right side are so badly 

 crushed and depressed that their arrangement cannot be made out. One can distinguish, 

 however, a longitudinal row of four or five overlapping plates, or fragments, that perhaps 

 correspond to the inner plates of the second and third rows on the opposite side of the 

 median line. 



The Median Plate is such an extraordinary structure that it deserves careful consider- 

 ation. It occupies an elongated triangular area between the first three rows of plates. It 

 lies considerably below the level of the left side, which has evidently been thrust ventrally 

 and toward the median line, thus making a conspicuous longitudinal fold which partly 

 overlaps the right side of the plate. The latter is triangular in outline with a smooth outer 

 surface like that of the other plates. Its posterior pointed end is slightly rounded; the 

 bounderies of its broad anterior end cannot be distinguished. A shallow groove runs along 

 the right margin while a rather prominent ridge or keel runs along its m.edian line. The 

 ridge is highest at about the middle point of the plate and on either side of this point is a 

 slight bulge in its lateral walls. A faint line, or groove, appears to cut off the right anterior 

 corner. Although these details are minute, I am confident they constitute normal structural 

 features of this plate. 



Unfortunately the relations of the median plate to the anterior margin of the dorsal 

 shield could not be determined with certainty. There is, however, little doubt that 

 it is attached to the median keel-like process on the under side of the dorsal shield. 



Rohon has figured and described a triangular plate found separately, which he sup- 

 posed might fit into this region (PI. I., figs. 12 & 13). I have in my own collection a plate 



