BROMINE IN SEA WATER. 85 



Then I ascertain the bromine in a given volume either by 

 argentic nitrate and chlorine gas or by titration with the 

 standard solution of hypochlorite of sodium, which has 

 been previously assayed for chlorine by one of the usual 

 methods. Deducting the volume used for assny I dilute 

 until the normal solution contains one milligram per cubic 

 centimeter. 



Although the standard solution of hypochlorite of sodium 

 is by no means unalterable, still it is much less changeable 

 than chlorine water. For instance, a solution prepared 

 three weeks ago and indicating 100 milligrams of bromine by 

 18.8 c.c, requires to-day 19.6 c.c, weakening by 4 per cent. 

 This weakening is, however, no matter, as it amounts to only 

 .01 c.c. per day for 25 milligrm. of bromine, and I recom- 

 mend, in all cases, to test its strength immediately previous 

 to the assay on the unalterable normal solution of bromine. 

 This is preferable to its test by an iodide of potassium, etc., 

 solution, which Fresenius [Quant. Anal., § 143, fi) recom- 

 mends, deviating in this one little instance from the funda- 

 mental principle of volumetric analysis expressed by him 

 elsewhere, that the test ought to be made, if possible, al- 

 ways in the same manner as the assay. 



The novelty of my method is a kind of fractional titration 

 which allows to foretell the final reaction on an aliquot por- 

 tion of the liquid, and — in the case of bromine — to compare 

 the color before and after addition of the chlorinated standard 

 solution. This kind of titration is advantageous also for 

 other determinations, especially those which have a very 

 sudden final reaction, as, for instance, the analysis of caustic 

 soda. For the last named kind of analysis I proceed as 

 follows: The apparatus consists of a flask provided with a 

 well-fitting cork with two holes. Through one of these 

 holes passes a little bent glass tube, stopping short at ihe 

 cork inside and connecting outside with a rubber tube and 

 clamp or "clip." Through the other hole passes a long 

 funnel to nearly the bottom of the flask. I prefer the kind 



