intersection-edges, in which therefore n' 2m' m'-m 9 to which among simple coefficients the forms 2-4 or 3-3 would answer. The face is too small and dull to allow of a measurement of angle; 2-4 appears however the more likely form, as chrysolites seem to have a preference for the ratio 1:2. Cleavages, three rectangular; O and i- very and almost equally eminent, with vitreo-pearly luster approaching the sub-adamantine; - splintery. Hardness=5.5-6. Sp. G.=395-4.08. The average of nine determinations with Jolly's spring balance gave 4.023. Color, dark green to black but eminently mottled, so that thin splinters or laminæ transmit a pale yellow light. Streak, light-yellowish-reddish-gray. The powder is slightly attracted by the magnet. BB. rather refractory, fusing at thin edges to a dull black slag. On charcoal gives a zinc coating, more distinctly on addition of soda. With the fluxes the usual reactions for silica, iron and manganese. The borax and microcosmic beads give in the O. F. the characteristic brownish purple color, indicating mixtures of iron and manganese, which becomes green in the reducing flame. With acids gelatinizes readily and completely. Some specimens leave a bright green undissolved residue, which I judge to be spinel both from its hardness, its not being attacked by fusion with soda, and complete decomposition by bisulphate of soda. In the following analyses the silica was separated in the usual manner, the filtrate from the silica neutralized by carbonate of soda, then acidified with acetic acid and a current of sulphuretted hydrogen passed through the solution, which separated the zinc as sulphid. This was filtered off, redissolved in HCl, and then precipitated from the uninterruptedly boiling solution by slowly adding NaC, and the ZnO finally converted into Żn by ignition. The filtrate after the separation of the zine was then boiled with the addition of 1KO,CIO, to sesquioxydize the iron, the iron precipitated in the usual manner as subacetate, redissolved and reprecipitated by ammonia, the manganese separated by bromine and determined as pyrophosphate with the precautions pointed out by Dr. Gibbs, (this Journal, No. xxxi, p. 216), and lastly the magnesia determined as pyrophosphate. I will here remark, that I did not succeed in separating the oxyd of zinc from the iron by the usual acetate of soda process, but that a great and often the greater part of the Zn went down with the iron, which I attribute to the necessary boiling of a dilute solution (vid. Johnson's Qual. Fresenius). Hence my former analyses were not correct, gave too little zinc and resulted in uni-oxygen ratio only on account of the near proximity of the equivalents of iron, manganese and zinc-oxyds. I may, however, not have handled the method correctly. The samples No. 1 and 2 were fresh pieces of cleavage crystals carefully examined by the lens so as to avoid all visible admixtures. No. 1 was lighter in color than No. 2, a and b, which latter are analyses of the same powder. No. 3 are analyses of two different powders of the massive variety. 99-35 99-90 100-11 100-80 100.52 16:37 16 15 15.99 16-10 15.98 The foregoing oxygen ratios make the mineral a unisilicate. The crystallization being orthorhombic with the parametric ratios of the chrysolite group, which is confirmed by the other physical and chemical characters; it is hence an iron-manganesezine chrysolite, the first, to the best of my knowledge, of the group, into the composition of which zinc enters as a constituent. It occurs, as before said, on Stirling hill, accompanied by Willemite, Franklinite, Jeffersonite and spinel. 2. Manganesian Dolomite. In the vast vein of Willemite, which is being worked on Minehill by the New Jersey Zinc Company, there occur small masses of a beautiful delicate pink mineral with a rhombohedral cleavage, which by their contrast with the purely apple-green Willemite make exceedingly pretty specimens. An analysis gave the following composition: FeC Mgö Insol. MnC 50.40 43.54 0.08 = 100.47 Specific gravity=3.052. Hardness = 4. The mineral differs from the known dialogites by its greater proportion of carbonate of lime, and may be considered either as a dialogite in which a little more than one-half of the Mn is replaced by lime, or as a dolomite in which about five-sixths of the magnesia is replaced by Mn. 3. A pseudomorph of opal after a micaceous mineral probably some chlorite. On Scotch mountain, Warren county, N. J., not far from New Village, among the Laurentian syenitic gneiss formation of that region, there occur, scattered over the ground, numerous masses of a white quartzose mineral apparently of agglutinated rounded granules of about inch diameter. Upon close examination, many of these granules show distinct cleavages, which exhibit a hexagonal outline. Searching the ground carefully I found wormlike contorted crystals, in shape like the similar forms of some chloritic minerals. The substance is distinguished from quartz by its low specific gravity =1.961, and inferior hardness (near 6). It is mostly soluble in caustic potash, leaving only 8 per cent insoluble, which seemed to consist, in part at least, of the original mineral. On ignition it loses 7-27 per cent water. It is therefore manifestly amorphous quartz or opal. Indeed small masses of unquestionable opal of various colors are found in the neighborhood. It hence appears, that micaceous structure is not, as is frequently assumed, the absolute closing scene of the metamorphism of minerals, but that the replacing power of silica is able to overcome the antimetamorphic energies of minerals even, which have arrived at the micaceous stage. Bethlehem, April 22, 1870. ART. VI-On the Size of Atoms; by Prof. Sir W. THOMSON, F.R.S.* THE idea of an atom has been so constantly associated with incredible assumptions of infinite strength, absolute rigidity, mystical actions at a distance, and indivisibility, that chemists and many other reasonable naturalists of modern times, losing all patience with it, have dismissed it to the realms of metaphysics, and made it smaller than "anything we can conceive." But if atoms are inconceivably small, why are not all chemical actions infinitely swift? Chemistry is powerless to deal with this question, and many others of paramount importance, if barred by the hardness of its fundamental assumptions, from contemplating the atom as a real portion of matter occupying a finite space, and forming a not immeasurably small constituent of any palpable body. More than thirty years ago naturalists were scared by a wild proposition of Cauchy's, that the familiar prismatic colors proved the "sphere of sensible molecular action" in transparent liquids and solids to be comparable with the wavelength of light. The thirty years which have intervened have only confirmed that proposition. They have produced a large number of capable judges; and it is only incapacity to judge in dynamical questions that can admit a doubt of the substantial correctness of Cauchy's conclusion. But the "sphere of molecular action conveys no very clear idea to the non-mathe * From Nature, No. 22, March 31. matical mind. The idea which it conveys to the mathematical mind is, in my opinion, irredeemably false. For I have no faith whatever in attractions and repulsions acting at a distance between centers of force according to various laws. What Cauchy's mathematics really proves is this: that in palpably homogeneous bodies, such as glass or water, contiguous portions are not similar when their dimensions are moderately small fractions of the wave-length. Thus in water, contiguous cubes, each of one one-thousandth of a centimeter breadth, are sensibly similar. But contiguous cubes of one ten-millionth of a centimeter must be very sensibly different. So in a solid mass of brickwork, two adjacent lengths of 20,000 centimeters each, may contain, one of them nine hundred and ninety-nine bricks and two half bricks, and the other one thousand bricks: thus two contiguous cubes of 20,000 centimeters breadth may be considered as sensibly similar. But two adjacent lengths of forty centimeters each might contain, one of them one brick and two half bricks, and the other two whole bricks; and contiguous cubes of forty centimeters would be very sensibly dissimilar. In short, optical dynamics leaves no alternative but to admit that the diameter of a molecule, or the distance from the center of a molecule to the center of a contiguous molecule in glass, water, or any other of our transparent liquids and solids, exceeds a ten-thousandth of the wave-length, or a two-hundred-millionth of a centimeter. By experiments on the contact electricity of metals made eight or ten years ago, and described in a letter to Dr. Joule, which was published in the Proceedings of the Literary and Philosophical Society of Manchester, I found that plates of zinc and copper connected with one another by a fine wire attract one another, as would similar pieces of one metal con nected with the two plates of a galvanic element, having about three-quarters of the electro-motive force of a Daniel's element. Measurements published in the Proceedings of the Royal Society for 1860 showed that the attraction between parallel plates of one metal held at a distance apart small in comparison with their diameters, and kept connected with such a galvanic element, would experience an attraction amounting to two ten-thousand-millionths of a gram weight per area of the opposed surfaces equal to the square of the distance between them. Let a plate of zinc and a plate of copper, each a centimeter square and a hundred-thousandth of a centimeter thick, be placed with a corner of each touching a metal globe of a hundred-thousandth of a centimeter diameter. Let the plates, kept thus in metallic communication with one another be at first wide apart, except at the corners touching the little globe, and let them then be gradually turned round till they are parallel and at a distance of a hundred-thousandth of a centimeter asunder. In this position they will attract one another with a force equal in all to two grams weight. By abstract dynamics and the theory of energy, it is readily proved that the work done by the changing force of attraction during the motion by which we have supposed this position to be reached, is equal to that of a constant force of two grams weight acting through a space of a hundred-thousandth of a centimeter; that is to say, to two hundred-thousandths of a centimeter-gram. Now let a second plate of zinc be brought by a similar process to the other side of the plate of copper; a second plate of copper to the remote side of this second plate of zinc, and so on till a pile is formed consisting of 50,001 plates of zinc and 50,000 plates of copper, separated by 100,000 spaces, each plate and each space one hundred-thousandth of a centimeter thick. The whole work done by electric attraction in the formation of this pile is two centimeter-grams. The whole mass of metal is eight grams. Hence the amount of work is a quarter of a centimeter-gram per gram of metal. Now 4,030 centimeter-grams of work, accord-ing to Joule's dynamical equivalent of heat, is the amount required to warm a gram of zinc or copper by one degree centigrade. Hence the work done by the electric attraction could warm the substance by only of a degree. But now let the thickness of each piece of metal and of each intervening space be a hundred-millionth of a centimeter instead of a hundredthousandth. The work would be increased a million-fold unless a hundred-millionth of a centimeter approaches the smallness of a molecule. The heat equivalent would therefore be enough to raise the temperature of material by 62°. This is barely, if at all, admissible, according to our present knowledge, or, rather, want of knowledge, regarding the heat of combination of zinc and copper. But suppose the metal plates and intervening spaces to be made yet four times thinner, that is to say, the thickness of each to be fourhundred-millionth of a centimeter. The work and its heat equivalent will be increased sixteen-fold. It would therefore be 990 times as much as that required to warm the mass by 10 cent, which is very much more than can possibly be produced by zinc and copper in entering into molecular combination. Were there in reality anything like so much heat of combination as this, a mixture of zinc and copper powders would, if melted in any one spot, run together, generating more than heat enough to melt each throughout; just as a large quantity of gunpowder if ignited in any one spot burns throughout without fresh application of heat. Hence plates of zinc and copper of a three-hundred-millionth of a centimeter thick, placed close together alternately, form a near approximation to |