No definite conclusions can be drawn from these neutralization experiments since it is well known that manganese oxides prepared as were those used in these experiments always contain some, though probably not a constant quantity of alkali. It is stated by Bemmelen[9] that manganese dioxide decomposed the salts of the alkalies with liberation of acid. [9] J. für Prak. Ch. (2) 23, 324-379. The Stability of Manganese Dioxide. Having found in the course of the work already described that manganese dioxide prepared in the wet way is much less stable than was supposed, it was decided to make some experiments upon the spontaneous decomposition which it undergoes. To this end a fresh sample was prepared in the following manner. 10 grams of potassium permanganate were dissolved in 500 c.c. of distilled water and the solution allowed to settle for one day. The liquid was then filtered through glass wool, heated to 65°C. and treated with 320 c.c. N/10 nitric acid, having also a temperature of 65°C. 20.5 grams of manganous sulphate dissolved in 2.5 litres of water heated to 65°C, were now added, with stirring, to the acidified solution of potassium permanganate. The precipitate was allowed to settle and the supernatant liquid which still retained the permanganate color, decanted. The residue was then treated with water, stirred and allowed to settle. The water was decanted and the oxide filtered. The filter used consisted of two platinum cones between which was placed a small paper filter which did not quite reach to the edge of the outer cone. This filter was placed in the bottom of a glass funnel and the oxide poured upon it. The oxide was repeatedly washed with distilled water and then transferred to a porous plate. It was afterwards heated for several hours to a temperature of 65°C. At this temperature it was found impracticable to bring the oxide to a constant weight, but this fact was no serious obstacle in the way of the subsequent work since it was only desired to ascertain what changes take place in the ratio of the manganese to the available oxygen. The manganese was determined by the method of Gibbs and the available oxygen by oxalic acid and potassium permanganate. An analysis of the oxide made shortly after its preparation in the manner described, gave {51.98 percent} Manganese { } = 51.86 pr. ct. {51.74 ” } {14.836 percent} Oxygen { } = 14.837 ” {14.839 ” } showing a ratio of manganese to available oxygen of 1.018 : 1.000. The above analysis was made May 1, 1892. The oxide was then placed in a glass-stoppered weighing bottle and allowed to stand until November. An analysis made on the tenth day of the latter month gave a ratio of manganese to available oxygen of 1.198 : 1.000 corresponding very closely to a composition MnO·5MnO₂, in which the calculated ratio is 1.200 : 1.000