482 TJ. S. BUREAU OF FISHERIES 



The "spiller" is another square compartment adjoining either end 

 of the pot (sometimes there are two spillers, one at each end), and is 

 simply a container for fish. A small tunnel leads the fish from the 

 pot into the spiller, whence the fishermen lift them out. This is 

 accomplished by closing the tunnel from the pot, after which the 

 ropes holding the front of the spiller are loosened and the net wall 

 allowed to drop almost to the level of the water. A steam or gasoline 

 tug then pushes a scow alongside the spiller and takes position on the 

 outside of this scow. From the deck of the tug a derrick is rigged 

 with a running line from the steam capstan through the block at the 

 top of the derrick. This line is attached to the far end of a net apron, 

 called a "brailer," which is heavily weighted by having chains along 

 each side and leaded crossways at several places. A small boat is run 

 inside the spiller, and the men in this draw the brailer across the barge 

 and let it sink in the spiller. The fish soon gather over it, when the 

 steam capstan ciuickly reels it in, the net folding over as drawn in 

 from its far side and spilling the fish out on the scow. Men on the 

 scow pick out and throw overboard the undesirable fish. The apron 

 is then drawn back across the pot and the operation repeated so long 

 as any fish remain. In this manner a trap with many tons of salmon in 

 it is quickly emptied. 



Traps, like nearly all other fixed fishing appliances, are built on 

 the theory that salmon, like most other fishes, have a tendency to 

 follow a given course in the water, whether a natural shore line or an 

 artificial obstruction resembling one; also that the fish very seldom 

 turns in its o^vn wake. The trap has taken advantage of these natural 

 tendencies of the fish, and is arranged so that, although the salmon 

 may turn, he will continually be led by the wall of net toward and 

 into the trap. 



If a trap is located in a place where fish play and where an eddy 

 exists, and the fish run one way with the incoming tide and the 

 opposite with the outgoing, it will fish from both directions; if located 

 where the fish simply pass by, as for instance, on a point or reef, it 

 will fish from one side only. 



A variation of the trap, to be used in places where piles can not be 

 driven, is the floating trap. An experimental trap of this variety 

 was used at Uganik, on Kodiak Island, Alaska, as early as 1896. 

 Its use was abandoned in 1897, not to be resumed until some years 

 later. A number of floating traps (of the type invented by J. R. 

 Heckman, of Ketchikan, Alaska) have been and are being used in 

 southeast Alaska, the first having been installed in 1907. The design 

 of this trap follows the shape of an ordinary Puget Sound driven trap. 

 It is constructed of logs, 20 to 26 inches at the butt, bolted and 

 braced together in one solid frame. Suspended from this frame 

 through the logs are 2}^-inch pipes extending down in the water 30 

 feet. Halfway down these pipes and also on the extreme lower ends 

 are ej'cbolts, to which the web is drawn down and fastened. Thus 

 the web is kept in place as well as if the pipes were driven piles. The 

 lead is also a continuation of large piles or logs bolted firmly together 

 with similarly suspended pipes and webbing. 



The so-called wooden traps on the Columbia River are essentially 

 weirs, being a modification of the brush weirs or traps used by the 

 Indians for the capture of salmon long before the advent of the white 

 men. They are built on shore, of piling and planks, the latter 



