300 BULLETIN OF THE BUREAU OF FISHERIES. 



scales from the same region of the same fish is approximately constant. He noticed that 

 the number of circuli increases where the scale becomes smaller and also that there is a 

 considerable difference between different specimens for the same year. In short, his 

 theory is that during favorable seasons when the food supply is plentiful, the fish grows 

 more rapidly than in seasons of poorer food supply, and that this difference of growth 

 is indicated by the circuli, they being closer together during seasons of slow growth than 

 in seasons of rapid growth. 



Johnston (1905, 1906, 1907, 1909), in Scotland, undertook to work out the difficult 

 life history of the salmon from its scales. He found the annuli and correlated them 

 with the known facts of the life of the fish in such a way as almost to estabUsh the con- 

 clusion that they are winter marks. He discovered the "spawning mark," which is 

 the worn or absorbed part of the scale that was the periphery at the time that the fish 

 went into fresh water to spawn. 



Thomson (1904) and Tims (1902, 1906) worked independently at about the same 

 time. The work of the former was done on the Gadidae and Pleuronectidae. Besides a 

 splendid review of the literature up to his time, of which liberal use has been made in 

 this paper, he gives many minute measurements to support the points he makes. He 

 is convinced that the annuli represent years and finds that the number of circuli in a 

 band of a given width varies but little. His statistics show that if the annuli do not 

 represent years there is a remarkable coincidence. 



Tims is skeptical of age determinations by this means. He admits that he is able 

 to follow Thomson through the first summer and winter bands, but that he can detect 

 no further alternations. In his reasons for his disbelief he points out Klaatsch's obser- 

 vation that scales do not appear simultaneously all over the body and not until the 

 fish is 3 to 4 cm. long, an objection which he admits, however, to be of little consequence. 

 He finds variations in the number of annuli on scales on different parts of the same fish. 

 Furthermore, the number of circuli varies indefinitely. And more important than all, 

 scales are lost and are replaced, and for this reason age determination by means of 

 scales is impossible. 



Thomson, referring to Tims's objections, admitted that age determinations in old 

 fishes are dilficult, and in some fishes, even in the young, probably more difficult than 

 in others, but asserted that in the cod the evidence is both plain and conclusive. He 

 includes in his paper many observations of variation, tending to corroborate his con- 

 clusions. 



Brown (1904) raises further objections to the theory supported by Thomson and 

 others. They are : 



1 . Gadoid fishes shed their scales immediately after spawning. 



2. After the age limit of spawning no further shedding takes place. 



3. The concentric rings of scales of fish do not represent annual increments, but 

 must have other causes. 



He finds scales on a 3-year-old cod with 30, 60, and 90 circuli, respectively, depend- 

 ing on the location on the body ; hence this method of determining age is of no value. 



Among other conclusions, Dahl (191 1) says that injuries or adverse conditions, 

 even in summer, will produce annuli. Further, he applied the method devised by 

 Johnston of calculating the length of the fish for each year by the proportionate width 



