102 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
young salmon in Hazel Creek continued to work their way upstream during Sep- 
tember, October, and November. 
This table shows an increase in the number of salmon in October. The average 
of two hauls in May was 17; one in July gave 55; six in August gave an average of 
19; two in September averaged 46; two in October 110; two in November 103. The 
simple fact of there being more fishes found on later dates would not indicate an 
upward movement; they might have come downstream. The larger percentage of 
marked fishes in the upper pool, however, would indicate an upstream movement. 
Several places above the upper pool were fished in October — one place within 15 
yards of it — but no marked salmon were found. This would indicate a lack of upward 
movement; but, all the data being considered, it is evident that there was at least 
a slight upward movement in October and November. 
In December the young salmon had all disappeared from the upper pool, and only 
1 were found in the lower. One of these was a specimen marked in July or August. 
That is, 1 out of 200 remained after the December rise. It is evident from this that 
practically all the young salmon left the creek between November 18 and December 
18. There was a heavy rain in the vicinity on November 28 and 29 (precipitation 
over 2 inches), and there can be little doubt that the salmon all left at that time. 
Migration during summer . — Pool B (see plate 11) of the river is separated from 
Pool A (at the mouth of Hazel Creek) by a rapid about 150 yards long with a fall of 
about 1 feet. It is over 6 feet deep, with large angular rocks along one edge afford- 
ing excellent hiding-places for young salmon. It required three men to seine this 
pool well, one to throw out the seine from a large rock at the upper end, and two to 
pull it in. When there were but two of us, one would pay out the seine from a 
riffle above the pool; the other would wade out as far as possible in the lower end 
of the pool and pull the seine down with a rope. When the seine was stretched 
through the pool it was pulled ashore. Obviously such work was not very satisfactory. 
The following table gives a record of the catch at each haul of the seine in Pool 
B in July and August. Four hauls were made in this pool in May, about 50 young- 
salmon being taken in one haul. The specimens caught in May represented two 
sizes, such as were found in Hazel Creek on the same date, but there were very few of 
the smaller size. These smallest were probably from a few late-spawning individuals. 
The table gives for each haul of the seine in July and August, (1) the date, (2) 
number of young salmon caught, (3) number previously marked and released in the 
pool in July, (4) time since the Juty-marked fishes were released, (5) number of July- 
marked fishes caught, (6) number previously marked in August, (7) time since the 
August-marked fishes were released, (8) number of August-marked fishes taken. 
These data may be used to estimate the number of young salmon in the pool by 
making the following proportion for any particular haul of the seine: The number 
of marked fishes taken is to the total number of marked fishes known to be in the 
pool (having just been released), as the total number taken is to the toted number 
in the pool. The results, of course, are variable, and it is only by a number of trials 
that Ave can get near the probable truth. The value of the estimates is not enhanced 
by there being no marked fishes taken at certain hauls. In such cases, however, 
there were but few of either kind. The estimate made from each seine haul' is given 
in the table. 
The freshly marked specimens could be distinguished from those marked a 
month previously by the latter having the scar healed. 
