12 William Patten. 



of the oral plates points to the conclusion that it lies in the centre of the oral region, heJiind 

 the anterior median plate. 



This plate clearly occupies a triangular depressed area between the first three or four 

 rows of paired plates, and the depression deepens at the narrow posterior end as though it 

 led into a small circular opening, like the mouth of a Limulus or a Scorpion. 



Indeed there is no trace whatever of any structural feature indicating the presence of 

 upper and lower mandibles or of an oral opening, like those in a true Vertebrate. On the 

 other hand the arrangement of the paired plates, the small central depression and the me- 

 dian triangular plate, so suggestive of the rostral plate of an Arachnid, indicate the pre- 

 sence of oral structures like those in an Arthropod. 



This view is strengthened by a consideration of the series of so called gill openings 

 that surround the oral plates. 



As is well known, the large compound plate of the fourth row is provided with a se- 

 ries of incisions that help to inclose a number of so called gill openings, in the manner de- 

 scribed by Rohon, and which my own observations amply confirm. The lateral corner of this 

 plate is broken off in the only isolated specimen known, and it is much obscured in the cast 

 which shows this plate in position. But enough is left to show that the most lateral incision 

 was much the largest and helped to inclose the third opening of the series. The first two 

 openings appeared to be inclosed respectively by the lateral plates of the first and second 

 rows. We have no certain evidence of this however, as the lateral margins of these plates 

 in the cast are partly concealed, and the margin of the dorsal shield opposite these plates 

 is only shown in specimens where the plates are absent. The plates in question however are 

 so near the lateral margin, that there could not be any other plates beyond them, unless 

 they were extremely small. The only doubtful points therefore are the position, and the out- 

 line of the lateral margin of the marginal plates of the first and second rows. It seems to 

 me very probable that they will be found with scalloped margins bounding the first three 

 openings, as the posterior lateral plate bounds the last six openings but one. 



The Appendages. What is the significance of the nine pairs of ventral openings? If 

 we regard Trcmataspis as a true Vertebrate, there is obviously no other course open than 

 to assume, as Rohon and others have done, that they are gill openings, for we known of no 

 other vertebrate structures with which they could be compared. But we may not complete 

 the vicious circle and claim, as some have done, that because Tremataspis has gill slits, it 

 is a true Vertebrate. The discovery of three new openings in front of the old ones places 

 the whole subject in another light. If we look at Tremataspis from the side, we see that 

 the most anterior opening is visible as a semicircular incision in the margin of the dorsal 

 shield, and dearly comes at the same point as the well known openings on the margin of Pter- 

 aspis, Oijathaspis and Tohjpaspis, which have been looked on as the openings for the lateral 

 eyes. But they cannot be eye openings in Tremataspis because the lateral eyes are already 



