CAyADT.iX FIS^HERIES EXPEDiriOX, lOl'rlo 93 



able, that in one and the same specimen may be found scales where the abnormal cen- 

 tral portion is (luite small, and others in which a larg-e portion, or the entire surface, 

 is of abnormal appearance. 



The same confusion noticed iii the relief of the outer surface is likewise encoun- 

 tered in the inner structure of such abnormal scales. It will be found impossible to 

 split them, for instance, as long' as the needle jxtint is introduced within the abnormal 

 central part, and if the covering' layer be removed, we do not find the regular alterna- 

 tion of radial fibrillar zones with those having- tangential fibrils; on the contrary, 

 the fibrils will be seen to intertwine, forming patterns similar to that shown in plate 

 VI, fig-. 17. If such abnormal scales be treated with silver solution as described be- 

 fore, it will be noticed that the division into zones, which otherwise makes itself so 

 distinctly apparent during- this process, is not discernible within a central portion 

 corresponding to the visibly abnormal portion of the surface layer. 



The most probable explanation as to the origin of these abnormal scales would 

 seem to be as follows. The scale originally set in the scale pocket where now the 

 abnormal one is found, must on some occasion have fallen out, this' taking place at a 

 time when the original scale was of a size corresponding to the abnormal portion of its 

 successor. Within a short time after the loss of the original scale, a new scale (the 

 abnormal portion) grows out, and then, having filled the vacant place, the scale pro- 

 ceeds to grow normal wise. 



Closer investigations of the scales of a herring, with this question in view, will 

 soon reveal the fact that the scales innnediately adjacent to an abnormal scale are, as 

 a rule, themselves abnormal, forming together a group. And outside these 

 again, or rather round about them, may frequently be found scales presenting another 

 form of abnormal structure than the first mentioned, the central portion correspond- 

 ing to the abnormal centre in the first being here either dislocated in relation to the 

 peripheral part, or at least separated from this by a line concentric with the peri- 

 phery, and more or less distinctly marked. Outside these scales lie the wholly nor- 

 mal scales, so that transition stages are seen to lie between the two extremes. In 

 order to make this clear, a group of entirely abnormal scales was sought for on the 

 body of a herring, and all the scales surrounding this group w^ere then picked off and 

 prepared, careful note being kept as to their respective positions. Plate VI, figs. 18a 

 to d are from photographs showing the central portions in a horizontal series of these 

 scales, the first giving a view of the last normal scale before the abnormal formation 

 begins, and the last illustration being that of the first abnormal scale. This series of 

 views, and similar series which I have had no difficulty in finding, distinctly suggest 

 that the abnormal scales are nothing but supplementary growths intended to replace 

 scales previously lost by accident. The scales immediately adjacent have become 

 slightly displaced from their natural position in the scale pockets, as shown by the 

 fact that the central portion forms, as it were, a scale upon the scale. It will also be 

 seen that the transition forms may well give rise to erroneous determinations of age, 

 the boundary line between the central portion and the periphery frequently bearing 

 more or less resemblance to a winter ring. This will be further referred to later on. 



IV.— THE SCALES AS Al^ INDICATIOX OF ACE. SOURCES OF ERROR 



IN ACE-DETERMI X.\ TIOX. 



We have no experimental proof of the fact that the so-called winter rings on the 

 scales of the herring actually are annual rings, i. e., that one such winter ring is 

 formed each year. On the other hand, statistical observations point so emphatically 

 to the correctness of this supposition, that it will hardly seem possible to otherwise 

 explain the regularity revealed by the observations. The observations in question 

 deal primarily with the Norwegian stock of herring, which have been under considera- 

 tion in this respect for several years. The proofs thus furnished can, strictly speak- 



