Iodometric Determination of Nitrates. 



121 



the nitrate there can be no reasonable doubt in view of the 

 proved behaviour of the manganese salt toward small amounts 

 of nitrates. To us it seemed probable that the explanation of 

 these results was to be sought in the failure of the titration 

 in alkaline solution to indicate completely all the final products 

 of the action of the nitrate. The formation, even in small 

 amounts, of compounds of nitric oxide with iodine analogous to 

 those which we know to be formed with bromine and chlorine, 

 or the partial reoxidation of the nitric oxides by the action of 

 iodine with the aid of water, an action which we recognize as 

 possible under certain conditions of dilution, would account 

 satisfactorily for the deficiency in the results of titration ef- 

 fected in alkaline solution. Upon this presumption the simple 

 and obvious modification of titrating in acid solution should cor- 

 rect the error. Accordingly in the following series of experi- 

 ments the plan of collecting the halogen and titrating in alkaline 

 solution was abandoned, and since the addition of an iodide to 

 the retort was no longer essential this practice was discon- 

 tinued. The products of the action of the nitrate upon the 

 manganese mixture — chlorine, nitric oxide and perhaps nitrosyl 

 chloride, — were received directly in potassium iodide, and the 

 iodine set free was titrated by sodium thiosulphate, itself stan- 

 dardized against iodine of known value with respect to a stan- 

 dard solution of decinormal arsenious oxide. 



With the abandonment of the plan of putting the alkaline 

 arsenite into the receiver the tendency of the iodine to pass 

 onward to the second receiver is augmented and the possible 

 action of the rubber stopper of the receiver becomes corre- 



