19] THE SKULL OF AMIURUS— KINDRED 19 



of the lateral septum semicircularis to the posterior end of the cranium, lateral 

 to the notochord. The sacculi communicate with each other across the 

 anterior ends of these grooves, above the tip of the notochord, by a transverse 

 canal sinus impar (ductus endolymphaticus, Wright, 1884), from the posterior 

 wall of which the sinus impar of the Weberian apparatus projects. This sinus 

 impar lies along the mid-dorsal surface of the cavum floor and is separated 

 from the laterally situated sacculi by a membranous V-shaped wall, the apex 

 of which is attached to the dorsal surface of the notochord (Figs. 17, 21). It 

 is separated from the cavum cranii by a membranous roof which continues 

 laterally as the roof of the saccular recesses and is attached to the cranial wall 

 at the jimction of otic capsule and parachordal cartilages. Posteriorly, the 

 membranous roof over the saccular cavities is replaced by cartilage which is 

 continuous medially with the ventral walls of the cavum sinus imparls and 

 laterally adjoins the otic capsule. The ventral floor of each recessus sacculi 

 is ver}^ thin (Fig. 8), but the posterior wall which marks the posterior extent 

 of the parachordal plate is thick dorso-ventrally (Fig. 21). The dorso-lateral 

 surface of this posterior part of the parachordal plate is separated from the 

 ventral end of the occipital arch posterior to the otic capsule (Fig. 17), and 

 through this space the sinus impar communicates with the Weberian ossicles 

 contained in the saccus paravertebralis. The first post-vagal nerve or hypo- 

 glossus passes out through this space, but does not touch the cartilage of the 

 posterior end of the parachordal plate which has narrowed considerably in this 

 region. 



The relations of the inner ear to the cranial floor have been described by 

 several investigators in those forms having a Weberian apparatus, but the 

 descriptions have been confined to adult conditions. In a later paper, I hope 

 to follow the developmental relations of the inner ear to the parachordals, if 

 I am fortunate enough to obtain the proper stages. 



There is no evidence of segmentation of the basal plate at this stage, such 

 as is found at the posterior end of the parachordal plate of Acanthias (Sew- 

 ertzoff, 1897). The distinct ridge of cartilage, called the ' Sattellehne, ' is 

 lacking in Amiurus because the trabeculae do not become attached to the 

 ventral surface of the parachordalia, but lie in the same plane with them. 

 Like the intercapsular floor of Acanthias, this region of Amiurus is solid, and, 

 although in the early stages of Acanthias the notochord projects into the basi- 

 cranial fenestra, it is later enclosed by cartilage as in Amiurus. The inner ear 

 relations have nothing in common as regards the parachordal plate, because 

 in the older Acanthias the cawmi of the ear is shut off from the cavimi cranii 

 by a wall of cartilage. 



As I was unable to find any statement concerning the later larval history 

 of the parachordalia of Amia, I found it necessary to study a series of trans- 

 verse sections through the head of a specimen 19 mm. long. I have referred 



