212 Cocker ell — Some Notes on Fish Scales. 



B. Scales cycloid, wholly without spines (Gadidse and 



Merlucciidse). 

 a. Scales without radii (Merlucciidse, Phycinse, Gaidrop- 

 sarinse). 

 Scales elongate, with a strong median transverse furrow, variously 

 developed, sometimes only near the margins, never crossing 

 the nucleus; circuli not angled . . . . Enchelyopus cimbrius (L.). 



Eel Pond, Woods Hole. 

 Scales larger, rather broad, without such a distinct median fur- 

 row, and with the basal circuli variously angled 1. 



1. Region above the nucleus (except in latinucleate* scales) con- 

 spicuously finely pitted Merluccius bilinearis (Mitchell). 



Woods Hole. Mass. 



Region above the nucleus not or hardly pitted 



Urophycis regius ( Walb. ). 

 Woods Hole. .Mass. 

 6. Scales with numerous radii (Gadinse, Brosminse). 

 Scales very long and narrow (about 4 nun. and 1% broad), the 

 central area usually free from sculpture and the tine radii 



broken Brosme brosme (Mull.) 



Scales not so long and narrow, usually with a fine reticulated 

 sculpture all over, though latinucleate scales of Microgndus re- 

 semble those of Brosme in structure though not in shape . . . 1. 



1. Scales comparatively large, about 7 to 8 nun. long and -I to 5 



broad Mel anogr animus seglefinus (L.). 



Scales smaller, less than 5 mm. long 2. 



2. Scales smallest (about 2 mm. long), narrow . Pollachius virens (L. ). 



Woods Hole. 

 Scales medium (between 3 and 5 mm.) .... Gadus callarias L. and 



Microgadus tomcod ( Walb. ). 



The scale- of the Macrouridse, except for the shape (round or trans- 

 versely elongate) and the spines, agree with those of the Merlucciidse and 

 Phycinse, having more or less angled circuli and no radii. Those of the 

 Gadinse and Brosminse are entirely different, and could not have been 

 derived from those of the others. According to the scales, some type 

 allied to the Macrouridse may have given rise to the different lines repre- 

 sented by Merluccius, Urophycis and Enchelyopus; but the Gadinse with 

 the Brosminse must have another origin, although no doubt all can be 

 referred to a common ancestor by going back far enough. There is no 

 essential difference between the Gadine and Brosmine scales, but in my 

 preparation of Brosme all the scales are latinucleate except one. 



* Latinucleate scales are those in which the nuclear region is broadly modified, the 

 normal sculpture largely suppressed or altered, and the radii more parallel. This 

 dimorphism is of general occurrence among fish scales, and appears to be due to regen- 

 eration after loss of the original scales. 



