142 



the same region. For example, this occurs at the posterior edge of 

 the operculum. Another striking fact, is that where continuity of the 

 two layers exists the large mucous cells of the ectoderm are cut some- 

 what obliquely and not meridionally. This is apparent in some of 

 Dr. Klaatsch's drawings (Figs. 5 and 6, PI. VIII). 



ng. a. Section through body wall of S. salar, 1,2 cm long. ect. ectoderm; m. m, 

 coraco-hyoideus ; br. apparent break in continuity of the basement membrane ; som. 



Fig. 3 

 co-h 

 somatopleure 



Moreover, these points at which ectoderm cells apparently crowd 

 into the lower layer are not confined to the vicinity of bone, as they 

 should be, were Dr. Klaatsch's view correct, but they may be found 

 in many other places, especially in the head where the complexity of 

 the folds of the skin gives ample opportunity for it to be cut obli- 

 quely. Figure 3 is a section through the anterior portion of the body 

 wall of the gill region, which cuts also the rudiment of the m. coraco- 

 hyoideus. The continuity of the muscle with the ectoderm is quite 

 as well marked here as in the case of any bone I have observed. 



The connection of the tooth-bearing bones with the ectoderm, 

 I believe to be of the same nature. At any rate it is quite impossible 

 that the dentale and max ilia re be formed through the coales- 

 cence of tooth-rudiments, for both of these bones arise deeply im- 

 bedded in the mesenchyme at a time when the modification of the 

 ectoderm covering them is but barely perceptible as a slight thicken- 

 ing of its deeper layer. The stages which Dr. Klaatsch has studied 

 and figured are much too old to have weight in support of his view. 



To recapitulate then, I believe to have shown, 1) that the apparent 

 disappearance of the basement membrane at certain points, and in- 

 crowding of ectodermal cells into the underlying tissue, is due merely 

 to the obliquity of the plane of section and the consequent overlapping 



