68 



BULLETIN OF THE BUREAU OF FISHERIES 



15 per cents of September 8 and 11 stand closer to the general average of this 

 group during that portion of the run. The general average for the 1,300 individuals 

 that we examined in 1925 is 17.3 per cent of 4 3 's. Had we attempted an estimate 

 of the 1926 run, based on the relative number of 4 3 grilse present in 1924 and 1925 

 and the ascertained run of 1925, the above figures would have warranted us in pre- 

 dicting a somewhat larger run in 1926 than we had in 1925, but not as large as the 

 1926 run actually proved to be. If the 1925 percentage of grilse had been 22 instead 

 of 17.3, this would have furnished us a basis for a fairly accurate forecast of the 

 1926 run. However, admitting the incompleteness of our data for both 1924 and 

 1925 and the improbability that even with adequate determinations of the numbers 

 of grilse we should have a basis for any closely accurate predictions, our experience 

 with the 4 3 grilse counts in 1924 and 1925 and the corresponding 5 3 returns in 1925 

 and 1926 is favorable to the theory that a fairly definite relation exists between the 

 number of younger fish occurring in the run of one year and the size of the run of 

 the following year. The experience of only two years, however, is insufficient founda- 

 tion for any high degree of confidence, and we must look to the results in 1927 and 

 susequent years either to confirm or to refute the theory we tentatively entertain. 



The run in 1927 will offer crucial evidence in this direction. As we have seen, 

 the spawning reserve of its principal brood year, 1922, was very limited and appar- 

 ently inadequate to produce a normal run. On the other hand, such superficial 

 observations of the fingerling migration of 1925 as were made gave favorable indica- 

 tions, although these were not of such character as to inspire confidence. We have 

 now to consider the 4 3 grilse run of 1926 as our third basis for prediction, and it may 

 be stated at the outset that it was almost nonexistent. Appreciating the significance 

 of these data, a much more extensive series of determinations was made in 1926 

 than in either of the two previous years. The run was sampled on September 1, 3, 

 4, 6, 8, 9, 10, 11, and 15, one thousand individuals being examined on each of those 

 dates, except the 3d, the 11th, and the 15th. Five hundred, each, were examined on 

 the 3d and 11th, and 700 on the 15th. The total number included in this test, there- 

 fore, was 8,700. It is believed this sampling was entirely adequate, especially in 

 view of the comparatively little variation in the results from day to day. On only 

 one day (the 15th) was the number of the 4 3 group present as high as 1 per cent of 

 the sample taken. On the other 9 days the percentages varied from three-tenths to 

 nine-tenths of 1 per cent, with no obvious trend. The total number observed was 

 48, of which 41 were males and 7 females. The average percentage of the 4 3 group 

 present in the run during the first half of September was thus only fifty-five one- 

 hundreths of 1 per cent. When these figures are compared with the 13.4 per cent 

 in 1924 and the 17.3 per cent in 1925 it is seen that the grilse of 1926 corroborate 

 the evidence derived from the limited spawning of 1922 and give no grounds for hope 

 of a normal return of 5 3 's in 1927. 



As we have stated above, such hope as exists must be based on the general 

 impression of an extensive fingerling migration in 1925. These fingerlings were 

 largely the progeny of the scanty 1922 spawning, and we have no knowledge of any 

 physical or other conditions in 1922 that could be considered extraordinarily favorable 

 for the production of a large crop of fmgerlings from a strictly limited number of 

 eggs. Nevertheless, such conditions may have existed. 



