CANADIAN FISUERIES EXPEDITION, lOl'rlo 



87 



along the greater part of the margin. In the case of hirge scales, the fibrils form a 

 highly complicated pattern farther in towards the centre. With small scales thus pre- 

 pared the patterns formed by the fibrils may be divided into two types, schematically 

 shown in fig. 5. a and h. The one of these types is characterized by a more or less 

 elliptical figure, and by the fact that the marginal fibrils run tangentially, the other by 

 a hyperbolar figure and radial marginal fibrils. There is thus no doubt that in addition 

 to elementary plates with tangential marginal fibrils, there must also be others with 

 radial fibrils, and the fact that these latter are rarely if ever found in the exposed sur- 

 face of a split scale must be due to the method employed. It is also obvious, that a 

 needle, thrust in with the point directed towards the centre, will find no hold until in- 

 comes 'into contact with transverse fibrils, i. e. until it reaches a plate with tangential 



fibrils. ^ £ iix n 



The scale may thus be considered as a greatly flattened cone composed ol tihnll- 

 ary plates, of which some have tangential fibrils, others radial. This cone is evidently 

 covered entirely by a non-fibrillary layer, on the upper side of which, however, is 

 found finely marked relief which gives the scales its characteristic appearance. If 

 this covering layer could be removed, the margin of each fibrillary plate would then 

 be visible, since'^each plate is larger than the next. A method by which large portions 

 at least of the covering layer can be removed is as follows : A drop of fish glue is 

 placed upon an object glass, and a scale set thereon, with the outer side downwards 

 in the glue, leaving, however, a small corner of the scale free, just large enough to 

 afl'ord a hold for the tweezers. As soon as the glue has dried the scale is damped very 

 slightly with an almost dry brush, the free corner is then grasped with the tweezers, 

 and the scale torn away. If the operator is fortunate, the covering layer of the scale 

 will then be found adhering to the glue on the glass, and the part detached by the 

 tweezers may be mounted for observation. Plate IV, fig. 6 gives a photographic repro- 

 duction of part of such a preparation. The curved boundary lines of the elementary 

 plates are distinctly visible, and it will also be seen that plates with tangential fibrils 

 alternate with those having the fibrils radially arranged. 



It will be easily understood that the arrangement of the fibrils in the elementary 

 plates as here described imparts to the scale a high degree of firmness in all direc- 

 tions, the fibrils in one lamella will, roughly speaking, form a considerable angle with 

 those of the two adjacent, so that the scale may, in a way, be compared with the com- 

 posite wooden plates which are made by gluing several thin sheets together, with the 

 grain of each perpendicular to that of the two adjacent. Fig. 5 c gives, purely schem- 

 atically, the arrangement of the fibrils in two plates, one with radial, the other with 

 tangential fibrils, showing the manner in which the direction of the fibrils in the one 

 lies transversely to that of those in the other, when two such plates are placed in 

 juxtaposition. 



Fig. 



Judging from what we have hitherto learned, the transverse section of a scale 

 should pl-eseut more or less the appearance shown in fig. 7. This is also, roughly 

 speaking, found to be case. The upper covering layer can be distinctly seen when the 

 section is stained with thionine, as shown in fig. 8, the layer in question then assuming 



