Later experiments in Canada indicate that they will do as well in salt water, but the construction and maintenance of inclosures is much easier when they are located above the reach of the tide, to say nothing of the proximity of suitable fresh water for the treatment of the eggs. In the precise location of the inclosures several changes have been made, but they have always been in fresh water, and within convenient distance (5 to 10 miles) of the place where the salmon were captured. In our experiments and routine work we have made use of four inclosures, which I will now describe. No. 1. In Craig's Pond Brook, a very pure and transparent stream, an artificial pond 40 square rods in area and 7 feet in extreme depth, was formed by the erection of a dam. The bottom of this pond was mainly a grassy sod newly flooded. About half the water came from springs in the immediate vicinity, and the rest from a very pure lake half a mile distant. The water derived from the lake was thoroughly aerated by its passage over a steep rocky bed. The transparency of the water in the pond was so great that a pin could be seen at the depth of six feet. This inclosure was a complete failure. The salmon placed therein were after a day or two attacked by a parasitic fungoid growth on the skin, and in a few days died. Out of 59 impounded not one escaped the disease and only those speedily removed to other waters recovered. Several, removed in a very sickly condition to the lake supplying the brook, recovered completely, from which it is safe to infer that the cause of the trouble did not lie in the lake water. Of the spring water I have some suspicions, and should not dare to inclose salmon in it again. No. 2. After the failure of the above experiment an inclosure was made in the edge of an ordinary lake by stretching a stout net on stakes. This water was brown in color, and objects 4 feet beneath the surface were invisible. The bottom was gravelly and devoid of vegetation. The depth was 7 and one half feet in early summer, and about 4 feet after the drought of August and September. The area inclosed was about 25 square rods in June, and perhaps half as much at the end of summer. This inclosure was entirely successful, very few salmon dying in it except those that had been attacked by disease before their introduction, and all the survivors were found to be in first-rate condition in November. This site was not afterwards occupied, because it was inconveniently located, and was exposed to the full force of violent winds sweeping across the lake, and therefore unsafe.