NATURAL HISTORY OF THE QUINNAT SALMON. 
105 
assuming that all the marked fishes remained in the pool, the result will be the fol- 
lowing proportion: 97, the number of marked fishes taken in September, is to 586, 
the number marked during August, as 178, the total catch in September, is to 1,075, 
the total number in the pool, which is remarkably close to 1,022. 
If there had been much of a migration between August 18 and September 18 we 
would have taken a larger proportion of unmarked fishes. If there were no migration 
we would expect to get marked fishes in proportion to the total catch as 586 to 1,022, 
which was the proportion of marked fishes in the pool in August. The record for 
September, 97 marked out of a total of 178, is very nearly that ratio, indicating 
that there was little or no migration between August 18 and September 18. 
August 17 we fished two pools, F and G, about a quarter and a half mile, 
respectively, below where any fish had been marked, catching 13 and 115 salmon. 
We also caught 57 in Pool G in September. Neither at this nor at any other 'time 
have we found marked salmon below the pool in which they were released, except 
in the case of Pool C, above referred to. The one marked salmon taken in the first 
haul in Pool E was released in the upper end of that pool. 
It is especially worthy of note that none of the 591 fishes taken in Pool E had 
been marked in the pools above the previous month. If any had left the upper 
pools, they had not stopped in the vicinity. 
Summary on number and movements. — The estimate of the number of young 
salmon in August in Pool B is 685, in C 861, and in E 1,022. Pool E is much longer 
than the others, and might very well have more fishes than either. From these 
'‘estimates it is probable that there were about 10,000 young salmon to the mile in 
the Upper Sacramento during the summer of 1898, or between a half and three- 
quarters of a million in all the headwaters of that stream. 
There is little migration of the young salmon between May and December. 
Where pools are separated by shallow riffles, no evidences of migration could be 
found. If connected by deep water, it was found that about 17 per cent of those in 
the upper passed to the lower. In Hazel Creek there was an upward migration of 
4 per cent during July, and a larger — about 12 per cent — during September. There 
was no diminution in number either in Hazel Creek or the river up to November — 
even an apparent increase. There was a slight migration, however, during the 
whole period, indicated by the disappearance of the larger marked specimens. 
This residence in the headwaters during the summer is probably due to low 
water. It has been noticed many times, both in the streams and in the hatcheries, 
that- young salmon dread going over a fall. When the river is very low, as it is 
every summer, the rapids become almost like waterfalls, thus preventing down- 
stream migration. A slight rise obliterates the fall and at the same time makes it 
difficult to find food; hence the decided migration in December. The abundance of 
food appears to be of some importance when we notice that there was a scarcity 
of food in September and also a slight increase in migration that month, although 
the water was the lowest of the season. 
Young salmon were reported abundant in the pools near Sims the 1st of May, 
1899, and they doubtless remained during the summer, as was found during 1898. 
Growth in fresh water. — It has been shown above (“Migration, general investi- 
gations of 1898”) that the size of young salmon found in May was the same for all 
parts of the river. This was true also in July for the portions of the river in which 
they were found. 
