246 



THE ROYAL SOCIETY OF CANADA 



From these results it would appear that the better results are 

 obtained in the more dilute solutions. This may be due to the solvent 

 action of the hydrochloric acid becoming less as it becomes more dilute. 

 In any case, however, the weight of iron is lower than it should be, 

 and the actual facts are probably worse than would appear from these 

 values, because, as shown later, the precipitate probably contained 

 some manganese. Some of the iron must be going through, either in 

 the mother liquor or wash water. 



It seemed necessary before making any more separations to 

 determine whether the low values for the iron were due to the solu- 

 bility of the precipitate in the mother liquor and wash water. With 

 this*point in view, the following series of estimations was carried out. 

 The solution of iron used was about double the strength of the previous 

 solution. From three 25 cc. portions the following weights of ferric 

 oxide were obtained by precipitating the iron as ferric hydroxide 

 with ammonium hydroxide. 



0-1373g 0-1374g 0-1372g 



The iron solution must have contained 11-154 grams of FeCls 

 to the litre. The following estimations were made with 25 cc. portions 

 of this solution. In each case 20 cc. of cone, hydrochloric acid were 

 added and the solution diluted to 150 cc. before adding the precipi-tat- 

 ing reagent. Approximately, 4 grams of cupferron were taken for 

 each precipitation, this portion being made up to a 6% solution, 

 filtered and cooled before using. The precipitate was washed with a 

 constant volume (75 cc.) of water, containing a varying amount of 

 hydrochloric acid, and finally with water (25 cc.) containing a con- 

 stant amount of ammonium hydroxide. 



The results were as follows: 



The solvent action of the acid wash is \ery apparent in the last 

 determination. It seemed necessary now to make a more extended 

 series of determinations, in which both the iron remaining in the 



