bottom. The mercury immediately rises in the tube; and by supplying its place in the cylinder and reducing the pressure in A or B, as the case may be, the whole or part of the gas may be transferred to the instrument. Of course in all these cases the india-rubber tubes must be moistened inside with solution of corrosive sublimate.

The gas, having been introduced in one way or the other, is brought all into the eudiometer q, mercury being allowed to run over from A into q until the whole capillary part, down to where it joins the eudiometer at g, is full of mercury. q is a short eudiometer with a scale of only six divisions, like the measuringtube in Frankland's apparatus*. It is immersed in the cylinder r, which is filled with water. The leg p is divided into millimetres, and is lowered or raised until the meniscus of the mercury stands exactly at one of the lines. The difference in height of the two columns is then measured by means of the millimetre-scale on p, and the temperature of the water in r ascertained. These, with the height of the barometer, give the volume of the gas. On the tuber marks are made (not shown in the drawing) corresponding to and on the same level with those on q. When the difference of level of the mercury in the two tubes is to be measured, p is applied to r and the level of the mercury in q read off, the marks on r giving the direction for the eye; that in p is then read off in the ordinary way, the difference giving the column of mercury to be added to or subtracted from the barometer-reading.

Let us suppose that we have a sample of air extracted from water. It consists of oxygen, nitrogen, and carbonic acid, the last of which is determined first by absorption with caustic potash. For this purpose mercury is run out of n by means of y, p is raised, and the stopcocks c, b, and a opened. The air is thus drawn over out of q into m, mercury being allowed to fill the capillary, when the cocks are again shut. The cup e is now filled with strong solution of caustic potash, and, the level in n being still kept low, the stopcock a is opened full, and then d very carefully, caustic potash being allowed to run down through a into m, where it meets the gas in the most advantageous way for quick absorption. When enough caustic potash has been allowed to enter, dis closed, some mercury poured into e, d again opened and the solution in the capillary replaced by mercury. When the absorption is finished, the level of the mercury in n is again raised and the stopcocks b and c opened. The stopcock a is now very carefully opened, the flow of the gas being further regu

The tube q in the figure is shown divided as stated in the text. I have found, however, that it is much more convenient to have it divided into millimetres; and the labour of calibration is no greater.

lated by raising or depressing p, and the gas allowed to pass over into q until the potash solution just touches the lower surface of the stopcock a, which, being open, has the position shown in fig. 3. The position of a is now shifted to that shown in fig. 2, when the potash solution is eliminated. The position of a is now brought back to that represented in fig. 3, and the gas remaining in the capillary swept out by mercury and the volume measured as above described.

If the oxygen is to be determined by absorption, the manipulations are exactly the same as in the case of carbonic acid, alkaline pyrogallic acid being used instead of caustic potash. If it is to be determined eudiometrically, then, after the carbonic acid has been absorbed, the gas remains in the eudiometer, the stopcocks c and b being shut. The stopcock dis now opened and a turned the reverse way to what it is in fig. 2—that is, with the side communication By communicating with the capillary above a. The capillary is then emptied of mercury and the hydrogenevolving apparatus connected with y by means of an india-rubber tube, and the hydrogen allowed to stream through a, d, e until all air is swept out; the stopcock d is then closed and a brought back to its position in fig. 2, when the gas enters m. When enough hydrogen has passed in, a is brought to its position in fig. 3 and the hydrogen-apparatus dispensed with. Mercury is then poured into e, and d opened and the hydrogen in the capillary driven into m. The hydrogen is now passed over into the eudiometer and exploded, the measurements being made as described.

When the analysis is finished the mercury is emptied out of the tubes, the parts A and B bolted each into its own side of the box, the screws x, x, x being lodged in the nuts w, w, w let into the board A. The box when closed measures 0.48 metre by 0.22 metre by 0-22 metre, outside, and is easily portable. Of course the apparatus is not meant exclusively for work on board ship, where indeed attempts at gas-analysis should be avoided, as at the very best the results must be very uncertain; but it makes a compact laboratory apparatus and is economical in mercury.

For the preparation of the electrolytic gases, either separate or mixed, in a way convenient for gas-analysis, I have had the apparatus represented in fig. 4 constructed. It is of the well-known lecture-apparatus form, and consists of two tubes, A and B, united at their lower extremities by the short tube C, which connects them at the same time by means of the tube D with the reservoir E. At their upper extremities A and B terminate in capillary tubes furnished with stopcocks, F, F. Communication with the reservoir can be made or interrupted by the stop

cock G. One of the tubes, A, is furnished with two platinum electrodes; the other, B, has but one. But delivery-tubes (not shown in the drawing) of the usual form fit upon the tubes above the stopcocks F, F. When about to be used, all the stopcocks are opened and diluted sulphuric acid poured in through the reservoir until it has eliminated all air and is running out itself at the delivery-tubes. The stopcock G is now shut and the battery connected, as circumstances may require, either with the two electrodes in A, or with one in A and one in B. Gas is allowed to escape freely until all dissolved air is eliminated. The stopcocks F F are then closed, G opened, and the liquid in A and B allowed to sink until it just covers the electrodes. G is then closed, and F F opened, when the gases may be introduced into the eudiometer in the ordinary way, or into the abovedescribed apparatus in the way there indicated. Connexion between the gas-generator, the tubes A and B, and the reservoir E being shut off by the stopcock G, there is never any difficulty in forcing the gas through the mercury into the eudiometer.

The whole apparatus is attached to a mahogany board, H, which fits into the box K, shown cut through the middle in the drawing, either as represented, when in use, or, when not in use, as a lid, with the apparatus attached to its inner side. The box thus fulfils the double purpose of a convenient stand and a safe packing-case. The only alteration which I should be inclined to make, would be to have the lid to which the apparatus is attached made of vulcanite, to avoid the risk of the wood's warping in damp weather.

II. On the Disintegration of the Electrodes in the Galvanic Arc of Light. By HERMANN HERWIG*.

THE

HE quantities of substance which are thrown off in fine powder from the electrodes of the voltaic arc have hitherto been determined only very uncertainly. To the statements hereupon, made chiefly by Van Breda† and Matteucci‡, a reduction to an otherwise well-known action of the current, perhaps to a simultaneous voltametric action of it, is wanting; and, besides, the experiments were made under circumstances which quantitatively can present only an indistinct picture of the disintegration, as will be seen from the following. Only a series of experiments by Groves on the present subject appears to wear a more precise * Translated from a separate impression, communicated by the Author, from Poggendorff's Annalen, vol. cxlix. pp. 521-533.

† Pogg. Ann. vol. lxx. p. 326.

Comptes Rendus, vol. xxx. p. 201; Ann. de Chim. et de Phys. S. 3. vol. xxxii. p. 350.

§ Phil. Mag. S. 3. vol. xvi. p. 478.

character; and it is just this series which has led to extraordinarily surprising results. Grove placed in the arc of light a positive electrode of zinc over against a negative one of platinum, and in his experiments took into consideration only the disintegration which took place at the former. Now this zinc electrode was of such dimensions, and the intensity of the current so regulated, that the air should not occasion combustion of the massive electrode, but only of the fine particles thrown off from it. The air present was an isolated quantity; and thus the amount of zinc thrown off could be calculated from the oxygen consumed, provided that all the disintegrated zinc, and this only, was oxidized. Grove believed that now every thing had been approximately fitted to the conditions of this experiment, and concluded from his results (which did not differ excessively from one another) that the quantity of zinc disintegrated was equivalent to that of the hydogen occurring in a simultaneously inserted vol

tameter.

The supposition herein expressed, of a surprising connexion between disintegråtion and electrolysis, has not, to my knowledge, been followed up by other investigators. Indeed the difficulty of satisfying, even in some measure, the conditions which Grove sought to realize in his experiment is extraordinarily great; and this circumstance may very well have prevented the repetition of the experiments, as it seems adapted not to permit too much reliance on his data. It must further be said that the idea which Grove pursued in these experiments rested on erroneous assumptions. Namely, Grove thought that, if the electricity of the current could only effect its passage between the electrodes in combination with electrode-substance, then the quantity of electricity corresponding to a definite current-intensity would probably always determine the transport of a definite quantity of substance of the electrodes. But, leaving altogether out of consideration the decided improbability of the law of condition herein contained, the electrode-substance thrown off and that transported between the electrodes are quantitatively quite different. Van Breda's experiments have shown, and it will likewise be found in what follows, that only a portion of the disintegrated substance takes the path to the opposite electrode, while another portion flies about in all directions. Grove's experiments, however, are independent of this point of view; for they would indeed express the quantity, not of the transported, but of the disintegrated zinc.

Independently, therefore, of the reason which induced him to make his experiment, it remains an open question whether there does not exist a simple connexion between the quantity of active. electricity (that is, the intensity of the current) and the quantity

of particles subjected to the action of disintegration, perhaps as signified by Grove's experiment. Only it would be desirable to approach this question less indirectly and under the necessity of observing fewer conditions.

In the first place, it seems requisite to produce the luminous arc in a medium not chemically proportioned, between two chemically different electrodes, and by weighing and quantitative analysis of the electrodes to be clear about the amount of disintegration. The loss of weight of an electrode during an experiment, increased by the quantity of foreign metal taken up by it, would express the amount of substance which it has thrown off. I took some pains to make experiments of this kind in hydrogen gas; but, with the battery I employed (50 Grove's elements of the usual size), only a very imperfect arc of light was obtained in hydrogen, and I did not arrive at any very exact results.

pur

I therefore replaced the hydrogen surrounding of the electrodes by as nearly perfect a vacuum as possible, in which the arc is obtained with decidedly greater facility. The electrodes, screwed upon strong iron rods, were led through two large caoutchouc stoppers, which tightly closed the two ends of a strong cylinder such as is used for lamps. A glass tube with a glass cock was also passed through one of the corks, and, for the pose of exhausting the cylinder, was connected with a Geissler air-pump. The air in the cylinder was thus reduced to 1 or 2 millims. pressure. The electrodes were placed at an average distance of 1 millim. At the commencement I introduced the luminous arc, as proposed by Herschel*, by means of a Leyden jar. Meanwhile it appeared that the operation could be as well performed by taking advantage of the elasticity of the caoutchouc and bringing the electrodes together by a momentary pressure with the hands. They spring asunder again immediately, before a perceptible quantity of hydrogen is developed in the voltameter through which the current simultaneously passes; and thus the bright arc is instantly formed with the fine metallic dust flying in all directions towards the sides of the cylinder. This spirting of metallic dust-particles (which, as said, directly commences) is characteristic for this, that the arc constitutes the only intermedium between the in every other respect separated electrodes, and therefore proceeds in the manner desired.

There is a series of metals which show themselves the least convenient for these experiments; they are the easily fusible ones, from which, with a somewhat continued action of the current, large portions readily melt off and so render difficult the determination of the quantities really separated in dust. It thus happened in experiments with zinc, cadmium, lead, and tin. * Pogg. Ann. vol. xlix. p. 122.