ordinary precaution being used to obtain accurate results, but which space will not permit us to detail.

Thus, in experiment 1, on a wire drawn from welded palladium, diameter 0-462 millimetre, and specific gravity 12.38; its normal length was then carefully measured, and then exposed by the negative electrode of a two-cell Bunsen battery, decomposing water for one hour and a half, a longer exposure being unnecessary. Then the wire was again measured, and the increase in length noted. The wire was bent, and introduced into a glass tube connected with the Sprengel's mercurial exhauster, and the wire distilled in vacuo. The hydrogen collected was carefully measured, and, for the sake of comparison, this and the following results were reduced by calculation to a barometric pressure of 760 millimetres, and temperature of 0° C. Thus the original length of the wire 609.144 m.m. (23.982 inches), and its weight 1.6832 grs. Distillation in vacuo proved that the wire contained 936 times its volume, or 128 cc. This charged wire measured 618-923 m.m., showing an increase in length of 9.779 m.m. (0-385 inch), or a linear expansion of 1·605 per cent. Assuming the expansion to be equal in all directions, this gives a cubic expansion from 100 to 104-908, therefore, supposing the two metals united, without any change of volume, the alloy may be said to be composed of

The density of the wire is reduced by calculation from 12.3 to 11:79, and as (the original volume) 100 is to (expanded volume) 4.91, so the volume of the palladium, 0·135 c.c., is to the volume of the hydrogenium, 0·006714 c.c., finally dividing the weight of the hydrogenium (140-01147 grammes) by its volume in the alloy, 0.006714 c.c., we find density of hydrogenium=1.708, or nearly the specific gravity of magnesium.

Then Mr. Graham describes one of the most singular and almost inexplicable facts developed by the research, after the expulsion of the gas by distillation in vacuo, the wire does not return to its normal length, but is actually contracted to about the same extent (on the minus side of zero) as it had previously been expanded by the presence of the hydrogen. Thus, the wire first measuring

609.144 m.m., increased with hydrogen 9.77 m.m., was ultimately reduced to 599.444 m.m., a contraction of 9.7 m.m.

That this retraction is in one direction only is proved by the fact, that the specific is slightly lowered-from 12:38 to 12:12. The result is the converse of the extension by wire-drawing.

It was found that the hydrogen might be expelled by simply exposing it as the positive electrode of a battery, thereby evolving oxygen on its surface, which united with the previously occluded hydrogen.

By alternate exposure to each pile, a wire, having an initial length of 609.14 m.m., was, after four days, actually retracted 24 millimetres.

An extremely beautiful experiment may be performed by simply heating one end of an hydrogenised palladium wire to redness, when the hydrogen burns along the wire, which is previously rubbed with magnesia to render the flame luminous.

Without describing the experiments in detail, it may be sufficient to give the general results :

The first and last experiments are probably exceptional, and it may be assumed that the density of hydrogenium is as nearly as possible twice that of water. The second portion of the investigation treats of the tensile strength of the palladium and hydrogenium alloy. A new palladium wire similar to the previous ones was broken in experiments made on two different portions of it, by a load of 10 and of 10-17 kilogrammes. Two other portions of the same wire charged with hydrogen were broken by 8·18, and by 8.27 kilogrammes. Hence we haveTenacity of palladium .

Tenacity of palladium and hydrogen

100.

81.29

The tenacity of the palladium is reduced by the presence of hydrogen, but not to any great extent, and it cannot be too strongly

urged that the decrease in tenacity would in all probability be much greater if the second element were not metallic.

3rd. The electrical conductivity next formed the subject of investigation. Compared with a wire of German silver, the following results were obtained at 10·5° C :

Dr. Matthiessen's researches have proved that a reduced conducting power is generally observed in alloys, and the charged palladium wire falls 25 per cent., but certainly, as in the case of the tensile strength, the diminution should be greater if the additional constituent were not metallic. The experiments were repeated in Dr. Matthiessen's laboratory, and confirmed by him.

4th. The magnetic state of the alloy is given, and as the details are important, we will consider them at some length. Faraday states as the result of all his experiments, that palladium is "feebly but truly magnetic," and this metal is placed among what are called the paramagnetic metals. The feeble magnetism of palladium did not extend to its salts. In repeating the experiments a horse-shoe electro-magnet of soft iron, capable of supporting 60 kilogrammes when excited by four large Bunsen cells; the magnet was placed with its piles directed upwards, and each supplemented by a block of soft iron terminating in points. The palladium under examination was suspended between these points in a stirrup of paper attached to three fibres of cocoon silk. A filament of glass was attached to the paper, and moved as an index on a graduated circle, the palladium in the form of a fragment of electro-deposited metal, 8 millimetres long and 3 millimetres in width, was allowed to take an equatorial position, or with its ends averted from the points of the magnet. The iron was charged by the battery, and a slight deflection of 10° only was obtained, the magetism acting against the torsion of the silk thread, the same fragment of palladium charged with 604-6 volumes of hydrogen was deflected by the electro-magnet through 48° when it set itself at rest. The hydrogen was then extracted by heating the metal in vacuo, then on replacing it between the poles, the magnet exerted no influence upon it.

Therefore it is evident that the addition of hydrogen adds manifestly to the magnetic condition of the palladium.

To have some term of comparison, the same little mass of

palladium was steeped in a solution of sulphate of nickel, under specific gravity 1.048, and by the action of the magnet a deflection of 35° was obtained, or less than that obtained with hydrogen (48°). The same palladium was afterwards washed, and then impregnated with a solution of protosulphate of iron, of which the metallic mass held 2-3 per cent. of its weight, the palladium gave a deflection of 50°, or nearly the same as with hydrogen; with a stronger solution of the same salt, of specific gravity 1.17, the deflection was 90°, and the palladium pointed axially.

These experiments were carefully repeated. Therefore from the above results it follows that hydrogenium is magnetic, a property which is confined to metals and their compounds. It may be urged that Faraday and E. Becquerel have proved that hydrogen as a gas is dia-magnetic, "but magnetism is so liable to extinction under the influence of heat, that the magnetism of a metal may very possibly disappear when it is vapourized, as appears with hydrogen in the form of gas."

Therefore on these grounds Mr. Graham considers that hydrogenium may be classed with the strictly magnetic metals, iron, nickel, cobalt, chromium, and manganese.

"The chemical properties of hydrogenium also distinguish it from ordinary hydrogen. The palladium alloy precipitates mercury and calomel from a solution of the chloride of mercury without any disengagement of hydrogen; that is, hydrogenium decomposes chloride of mercury, while hydrogen does not. This explains why M. Stanislas Meunier failed in discovering the occluded hydrogen of meteoric iron, by dissolving the latter in a solution of chloride of mercury; for the hydrogen would be consumed, like the iron itself, in precipitating mercury. Hydrogen (associated with palladium) unites with chlorine and iodine in the dark, reduces a persalt of iron to the state of protosalt, converts red prussiate of potash into yellow prussiate, and has considerable deoxidising powers. It appears to be the active form of hydrogen, as ozone is of oxygen.

"The general conclusions which appear to flow from this inquiry are-that in palladium fully charged with hydrogen, as in the portion of palladium wire now submitted to the Royal Society, there exists a compound of palladium and hydrogen in a proportion which may approach to equal equivalents;* that both substances are solid, metallic, and of a white aspect; that the alloy contains about 20 volumes of palladium united with one volume of hydrogenium; and that the density of the latter is about two, a little higher than * "Proceedings of the Royal Society," lxviii., p. 425.

magnesium, to which hydrogenium may be supposed to bear some analogy; that hydrogenium has a certain amount of tenacity, and possesses the electrical conductivity of a metal; and, finally, that hydrogenium takes its place among magnetic metals. The latter fact may have its bearing upon the appearance of hydrogenium in meteoric iron, in association with certain other magnetic elements."

In conclusion, it may be of interest to mention another beautiful experiment that may be performed at a moderate cost, notwithstanding the high price of palladium (£8 per ounce).

If two wires of the metal be placed side by side in acidulated water, and connected with each pole of a battery, oxygen is freely evolved from the positive pole, while the wire from which hydrogen should be liberated is perfectly quiescent, owing to the absorption by the metal; when the surface is saturated there is a rapid evolution of gas.

For lecture demonstration the following experiment will be found the most striking :-Provide a wire of palladium not less than three feet in length. Attach this to a lever of the third order, as indicated by the diagram. Mark the normal position of the index, B, a light straw, then remove the wire, and place it as the negative electrode of a battery decomposing acidulated water for thirty-five to forty minutes. Replace it in its former position, the wire having expanded, the index will fall, C. Then heat the wire with a spiritlamp, the hydrogen escapes, burning along the wire, at a short distance from the flame of the lamp, the index slowly rising past the normal position, setting itself at rest at A an equal distance on the other side of zero to that determined by the expansion.

A

B

C