280 BULLETIN OF THE BUREAU OF FISHERIES 



If the number of scales in the horizontal rows is less in the adults than in the 

 juveniles whose scalation is complete, the ratio of the body-scale growth would 

 decrease with age, as the scales would grow relatively faster than the body. When 

 an adult fish loses a scale it is replaced by a "regenerated" scale, and the number of 

 scales is not altered; but there appears some evidence that scales may be crowded out 

 and covered by others. Both Brown (1904) and Thomson (1904) discovered minute 

 scales under and between the large ones in species of Gadidee. The former author 

 concluded from this fact that scales are shed after spawning and are replaced by the 

 small underlying scales, while the latter stated that these juvenile scales, which do not 

 possess many lines of growth, are crowded out and covered by the larger ones and 

 finally disappear. Thomson's interpretation is based partly on his belief "that the 

 exact number of scales in a row on the fish has been regarded as sufficiently constant 

 for use in the determination of species [p. 58]." He states further that Klaatsch 

 found the same thing to occur in the trout. Klaatsch found that "between such 

 large scales as already partly cover one another, small scales are very frequently 

 found which are in the earliest stages of development. In older animals such an 

 irregularity does not occur." And "As in Elasmobranchs, new scales originate in 

 the trout between the well-developed scales; thus one finds lying between the older 

 scales of the trout even in later stages quite young scale foundations. This irregular- 

 ity in the early development soon ceases in the trout [p. 58]." 



In the lake herring the presence of minute scales and scales strikingly smaller 

 than the neighbormg or contiguous, well-developed scales is not uncommon. These 

 minute scales usually are covered by the normal ones; but the small scales are not, 

 as Klaatsch and Thomson say of their material, "juvenile." They are abnormal, in 

 that they are stunted in growth; but they possess as many annuli as the normal con- 

 tiguous scales and are as old. 



These facts indicate that all the scales formed during the first year of life do not 

 necessarily grow proportionately with the body of the fish, and that in some instances 

 the number of normal scales in a horizontal row is reduced as the adult stage is at- 

 tained by the fish. Whether or not the abnormal minute scales occur generally, 

 their existence in some forms throws doubt on the assumption tliat the number of 

 scales remains constant with age. Especially is this true when it is known that this 

 assumption has never been supported by critical data. 



To test the question of constancy in the number of scales throughout life, I have 

 enumerated the scales in the lateral line of the lake herring (Leucichthys artedi) collected 

 at Bay City, Mich., October 26, 27, and 29, 1921, November 1, 1922, and November 

 12, 1923. The fish of October 26 and 27, 1921, were taken in different pound nets 

 set in Saginaw Bay. Each of the remaining collections forms a unit, as each repre- 

 sents the partial catch of 1 pound net. These unit collections were taken from the 

 same fishing grounds in Saginaw Bay about 3 miles west of the mouth of Saginaw 

 Eiver. AO the herring were taken on or near their spawning ground and were nearly 

 ready to spawn. There can hardly be any question, therefore, as to the homogeneity 

 of the materia] taken in the same haul. As individual herring are not known to return 

 to the same spawning ground each year, the possibility exists that the fish collected 

 in different years may belong to different races, even though taken on the same 

 spawning grounds. 



