NEW FACTS CONCERNING BOTHRIOLEPJS. I 23 



ings, evidently comparable with the openings to the so called 

 endolymphatic ducts of Trematdspis. 



In order to understand the morphology of the cephalic region 

 of the Ostracoderms it is essential to know the location of the 

 gills. Certain problematical structures found in the chamber in 

 front of the ridge on the inner surface of the anterior ventro- 

 laterals I formerly regarded as gills. I have not found any ad- 

 ditional evidence in support of this conclusion. The presence 

 of gills at this place would indicate that the extra-lateral was an 

 operculum guarding the exit to the gill chamber, for it is the 

 only cranial plate that could possibly act as such. 1 But its move- 

 ments must have been very restricted and the place left for the 

 passage of water extremely narrow, for both dorsal and ventral 

 margins are articulated to the margins of the laterals and to the 

 anterior ventrals respectively. The movement, such as it is, ap- 

 pears to be merely a tilting or partial rotation of the plate on its 

 long axis in order to accommodate the slight dorso-ventral move- 

 ment of the front part of the head, which would be impossible 

 without some such arrangement. 



On the other hand several facts indicate that the gills may have 

 been situated in the post-cephalic portion of the buckler. These 

 facts are(i) : The presence of the superimposed lamellae covered 

 with what appears to be a continuation of the integument. These 

 folds are more suggestive of gills than of any other known organs, 

 and are almost invariably distinctly preserved, even where there 

 is no trace of a notochord, showing that they must have had 

 considerable power to resist decay, more, for example, than the 

 folds in the mucous membranes of the viscera, the only other 

 structures with which they might be compared. Moreover, it 

 the walls of the viscera are as well preserved as this, then other 

 internal organs should be visible in sections also, but as they are 

 not, we are obliged to conclude that the folds in question are not 

 produced by intestinal folds, or by any other structures of that 



1 The rather large opening in the sides of the anterior ventrals beneath the base 

 of the pectoral spines, Fig. I, could have hardly served any other purpose than for 

 the passage of blood vessels and nerves to the pectoral spines. 



The structure of the proximal end of the pectoral appendage indicates that some of 

 the muscles moving the appendage were attached to the outer surface of the base of 

 the appendage and to the outer surface of the anterior ventro-laterals. 



