When the electrized gas was passed through a solution of hyposulphite of soda, it was found to undergo a diminution in volume equal in amount to twice the "iodine-titre" of the gas, the mean of twelve experiments giving for this value 1.9, the "titre" being 1. The oxidation also effected in the solution of hyposulphite of soda was equal to three times the "titre" of the gas, the mean of a set of five experiments giving for this value, according to one method of estimation, 2-72, and according to a second and more exact method 3.01. Again, the oxidation effected by the electrized gas in a solution of protochloride of tin was equal to three times the "titre" of the gas, the value found as the mean of six experiments being 2.77. In two experiments also, in which both the oxidation and contraction were respectively estimated, 2-75 and 2.69 were the values found for the oxidation, and 1-9 and 1-8 the values found for the contraction, the "titre" of the gas being in all cases assumed as 1. These experiments conclusively demonstrate the identity of the ozone formed by the action of electricity upon carbonic-acid gas with the ozone similarly generated from pure oxygen.

Various experiments were instituted with the view of ascertaining the conditions most favourable to the conversion of a large proportion of this oxygen into ozone. The conclusion at which I arrived was that the greatest proportion of ozone in relation to the total oxygen was produced when a rapid current of dry carbonic-acid gas was acted upon at a low temperature by electricity of feeble tension. When these conditions were realized, it was found that 75 per cent. of the total oxygen eliminated in the induction-tube by the decomposition of the carbonic-acid gas could readily be converted into ozone, but that it was difficult to pass this limit. This amount corresponds with the formation of a gas constituted of ozone and oxygen in the proportion of two units of ozone to one unit of oxygen, the matter of the oxygen being thus distributed 23 +2. In several experiments, nevertheless, this limit was considerably exceeded; and a table is given of nine experiments, in which the amount per cent. of ozone ranged from a minimum of 76.6 per cent. to a maximum of 85.5 per cent. This latter amount corresponds with a gas constituted of ozone and oxygen in the proportion of four units of ozone to one of oxygen, thus 43 +2. this case the ratio of the total oxygen present to the titre of the gas is 3-5; if the whole oxygen were converted into ozone, the value of this ratio would be 3.

In

These experiments, taken in connexion with those described in my previous communication, leave no room for reasonable doubt as to the composition of the unit of ozone, which is the triad form of the element oxygen, and is to be symbolized as 3, an hypothesis henceforth to be regarded as conclusively demonstrated. This remarkable discovery has many important bearings upon chemical theory, especially in relation to the problem elsewhere raised by me of the true nature of the elemental bodies and the constitution of the unit of the element chlorine, which ozone so closely resembles in its chemical properties, and which is also a triad element symbolized as ax2.

Nov. 27.-W. Spottiswoode, M.A., Treasurer and Vice-President, in the Chair.

The following communication was read:

"On the Quantitative Analysis of certain Alloys by means of the Spectroscope." By J. Norman Lockyer, F.R.S., and William Chandler Roberts.

The authors, after referring to experiments which showed clearly that the spectroscope might be employed to detect minute differences in the composition of certain alloys, proceed to give an account of the researches which they had instituted with a view to ascertain the degree of accuracy of which the method is capable.

The image of an electric spark passing between the unknown alloy and a fixed electrode being thrown by means of a lens on the slit of the spectroscope, the phenomena observed were found to vary with the composition of the alloys; and further, by arranging them together with known check-pieces on a suitable stand, and bringing them in turn under the fixed electrode, the composition of the unknown alloys was determined by comparison with the known check-pieces.

The shape of the electrode ultimately adopted was that represented in the sketch; and these pieces were held in their places by suitable metallic clips. Special attention was then directed to the adjustment of the length of the spark, which was found to materially influence the phenomena. The method adopted consisted in placing the variable electrode in the field of a fixed microscope having a 3- or 4-inch objective, and adjusting the summit of this electrode to coincide with the spider-lines of the eyepiece. After a series of experiments on alloys of zinc and cadmium of various compositions, the results of which were shown on a curve, more extended trials were made with the gold-copper alloy employed in coinage, which was peculiarly suited to these researches in consequence of the known method of assay having been brought to so high a state of perfection (the composition being determined with accuracy to the part of the original assay-piece of about 7 grains), and from the fact that reliance can be placed on its homogeneity. The paper is accompanied by a series of four curves, which show the results of experiments, and in which the coördinates are given by the ordinary method of assay, and by the spectroscopic readings.

1 10,000

The chief practical advantage which appeared to flow from this inquiry was that, if it were possible to replace the parting assay by the spectroscopical method, a great saving of time in ascertaining the value of gold bullion would be effected.

Dec. 18.-Joseph Dalton Hooker, C.B., President, in the Chair. The following communication was read:

"On the Period of Hemispherical Excess of Sun-spots, and the 26-day Period of Terrestrial Magnetism." By J. A. Broun, F.R.S. It appears from the interesting communication to the Royal

Society, June 19th, by Messrs. De La Rue, Stewart, and Loewy*, that the difference of the area of spots on the visible northern and southern quarter-spheres of the sun seems, during periods of considerable solar disturbance, to obey a law such that the difference is a maximum in the same quarter-sphere during several successive rotations of the sun, the difference being a maximum alternately in the northern and southern hemispheres-the time from maximum to maximum, for the same hemisphere, being variable between 18 and 32 days, but having a mean value of about 25-2 days.

It occurs at once that if the variations of the mean terrestrial magnetic force are connected in any way with the solar spots, or the causes which produce them, we might here find some explanation of the magnetic period of 26 days, the difference of spot-area in one hemisphere from that in the other being related to à difference of the solar magnetic action.

In order to determine whether such a connexion existed, I projected first the curves of excess of spot-area given in the paper of Messrs. De La Rue, Stewart, and Loewy, and below them the daily mean horizontal force of the earth's magnetism during the same periods. The conclusion from these projections is, that there is no relation whatever between the two classes of curves. The maxima and minima of the one agree in no way with those of the other: the greatest excesses of sun-spot area in the one hemisphere over those in the other occur when the earth's magnetic force is the most constant; the greatest variations of the earth's magnetic force from the mean occur in several instances when the sun-spot area is equal in the two visible quarter-spheres.

It should be remembered, in considering the curves of sun-spot excess, that the minima and maxima are in some cases only relative, -sometimes the one, sometimes the other being really cases in which there is neither maximum nor minimum-that is to say, cases in which the sun-spot area is equal, or nearly so, in the two visible quarter-spheres.

It would be hasty to conclude from this comparison that the variations of the mean magnetic force are really unconnected with the mode of distribution of the sun-spots. Other methods of grouping the spots may perhaps be employed with advantage relatively to this and other questions; for example, were the position of the centre of gravity of the sun-spots determined for the visible quarter-spheres and hemisphere, giving each spot a weight in proportion to its area, the variation of these positions in latitude and longitude, and their weights, might give a more satisfactory base for this comparison and for other deductions.

It will be obvious also that this investigation refers only to the visible hemisphere of the sun; an approximation to the spot-distribution on the other hemisphere, however, will be frequently possible.

*Proc. Roy. Soc. vol. xxi. p. 399.

GEOLOGICAL SOCIETY.

[Continued from p. 235.]

April 30, 1873.-Joseph Prestwich, Esq., F.R.S., Vice-President, in the Chair.

The following communications were read:—

1. "On the Permian Breccias and Boulder-beds of Armagh." By Prof. Edward Hull, F.R.S., F.G.S., Director of the Geological Survey of Ireland.

In this paper the author described certain breccias occurring in the vicinity of Armagh, which he referred, both on stratigraphical and physical grounds, to the Lower Permian series, considering them to be identical with the "brockram" of Cumberland and the breccias of Worcestershire and Shropshire. The lower beds rest on the denuded surface of the Lower Carboniferous Limestone, and consist of a breccia of limestone pebbles in a reddish sandy paste, sometimes becoming a bedded calcareous sandstone with pebbles. These beds, which are 10-12 feet thick, are overlain by soft rudely stratified conglomerate and Boulder-beds of subangular and rounded blocks of grit, felspathic sandstone, vein-quartz, and limestone. In some places the breccia graduates up into the overlying Boulder-beds; but sometimes its surface is eroded. These beds are overlain by Boulderclay of the Drift period. Within the city of Armagh well-borings &c. have revealed the existence of New Red Sandstone above the Breccia and Boulder-beds. In the author's opinion the only agent which could have brought the blocks of Silurian and Old-Red-Sandstone age found in the Boulder-beds from their place of origin is floating ice. The author further referred to the extensive denudation which the Carboniferous beds have undergone in Armagh, and also alluded to the occurrence of beds of Permian age near Benburb, between Armagh and Dungannon.

2. "Geological Notes on Griqualand West." By G. W. Stow, Esq., F.G.S.

The geological results of a journey made by Mr. G. W. Stow and Mr. F. H. S. Orpen from the Orange Free State into Griqualand West are communicated by Mr. Stow in this paper, with numerous carefully executed sections and a geological map based on the survey map prepared by Mr. Orpen for the Government. From the junction of the Riet and Modder rivers (south of the Panneveldt Diamond-fields) westward to the junction of the Vaal and the Orange, over the Great Campbell Plateau to Griquatown, Ongeluk, Matsáp, Potgieter, the Langeberg, Witte Zand, and to Kheis and the Schurwe Bergen, the track traversed three degrees of longitude, but nearly 300 miles of road. The return route north-east to Mount Huxley and Daniel's Kuil, and eastward to Likatlong, on the Hart or Kolang river, was nearly as long; and its results form part of the present communication. A subsequent portion of the journey up the Hart, across to the Vaal, down the valley by the diamond-diggings of

Hebron and Klipdrift to the Panneveldt, will be treated of in another paper.

From the Modder, first south-westward and then westward, to the junction of the Vaal and Orange, the olive shales of the Dicynodon or Karoo series, traversed frequently by igneous rocks, form the country, and are seen in some places to lie unconformably on older rocks. The shales reach to the edge of the Campbell Randt, on the other side of the Orange River, and have been, it seems, to a great extent, formed of the débris of those old hills. The oldest rocks of the locality are seen cropping out here and there in the gorges at the foot of the Randt, and consist of metamorphic rocks, greatly denuded, on which the massive and extensive siliceo-calcareous strata of the Great Campbell Plateau lie unconformably. These latter and the breccias of their slopes are coated thickly with enormous travertine deposits. Beyond the Plateau, at Griquatown, a long parallel range of jaspideous rocks comes out from beneath the Campbell Plateau, presenting a wonderful group of yellow, brown, chocolate, and red jaspers, with magnetic and other ironstone, and beautiful seams of blue and yellow crocidolite. The southern portion of this range has long been known as the "Asbestos Mountains" and the "Doornberg." Igneous rock-masses occur around Ongeluk, west of the Jasper range and then bright-red jasper rocks crop up near Matsáp, succeeded to the west by the parallel quartzite range of Matsáp, and again by other bedded jaspers, which seem to lie in a synclinal of the quartzite rocks, which come up again in the Langeberg. These are succeeded by lower rocks, consisting largely of sandstone, grit, and quartzite, with more or less pervading mica, as far as the journey extended in the Schurwe Bergen, also parallel to the former ranges. The maximum thickness of the successive strata is calculated by the author at 24,000 feet; allowing for possible reduplications, the minimum is regarded as not less than 9000 feet. The details of stratification, successive upheavals, denudation, nature and origin of the salt-pans, escarpments, river-valleys, and other features were treated of by the author, who has supplied a very large collection of specimens illustrative of the phenomena observed in the line of march, and of his numerous sections, maps, and sketches.

3. "On some Bivalve Entomostraca, chiefly Cypridinidæ, of the Carboniferous formations." By Prof. T. Rupert Jones, F.R.S., F.G.S. The larger forms of bivalved Entomostraca are not rare in the Carboniferous Limestone; and some occur in certain shales of the Coalmeasures. Professors Phillips, M'Coy, and De Koninck have figured and described some specimens. The collections made by M. J. Bosquet, F.C.G.S. (Belgium), Mr. Joseph Wright, F.G.S. (Cork, Ireland), and Mr. J. H. Burrow, M.A. (Settle, Yorkshire), and submitted to the author and his friend Mr. J. W. Kirkby, have elucidated many obscure points in the history of these old Entomostraca. Other specimens from Scotland, Yorkshire, Derbyshire, and the Isle of Man, communicated by many friends, were treated of in this paper.

The Entomoconchus of M'Coy proves to have the Cypridinal cha