times that of the metal. Submitted, at a dull-red heat, to the action of a current of hydrogen, it reabsorbs a volume sensibly equal to that which it had discharged. The metal is again pyrophoric after the hydrogen is expelled.

II. Cobalt.-An ingot, cast in lime, of pure cobalt was submitted to a red heat for twenty-four hours in a current of hydrogen, and then cooled slowly in that gas. The volume of hydrogen extracted from it in vacuo at a red heat was only one tenth of that of the metal.

Lamine of cobalt, obtained by galvanic decomposition of the double sulphate of cobalt and ammonia, were heated in vacuo to 200°. They gave up thirty-five times their volume of hydrogen. Heated subsequently in a current of hydrogen to about 200°, and again cooled slowly in the same gas, they absorbed twenty-four times their volume of it, which they set free again in vacuo at 200°. The same laminæ, placed during twenty-four hours at the negative pole of a voltameter, absorbed seven times their volume of hydrogen.

Pyrophoric cobalt loses its hydrogen in vacuo still more readily than nickel. Instead of making a vacuum, the condensed gas can be expelled by putting the metal into a small balloon furnished with a discharge-tube and filled with water exhausted of air. Heated to 100°, all the gas is disengaged in a few hours. The volume of the gas thus collected is about one hundred times that of the metal; the cobalt, too, is again pyrophoric after complete expulsion of the hydrogen. Submitted at a dull red heat to the action of a current of hydrogen, it reabsorbs a volume equal to that which it has set free.

III. Iron.-We have previously proved that 1 kilogramme of soft iron in the form of an ingot can dissolve at about 800°, and afterwards set free in vacuo at the same temperature, 20 cubic centims. of hydrogen, or one sixth of its own volume. Under the same conditions 1 kilogramme of grey pig-iron, wood-cast, dissolves 88 cubic centims. of hydrogen, or more than the half of its volume‡.

It is known that the iron obtained in decomposing by the pile chloride of iron in the presence of sal ammoniac, when plunged into hot water disengages hydrogen and at the same time a small .quantity of ammonia, as was proved by MM. Meidinger§ and Kroemer. M. Cailletet has recently obtained in this way a volume of hydrogen equal to 260 times that of the metal.

* Analysis of the gas did not show any sensible quantity of nitrogen. Some lamina prepared in the same manner, then washed and dissolved in chlorhydric acid, gave, like the nickel, traces of ammonia. + Comptes Rendus, vol. lxxvi. p. 562.

We have since ascertained that iron wire, hardening slightly by steeping, dissolves at a red heat nearly one fourth of its volume of hydrogen; the same wire, after cémentation, dissolved one third of its volume of the gas. The solubility of hydrogen in steel increases, therefore, with the amount of carbon contained in the latter.

§ Dingl. Polytech. Journ. vol. clxiii. p. 283. Arch. Pharm. 2nd Series, vol. cv. p. 284. Comptes Rendus, vol. lxxx. p. 319.

Pyrophoric iron obtained by reducing at a low temperature either the sesquioxide alone or a combination of oxide of iron and alumina (precipitated from their chlorides by ammonia), gives up, like pyrophoric nickel and cobalt, all its hydrogen in vacuo, and, like them, retains the property of ignition at a low temperature in air. As to the quantity of hydrogen which can be fixed by pyrophoric iron, its determination presents special difficulties. The cold metal loses in vacuo a portion of the gas which it absorbed. Boiling water, by the use of which we succeeded in obtaining the hydrogen dissolved in pyrophoric nickel or cobalt, gave with iron entirely different results. In fact, the pyrophoric iron which comes from the reduction of the combined oxides of iron and aluminium, put with air-exhausted water into a small balloon furnished with a discharge-tube, gave a continuous disengagement of hydrogen when heated: thus I grm. of iron liberated 10 cubic centims. of gas per hour; and the liberation went on until the iron was almost completely oxidated. The water was therefore decomposed at about 99° by the minutely divided iron. Pyrophoric iron from the reduction at a low temperature of the hydrate of the sesquioxide alone, decomposes water with a rapidity nearly equal to that of the metal combined with alumina. With respect to the pulverulent iron less minutely divided which is obtained on reducing by hydrogen the sesquioxide resulting from the calcination of the nitrate, it also decomposes water at about 99°, but the decomposition is much slower. The reduced iron of commerce and the spongy iron obtained by the galvanic pile behave in the same manner*.

Not being able to determine by immersion in boiling water the volume of the gas condensed in pyrophoric iron, we essayed to determine it by keeping the iron in cold water; but here again we had to recognize the decomposition, though slower, of the water: 1 gramme of pyrophoric iron, kept in water deprived of air and at 15°, liberated hydrogen regularly during two months.

In brief, iron, nickel, and cobalt directly absorb hydrogen; but we cannot affirm that combination takes place: this is what we have already proved for lithium and thallium. Pyrophoric iron, nickel, and cobalt condense a greater quantity of gas than the compact metals; but the whole of the gas is liberated below a red heat, and the metals deprived of hydrogen continue to be pyrophoric: this property therefore does not depend on the presence of condensed hydrogen. Finally, iron in a state of minute division exhibits a property which is not found in either nickel or cobalt; it decomposes water slowly at ordinary temperatures, and rapidly at about 100°: in this respect it approaches manganese, of which we shall shortly make known some new properties.-Comptes Rendus de l'Acad. des Sciences, vol. lxxx. pp. 788-791.

*The vapour of water, under the tensions comprised between 5 and 25 millims. is likewise decomposed by iron at the temperature of 100°, as results from experiments by H. Sainte-Claire Deville.

ON A LECTURE-ROOM APPARATUS FOR THE DETERM THE MECHANICAL EQUIVALENT OF HEAT. BY ASSISTANT IN THE IMPERIAL MARINE ACADEMY,

The apparatus is of very simple construction, con tially of a calorimetric and a dynamometric portion, nected with an oscillation-machine such as is found in cal cabinet.

The calorimetric part is formed by two hollow tru of cast iron, one fitting into the other, the inner not o to the bottom of the outer, and appearing a little The outer cone can be fixed coaxially in the vertical machine. The inner cone contains mercury. If the m in motion and the inner cone fixed, heat is generated b of the surfaces.

An arrangement which is an inversion of Prony's for the measurement of the work which is converted in the wooden lid of the inner cone a light wooden bea horizontal. A perforation passing through the beam a the thermometer. At some distance from the beam ley on a level with it, over which a thread, to which pended, is slung and is fastened to the end of one arm the other arm serves as a counterpoise. When the in motion, the interior face of the outer cone rubs ag face of the inner one, and tends to turn the beam whi to the latter in the direction of the motion. With a the horizontal part of the thread and the axis of the clude a right angle. From the length of the beam-a the load, and number of rotations the work converted from the water-value of the calorimeter and the rise o the quantity of heat generated, can be calculated.

The memoir contains also the development of the apparatus, taking into consideration the heat radiated lorimeter, and, finally the numerical calculation of th mechanical equivalent of heat from 28 experiments. those numbers, 425.2, with the mean error ± 5.4, accordance with Joule's result, 424-9, and may be rega as a fresh corroboration of it, but also as a measure o with which the experiments can be conducted by simple apparatus; and these occupy but very little t riment proper lasting but 30-60 seconds, on which a paratus may be recommended for lecture-experime Akad. der Wissenschaften in Wien, math.-naturw. Cl 1875.