360 Wiscojisin Academy of Sciences^ Arts, and Letters. 



quantities usually taken for the analysis, and the amounts finally 

 weighed up. Only the most painstaking conduct of a chlorine- 

 deteraiination could keep the errors liable to arise during the 

 process of the analysis within one milligram. Many of our good 

 natural waters, however, carry less than one part of chlorine to 

 the million of water ; take Lake Superior water, e. g., with 1.2 

 parts off-shore, and less than one part some distance out in the 

 lake; or the waters from some of the natural streams from the 

 upper part of this state above the region of the limestones, as^ 

 from the Wisconsin or the Chippewa Rivers, with chlorine gen- 

 erally less than one part to the million, or less than one milli- 

 gram of chlorine to ihe kilogram of water. Such minute quan- 

 tities require the most careful manipulation to insure the re- 

 quired accuracy. 



Water analysts as a rule have adopted Mohr's method, viz. : 

 the silver method, using potassium chromate as the indicator. 

 This method is convenient, is easily applied,^ and has much to 

 recommend its wide use at the present time. The details of the 

 manipulations as ordinarily applied are knowTi to all analysts. 

 There are, however, some errors possible in the application of 

 this method u-'hich can seriously affect the results obtained. The 

 thought that out of an experience with waters from various 

 sources and of different characters. I might make some sug- 

 gestion to those having to do with this class of analytical work 

 and thus assist them to reduce, if not wholly to eliminate the er- 

 rors so aj)t to creep in despite all care, prompted this brief paper. 

 Prominent among such difficulties are : 



1.) The indefiniteness of the statements concerning the prep- 

 aration and proving of the standard solutions employed 

 and the details of manipulation of the processes, even 

 in our best works on the subject; to such works, ally 

 whether teacher or student, must refer, and unless one 

 has had a considerable experience or unusually good 

 laboratory instruction, the description of methods is 

 liable to mislead. 

 2.) The ability of the analyst to recognize colors ; and in par- 

 ticular to discriminate between a yellow and a reddish- 

 yelloiu and thus to define sharply the end-reaction. 



