ble. The sole use of these stopcocks was, to prevent the admission of any considerable quantity of air during the preparation of the tube for filling it again with hydrogen.

To the tube of infusible glass a, a, therefore, was fitted by a ground joint the stopcock c, ending in another ground joint. When the tube was to be filled with hydrogen, a connecting tube with a ground joint was fitted to the second joint, d, with paraffin, and connected by fusion to the source of hydrogen. The connecting tubes were then freed from air by repeated exhaustion. The stopcock e was then opened, the palladium was heated, and the exhaustion continued. After cooling the palladium, hydrogen was admitted to saturation. In these two experiments,

any nitrogen which might possibly have accompanied the hydrogen was removed, or, rather, an attempt was made to remove it, FIG. 34.-Palladium tube protected against leakage by repeated exhaustions of the tube at the end of the admission of hydrogen.

through stopcock.

When the tube was charged with hydrogen, the pressure in the tube was a little greater than that of the atmosphere. No leakage was therefore to be feared during the moment required to remove the ground joint which connected the, palladium tube with the source of hydrogen, its stopcock being closed, and to replace it with the joint f, 7, Fig. 34, which had been prepared beforehand. The tube g was then connected with the air pump, and the space between the stopcock and the plug of fusible metal was repeatedly exhausted and filled with hydrogen. When it was exhausted for the last time, the connection with the pump was closed by fusion at g.

The space between the fusible metal plug 7, and the point 7, had been exhausted before the point / was sealed. The apparatus, therefore, contained no gas but hydrogen.

The volume of the two tubes was now determined by a hydrostatic weighing of both at once. A counterpoise was made of the same two kinds of glass, and of nearly the same volume and weight.

When the hard glass tubes containing the palladium were used, they were heated so gently that it was obvious that

no change of volume was to be feared; it was therefore unnecessary to determine their volume after the expulsion of the gas by a second hydrostatic weighing.

a

m

a

FIG. 35.-Palladium tube, second form.

Procuring some tubes large enough to contain the whole of the palladium, it was concluded to use them, although they were of soft glass.

This tube took the form shown in Fig. 35. It was now not safe to assume that there would be no change of volume on heating to expel hydrogen. Accordingly, a mark was etched at e. After determining the volume of the whole tube by hydrostatic weighing, the volume below this mark was also determined. After the expulsion of the gas, any change of volume was determined by immer sion to this mark. The stopcock was therefore not wet during the progress of the experiment. The effect of the immersion of the stopcock during the preliminary work of reconstructing the tube before the charging with hydrogen was eliminated by waiting a sufficient time before the weighing.

One important advantage was gained by the use of soft glass. It was now possible, when the absorption was ended, to continue the current of hydrogen, by breaking off the point m of the tube, so as to expel any nitrogen which might have accompanied the hydrogen and accumulated in the tube. After this was accomplished by a current lasting three or four hours, the point was closed by fusion, and the tube prepared for weighing in the same way as before.

But a certain difficulty remained. The stopcock had to be immersed for at least one hydrostatic weighing for each experiment, and this involved waiting a day or two in order to be sure that the lubrication had come again to a constant A third apparatus was therefore constructed, in which no stopcock was used. This is the form always used afterwards. It has been shown in Figs. 25, 26, and 27, and its manipulation has been sufficiently described.

state.

5. SYNTHESIS OF WATER.

gases

APPARATUS FOR PRODUCING AND WEIGHING WATER.

The apparatus in which the gases were combined and in which the water produced was weighed is shown in Fig. 36. At a are two platinum jets from which the issue into the combustion chamber. Two tubes filled with phosphorus pentoxide, kept in place by rolls of asbestos, are shown at b, b. At c, c, are joints ground to fit corresponding joints by which the apparatus may be connected to the sources of hydrogen and of oxygen. These tubes are closed by fusion at d and e, and notches are made where they are afterwards to be broken off. Wires ƒ, f, f, provide for the inflammation of the jet of hydrogen, or of oxygen, as the case might be. The ends of the apparatus have hooks facilitating the hydrostatic weighing which determines its volume. The whole apparatus was exhausted to a ten-thousandth of an atmosphere and sealed off from the pump at g.

A counterpoise of suitable weight and volume was then made. Since the apparatus had to be reconstructed for each experiment by providing it with new drying tubes, a new counterpoise had also to be made. To facilitate this, a series of volumes of 50, 20, 20, 10, 5, 2, 2, and 1 cubic centimetres were prepared and the vol

ume of each determined. Then a series of tubes having all nearly the same volume, and weights of 55, 60, 65, 70, 75, and 80, grammes were made, and their volumes determined. By the aid of these a counterpoise of a given volume and nearly a given weight could be made up in a moment.

The combustion apparatus was then weighed by reversal, as in the case of the palladium and of the globes containing oxygen. But few weighings were required for a small and light object like this.

When the weights of the globes, of the tube containing palladium, and of the combustion apparatus, had been determined, the apparatus was put together for a synthesis. First, the combustion apparatus had one of the points d, Fig. 36, broken

off while it was enclosed in a tube which delivered dry air. The ground joints which fitted c, c, formed the ends of two drying tubes filled with phosphorus pentoxide. One of these tubes was now put on c. The other point e was broken off and the other drying tube also put in place; both joints were of course made tight with paraffin. The object of these tubes was, to prevent the admission to the weighed tube of any moisture produced in the glass-blowing operations required in putting the apparatus together. It is now seen that the manipulation might have been made simpler; but not more accurate.

The combustion apparatus with the addition of the phosphorus pentoxide tubes was now placed in the condenser h, h, Fig. 37. This consisted of a cylinder of glass, having a brass bottom with two tubulatures; into one of them the combustion apparatus was placed, and made tight with a rubber tube.

7. SYNTHESIS OF WATER.

MANIPULATOR FOR REGULATING THE PROGRESS OF THE

COMBUSTION.

The condenser with its contents is next placed in the manipulator shown in Fig. 37. At and are the tubes by which oxygen and hydrogen are to be introduced; 7 leads to the Toepler air pump. The stopcocks m and n serve to control the admission of the gases into the combustion chamber. The tubes o and P which are connected directly to the tubes i and k, show the pressure of the oxygen and hydrogen at their sources. The tube g, interposed between the stopcock m and the combustion chamber, shows the pressure of the gas therein. The ends and s are now fused to the somewhat flexible connections t and u. As said before, the drying tubes a and y prevent the entrance into the weighed drying tubes of the

water formed in this process.

The globes containing oxygen are now fitted into the corresponding ground joints, and these are fused to the branches leading from the tube i, as shown in Fig. 38. The point of the palladium tube is broken off and a joint is fitted on and connected by fusion to the tube k.

The five stopcocks m, n, 1, 2, 3, being open, the apparatus is exhausted to one part in ten thousand, and the pump is shut off by closing the stopcock 2. The stopcock 3 is opened, when mercury fills the tubes 4, 5, closing all connection between the hydrogen and the oxygen sides of the apparatus. The stopcocks 1 and 3 are now closed, preventing disturbance of the level of the mercury in these tubes. The stopcocks m and n are next closed, and the stopcock of one of the globes of oxygen is opened, when the gauge o shows the pressure of the oxygen in it; the other globe is also opened. The fusible metal plug is then fused, when the gauge p shows the pressure of the hydrogen from the palladium.

The stopcock m is opened so as to admit say half an atmosphere of oxygen to the apparatus, after which it is closed. Sparks are passed between the points f, f, and stopcock n is cautiously opened. Hydrogen enters and mixes with the oxygen which already fills the apparatus, and the mixture soon comes to the jets; it takes fire, and for a moment burns within the asbestos at 6. Soon the gas burns quietly at one of the jets; heat is applied to the palladium, and the pressure of the gas from the palladium is watched by means of the gauge. The stopcocks m and are regulated by the indications of the other gauges and the appearance of the flame. The combustion chamber is kept cool by renewing the water surrounding it.

n

TO DIMP

nomenon.

FIG. 38.-Manipulator, rear view, with globes and palladium tube connected.

The combustion of pure hydrogen in pure oxygen is a very interesting pheSince it was convenient to have a visible indication by which to regulate the admission of gas, I commonly adjusted so as to produce vivid incandescence of one or the other of the platinum jets. In this case, nothing else could be seen unless for a moment the heat was sufficient to volatilize something from one of the wires used to ignite the gases. But often, when the current was made more rapid, the incandescence of the jet ceased, and then there was a small pointed flame, invisible except on careful exclusion of other light. But the risk and consequent anxiety attending the appearance of this flame prevented minute and trustworthy observation. The risk was due to the fact that if the flame were