ON BOILING W ATE It. 109 
phuretted hydrogen was given off, amounting in several experiments to more than a 
cubic inch. 
A similar experiment was made with melted sulphur, and sulphuretted hydrogen was 
given off, but not in such quantities as the phosphuretted hydrogen. I tried in vain to 
carry on these experiments beyond a certain point; the substance became pasty, mixed 
with platinum from the arc, and from the difficulty of working with the same freedom 
as when they were fresh, the glass tubes were always broken after a certain time. Had 
I time for working on the subject now, I should use the discharge from the Ruhmkorf 
coil, which had not been invented at the period of these experiments. At a subsequent 
period, when this discharge was taken in the vacuous receiver of an air-pump from a 
metallic point to a metallic capsule containing phosphorus, a considerable yellow deposit 
lined the receiver, which, on testing, turned out to be allotropic phosphorus. No gas is, 
however, given off. I had an air-pump (described Phil. Trans., 1852, p. 101) which 
enabled me to detect very small quantities of gas, but I could get none. It was in ma¬ 
king these experiments that I first detected the striae in the electric discharge, which have 
since become a subject of such interesting observations, which are seen, perhaps, more 
beautifully in this phosphorus vapour than in any other medium, and which cease, or 
become very feeble, where the allotropic phosphorus is not produced. 
I tried also phosphorus highly heated by a burning-glass in an atmosphere of nitro¬ 
gen, but could eliminate no perceptible quantity of gas, though the phosphorus was 
changed into the allotropic form. 
It is not difficult to understand why gas is not perceptibly eliminated in the last two 
experiments; the effect is probably similar to that described in my paper on the “De¬ 
composition of Water by Heat,” where, when the arc or electric spark is taken in aqueous 
vapour, a minute bubble of oxyhydrogen gas is freed and disseminated through the va¬ 
pour, recombination being probably prevented by this dilution ; but, however long the 
experiment may be continued, no increased quantity of the gas is obtained, all beyond this 
minute quantity being recombined. If, however, the bubble of gas be collected, by al¬ 
lowing the vapour to cool, and then expelled, a fresh portion is decomposed, and so on. 
So with the phosphorus in the experiments in the air-pump and with the burning- 
glass; if any gas is liberated it is probably immediately recombined with the phosphorus ; 
possibly a minute residuum might escape recombination, but the circumstances of the ex¬ 
periment did not admit of this being collected, as the gas was with the aqueous vapour. 
When, on the other hand, the gas freed is immediately cut off from the source of heat, 
as when the spark is taken in liquids, an indefinite quantity can be obtained. 
Decomposition and the elimination of gas may thus take place by the application of in¬ 
tense heat to a point in a liquid, or also in gas or vapours ; but, in the latter case, it is more 
likely to be masked by the quantity of gas or vapour through which it is disseminated. 
I believe there are very few gases in which some alteration does not take place by the 
application of the intense heat of the voltaic arc or electric spark. If the arc be taken 
between platinum points in dry oxygen gas over mercury, the gas diminishes indefinitely, 
until the mercury rises, and by reaching the point where the arc takes place, puts an end 
to the experiment. I have caused as much as a cubic inch of oxygen to disappear by 
this means. I at one time thought this was due to the oxidation of the platinum ; but 
the high heat renders this improbable, and the deposit fonned on the interior of the 
glass tube in which the experiment is made has all the properties of platinum-black; so 
if the spark from a Ruhmkorf coil be taken in the vapour of water for several days, a 
portion of gas is frped which is pure hydrogen, the oxygen freed being probably changed 
into ozone, and dissolved by the water in this case, while in the former it combined with 
the mercury. 
I have alluded to the eudiometer by which I analysed the gases obtained in these ex¬ 
periments ; it was formed simply of a tube of glass, frequently not above 2^ millimetres 
in diameter, with a loop of wire hermetically sealed into one end, the other having an 
open bell-mouth. By a platinum wire a small bubble of the gas to be examined could 
be got up through Avater or mercury into the closed end of the tube, and by the addition 
of a bubble of oxygen or hydrogen gas, a very accurate analysis of very minute 
quantities of gas could be made: I have analysed by this means quantities no larger 
than a partridge-shot. 
I need hardly allude to results on the compound liquids, such as oils and hydrocarbons, 
as the fact that permanent gas is given off in boiling such liquids Avould not be unex- 
VOL. VI. K 
