124 BULLETIN OF THE 



tubules (p. 138). I have found by diligent searching a few cases in 

 which, as stated by the latter author, cells seemed to be present in the 

 lumen of the tubes. I have, however, observed other cases in which a 

 flattened cell lay just at the side of the tube, but separated from its 

 lumeu by a thin layer of secreted material. T am therefore led by this 

 and by the infrequency of the cases to believe that the cells apparently 

 occupying the tube have a similar position, but that since they lie in 

 radii of the tube perpendicular to the plaue of the section they have the 

 appearance of being in the lumen of the tube. Figure 8 will make this 

 clear. At the opening of the tube is seen a cell which is apparently 

 about to be enclosed between the secretion of the large odontoblast cell 

 and the present wall of the canal. At a little later stage this, if seen in 

 a plane at right angles to the present, would appear to lie in the lumen. 



Each of these tubes is lined by a very thin secretion from the cell 

 which extends into it, — a secretion of the same character as the material 

 of the scale. In later stages, after the ganoin layer has begun to be 

 formed, this sheath persists even after decalcification, when the ganoin is 

 entirely destroyed (Plate IV. Fig. 23). 



The scale increases both in thickness and in lateral extent by the 

 deposition of new layers of material secreted by the surrounding sclero- 

 blasts. Thus the edges of adjacent scales come together, and finally 

 overlap one another. Owing to the slight elevation of the posterior 

 margin of the scale, previously mentioned, this edge of each scale over- 

 laps the anterior edges of the adjacent posterior scales ; for the same 

 reason, each scale is in turn overlapped on its anterior edge by the poste- 

 rior edges of the scales in front of it (Plate IV. Fig. 31). 



Increase in thickness takes place upon both the outer and the inner 

 surfaces, though much more rapidly upon the inner one. As the scale 

 thickens inwardly, fibres which lay irregularly disposed beneath it in 

 the spongy dermis are enclosed by the secretion. As the calcification 

 extends deeper, the felted regularly arranged bundles of connective-tissue 

 fibres of the deeper dermis are also enclosed, and help to make up the 

 deeper part of the scale. These fibres extend across from one scale to 

 another and so form a strong but flexible connection between adjacent 

 scales. They are the " Schuppenligamcnte " of Hertwig. The ratio of 

 outward to inward growth is shown by the arrangement of the layei'S of 

 the scale (Plate II. Figs. 15, 16, and Plate III. Fig. 21). The oldest 

 part of the scale is that in which the dentinal tubules from the opposite 

 surfaces meet and break up into minute branches (Plate III. Fig. 21). 



In their immature state, and before the outer layer has begun to be 



