278 BULLETIN OF THE BUREAU OF FISHERIES 



ander entfernt und in geringerer Zahl im Verhaltniss zur Gesammtlange des Kor- 

 pers." According to Arwidsson (1910), Hofer (1895), however, refutes Fritsch's 

 conclusion that the age of a salmon can be ascertained from the number of concentric 

 ridges on its scales. 



Petersen (1895) referred to the zones on eel scales as "growth streaks which 

 possibly correspond in number fairly exactly to the years passed." 



Smitt (1895, p. 957), in delineating and describing a marine herring scale, writes: 

 "In the striature the growth rings of the scale also appear as concentric lines." The 

 accompanying figure indicates clearly that by "growth rings" the author referred 

 to the annuli. 



It was not until 1898 that the scale metliod of age determination was tested criti- 

 cally. In that year Hoilbauer published a preliminary paper, in which he set forth 

 the true character of the annidi, supporting his views by experimental evidences. 

 These are discussed elsewhere (p. 294). 



Thus far I have traced the development of the age hypothesis based on the 

 structures of scales from its inception in 1686 to the end of the nineteenth century. 

 During this period all the fundamental structural phenomena utilized in the scale 

 method were discovered and described; but the exact relation of these structures 

 to the life history of the individual remained imdetermined, in spite of the fact that, 

 peculiarly enough, the first presentation of the theoiy was in all probability the cor- 

 rect one. In 1898 the correct hypothesis was rediscovered and critically tested by 

 Hoffbauer, and the second period of scale study was ushered in. During this second 

 period the scale method was established firmly, elaborated greatly, and applied 

 extensively. It need be said here only that publications appeared in England, Scot- 

 land, Norway, Denmark, Sweden, Russia, Holland, Germany, France, America, 

 and other countries. In these studies more or less elaborate life histories based on 

 scales have been worked out for the whitefishes, salmon, trout, marine herring, hali- 

 but, plaice, flounder, sole, smelt, mackerel, muttonfish, sardine, eel, hake, haddock, 

 cod, squeteague, perches, and other fishes. The literature of this period is too 

 voluminous to review. The more important papers will be considered under the 

 various subtopics to which they refer. 



Part I— CRITIQUE OF SCALE THEORY AND METHOD 

 ASSUMPTIONS OF THE METHOD 



The soundness of the scale method of detennining the length of a fish at successive years of its 

 life and its annual growth increments depends on the validity of the following propositions: 



1. That the scales remain constant in number and [retain their] identity throughout tlie life 

 of the fish. 



2. That the annual increment in tlie length (or some other dimension wliich must then be used) 

 of the scale maintains, througliout the life of the fish, a constant ratio with the annual increment 

 in body length. 



3. That the annuli are formed yearly and at the same time each year [or that some other dis- 

 coverable relation e.xists between their formation and increment of time]. 



Incidentally the following questions are raised, but the validity of the scale method of compu- 

 tation is not affected by them: 



4. Whether the annuli represent periods of retarded or arrested growth of the scale? 



