214 Qoocli and Austin — Estimation of Manganese, etc. 



In the next step, this oxide was further oxidized by moisten- 

 ing it with nitric acid and heating the residue gently until the 

 evolution of fumes ceased, the containing crucible being placed 

 well above a porcelain crucible used as a radiator and heated so 

 that only the bottom showed a faint red heat. In this process 

 the attempt was made to arrest the ignition at the point where 

 the anhydrous dioxide was produced. As the table shows, and 

 as would be expected, this attempt was only occasionally and 

 partly successful. 



The residue of the last process was then submitted to a 

 higher heat. The platinum crucible containing the oxide was 

 placed within and touching the bottom of a larger porcelain 

 crucible which was heated to redness. Under these conditions 

 the temperature should not be too hot, and the products of 

 combustion should naturally be thrown so far away from the 

 oxide undergoing ignition that circumstances should be favor- 

 able for the formation of the oxide Mn 2 3 . The event proved 

 that the attainment of the exact condition corresponding to the 

 symbol Mn 2 3 is a matter of some uncertainty. 



Next, the oxide was subjected to the highest heat of a strong 

 Bunsen burner (or in some cases, the broad flame of a blast 

 lamp), the crucible being well surrounded by the products of 

 combustion. The results of this treatment, it will be seen, 

 agree, with a single exception out of ten experiments, reason- 

 ably well with the theory for Mn 3 4 . 



By treating the final oxide with nitric acid and repeating the 

 cycle of operations described, the observations of the phe- 

 nomena were multiplied, until finally, the oxide formed last 

 was treated with sulphuric acid, ignited in the manner previ- 

 ously detailed, and weighed as the anhydrous sulphate, thus 

 showing that no significant loss of material had taken place in 

 the series of manipulations. The table comprises the results of 

 these experiments. The Roman numerals indicate the order of 

 treatment. 



The inference is plain that the estimation of the manganese 

 in the form of the manganoso-manganic oxide, Mn 3 4 , is by no 

 means to be considered utterly untrustworthy when the process 

 is conducted in the manner described, though it must be recog- 

 nized that an irregular result may occur occasionally. The 

 danger of accepting such an irregularity as a correct indication 

 may be eliminated to a very considerable extent if the precau- 

 tion is taken invariably to moisten the ignited oxide with nitric 

 acid, and ignite again. The indications of harmonious results 

 thus got may be taken with a fair degree of confidence. How- 

 ever, it is, in our judgment, by far the wiser and simpler plan 

 to convert an oxide of manganese obtained in course of analysis 

 into the sulphate and to weigh the manganese in that form. 



