114 Mr T. H. Easterfield, Some observations on [Oct. 31, 
greatly with the temperature and the proportions of the ingre- 
dients employed in the retort. Rapid removal from the retort by 
a current of dry air appears to favour the production of the gas. 
Below 50° the gas does not appear to be given off, or at any rate 
only in minute quantities. Above 70° the gas obtained is nearly 
pure oxygen with very little of the manganese compound. If the 
volume of pure sulphuric acid employed is only about twice that 
of the permanganate, a violent explosion is almost certain to occur 
when the temperature reaches 60°. If still less acid be employed 
the risk of explosion appears to be increased. By increasing con- 
siderably the proportion of acid all risk of explosion may be 
obviated. The best yield of gas was obtained in an experiment 
where the proportions were so adjusted that a very shght explosion 
took place when the temperature reached 65° by allowing the 
temperature to fall to 60° and then passing air through the retort 
as if no explosion had occurred. The exact proportions of acid 
and permanganate had unfortunately not been previously deter- 
mined. 
When the spent liquor from the retort was dropped into water 
a copious evolution of the pink vapours occurred. This method 
was tried as a means of obtaining crystalline permanganic acid, 
and gave satisfactory results. The following experiments were 
subsequently tried with the gas :— 
The vapour density bulb was removed and a large flask, well 
dried by the passage of dry air through it, substituted. The flask 
had a well fitting cork, through which passed three tubes. Two 
of these reached nearly to the bottom of the flask, one for the 
delivery of the manganese gas, the other for that of any other gas, 
whilst the third served as an exhaust pipe. ‘he flask was now filled 
with air which had passed through the generating retort, and 
which carried with it the manganese gas as usual. The flask 
remained colourless as when full of dry air. Air saturated with 
aqueous vapour by passing through three wash-bottles containing 
warm water was now sucked into the flask. A change in colour 
at once became evident. An opaque brown film, insoluble in 
water, gradually covered the sides of the flask, except immediately 
beneath the delivery tube, where a purplish deposit accumulated ; 
this was partially soluble in water, with formation of a pink solu- 
tion. It has already been stated that a trace of water seems to 
decompose the gas, whilst an excess of water dissolves it, which 
probably accounts for the brown deposit. 
A second flask, similar to the first, was now filled in the same 
way with air which had passed through the retort; then dry 
ammonia gas was passed in. At first no action seemed to occur, 
but in a few minutes a brown deposit formed on the sides similar 
