62 Cocker ell — The Scales of the Clupeid Fishes. 



upwards to reach the margin at an acute angle. At first sight 

 the nature of these structures seems wholly ohscure, but in a 

 longitudinally oval scale, probably from the caudal peduncle, 

 it is seen that they pass gradually into apical radii. The first 

 stage of modification is that in which the apical radii on each 

 side of the middle become elbowed or curved at the base, form- 

 ing a sort of J. In Catostomus the basal radii show this condi- 

 tion, and the tendency is for the curved part of tlie J to 

 disappear, leaving radii which no longer point to the nuclear 

 area. In Alosa this curving of the apical radii continues until 

 a U rather than a J is formed, one arm of the U now pointing 

 apicad to the nuclear area. Then the inner arm loses all con- 

 nection with the nucleus, and those of opposite sides uieet at 

 an acute angle, forming a sort of reversed V. From this it is 

 a comparatively short step to a single line running transversely 

 across the scale. All this is complicated, especially in some 

 scales, by various degrees of anastomosis, and the frequent 

 disappearance of the inner part of tlic line. Nevertheless, by 

 taking different scales from a single example of the Alosa, it is 

 possible to demonstrate every transition from apical I'adii to 

 transverse lines below the nuclear area; the conclusion being 

 that these lines, which I have found only in Clupeidte, are 

 really greatly modified apical radii. Dr. Evermann kindly 

 sent me some scales of very young Alosa sapldlssima from the 

 fish ponds at Washington, D. C. These scales, only about 2 

 mm. diameter, do not look like those of the adult, but they 

 correspond exactly with tlu- nuclear region of the latter. They 

 show a strongly differentiated sculptureless apical li<ld ; the rest 

 of the scale is covered l)y very fine whoUy transverse circuli, and 

 exhibits two or three of the lines representing modified radii, 

 which are transverse, more or less bent apicad and obtusely 

 angled in the middle. Thus the young scale does not throw 

 any light on its evolution, and is, taken by itself, much more 

 difficult to interpret than that of the adult. I am indebted to 

 Dr. B. W. Evermann and Dr. S. Graenicher for other Clupeid 

 scales, which all possess the same essential features as those of 

 Alosa. They may be described as follows: 



(1.) Clupea hareiKjus L. Sandy Island. Scalus about 8 mm. long and 

 73'3 broad; strncture as in .l/uaa, tht' ai)ie'al radii feeble, the trans- 



