232 HENRY C. TRACY 



ration, and hence is always seen pulled off from one or both of 

 its surfaces of attachment. There are often, however, attach- 

 ing strands which extend down to the surface of the macular 

 cells (particularly in the case of the anterior division). 



The relation of the otolithic membrane to the labyrinth epi- 

 thelium and to the endolymph at once suggests a comparison 

 with the tectorial membrane of the mammalian ear. According 

 to the conclusions of Prentiss ('13), the latter is primitively a 

 cuticular organ of a chambered structure which is secreted 

 between and at the ends of the cells composing the basal epi- 

 thelium of the cochlea. The otolithic membrane in Clupeoids 

 is probably formed by an analogous process. 



This chambered structure appears in sections of the organ in 

 all specimens from about 6 mm. up. In figure 20 is shown a 

 section through that part of the otolithic membrane which 

 covers the middle division of the macula; the section passes 

 nearly parallel to the surface of the latter, and hence at right 

 angles to the chambers. The section appears as a plate perfor- 

 ated with round holes. Near its margins the holes become oval 

 and then elongated and less distinctly visible, due doubtless to 

 the more oblique angle at which the chambers are sectioned. 

 When studied by means of the high-power binocular microscope 

 with oblique substage illumination, a dark rim always appears on 

 one side of each hole. By rotating the decentered substage 

 diaphragm the dark rim is seen to move evenly around the mar- 

 gin of each hole. The only probable interpretation seems to be 

 that the dark rim is a refractive phenomenon and the holes are 

 circular, each with a smooth even circumference. I was unable 

 to demonstrate sections of cilia in these holes. If account is 

 taken of the amount of shrinkage which the membrane under- 

 goes in preparation, it is fair to assume that these chambers 

 have been considerably enlarged by the shrinkage process. This 

 consideration and the circular shape of the sections of the cham- 

 bers suggest that they may in life be just large enough to accom- 

 modate the cilia. Probably, then, the substance of the mem- 

 brane begins as a cuticulum, and is further developed by secre- 

 tion from the tops of the macular cells, and in this secretion the 



