FISHERIES OF ALASKA, 1908. 39 



easy accomplishment of tliis object. Where the river in its upper 

 reaches is narrow enough to rack, the current is too swift or the bottom 

 not suitable for driving piles. The river is narrowest at its origin at 

 the foot of Lake Aleknagik. On May 31, 1908, at tliis point, which is 

 at the Indian village, the width was very nearly 275 feet at liigh tide. 

 The highest tides affect the level at this point but a few inches, or 

 perhaps a foot, and the differences in the width of the stream due to 

 tidal influence are very small. Seasonal changes, however, make a 

 difference of several feet in the width. This place was expected to be 

 the most feasible for the counting experiments, but when examined on 

 May 31 the depth in the channel was 16 feet and the current 4 to 5 

 miles an hour. These conditions, in connection with the probable 

 difhculty of driving piles in the hard bottom, led to the abandonment 

 of this site and the selection of the head of the so-called lagoon which 

 constitutes the lower end of the lake. This lagoon is a circular body 

 of water about 1 mile in diameter. Its junction with the main body of 

 the lake makes a comparatively narrow constriction rather more than 

 200 yards wnide, with a greatest depth on May 31 of about 12 feet. 

 The constriction is marked by two distinct gravel spits making out 

 from shore on either side. The rack was placed here and was thus 

 really across the lake at its narrowest point instead of across the 

 river. 



On tliis date, May 31, the lake was almost entirely covered -with a 

 layer of greatly honeycombed ice 4 to 6 inches in tliickness. The 

 lake was open only for a short distance above the lagoon. The tem- 

 perature of the lake near the surface in the lagoon was 38° F. at 

 8 a. m. On June 12 about two-tliirds of the lake was open, but the 

 ice still covered all that portion between the main inlet and the head 

 of the lake. 



Work was begun June 14 and the pile driving was finished June 16. 

 A somewhat curved line of piles was driven diagonally across the 

 channel from one ^hore or spit to the other, with the convexity of 

 the curve upstream. Ordinary trap web was hung on the upper side 

 of these piles and anchored to the bottom of the lake by means of a 

 heavy chain running along and lashed to its lower edge. At the 

 north end of the rack and close to the left bank a trap or pot was 

 constructed similar to the pot or pound of a fish trap, but without a 

 web bottom, the web of the sides reaching to the bottom of the water. 

 The trap was somewhat bottle-shaped, about 40 feet long and 25 feet 

 wide, tapering at either end to an entrance and outlet gate, each 

 4 feet wide. (Both these gates were afterwards modified as a result 

 of experience with the running salmon.) One hundred and forty-five 

 fathoms of web were used in the trap and rack, and tliis was sup- 

 ported by about 50 piles. At the highest stage the water was 5i feet 

 deep at the upper end and 6^ feet at the lower. 



