dition alone will his representation be true and faithful. How far has this condition been fulfilled in the most celebrated pictures? in other words, how far are the master-pieces of art true to nature? This is the problem M. Jamin has proposed for solution by the aid of optical science. More fortunate than the painter, the optician, knowing the imperfections of the eye, has invented photometrical apparatus, by which he can compare the brilliancy of neighbouring objects, and numerically express their relative illumination. By the aid of such apparatus, for instance, he ascertains that the shadow of a stick cast upon white paper has a twentieth of the brilliancy of the portions directly illuminated by the sun. M. Jamin himself has invented one of these precious instruments, of which we shall try to give a general idea. Imagine a small telescope like a single-barrelled opera glass. By putting the eye at the front we see that its interior is divided by a partition. On looking at an object through one of the compartments, a neighbouring object can be seen through the other; and by turning the tube upon itself, the partition may be made to coincide with the line of separation of the two objects. Close to the eye the instrument carries a movable graduated circle. If, continuing to regard the two objects, you turn this circle, you will remark that one becomes more distinct, while the other darkens. Soon the darker object becomes extremely black, while the other attains its maximum brilliancy. The graduation of the circle is so arranged as to show the difference in brilliancy of the two objects by the number of divisions which this circle has to be turned from the zero (found as above) until the objects appear in the field of view with an equal degree of brilliancy. To understand this better, conceive the shadow of a house cast on a white wall. Let us direct the telescope on the boundary line of the shadow; we see in one compartment the brilliant surface, and in the other the shadow; let us now turn the circle until the two parts acquire the same brilliancy, or until we see an equally illuminated surface. The divisions in the circle will then show that the mark which stood at zero at the commencement of our experiment has moved to 20, showing that the illuminated part of the wall is twenty times more brilliant than the shaded portion. Had the wall been yellow, blue, or any other colour, we should have found the same result. Instead of the wall and the shadow of the house, we might consider the ground and the shadow of a tree, a sunbeam and a shadow cast anywhere, the lines of separation of light and shade in a landscape, of a building and the sky, of blue sky and a cloud, etc.-in every case we would have obtained numbers expressing the relative brilliancy of objects contiguous to the field of vision, provided always that the photometer be suitably modified, not only according to the brilliancy, but the tints or colourings of the contiguous objects.

Let us now suppose an artist to have reproduced in a landscape a wall with a shadow, a piece of ground with the shadow of a tree, etc., and let us try to investigate the truth of his representation. The operations are precisely similar to those already described when examining the relations of the objects themselves.

M. Jamin states that after having submitted to the test of his photo

meter a great number of pictures, he has arrived at the unforeseen result, that in almost all, the proportional relations of the lights differ from those of nature. Always, or almost always, the shadows are not sufficiently deep; light and shade in pictures have also different colouring, so that the photometer, such as described, cannot, as in nature, render the apparent brilliancies of objects exactly equal. A twofold incorrectness is thus everywhere indicated, incorrect proportions of lights, false imitations of tints. Had these deviations from nature been trifling, painting might be admitted to be an approximate imitation of nature; but they are on the contrary very considerable. In the simple case of a body illuminated by the sun, and a shadow cast upon it, the relations found in summer, winter, different hours of the day, fine and bad weather, have been extremely varied. In general the minimum value of the relation of light and shadow is 10, its maximum value 20. But when sunbeams in pictures are examined, we find the above relations comprised between 2 and 4, so that the brilliancy of the sun-light is incomparably weaker in the pictures than in the true landscapes. It is difficult to conceive how the eye can

tolerate such considerable inaccuracies. Still, all landscape painters do not deserve this reproach in the same degree; the modern school has made great progress towards exactness; every one may remark that their pictures have deeper shadows and brighter lights; some pictures of Decamps, for instance, present luminous effects comprised within the limits assigned by nature.

The discordance between nature and art in night pictures is not less remarkable. If in one of these pictures, usually lighted by a murky lamp, we compare the light of the lamp with the best illuminated parts, we shall find a relation comprised between 20 and 30. By placing in a room a lighted candle and a sheet of white paper, the ratio of the light of the candle to that reflected from the paper, will be found to be 1500; the candle flame is thus 1500 times as luminous as the paper, while in a picture it is made scarcely 30 times as luminous.

In the most celebrated interiors of Granet, the sky is 4 or 6 times brighter than the window-sashes of the rooms. To test this relation, M. Jamin selected a room with newly-painted sashes, which presented some similarity with those represented in Granet's pictures. By placing his photometer before the window, he found the sky 400 times brighter than the sashes. M. Jamin admits from trial the impossibility of imitating nature in this matter.

Let us now consider a complete landscape: in the foreground, masses of earth, trees, or buildings; in the middle distance, similar objects, seen through a stratum of air, which forms a kind of luminous veil, and increases their brilliancy; in the background, mountains, which blend themselves with the sky; the clouds, whose light far surpasses that of terrestrial objects; the sun, finally, whose dazzling splendour no eye can bear. Measured by the photometer, the luminous intensity of the clouds is several thousand, sometimes several million, times as great as a tree close to the observer. What can the painter do to imitate the infinite gradations in such a scale, when his brightest white has only a

relation 80 times as great in the photometer as ivory-black? If he would adhere to the truth, he would be forced to recognize the existence of scenes which he should not attempt to paint; he should banish the clear sky from his pictures, and never try to represent brilliant clouds. But he can do better by consulting his imagination more than his eye, his interpretation of the reality rather than reality itself, and he will produce a picture possessing indeed only a fictitious reality, but still, charms of life and spirit that will render it acceptable and admired.

To recapitulate, painting is not, as too often supposed, an imitation of nature, but an admitted fiction whose productions do not possess physical reality. Moreover, were it attempted to give the art this character of reality which it wants, insurmountable material obstacles would arise. So that, of all schools of art, the least rational is the realist. Assuming to be accurate, the realists should be guided by photometry; but the photometer proves that they have not even approximated to nature in the relations of their lights. M. Jamin concludes by expressing his satisfaction that physical science should recall to painting its spiritual tendency, which seemed of late about to be forgotten.

CHEMISTRY.

5.-Electro-Chemistry.

Professor Miller, of King's College, London, in presenting the first part of a "Report on the recent progress of Electro-Chemical Research", made some observations bearing upon the binary theory of salts, which are especially interesting and important, coming as they do from one who was hitherto one of the ablest supporters of the theory. According to Dr. Miller, the inquiries made of late years in the field of electrochemistry were characterized rather by modifications of the laws previously admitted, than by any striking or important additions to the stock of scientific truth. Adverting to Faraday's observations of the exceptional conducting power of solid sulphide of silver and one or two other bodies, he stated that it had been shown by the researches of Beetz and Hittorf, that in these cases a true electrolytic decomposition occurred, a circumstance rendered possible by the somewhat viscous condition of the substances which exhibit this anomalous character. The true electrolytic nature of the decomposition was proved, firstly, by the rise of conducting power occasioned by rise of temperature (whereas in metals the effect of heat is exactly the reverse); and secondly, by the effects of polarization observed upon the electrodes between which such bodies are placed.

Allusion was next made to the insulation of metallic bodies by Bunsen, who had shown that in many instances, as in the decomposition of a solution of sesquichloride of chromium, the deposit upon its negative pole

could be made to assume the reguline form, by reducing the surface of this plate to dimensions considerably smaller than the positive plate; a result probably owing in part to the secondary decomposition produced in the limited portion of liquid around the wire, whereby the sesquichloride was reduced to the protochloride of chromium, and subsequently the metal itself was deposited. This view was rendered probable by observing the effects obtained during the electrolysis of sesquichloride of iron, in which these successive stages could be distinctly traced. In cases in which, like the chloride of manganese, the compound was already in the form of protochloride, it was a matter of slight importance whether or not the negative electrode presented a smaller area than the positive electrode. Attention was called to the fact pointed out by Faraday, of the non-existence of more than one electrolyte in multiple series. In the case of the protochloride and of the bichloride of tin, the protochloride only is an electrolyte while in the anhydrous condition. The bichloride is not an electrolyte. Yet, when dissolved in water, itself also not an electrolyte, the solution conducts freely, and a similar result is observed in other analogous cases.

Referring to the decomposition of salts in solution, the bearings of electrolysis upon Davy's binary theory of the composition of salts, was briefly alluded to, and some of the difficulties attending the adoption of this theory in the case of the subsalts were mentioned; these facts, taken in connection with those already alluded to in the case of the bichloride of tin, leading the author rather to the view that a salt is to be regarded as a whole, susceptible of decomposition in various modes, and therefore admitting of representations under two or three different rational formulæ, each of which may, under particular circumstances, be advantageously made use of.

In the discussion which followed the statements of Dr. Miller, Dr. Apjohn observed that the advantages in certain cases, of reducing the dimensions of the anode had been well understood previous to the experiments of Bunsen. It was well known that when Wollaston decomposed water by a succession of electric sparks, he employed this expedient, no doubt because he had ascertained that it facilitated the electrolytic action of the interrupted current on the water.

6.-On Proto-Sulphide of Carbon. By M. E. BAudrimont.

The proto-sulphide of carbon, CS, may be obtained by any of the following reactions: 1. In decomposing the vapour of the bisulphide of carbon, CS2, by spongy platinum or pumice stone heated to redness; under those circumstances CS2 is decomposed into CS and sulphur, which deposits on the sponge and obstructs its further action; 2. it is obtained during the preparation of bisulphide of carbon, and simultaneously with it; 3. by the decomposition of the vapours of CS at a red heat, in contact with pure lampblack, wood charcoal, and especially animal black in fragments; 4. by the decomposition at a red heat of the vapours of CS, by hydrogen; 5. by the calcination of sulphide of antimony with

excess of carbon; 6. by the reaction, at a red heat, of oxide of carbon on sulphide of hydrogen CO+HS=HO+CS; 7. by the reaction of sulphurous acid on olefiant gas at a red heat; 8. by the reaction of olefiant gas on chloride of sulphur at a red heat; 9. by the decomposition of sulphocyanogen by heat, etc.

The first process gives the gas sufficiently pure; the other methods give it mixed with sulphide of hydrogen and oxide of carbon. It may be purified by passing it rapidly through a solution of acetate of lead and protochloride of copper dissolved in hydrochloric acid, then drying it, and receiving it over mercury. It is a colourless gas, having an odour which reminds one of the common bisulphide of carbon, but not disagreeable, and strongly etherial. It burns with a beautiful blue flame, producing carbonic acid, sulphurous acid, and a little sulphur. Its density is a little greater than that of carbonic acid. It is not liquefied by the cold produced by a mixture of ice and salt. Water dissolves about its own volume of the gas, but it soon decomposes it into sulphide of hydrogen and into oxide of carbon HO+CS=CO+HS. It is scarcely more soluble in alcohol or ether. It is not absorbed by a solution of protochloride of copper. A solution of acetate of lead is not immediately blackened by it, but is after several hours' contact with it, and in the course of several days the gas is completely decomposed, oxide of carbon and sulphide of lead being formed. It is rapidly decomposed in contact with alcaline solutions. With lime-water, for example, the reaction gives sulphide of calcium and an equal volume of oxide of carbon, CaO+CS=CaS+CO. No carbonate of lime is produced in this reaction. At a red heat it is slightly decomposed: 1. by spongy platinum; 2. by the vapour of water into HS and CO; 3. more easily by hydrogen into HS and carburetted hydrogen; 4. by copper into sulphide of copper and carbon (graphitoïdal); 5. lastly, mixed with an equal volume of chlorine, a partial condensation takes place, the products formed being the subject of present research by the author.

On analysis made with oxygen in the eudiometer, it yielded equal volumes of carbonic and sulphurous acids, from which the formula CS is deduced. Its composition is, however, perfectly established also by its reaction with lime water, mentioned above. This is also corroborated by the determination of the quantities of carbon and sulphide of copper obtained when the gas is made to act upon that metal. The author promises a complete study of this body, which, he says, several chemists attempted to obtain, but heretofore without success, probably in consequence of water and alkaline solutions transforming it into oxide of carbon and sulphide of hydrogen. M. Persoz has, however, since drawn attention to a passage in his work, Introduction à l'Etude de la Chimie Moléculaire (Strasburg, 1837-38, page 117), in which he distinctly points out the fact of the formation of the compound CS during the preparation of bisulphide of carbon, especially if the vapour of sulphur be not rapidly and abundantly produced. He did not, however, examine its properties very minutely.-Comptes rendus, t. xliv. (11th of May, 1857), p. 1000.