780 
MR. BRODIE ON THE CONDITION OF CERTAIN ELEMENTS 
experiment, nor is any variation of the action to be feared. I consider therefore this 
method to be as accurate as can be desired. 
Experiments with Metallic Oxides. 
The method of weighing the loss of oxygen, which succeeds very well for deter- 
mining the amount of oxygen given off in the preceding experiments, cannot be em- 
ployed with advantage for determining the amount given olf by a metallic oxide in 
contact with the peroxide of barium. The evolution of gas is not rapid, continuing, 
at ordinary temperatures, sometimes for several days before all action ceases, during 
which time the apparatus is liable to alter in weight ; and also more water is required 
to dissolve the baryta than can be introduced into the bulb of the apparatus, or con- 
veniently and accurately be weighed. In these experiments, therefore, I determined 
indirectly the amount of oxygen given off by ascertaining the amount of metal 
reduced. 
The form' of the experiment was very simple. The weighed peroxide was intimately 
mixed in a flask with the metallic oxide, and a certain portion of water, about twelve 
ounces, poured upon the mixture. In the following experiments, those at the tempe- 
rature of the air were placed in a cellar where the temperature was uniform, so that 
they might be comparable with one another, and the action was allowed to continue 
until all evolution of gas had ceased. In the experiments at 100° and at other tem- 
peratures, the water was first brought to the temperature required, then poured on 
the mixed substances and the mixture kept at the same temperature, either boiling or 
in a water-bath, during the action. In the experiments with oxide of silver, to deter- 
mine the silver reduced, the residue, after the evolution of the gas had ceased, was 
treated with very dilute hydrochloric acid, with which it was allowed to remain in 
contact at least twelve hours. All oxide was thus converted into chloride. The fluid 
was now filtered from the residue and the hydrochloric acid washed out, the bulk of the 
solid substance being still kept in the flask in which the experiment was made. From 
this residue, consisting of chloride of silver and metallic silver, the silver was extracted 
by boiling with nitric acid, the substance on the filter being added to the rest and 
the filter also treated with acid. The solution was diluted and the silver dissolved, 
which was in fact the silver reduced, estimated as chloride. It is plain, that whether 
the oxide were reduced to metal or to a lower degree of oxidation, this silver must 
still be a correct measure of the oxygen lost. This method of determining the loss of 
oxygen I followed in all the experiments with silver compounds. 
My first experiments were made before I had ascertained all the points necessary 
for the careful preparation of the peroxide, and the preparations used for these con- 
tained considerably less per-centage of oxygen than that which I afterwards readily 
procured, and contained also impurities from which the other preparations were 
free ; in these experiments also the oxygen was determined only with bichromate of 
potash and hydrochloric acid. These circumstances in no way affect their general 
