DEVELOPMENT AND GROWTH OF SALMON-FRY. 553 



error, but to place the young fry in circumstances as nearly resembling the state 

 of nature as was consistent with their preservation. 



The ponds, which are three in number, are two feet deep, and thickly em- 

 bedded with gravel, while they are at the same time supplied with a small stream 

 of spring- water in which the larvae of insects abound. Pond No. 1 is 25 feet in 

 length by 18 in breadth, and is fed by the stream, which debouches into it at the 

 fall F. Pond No. 2 is 22 feet in length by 18 in breadth, and is fed from pond 

 No. 1 at G, where the communication is carefully grated with wire. Pond No. 3 

 is 50 feet in length by 30 in breadth, and is fed by the stream at F, having no 

 communication with either of the other ponds. The waste water from pond 

 No. 1 is conducted into pond No. 2, through a square wooden pipe covered at the 

 mouth with a wire-grating, the bars of which are about one-eighth of an inch 

 apart. The waste water from pond No. 2 is conveyed under ground to the dis- 

 tance of 20 feet in a square wooden pipe grated in the same manner as the former. 

 The waste water from pond No. 3 passes down a square wooden pipe 2 feet deep 

 covered at the top with wire-gauze, and is conveyed under ground in a smaU 

 covered drain to the distance of 20 feet from the pond. The water of the whole 

 is then left to find its way to the river. 



To prevent any communication arising from an accidental overflow of the 

 ponds themselves, I raised embankments upon the intersecting walks of 2 feet in 

 height, so that the several families of fish which the ponds contain can have no 

 access, du-ect or indirect, to each other. Where the rivulet is divided for the 

 purpose of supplying the several ponds, I have formed an artificial fall in each 

 stream, of a construction to prevent the fish from ascending one stream and de- 

 scending another. Finally, where the water discharges itself from the ponds, 

 the channels are so secured by wire-grating that it is as impossible for the young 

 fish to escape as for any other fish to have access to them. The whole occupies 

 an area of nearly 80 feet square. 



My experimental basins being thus prepared, my next object was to secure 

 the fish, the progeny of which were to form the subject of experiment. With 

 the view, therefore, of securing two salmon, male and female, while in the very 

 act of continuing their kind, I provided myself with an iron hoop 5 feet in dia- 

 meter, on which I fixed a net of a pretty large mesh, so constructed as to form a 

 bag 9 feet in length by 5 in width. I then attached the hoop and net to the end 

 of a pole 9 feet long, thus forming a landing-net on a large scale. The weight of , 

 the net with its iron hoop being upwards of 7 lb., it instantly sunk to the bottom 

 on being thrown into the water. 



Being thus prepared with all the means of carrying my experiment into 

 practice, I proceeded to the river Nith on the 4th January 1837, and readily dis- 

 covered a pair of adult salmon engaged in depositing their spawn. They were in 

 a situation easily accessible, the water being of such a depth as to admit of mv 



