Ball No. 2. was put into 186 volumes of the same mixture, which was allowed to remain in the graduated tube for half an hour, when the diminution amounted to 4 volumes. 2.6 volumes of hydrogen were therefore indicated, while only 1.86 were present.

Though there is a very slight error of observation in both these experiments, we may nevertheless safely infer that hydrogen may be detected in oxygen gas, and its quantity ascertained with accuracy by platinum, when it does not exceadth of the whole volume. To detect any unforeseen source of error, I performed the following experiment. No. 2. gently warm was put into 187.5 volumes of atmospheric air, dried by fused potash, and was left there three quarters of an hour; after this interval the air measured 189, which is 187.5+1.5 the space occupied by the ball. I had ascertained, by some preliminary trials, that platinum, though strongly heated, could not cause the combination of oxygen and nitrogen, as electricity does; and this is confirmed by the foregoing experiment. It shows, too, that atmospheric air contains either no hydrogen, or so very small a quantity as not to be rendered visible by the action of platinum.

Different portions of oxygen were added to hydrogen gas, and the electric spark passed through the mixtures. When 9 volumes of hydrogen were mixed with 1 of oxygen, electricity occasioned detonation; but when they were mixed in the proportion of 11 to 1, a strong charge from a Leyden-jar caused neither detonation nor diminution. Platinum, however, caused immediate formation of water, and the quantity of oxygen was indicated by the diminution.

I mixed 3 volumes of oxygen with 300 of hydrogen gas, with which a platinum-ball had been in contact for some hours. The ball No. 2., recently ignited, but quite cold, was put into 112 volumes of this mixture. In half an hour there was a diminution of 4 volumes, which indicates the presence of 1.3 oxygen, while 1.1 was the quantity present. The same ball was put into 154 volumes of the same mixture. In half an hour there was a diminution of 5.5 volumes, which indicates 1.8 oxygen, while 1.5 was the quantity present.

From these experiments, it is apparent that the presence of hydrogen gas in oxygen or atmospheric air, or of oxygen gas in hydrogen, may be detected, and the exact quantity of either indicated by the action of platinum. It is clear, likewise, that analytical chemistry has hitherto possessed no agent of equal delicacy for these gases, and, therefore, that platinum may be used with great advantage in eudiometry.

Το put this conclusion to the test of experiment, I now proceeded to an analysis of atmospheric air. A jar full of the air of my apartment was dried over mercury, and deprived of carbonic acid by fused potash; known quantities of this air, mixed with hydrogen, were exposed to the action of an active platinum-ball.

1st. Exp. The result was 21.8 of Oxygen per cent.

2d Exp.

3d Exp.

22.3

21.7

The two first were performed in the narrow graduated tube, and lasted half an hour; the third was conducted in a capacious jar, and was over in five minutes.

The large quantity of oxygen indicated by these experiments surprised me much, and I was at a loss to conjecture whether this arose from the quantity of oxygen in the atmosphere being actually greater than was generally supposed, or whether there was not some source of error common to all the three experiments. It occurred to me, that the oxygen of the air contained in the hydrogen, as impurity, might perhaps be made to combine with a portion of hydrogen during the experiment, and thus occasion a greater diminution than there really ought to have been. It will be readily inferred from some preceding experiments that this did occur. I was not, however, aware of the circumstance on the present occasion, for this analysis was made prior to the experiments to which I allude.* Acting upon this supposition, I left an active ball during the night in contact with the hydrogen which I intended to use on the following day. I now made six experiments with great care, allowing the action to go on at one time in the graduated tube, at another in a capacious jar. In the latter case the diminution had attained its maximum in five or ten minutes, while half an hour was necessary in the former. The residue was dried by fused potash before being measur ed, and due correction made for change of temperature.

Exp. 1. indicated 203 per cent. of Oxygen.

Exp. 2.

20.3

Exp. 3.

20.7

Exp. 4.

21.0

The mean of these six experiments is 20.88, so that we may

To put this beyond a doubt, a platinum ball was put into a known quantity of hydrogen. It did at first occasion a slight diminution of volume, proportionate to the quantity of air present, but afterwards had no farther action on the gas.

safely assume 21 to be the correct number, which agrees very accurately with our best analyses upon this subject.

The action of platinum affords a neat and expeditious method of ascertaining the purity of hydrogen or oxygen. It is easy, too, to prepare nitrogen of great purity, by adding just sufficient hydrogen to combine with all the oxygen of a known quantity of air, and putting a platinum-ball into the

mixture.

It appeared to me at a very early period of this investigation, that spongy platinum might be expected to produce the composition and decomposition of gases, whenever the electric spark succeeded in doing so. Professor Dæbereiner had himself found, that mixtures of carburetted hydrogen and carbonic oxide with oxygen gas were made to combine by this agent. MM. Dulong et Thenard have observed, that, in a mixture of hydrogen and nitrous gas, spongy platinum occasioned the decomposition of the latter, with formation of water and ammonia; and that it acted likewise on a mixture of hydrogen and nitrous oxide. My friend Mr. Blundell, a most intelligent and promising student of this University, has likewise noticed some interesting facts of the same nature. finds that platinum causes hydrogen to unite with chlorine, and with the elements of euchlorine gas; and has likewise observed other facts of a similar nature. In a few observations, read some weeks ago before the Royal Medical Society, I suggested the probability that the same agent would make iodine and hydrogen combine; and though I have not myself had leisure to examine this subject with care, yet Mr Blundell informs me he has succeeded in forming hydriodic acid in this way.

He

Several of these experiments I have myself repeated, and found perfectly correct. My attention, however, was chiefly directed to the action of platinum on mixtures of oxygen with olefiant gas, with coal gas, and with carbonic oxide. Notwithstanding the ingenious researches of Dr Henry, a method of separating hydrogen, light carburetted hydrogen, and carbonic oxide, from one another, is still a great desideratum in analytical chemistry; and I entertained some hope that platinum might prove useful in this point of view. My attempts to apply it to the analysis of these gases have failed, but as I have examined this subject with considerable care, it may not be uninteresting to relate some of the experiments.

A jet of coal-gas, procured from the city gas-pipes, was thrown upon freshly ignited spongy platinum. At the first

moment a particle of the metal became luminous, but the light disappeared on the instant, nor have I been able, upon any subsequent occasion, to reproduce the same phenomenon, though platinum, freshly reduced, and in a state of great activity, was employed for the purpose. A jet of coal-gas and oxygen from separate vessels was thrown upon spongy platinum, without the least luminous appearance whatever; but if the metal be first strongly heated before the blow-pipe, and a jet of coal-gas be thrown upon it immediately after all light has disappeared, the platinum quickly becomes red, and will continue so for any length of time. A vivid light is emitted, but the gas itself is not enflamed. This result would, a priori, be expected, because the heat necessary to inflame an explosive mixture of coal-gas and atmospherical air, is greater than the white beat of solid bodies.

Coal-gas was mixed with rather more than twice its volume of oxygen. I heated a piece of spongy platinum to vivid redness; put it quickly on a little basin of platinum-foil, floating on the mercury, and covered it with a jár full of the explosive mixture. Watery vapour condensed on the surface of the mercury, the platinum became red, and rapid diminution followed: the redness, however, speedily disappeared, and the progress of the operation was arrested long before. either gas was consumed. The platinum was left two hours after this in contact with the residue, but did not cause any marked reduction of its volume. The carbonic acid was absorbed by fused potash, and the remainder detonated strongly with the electric spark. This experiment was several times repeated with a similar result. It is curious that the process was in each case so quickly interrupted, for it is obvious that heat enough would be generated for its continuation. The cause is to be sought in the atmosphere of carbonic acid gas, which, as it forms, collects around the platinum, and thus prevents an adequate supply of the explosive mixture from reaching the metal, till its temperature has fallen below the point at which the combination can go on. I heated the balls No. 1. and 2. in a similar manner, and covered them with a jar full of the explosive mixture of coal-gas and oxygen. Rapid action ensued, without any visible redness; the proeess ceased when much of the mixture remained, and after the carbonic acid had been absorbed by potash, a spark of electricity exploded the residue.

A mixture of coal-gas and oxygen was divided into three portions. Pure spongy platinum was put into one portion,

and the balls No. 1. and 2. into the others, all being cold; no immediate action ensued, and the diminution after twelve hours was very slight; traces of carbonic acid were discoverable by lime-water, which shows that some chemical change had been occasioned. The experiment was repeated, with this difference only, that the mixtures were heated to near the boiling point of mercury. The diminution was now greater than before, and more carbonic acid had formed, but the residue was still explosive.

Coal-gas and hydrogen were mixed in various proportions, and to each mixture sufficient oxygen was added for complete combustion.

When the coal-gas was to the hydrogen at 3: 1, No. 2. had no immediate action, at common temperatures; but if ignited before the blowpipe, and then plunged rapidly into the mixture, just after it had ceased to be red, an immediate diminution of volume followed, and the ball became luminous.

When the coal-gas and hydrogen were as 2:1, the ball No. 2. acted precisely as in the preceding experiment. The same occurred when the gases were as 3:2. When they were in the proportion of 2: 3, and the ball was heated so as just to be borne on the hand, a rapid diminution of volume succeeded, without emission of light. The ball likewise acted, though with far less energy, when cold.

When the coal-gas and hydrogen were as 1 : 2, the ball acted nearly in the same manner as before, though with more energy.

In all these experiments there was copious production of carbonic acid, but the residue still contained an explosive mixture. The coal-gas, as obtained from the pipes, always contains some carbonic acid, which, of course, was previously removed by potash.

Olefiant gas, carefully prepared and well dried, was mixed with three times its volume of oxygen. Spongy platinum, as well as the balls, had hardly any action upon this mixture, when cold or gently warm. Heated to near the boiling point of mercury, a partial action succeeded, with production of carbonic acid. A platinum ball was heated to vivid redness before the blowpipe, and then introduced quickly into the mixture; it acted with energy, for there was on the instant a copious production of water and carbonic acid, but still the residue contained some explosive mixture.

Olefiant

gas and hydrogen were mixed in various propor