42 FISHERIES OF ALASKA, 1908. 



Up to and including the 9th of July all salmon were actually 

 counted. This was done with the aid of tally registers such as are 

 used by the tallymen in counting salmon from the fishing boats 

 into the lighters. All counting was done at the outlet gate of the 

 trap during this time. The inlet tunnel to the trap was kept open 

 almost continuously, and at times during the process of tallying the 

 fish passed directly through the trap without stopping within it. 

 At the bottom of the gate and just outside the sill was placed a 

 board about a foot wide and covered with white cloth to serve as a 

 background against which the dark outlines of the passing salmon 

 could be plainly distinguished. This made the operation of tallying 

 very simple, since every individual fish could be plainly seen, and the 

 stream of fish had only to be so regulated by the gate aperture as 

 not to come too fast to tally. 



The first 84,000 salmon were thus counted individually. On 

 July 10, however, the fish massed in such quantities below the web 

 that it was considered inadvisable to hold them back for an indi- 

 vidual tally and it became necessary, as was anticipated, to make a 

 basis for a more wholesale count. The method afterwards used as 

 long as the run remained heavy was developed by Mr. Wallich as a 

 logical and natural result of *the behavior of the schools of salmon 

 in passing the rack through the gate and tunnel. Contrary to expec- 

 tation, salmon, even when pressed by hundreds of thousands in the 

 rear, passed the gate in an orderly and almost regular way, and at 

 a rate which decreased or increased gradually and never suddenly. 

 They did not become panic stricken and wedge themselves in the 

 opening, nor otherwise interrupt a fairly uniform flow of fish com- 

 parable to the flow of a streani. This uniformity afi^orded a basis 

 for accurate estimation of numbers without the actual tally of every 

 individual. The method adopted consisted of an actual tally of 

 every individual during a period of one minute, repeated at short 

 regular intervals, as every quarter hour. The number passing during 

 this minute was regarded as the average for fifteen minutes. A 

 sheet with the whole day divided into quarter hours was kept ready 

 at the gate and the number for one minute as taken from the tally 

 register was immediately entered thereon by the attendant who 

 made the tally. From these figures the total for the day was obtained. 



This method was checked several times by making the one minute 

 count at more frequent intervals, but no additional accuracy resulted. 

 The method is best applicable to a continuous stream of fish and 

 was mainly used therefor when there was a press of salmon sufficient 

 to keep up an uninterrupted flow. When their numbers became so 

 reduced that they passed out intermittently it was then easy to 

 make an actual tally of all passing, or the outlet gate could be 

 closed until the trap became again well filled with fish. The 



