194 MARINE BIOLOGICAL LABORATORY. 



Each of these canals opens into the main sac of the 

 ear by its two ends, though this usual relation is not 

 adhered to by all vertebrates, since some of the cartilag- 

 inous fishes show the posterior canal separated almost 

 completely from the main sac, communicating with it 

 only by a small, much reduced canal, the two ends of 

 the semicircular canal having been brought together 

 and its two terminal openings fused into one. 



In Torpedo the proximal (superior) ends of the canals 

 (Fig. I, near co), are brought together and empty into a 

 common but very short tube, which serves to maintain 

 the communication with the main sac. Before uniting, 

 however, the external and posterior canals unite by 

 communicating with a common ama,^ while the anterior 

 canal swells out into an ama of its own. There is no 

 constancy in the relation of the ends of the canals to 

 each other or to the main sac, and, as we shall see later 

 on, the matter has no great morphological or physiologi- 

 cal importance. The enlargements at the proximal ends 

 of the canals are found among most of the lower forms 

 and occur in the divslopment oi the higher groups, usually 

 disappearing before maturity is reached. The distal 

 (inferior) ends of the canals are swollen into globular 

 bodies, the ampullae, which are usually flattened on one 

 side by the entrance of the nerve branch, which breaks 

 up into a brush of fine fibrils as it enters the ampullar 

 wall. The nerve fibrils are distributed to the sensory 

 cells of the crista acustica or transverse ridge, which 



^ I shall call the superior terminal enlargements of the semicircular 

 canals amae, from the Greek a?na, — ?i water-vessel, — since, unhke the 

 ampullae or the inferior terminal enlargements, they contain no nerve end 

 organs and are mere fluid-holders. Their significance is unknown to us. 



