8 IRVING HARDESTY 



easily far better than mine, one is disappointed in that he gives no 

 description whatever of his more favorable methods and none of the 

 results obtained with them. Through the pages of his paper, he pro- 

 gressively describes the structure of the tectorial membrane as at- 

 tached by 'dehcate threads,' as composed of 'lamellae' (supporting 

 Shambaugh), "parallel fibers or lamellae," "delicate parallel plates," 

 'a reticulum,' and finally advances the one new claim of his paper, 

 namely, that the membrane is a honey-comb or chambered structure, 

 each chamber corresponding to and produced by a separate cell and 

 that thus the transverse dimensions of the chambers are the same as 

 those of the cells of the greater and lesser epithelial thickenings or 

 ridges of the earlier fetal stages, both of which ridges he claims pro- 

 duce the membrane. The producing cells being situated on the basal 

 side of the membrane, the lengths of his chambers, curving axis ward, 

 must extend through the thickness of the membrane. He concluded 

 that the content of his chambers (matrix, Zwischensubstanz, etc., of 

 others) is a fluid "resembling the endolymph." 



A claim suggesting Prentiss' contention as to the origin and struc- 

 ture of the membrane was advanced by Coyne and Cannieu ('95), 

 and Prentiss states that I misquoted these authors when I cited their 

 paper as one of those in which a fibrous character of the membrane 

 was suggested. In citing Coyne and Cannieu at the place in question, 

 I had no reference whatever to their idea, which I thought quite er- 

 roneous, as to the process by which the membrane is produced by the 

 (;ells basal to it, but merety to the fact that they considered the struc- 

 ture of the membrane as "im reseau." Vasticar ('11) quotes directly 

 a later statement by Coyne in which "'d' aspect areolaire^' and "la forme 

 d^un reseau'" are again applied to it. I translated reseau as meaning 

 reticulum. A reticulum is a net and, in anatomy, is a network whose 

 meshes extend in all planes. A net is composed of threads, not plates 

 nor lamellae, and the word does not convey the idea of walled pockets 

 or alveoli, nor that of a honey-comb structure. Whether, as Coyne 

 states, the meshes (mailles) of the reticulum conform to the sizes of 

 the cells of the organ of Corti (cellules sensorielles ciliees), which he 

 assumed secrete the membrane, was a question, however doubtful, 

 not in mind at the time. Prentiss uses the paper of Coyne and Can- 

 nieu in support of his claim of the chambered structure. He is rigltt 

 in showing that my citing the paper at all in the relation I did was un- 

 fortunate. Support of the fibrous character of the membrane alone 

 was sought, and while I find it is fibrous, the normal arrangement of 

 the fibers is not in the form described in the paper. 



Prentiss describes the walls of his chambers as coinciding with and 

 continuous with the sides of the producing cells, the cavities of the 

 chambers with the bodies of the cells. His figure 5, drawn from a 

 section of the actual specimen in which the producing cells are cut 

 longitudinal^, is given to prove his chambered structure, but in this 

 figure nearly half of the filaments of the membrane are given off from 

 the ends of the cells instead of from their sides. Figure 9 of my paper 



