222 Froceedrngs of the Royal Society of Victoria. 



(iii.) Estimation (volumetrically) of the amount of 

 manganese peroxide precipitated. 



The first was not made use of, as the volumetric method, 

 besides being quicker, would give quite as accurate results with 

 the dilute solutions employed. 



The method of conducting the experiments was as follows : — 



N N 



Generally 50ccs. -- KMn04 and 5-30ccs. — — sugar were used, 

 ■^ 50 500 ® 



N 

 and the desired amount of ,^ H.SO^ added to the same flask. 



The flask was now heated for the time required at some 

 particular temperature and then rapidly cooled. 



The solution could now be filtered through asbestos and the 

 " active oxygen " in either the filtrate or the precipitate or both, 

 estimated by titration with a standard solution of ferrous ammo- 

 nium sulphate. 



In some experiments the solution was not filtered and the 

 total " active oxygen " remaining in the liquid determined in the 

 same way. 



The ferrous solution mentioned contained about two equi- 

 valents of free sulphuric acid to each equivalent of the salt, in 

 order to render it capable of readily dissolving manganese oxides, 

 and is subsequently denoted by [Fe]. 



It was soon evident that, to obtain accuracy, attention must 

 be paid to 



1 Purity and exact standardization of reagents. 



2 Careful regulation of the duration and temperature 



of experiments. 



1. The reagents employed were 



(i.) Cane Sugar : — 



Ordinary white " brewer's ci'ystals " were used 

 as being almost chemically pure. 



At first trouble was experienced with an 

 organic growth in the dilute solutions used ; 

 this was effectually prevented by previously 

 rinsing out the well-stoppered bottle with 

 mercuric chloride. 



