BULLETIN OF THE UNITED STATES FISH COMMISSION. 381 



at home in fresh water and having there their feeding grounds they 

 continue to feed until the close approach of the spawning time, and 

 hence they could not be penned up in the summer without some pro- 

 vision for an artificial supply of food, which would probably involve a 

 great deal of expense and trouble. Moreover, the necessity of collect- 

 ing breeding fish early in the summer does not exist, because they are 

 at no time more congregated and easy to catch than at the spawning 

 season. 



The capture of the fish is easily effected by stretching a net across the 

 outlet of the lake and leading them through a tunnel-formed passage 

 into an inclosure of netting. There happens to be at this point a wid e 

 surface of smooth bottom, with water from 1 to 3 feet deep, affording an 

 excellent site for spacious inclosures not only for entrapping but for 

 assorting and storing the salmon during the spawning season. Nets 

 are generally stretched across the stream (to keep the fish back in the 

 lake) immediately after the beginning of the close season, September 15. 

 The very earliest of them begin to spawn before the end of October, but 

 the actual inclosing of the breeding stock is deferred until the early days 

 of November. The taking of spawn generally begins about November 

 6, and continues for two or three weeks. Commonly by November 20 

 or 22 this work is completed, and the breeders are carried a mile or two 

 up the lake and liberated. 



The method of manipulation is the same employed at the Penobscot 

 station, and is not supposed to differ materially from that adopted by all 

 American breeders of Salmonidae. The results in the impregnation of 

 the spawn are not so uniformly satisfactory as at the Penobscot station. 

 There appears to be a greater prevalence of ovarian disease than among 

 the migratory salmon. The occurrence of white eggs among the nor- 

 mally colored and healthy ones, as they are yielded by the fish, is very 

 common, and occasionally the entire litter is found to be defective. It 

 is not improbable that there are some eggs that are incapable of impreg- 

 nation, though exhibiting no visible symptoms of disease. However, 

 the general result is satisfactory, the ratio of impregnated eggs being 

 from 93 to 95 per centum. 



The facilities for developing and hatching the eggs are rather poor. 

 No good site could be found by the side of the stream, no suitable brook 

 could be found near enough to the fishing grounds, and the neighboring 

 springs lacked either volume or facilities for utilization. At present 

 there are three hatcheries in use, two of them using spring water exclu- 

 sively and one of them lake or stream water exclusively. The lake 

 water would be preferred, but unfortunately it can only be used for the 

 slow development of part of the eggs, circumstances connected with 

 the floating of timber down the stream compelling the evacuation of 

 that hatchery in March. The main hatchery is very favorably located 

 and served, except that the water is all spring water, and this unfavor- 



