40 FISHERIES OF ALASKA, 1908. 



On June 21 the web had all been hung, but the attachment of the 

 chain had not been finished. The construction or trap crew was 

 taken away and the work finished with the help of native labor, but 

 this was not done mthout considerable difficulty. In the channel 

 the width of web proved too short to reach bottom by some 3 feet. 

 While the soundings made May 31 and June 11 had shown only 12 

 feet on a straight course between the two spits, the web was found to 

 be in 18 feet of water in the deepest place. The fact was that the 

 line of piles did not follow the line of soundings but was considerably 

 curved upstream, and the difference was sufficient to add 6 feet to its 

 deepest water. The lack was made up by attacliing a strip of web at 

 the top and lowering the whole. 



The piles closest to the south shore, at the opposite end from that 

 of the trap, struck either rock or frozen ground, for the first two of 

 them gave way and were found to have their lower ends mushroomed 

 without having penetrated far. One gave way June 21, the other 

 June 23, and their places were taken by cross shear legs cut from the 

 neighboring timber and guj^ed by lines to the shore. The third pile 

 bent over and had to be strongly guyed. These repairs held the web 

 up efficientl}^ until July 17, when, during an attempt to strengthen 

 the guy ropes, the whole structure cf shear legs collapsed and required 

 all hands during the next few hours to repair. 



After the rack was supposedly finished it was found that the depth of 

 water made it impossible to tell whether or not, in the deepest places, the 

 chain was securely on bottom and the structure impassable to salmon 

 save by \\'a,j of the gate. Even in the shallow places it was not easy to 

 see the exact condition of the web and chain, since the slightest ripple 

 on the surface confuses the picture of the object beneath. The 

 current was seen to exert a strong pressure on the web, increased by 

 a certain amount of drift caught b}^ the meshes. Nowhere did the 

 web hang perpendicularly, but between the piles was bellied but 

 downstream by the steady pressure of the current. This tended to 

 lift the chain off the bottom, especially midway between the piles. 

 A crude form of water glass, or submarine telescope, was made by 

 inserting a pane of window glass in the bottom of an empty keg. 

 With the aid of this the bottom could be distinctly seen, in bright 

 sunlight, even in the deepest places, and several places were found 

 through which it was possible for salmon to pass. These were chiefly 

 due to the raising of the chain from bottom, but in a few places holes 

 had been excavated by the current at the base of some of the piles. 

 Closure of all these openings was accomplished chiefly by burlap salt 

 bags filled with gravel. These were sunk on the bights of the web 

 close to or overlying the chain, where by their weight they held the 

 chain in position. They were also dropped into the excavated holes. 



