82 CALIFORNIA ACADEMY OF SCIENCES. 



witli the liquid. It is advisable to loosely cork the flask. 

 It will be found that the precipitate of mixed cuprous chlo- 

 ride and bromide is an unusually heavy one, settling rapidly 

 from a perfectly clear liquid, from which any little particles 

 floating on the surface can be made easily to sink (by a 

 glass rod or judicious shaking), so that the liquid may be 

 decanted to almost the last drop after short settling. 

 The solution has become strongly acid by the sulphuric 

 and sulphurous acid formed through the reaction between 

 cupric chloride and sodium sulphite. For the 250 c.c. of 

 sea-water employed about 3 grm. of copper have been sepa- 

 rated, varying in amount somewhat according to the tem- 

 perature after heating and after cooling. Heating to the 

 boiling point only increases unnecessarily the copper sepa- 

 rated. The second addition of sodium sulphite I found 

 necessary for removing the last trace of bromine. Now, in 

 order to test if all bromine has been separated, I add to the 

 decanted liquid another portion of sodium sulphite, say 

 as much as the second time, heat and cool as before, treat 

 the precipitate by zinc and hydrochloric acid, as will be 

 described hereafter, filter some of the solution of zinc chlo- 

 ride obtained into a test tube and test by one drop of the 

 standard chlorine sohition and chloroform, whether after 

 shaking, the chloroform turns yellow. If it does not it is 

 certain there is no trace left in the decanted liquid. 



By these operations bromine can be very conveniently 

 and completely separated in about ten minutes. Enough 

 cupric sulphate ought to be added to form cupric chloride 

 also from the sodium sulphite, as this is partially converted 

 into sodium chloride by the formation of cuprous salts. In 

 analyzing saline solutions other than sea-water, I ascertain 

 the specific gravity and the corresponding percentage in 

 sodium chloride from the tables, and reckon three parts of 

 crystallized cupric sulphate for one part of sodium chloride, 

 assuming all dissolved salts to be sodium chloride, as an 

 excess of copper does not matter. 



1 



