ATKINS THE SALMON AND ITS ARTIFICIAL CULTUEE. 247 



was with the intention of escaping, but another salmon was shortly 

 after seen to jump into the enclosure over the same hedge ; and I presume 

 both leaps were made at random. 



At times during the summer the salmon would for a day or two al- 

 most entirely disappear. This happened on the 1st and 2d days of July, 

 and an examination of a weather record shows that both days were 

 warm, with very light and variable winds, and a clonded sky. But they 

 re-appeared a few days later as active as ever. 



This extraordinary activity soon diminished, and by the middle of 

 August had decreased 80 per cent., bnt there was hardly a day during 

 the whole season, up to the spawniug time, when there were not some 

 salmon leaping. Meanwhile they had ceased to frequent the shores, and 

 had become so wary that it was no longer possible to approach within 

 arm's-length of them, as miglit easily be done during the first days of 

 tlieir confinement. 



Late in August, being now confident of the survival of a large part of 

 the salmon, and of a correspondingly large yield of eggs, I began prep- 

 arations for the spawning season. A site for a hatching-house had been 

 selected on the brook, about GOO feet below the outlet of the pond, and a 

 substaiitial dam had been built here by Mr. Swazey, the owner of the 

 premises, from whom I had secured a lease. The only source from 

 which to draw a water-supply was the brook itself, there being no spring 

 of considerable size in the vicinity; and the house was accordingly so 

 ocated as to take water conveniently from the dam. The main building 

 was made 70 feet long and 28 feet wide, one story high,* and contained 



* It was built entirely of wood, the walls being boarded on both sides of the studding, 

 and the space between filled with sawdust, as a precaution against cold. With the 

 sauie object in view, the exterior was covered with sheathing-paper and then shingled. 

 Two stoves, oue for wood and the other for coal, were provided, with the anticipation 

 that it might prove a difficult task to keep the water in the hatching-troughs from 

 freezing. Happilj^, the fact proved otherwise. Across the upper or north end of the 

 building ran a feed-trough 11 inches deep and 12 inches wide, wbich received water 

 from a conduit leading from the dam, and discharged it into fifteen hatching-troughs 

 which ran lengthwise with the building. Thirteen of these troughs were each 60 feet 

 long, 1 foot wide and 6 inches deep inside. The other two troughs were of the same width 

 and depth, and 56 feet long. The trongh-room, therefore, measured 892 square feet- 

 A deduction of 18 inches per trough, as an allowance for waste room at each end, 

 leaves 869.5 feet as the total available space for the deposit of eggs. Each trough was 

 fed by a short spout an inch in diameter, and, when in actual operation, was found to 

 use about 300 gallons of water per hour, making 4,500 gallons per hour as tlie total 

 amount used in the building. The troughs were made of unseasoned spruce boards 

 and put into use without paint or varnish of any kind, or even planing. They were 

 set on the floor nearly level, and had no divisions to check or break the water. 



The liability to an excess of sediment from the water of the brook was so great that it 

 was not considered prudent to place the eggs on gravel, and a sufficient number of wire- 

 cloth trays was provided to hold them. The trays were made of iron wire cloth, with 

 wires an eighth of an inch apart, tacked to light wooden frames, two feet long, one foot 

 wide, and seven-eighths of an inch deep; they were varnished with a so-called paraffine 

 varnish to prevent rusting, and were furnished with legs five-eighths of an inch long 



