The author next proceeded to notice the raised beach at the Bill of Portland, in which he had, with the assistance of Mr. Jeffreys, determined 26 species of shells, two of them not now living in the British Channel, and one new. This beach contains pebbles of the

Devonshire and Cornwall rocks.

The raised beach Mr. Prestwich found to abut against an old cliff that had been swamped at a later geological period by a land-wash, which had levelled it and the old sea-land with the adjacent landsurface. The mass which had thus swamped the cliff and buried the beach consisted of loam and angular débris, the latter being in larger proportion at top. In the loam he found several species of land and freshwater shells and fragments of bones. The angular débris consisted of pieces of the local rocks, together with a number of specimens which by their organic remains were shown to belong to the Middle Purbecks, a part of the series not now existing in Portland. A similar bed, but much thicker, was then described at Chesilton, in the north of the island. It is there 60 feet thick, and contains large blocks of Portland stone and Portland chert, the greater number of which are in the upper part of the deposit, which is here on the sea-level, and 400 feet lower than the Portland escarpment which rises above it. This loam and angular débris the author was disposed to attribute to a temporary submergence of the land to a depth exceeding the height of Portland, by which the land as it emerged was swept and its débris carried down to the lowest levels, with the remains of its land-animals and land and freshwater shells, which latter, where protected by large masses of loam and suddenly entombed, have been preserved uninjured. To this deposit, which is common over the raised beaches on the south coast, the author proposed to apply the term "Land-wash."

The paper concluded with a short notice of the drift-beds formed subsequently to the denudation of the Weymouth district, and therefore never on the high-level Portland drift. Among these was one near Weymouth of singular character, consisting almost entirely of subangular fragments of Greensand chert, which could not have been derived from beds nearer than Abbotsbury. The lower drift of the district is the valley-gravel of Upway and Radipole, in which the remains of Elephas primigenius have been found.

2. "On the Character of the Diamantiferous Rock of South Africa." By Prof. N. Story Maskelyne, F.R.S., F.G.S., Keeper, and Dr. Flight, Assistant, of the Mineralogical Department, British Museum.

In this paper the authors confirmed certain statements made by one of them from a superficial examination of specimens brought to this country by Mr. Dunn. The specimens examined and analyzed by Dr. Flight were obtained from various diggings and from different depths, down to 180 feet in the case of one mass from Colesberg Kopje. Their characters throughout are essentially the same.

The rock consists of a soft and somewhat pulverulent groundmass, composed of a mineral (soapy to the touch) of a light yellowish

colour in the upper, and of an olive-green to bluish-grey colour in the lower parts of the excavations. Interspersed in the mass are fragments of more or less altered shale, and a micaceous-looking mineral of the vermiculite group, which sometimes becomes an important constituent of the rock, which also contains bright green crystals of a ferruginous enstatite (bronzite), and sometimes a hornblendic mineral closely resembling smaragdite. A pale buff bronzite occurs in larger fragments than the green form of the mineral; and in the rock of Du Toit's pan an altered diallage is present. Opaline silica, in the form of hyalite or of hornstone, is disseminated through the greater part of the rock-masses; and they are everywhere penetrated by calcite.

The analyses of the component minerals (given in detail in the paper) show that this once igneous rock is a bronzite rock converted into a hydrated magnesium silicate, having the chemical characters of a hydrated bronzite, except where the remains of crystals have resisted metamorphism. Except in the absence of olivine and the small amount of augitic mineral, it might be compared with the well-known Lherzolite rock.

The diamonds are said to occur most plentifully, or almost exclusively, in the neighbourhood of dykes of diorite which intersect the hydrated rock, or occur between it and the horizontal strata through which the igneous rocks have been projected. The authors compare the characters of the diamonds found in different positions, and come to the conclusion that their source is not very remote from that in which they are now found.

The mineral above-mentioned as resembling vermiculite is described by the authors as a new species under the name of Vaalite.

I

XLVII. Intelligence and Miscellaneous Articles.

ON THE SPECTRUM OF THE AURORA.

To the Editors of the Philosophical Magazine and Journal.
GENTLEMEN,

WISH to point out a slight error in a communication from Mr. J. R. Capron on the Spectrum of the Aurora, which appears in the last Number of the Philosophical Magazine. Mr. Capron writes as follows:

From Dr. Watts's 'Index of Spectra' I have extracted the three principal carbon-tube bands or lines; and they compare with Dr. Vogel's oxygen-tube as under:--

Dr. Vogel's oxygen-lines

Yellow. Green. Blue. 5603 5195 4834

Dr. Watts's carbon-tube bands or lines. 5622 5189 4829

"The 5622 for the yellow line of the tube must be an error. 5608 seems to me, from my own observations, nearer its place; and I calculate 5193 and 4825 for the other lines."

The number 5622 Mr. Capron has taken by mistake from the wrong series of numbers-from spectrum No. I. instead of from spectrum No. II. The other two numbers, 5189 and 4829, I am not able to find in the Index of Spectra' at all. The wave-lengths actually given for the three lines of the tube-spectrum are as follows:

6

I have satisfied myself that these lines are due to a compound of carbon and oxygen; and their production in tubes supposed to contain oxygen is not at all surprising. The results obtained by the use of vacuum-tubes must always be received with caution, since it is almost impossible to form any certain conclusion as to the real nature of the small quantity of gas remaining in the tube, even when one has filled the tube one's self; and in the case of tubes purchased from makers of such apparatus, the nature of the contents is simply a matter of conjecture.

Giggleswick Grammar School,
April 14, 1875.

I am, &c.,

W. MARSHALL WATTS.

EXPERIMENTS WITH THE HOLTZ MACHINE.

BY F. ROSSETTI*.

In a recent series of experiments upon the Holtz machine, first pattern, M. Rossetti sought to determine in what measure the intensity of the current produced depends on the velocity of rotation of the machine, on the work expended, and on the humidity of the air, to estimate its electromotive force, internal resistance, &c. results at which he arrived on these points are the following.

The

In one and the same series of experiments, the intensity of the current is very sensibly, but not exactly, proportional to the velocity of rotation of the plate; the intensity increases a little more rapidly than the velocity of rotation.

The effect is modified by the humidity of the air, the velocity necessary to produce a certain intensity being greater in wet than in dry weather.

The work expended for the production of the electricity is exactly proportional to the intensity of the current. It was measured by the difference of the weights necessary to impress a certain velocity on the plate, according to whether the machine was charged

or not.

* Nuovo Cimento, Ser. 2, vol. xii. p. 89.

The ratio between the work expended and the intensity of the current diminishes when the humidity augments-in such a manner that, to obtain a current of given intensity, a greater velocity of rotation is requisite in wet than in dry weather, but a less expenditure of work. The Holtz machine is therefore more economical in wet weather than in dry.

The distance between the two disks of the machine has also some influence on the intensity of the current: the less the distance the stronger the current, and the greater the amount of the work.

The Holtz machine, like the voltaic couples, possesses electromotive force and internal resistance. The electromotive force is independent of the velocity of rotation; but it diminishes as the degree of humidity increases. The effective motor weight (difference between the weights necessary to turn the machine charged and not charged) is proportional to the electromotive force produced. This is very great in comparison with the electromotive forces of the most energetic voltaic couples; in fact it has been found to amount to 433000 Siemens units with a relative humidity equal to 0-69, to 599000 with the humidity 0-35-the Daniell couple giving E=11.57, and that of Grove E=19.98. The electromotive force of the Holtz machine is therefore about 50000 times as great as that of the Daniell couple, and 30000 times as great as that of the Grove couple.

The internal resistance of the Holtz machine is independent of the hygrometric state, but varies with the rotation-velocity, diminishing more rapidly than the velocity increases. It is very great: the lowest resistance (which corresponds to the greatest velocity attainable, or 8 turns per second) is equal to 570 million Siemens units; for a velocity of 2 turns per second it is 2810 millions of the same units.

Under these conditions a resistance inserted in the outer circuit must be very considerable in order to exert any sensible influence on the intensity of the current. It was because the resistance employed by Poggendorff was too feeble that he could not verify an effect of this kind. M. Rossetti, on the contrary, by interposing in the circuit a column of distilled water of greater or less length, has ascertained that the current of the Holtz machine is susceptible of being very notably weakened by augmentation of the external resistance, following Ohm's law in this equally well with the ordinary galvanic currents.

The author has deduced from his experiments a measure of the mechanical equivalent of heat, by comparing the serviceable work expended for the production of the electricity with the total heat which could be developed by the current obtained. He found for that equivalent the number 428.-Bibliothèque Universelle, Archives des Sciences, March 15, 1875, pp. 250–252.

ON THE SOLUTION OF HYDROGEN IN METALS, AND THE DECOMPOSITION OF WATER BY IRON. BY MM. L. TROOST AND P.

HAUTEFEUILLE.

In previous researches on the alloys formed by hydrogen*, we have pointed out the characters by which these definite combinations may be distinguished from the solutions of hydrogen in the metals. We have seen that besides potassium, sodium, and palladium, which can combine with hydrogen, there are other metals which simply dissolve this gas. The number of those which possess this last property appears to be considerable.

We shall see that iron, nickel, cobalt, and manganese, which are united by the analogy of their chemical properties into a natural group, present great similarities in their behaviour in the presence of hydrogen at various temperatures. As the facility with which they absorb or give out hydrogen depends largely on their physical state, it is necessary, in order to account for the differences observed, to investigate these metals successively in ingots, in thin plates, and in the pulverulent condition.

I. Nickel.-An ingot of pure nickel, cast in lime, was submitted for twenty-four hours, at a red heat, to the action of a current of hydrogen gas, and then cooled slowly in the gas. The volume of hydrogen extracted from it at a red heat in vacuo was one fifth of the volume of the metal.

Some lamina of nickel, obtained by decomposing with the pile the double sulphate of nickel and ammonia, were heated in a vacuum to 200° C.; they gave out forty times their volume of hydrogent. On afterwards being heated to near 200° in a current of hydrogen and slowly cooled in this gas, they absorbed sixteen times their volume of it, which they gave up in vacuo at 200°. The same laminæ, placed for twenty-four hours at the negative pole of a voltameter, absorbed about ten times their volume of hydrogent.

The pulverulent nickel was obtained by reducing the oxide or a mixture of oxide of nickel and alumina by means of hydrogen at 300°. Nickel thus prepared is pyrophoric, as Magnus has shown §. In a vacuum it gives up a certain quantity of hydrogen at the ordinary temperature; but to expel this gas entirely a dull red heat is requisite. The total volume of gas discharged is about one hundred *Comptes Rendus de l'Acad. des Sciences, vol. lxxviii. pp. 686, 807. Phil. Mag. S. 4. vol. xlvii. p. 397.

†The gas analyzed did not give any perceptible quantity of nitrogen. Some lamina prepared in the same way, then washed and dissolved in chlorhydric acid, gave traces of ammonia.

M. Raoult states (Comptes Rendus, vol. lxix. p. 826) that the impure porous nickel cubes of commerce, when placed at the negative electrode of a voltameter, absorb 165 volumes of hydrogen, which they gradually disengage at the ordinary temperature. The same cubes electroplated with pure nickel did not appear to him to disengage any appreciable quantity of gas. Annales de Chimie et de Physique, 2 série, vol. xxx. p. 103. Phil. Mag. S. 4. Vol. 49. No. 326. May 1875.

2 F