326 U. S. BUREAU OF FISHERIES 



ascertaining the amount of chlorine introduced were used. First, 

 measured quantities of the concentrated chloriue sohition were added 

 to 10-liter batches of water, and the chlorine present was determined 

 by comparison with standards •'^ using orthotolidine reagents as de- 

 scribed by EUms and Hauser/ This was not found to be as satis- 

 factory as the method later used — namely, the analysis of the strong- 

 chlorine solution using 0.1/N-sodium thiosulphate with the starch 

 iodide indicator and the addition of known amounts of this solution. 

 In all cases the analysis of the strong chlorine solution was made im- 

 mediately before using. Ten liters of water were used in each trial. 

 Chemicals were added to the water and stirred a very short time. 

 Five fingerling trout were then added. No food was given the tish 

 during any one experiment, and new, healthy fish were usecl for each 

 experiment. This was to prevent the results from being aiiected by 

 unknown factors, such as water pollution or condition of fish. 



Inasmuch as the chlorine Avas added in quantities much greater 

 than those encountered in city waters, and as the amounts of chemicals 

 used to counteract this are also correspondingly larger than will ordi- 

 narily be needed, it is felt that the results obtained should constitute 

 an excellent indication as to what will happen to the fish when ex- 

 posed to smaller quantities of these chemicals for longer periods of 

 time. Also, as the rainbow trout seems to be most susceptible to 

 poisons of this sort, it is reasonable to assume that the results ob- 

 tained will apply to other species of fish. 



The experimental results may be summarized as follows : Except- 

 ing ferrous sulphate, none of the reagents used for removing chlorine 

 were toxic to the fish. Activated charcoal will remove chlorine, but 

 due to the fact that the removal is a surface phenomenon it will not 

 be complete unless special care is taken. This seems to render char- 

 coal undesirable. Silica gel would be open to the same difficulties. 

 The chemical reducing agents tried are efficient in the following 

 order: Sodium bisulphite, sodium thiosulphate, and sodium nitrite. 

 Although there were fewer experiments carried out using sodium 

 bisulphite than with the other two reagents, it has been placed first 

 not only because it is the cheapest of the three salts but also because 

 other obserA^rs * have found this to be efficient. 



The maximum daily dose of chlorine introduced into the water 

 supply of Washington during the past two years was. according to 

 the waterworks records, 0.32 part of chlorine per million parts of 

 water. Considering sodium bisulphite as the chosen antichlor, about 

 0.5 part of this chemical per million parts of water avouIcI react with 

 all the chlorine introduced. However, it is certain that this quantity 

 is not needed, as a portion of the chlorine is ahvays consumed in 

 oxidizing organic matter before the Avater has been carried A^ery far 

 from the chlorination plant. NcA^ertheless. in order to care for acci- 

 dental surges, it is thought best to use one part of sodium bisulphite 

 in a million parts of Avater. This is still only one-tenth of the quan- 

 tity used in the test experiments. If sodium thiosulphate is used, 

 it is best to use a concentration of two or three parts per million of 



' An oxprpssion of appreciation is due Mr. I.autcr, of tlie Wasliington waterworlis, for 

 lending iiis standards to us for use in tliis worlc. 



' .Tour. Ind. and Eng. Chem., vol. 5. No. 11, November, 1913. 



* " The effect of free chlorine on the gudi?eon." Anonvmous. L'Eau, 16, No. 7, p. 81, 

 1923. Also Chem. Abs. 17, No. 21, p. 3559,' 1923. 



