64 BULLETIN OF THE BUREAU OF FISHERIES. 



that there may be a difiference of several months in the time at which the yolk sac will 

 be fully absorbed and the fish begin an active existence. It can not well be doubted, 

 then, that the earliest fry to migrate have been hatched in the lower tributaries, and 

 it seems reasonable to assume that, in a general way, at least, the successively later 

 migrants have come from successively higher tributaries. 



The abnormally large proportion of smaller fish found in the lower part of the river 

 during the spring and early summer, which causes the "skewing" of the frequency 

 cur\'e of length noted on page 8, may also indicate that the height of the migration 

 has not been passed and that the smaller fish entering from above are doing so in 

 constantly increasing numbers. After the height of the migration the skewing effect 

 of the constantly decreasing numbers of smaller fish would not be noticeable. This 

 skewing of the curves of length is not found to any noticeable degree after the early 

 part of the summer, a fact which seems to give additional evidence that the height of 

 migration comes, in the lower part of the Columbia River, during the latter half of the 

 summer or early in autumn. 



The migration of fry in the Sacramento River has been given in detail by Rutter 

 (1903). He found that fry were migrating in the lower part of the river during the 

 months from January to May, inclusive, and that they started the migration from 

 the streams in which they were hatched as soon as the yolk sac was absorbed, as early 

 as October. This migration is much earlier than that observed by the author in the 

 Columbia River, a fact associated with the earlier hatching of the eggs and the more 

 rapid development of the fry in the warmer water of the southern stream. The data 

 presented in this study add nothing to Rutter's conclusions on this point. No migrating 

 yearlings were taken by Rutter (1903) nor by Scofield (1898) in their work on the lower 

 river, but, as no collections were made during the fall and early winter, it is quite 

 possible that there is a migration of the older fish at this time of the year. It is 

 possible that yearlings migrating in the spring are so scarce that none were captured. 

 It has been shown (p. 36) that the new growth of the second year may begin in the 

 case of the young chinooks in the McCloud River as early as September, varying, how- 

 ever, in different years. It has also been shown (p. 48) that in some cases, at least, 

 there is a tendency for the older fry or yearlings to migrate soon after beginning the 

 new growth of the second year. Consideration of these two facts lends considerable 

 probability to the theory that there is a fall migration of older fry in the Sacramento 

 River. An investigation of this matter would be pertinent, since a distinct difference 

 in the scale growth between fry migrating in the spring and those migrating in the fall 

 would be expected. The relation between the young migrating at these two periods 

 (granting that such a later migration takes place) and the adults comprising the sharpl\' 

 separated spring and fall runs of spawning fish might well prove to be of considerable 

 practical importance. 



