10 THE TERIODIC GltOWTH OF SCALES IN GADID.E 



grooves ami longitudinal ridges corresponding to tlie concentric ridges and 

 longitudinal grooves on the scale. Tliis membrane consists of thin crossed 

 fibres, the intercrossing of which results in the indentations of the concentric 

 ridges. The constituent fibres swell strongly under the action of acetic 

 acid, a character which evidently belongs to fibres of connective tissue. The 

 superior or external portion of the scale shows, moreover, a very fine in- 

 separable layer, which shows the presence of fibres under the action of acetic 

 acid, and which it is difficult to destroy by combustion. Tlicse intimate con- 

 nections between skin and scale enable one to see how, during growth, the 

 appearance of the scale surface may be modified without exfoliation taking 

 place. 



According to Peters, the scale is not formed in the epidermis, but in the 

 skin itself ; in that case the scale cannot be simply a horny secretion of the 

 epidermis. 



Peters agrees with Mandl in admitting the existence in all scales of a very 

 soft lamellated inferior layer consisting of fibrous cartilage ; he disagrees with 

 Agassiz as to the number of lamella? corresponding to the number of concentric 

 striae on the upper surface. He held that Agassiz had not sufiicient proof of 

 the non-existence of cartilage in scales, and did not believe in his statement 

 as to the horny nature of scales. 



As to the corpuscles, he maintained that these were found, not on both 

 surfaces of the scale, as stated by Agassiz, but only on the inferior surface. 

 He regards the corpuscles as special elements, and not as being due to incom- 

 pletely formed lamellte or to the wearing down of these thin layers. Corpuscles 

 of some solidity show a granular appearance towards the border of the scale, 

 and give rise to the asperities existing on the posterior border of many scales 

 (Perch). Towards the centre of the scale one finds beneath the elliptical 

 forms quadrangular corpuscles which are arranged in regular series, and give 

 rise to spines. These spines are not, as Mandl supposed, comparable to true 

 teeth. Peters believed that scales could not afford a proper basis for a rational 

 classification, showing that two kinds of scales, cycloid and ctenoid, occur in 

 the same fish {Pelamys sarda). 



As to the superior or external layer of the scale, Peters realised much 

 difficulty, especially in attempting to explain the origin and meaning of the 

 concentric lines and radiating canals. 



The superior layer, he said, does not usually show any distinct elements, 

 though sometimes one can recognise in it the same fibres and the same corpuscles 

 as in the inferior layer ; but never with the same degree of distinctness and 

 clearness. He did not believe that the concentric stri.T? represented the borders 

 of superimposed lamelkB or plates of the scale, in as much as the striic are not 

 always i)arallel to the free border, but are sometimes perpendicular to it. 

 For example, in Alepocephalus rosfratus the stria? are only disposed con- 

 centrically in the posterior third of the scale, while in the remainder of their 

 extent they run straight forward, parallel the one to the other. He agrees 

 with Agassiz in refuting the statement of Mandl in regarding the radiating 

 canals as serving for the nutrition of scales, and regards them rather as 

 sutures rendering growth possible in all directions. He adds that these 

 sutures are not only found running out from the centre in the direction of the 



