86 



DEPARTMENT OF THE NATAL SERVICE 



simple methods, the results of which will be found to agree with those obtained by 

 more complicated experimental processes. 



With the aid of a needle and tweezers, a herring scale may be split up into 

 several thin plates, all exhibiting a contour identical in form with that of the scale 

 itself. The scale can only be thus split by inserting the point of the needle in the 

 outer surface, and the size of the flakes thrown off will depend upon whether the cleav- 

 age is commenced near the edge of the scale or farther in towards the centre. By 

 commencing near the margin, two large plates will be obtained, starting from a point 

 farther it will give one large and one small. Plate III, fig. 4 shows the two portions 

 of a scale thus split. The central portion here thrown otf presents the appearance of 

 a small scale, with the margin somewhat torn, the remainder, constituting the larger 

 part, exhibits a transparent " sore " in its central portion corresponding to the smaller 

 flake detached. From such experiments it may be seen that the scale is built up of 

 thin plates, similarly formed, each somewhat smaller than the next. The layers must 

 have been deposited in order of size, forming something approaching a very low cone, 

 the base of which is represented by the inner surface of the scale, and the apex by the 

 centre. The exact number of layers of which a sca^e is composed cannot be deter- 

 mined by mechanical means, as it is impossible to say whether each flake detached 

 comprises but a single layer, or is itself composed of several adhering together. 



In a scale thus split, the surface exposed reveals a kind of pattern formed by fine 

 fibrils curving in prcs and whorls (vide plate III, fig. 4.) In most cases it will be 

 found that the fibrils nearest the edge of the exposed portion run parallel 

 with the same, while those farther in, on the anterior part, form a kind of elUptical 

 figure, and those in rear, on the posterior part (viewing the scale as in its normal 

 position on the body) form irregular whorls. From the pictures thus presented, 

 which are almost always to be found in such exposed portions where flakes have been 

 detached, it might well seem that the fibrils in each elementary plate were arranged 

 in this manner, with tangential marginal fibrils. This is, however, not the case, for 

 on studying the arrangement of the fibrils in scale preparations obtained by other 

 methods one may sometimes find the pattern above described, and in other cases pat- 

 terns of altogether different character. 



The inner side of the scale is formed by a fibrillar plate, and, with a fairly 

 strong lens, the fibrils in this may easily be seen, especially if the scales are first 



a 



Fig. 5. 



treated with nitric acid, and, in mounting, placed with the outer side down in a drop 

 of glycerine, leaving the inner side upwards, exposed to the air. On examining a 

 number of such preparations, not a few will be found in which the fibrils, instead of 

 running tangentially to the edge of the scale, fall perpendicularly to this, at any rate 



