894 BULLETIN OF THE BUREAU OF FISHERIES. 



Rain water was tried in substantially the same way and with mummi- 

 chogs, sunfish, perch, and trout, with the same result. The mummichogs were 

 the most resistant, living 41 hours. 



One is led to conclude from this that foreign matter other than dissolved 

 air is a necessary accompaniment of water that supports fish life, and that 

 water can be too pure for fishes. The law is probably of wide application, for 

 low forms of life are known to die readily in distilled water. It is natural to 

 infer also that death is brought about in these cases by some osmotic reactions 

 through the gills, which bring the blood, known to contain various salts essen- 

 tial to the life of the fish, into intimate relation with the water. It is an assump- 

 tion open to several objections to explain the death as due to the dissolving 

 out of salts or other substances from the blood. Certain obscure poisonous 

 products are believed to be generated in the distillation of water and, conceivably 

 even in rain water, these may have a toxic action on fish. If so, their toxicity 

 is neutralized by contact with many simple substances. It is known that some 

 toxic principles in ordinary water are thus neutralized, as will appear later. 



For practical fish-cultural purposes it may be assumed that a certain mini- 

 mvun of dissolved solids is necessary to water before it is suitable for fishes, and 

 no doubt there is also a maximum which should not be exceeded, though a wide 

 adaptability must exist, as some fishes can frequent both fresh and salt water. 

 Where either of these limits lies can not be at present stated. Of course natural 

 waters which contain fishes furnish the safe conditions both as to quantity and 

 quality of these necessary impurities, which are common substances — carbonates, 

 sulphates, chlorids, in combination with calcium, magnesium, sodium, and other 

 common metals. Potomac water had in October, 1905, 240 parts per milHon; the 

 spring water at the White Sulphur Springs (W. Va.) hatchery had in February, 

 1906, 484 parts. Neither of these amounts is objectionable so far as known. 

 Many waters have a total solid content below 50 parts per million, and fishes 

 inhabit waters containing no more than 20. It is perhaps true that water with 

 much less solid matter than this would support fishes. It would be interesting 

 to find if possible some natural water fatal to fishes solely on account of its 

 high purity. 



It seems that these considerations about the quantity of dissolved solids 

 may become of some practical importance when fish are transferred from one 

 water to another, as from one high in total solids to one low in total solids. 

 Possibly one water may differ so greatly from another in this respect alone that 

 a gradual transfer by slowly mixing the two waters is advisable, in order that 

 the fish may adjust itself from the one to the other. Trout, for instance, do not 

 always thrive after transfer, even when both waters seem admirably adapted to 

 the trout already in them. 



