HISTORY OF THE PACIFIC SALMON. 585 



long when recaptured. The other fish, marked 1180, was 

 26| inches long as a kelt and 36 inches long when recaptured. 

 The measurements on the photographs of the scales from the 

 centre to the anterior edge, and from the centre to the spawning 

 mark are 130 and 116 for No. 9194, and 198 and 175 for No. 1180, 

 the unit of measurement being ^^^^th inch. This would give the 

 kelt measures as 27| inches and 32 inches respectively. The 

 former measure is only ^ inch wrong, but the latter is nearly 

 <5 inches wrong, and shows either that the scale is abnormal or 

 that Dahl's system of measurement is not applicable to a fish 

 that has spawned. 



Another criticism is that measurements from several different 

 scales of the same fish seldom all agree exactly, and I have there- 

 fore come to the conclusion that it is very unsafe to rely on the 

 measurement of one scale in estimating the size of the fish at 

 various ages. The reason may be that it is not easy to recognize 

 •either the exact centre of growth, or the exact limits of the 



Text-fia-. 102. 



'^. 



iA 



J^\ 



■Quiimat (O. tschaioytscha). 25 lbs. 13th October, 1911. Shuswap, 



South Thomson Kiver. 



The black lines show the variations in the long axis of the scale. 



various bands, but I do not think that the scales grow quite 

 equally. Still the idea seems so well founded in theory, and to 

 have worked out so exactly in practice on a large scale in Norway, 

 that I believe I may consider myself justified in drawing con- 

 clusions from the average measurements of a considerable number 

 ■of scales taken from the same fish, and whenever possible, I have 

 measured 20 scales. A smaller number might sufl&ce in the case 

 of a true salmon, but the scales of the Pacific salmon are more 

 difficult to read. As a rule, the limits of the bands are less 



