8 William Patten. 



The first specimen carefully studied, PI. I fig. 1 shows three somewhat broken aud distorted open- 

 ings g, k, i,j, on the left side and in front of the series described by Rohon. The figure, although 

 carefully drawn, hardly does justice to a preparation so difficult to reproduce faithfully and 

 accurately. Two other preparations at Dartmouth show substantially the same condition. 

 We must therefore now recognize, certainly eight, probably nine pairs of openings, arranged 

 in nearly a semi-circle around the oral region. The great difficulty in determining tlie exact 

 number is in obtaining a preparation showing the whole series in place. It is possible, but 

 I do not regard it as probable, that the most posterior of the three new incisions coincides 

 with the most anterior one described by Rohon. The incisions gradually increase in size 

 from behind forwards, the first five being on the anterior margin of the ventral shield, the 

 next three on the median side of a narrow strip that may be regarded as a continuation for- 

 wards of the ventral shield ; while the most anterior and largest opening is cut out of the 

 lateral margin of the dorsal shield. The position of this last opening is clearly shown in the 

 side view of. the model. 



The Oral Plates. That the large ventral opening in the head of Tremataspis is in part 

 tilled by a number of remarkable plates was first shown by Rohon in 1892, in his description 

 of a remarkable specimen preserved in the Petersburg Academy, the only specimen in which 

 such plates have been found in place. 



When it was my privilege to examine this specimen for the first time, I had some 

 doubts whether it had the value and importance attributed to it. Further study, however, 

 convinced me that its value had been underestimated, rather than otherwise, and I began to 

 make as careful and painstaking a study of all its details as lay in my power. I am conse- 

 quently now able to add new and important data to our knowledge concerning it and to put 

 upon it an interpretation that diflfers in some essential respects from that of Rohon. 



Unfortunately, in this specimen, the anterior margins of the dorsal sliield were either 

 destroyed or absent, so that the relation uf the plates to them and to the three anterior pairs 

 of ventral openings could not be determined. 



In constructing the model from which the figures were made the boundaries of the 

 oral region were first reproduced from several different specimens showing these parts. Then 

 from the original cast of the fossil, several impressions were taken iu dentist's wax, which 

 gave a faithful reproduction of the original fossil. The impressions were then modelled in 

 bee's wax on an enlarged scale and finally copied on the open area of the model of the 

 whole head. It then required only trifling alterations of some lines to make the whole series 

 of plates fit into the area that had been left open for them. This was true, however, of the 

 left side only. The plates on the right side were considerably crushed aud iu part absent, 

 so they were simply made to duplicate those of the opposite side, but without the details due 

 to distortion or the fracture of the plates. In other words, the oral plates of the left side are 

 copies of those in the original fossil, while the right side is largely a duplication of the left. 



