extent supplied the deficiency, and in the case of trichloracetic acid and orthonitrobenzoic acid I have experimentally verified the validity of the dilution law = a2 k. An account of these investigations will be given in the Journal (1-a)v of the Chemical Society' and in the Zeitsch. f. physik. Chemie.' 7. On the Effect of Dilution upon the Colours of Salt Solutions and the Measurement of this Effect. By WYATT W. RANDALL, Ph.D. The writer called attention to the conclusions deduced by Ostwald from the hypothesis of electrolytic dissociation with regard to the source of the colour of salt solutions, and to the work of Knoblauch, Krüss, Traube, Arrhenius, Magnanini, Ostwald, Wagner, and others upon this subject. The experiments and conclusions of Vernon were then more particularly discussed, to show the inaccuracy which the author believes characterised them. A drawing was then shown of an apparatus, through the use of which the author hopes to be able to determine with comparative accuracy the effect of dilution upon the colours of salt solution. Since, according to the theory of electrolytic dissociation, the colour of a salt solution is due to the presence in it of free coloured ions, it follows that with increased dissociation the colour must become proportionally more intense. One of the simplest ways to produce more complete dissociation is by dilution. Hence a dilute solution, ceteris paribus, should, in proportion to the amount of salt contained, show a more intense colour than a concentrated one. In the apparatus shown the author examines the relative length of column of coloured solutions of different concentrations which show the same intensity of colour, in order to determine whether or not the colour remains proportional to the concentration. Several advantages claimed for the apparatus were pointed out. It is, for example, so arranged that the light transmitted through the stronger solution also passes through a column of water equal in length to the difference between the lengths of column of the concentrated and the dilute solutions. By this means any error due to the colour, &c., of the water of the solutions is presumably eliminated. The tubes through which the light is transmitted are silver-plated, and the measurements are made in a dark room, in order that all error due to diffused light may be removed. The results, so far as they have gone, indicate that the colour remains practically proportional to the concentration, whereas in the solutions examined the dissociation varied in amount from about 25 per cent. to nearly 60 per cent. The author, however, desires his results thus far to be considered as merely preliminary. The behaviour of one solution examined suggests that the effect produced upon the colour by varying the rate of dilution, which Vernon claims to have noticed in the case of certain compounds of chromium, may be much more general than that investigator imagined. This point will in the future receive special attention. 8. On the Distinction between Mixtures and Compounds. The distinction between mixtures and compounds, as it is now understood, dates from the controversy between Berthollet and Proust at the beginning of the century. Most text-books state that Proust showed that the same compound always contains the same elements united in the same proportions,' and imply that the statement does not hold good for the same mixture.' Interpreted literally, the statement is a mere truism, and applies equally to both classes of substances. By altering its form somewhat we are led to the postulate: 'Substances in other respects alike possess the same quantitative composition'-a postulate daily made use of in the laboratory, though it is not to be found in the text-books. This postulate was tacitly accepted by Berthollet as well as Proust, and has nothing to do with the distinction sought for; we see, therefore, that the points at issue between the two men have been misrepresented. The author has shown elsewhere' how this misrepresentation has arisen. Proust was unable to furnish the experimental distinction between compounds and mixtures which Berthollet demanded again and again; nor until lately has it been possible to establish one. It is possible that such a one may be based on the recent work of Raoult, who has shown that the melting-point of a pure compound is always lowered and its boiling-point raised by the addition of a small quantity of some dissimilar substance. But it is important to realise that a satisfactory experimental distinction is still a desideratum, and that the only definition that we can give of a compound to distinguish it from a mixture is a theoretical one based on the consideration of molecules. 9. The Atomic Weight of Carbon.2 By Professor J. A. WANKLYN. Members of this Section who, like myself, a third of a century ago were at that time charged with the responsibility of teaching chemistry to the students of a University will have a lively recollection of the incidents attendant on the change of notation at that period. The controversies of that day evolved, as will be remembered, a short and easy method of arriving at the molecular weight of a chemical substance, and likewise a short and easy method of finding the atomic weight of an element. : In Kekulé's words, very slightly modified, these methods were as follows:'Defining standard volume (or, as it was called, the standard two volumes) as that volume which is occupied by two grammes of hydrogen at a given suitable temperature and pressure, we were told that, if we would know the molecular weight of any chemical substance, we must ascertain how many grammes of the substance were required to fill the standard volume with the vapour of the substance, and that that number was the molecular weight. And the atomic weight of an element was to be found by observing what was the very least quantity of that element ever entering into the standard volume filled with a compound of that element.' Having laid down the law much in that style, advocates of the new notation forthwith proceeded to make the practical application by noting that the least number of grammes of carbon ever occurring in the standard volume of any carbon compound was 12-ergo, the atomic weight of carbon is 12. It was at the same time incidentally noted that in all those cases where more than 12 grammes of carbon was found in the standard volume, the number was either 24 or 36, or some other multiple of 12. And so the matter has rested until the present day. I have now to announce that, as the result of most laborious investigation carried on conjointly with my friend and colleague, Mr. Cooper, there exists a multitude of carbon compounds wherein the quantity of carbon in the standard volume is not a multiple of 12, but is a multiple of 6. And the consequence follows that the atomic weight of carbon is 6, as was commonly believed by chemists a third of a century ago. 10. Popular Method for the Estimation of Carbon Dioxide in the Air. By J. B. COHEN, Ph.D., and G. APPLEYARD, Yorkshire College. The method consists in determining the time required to precipitate the lime in dilute lime water containing an insufficient quantity of lime to unite with all the CO2 present. Phenolphthalein is used as indicator, and the end of the reaction is determined by noting the point at which the liquid becomes decolourised. A 22-ounce stoppered bottle is used with 10 c.c. of lime water made from saturated lime water, diluted 100 times with distilled water. One-third of a c.c. 1 Nature, June 14, 1894. 2 See Phil. Mag., May 1894, p. 495. of phenolphthalein solution is added, which is prepared by dissolving 2 grm. of phenolphthalein in 100 c.c. of equal volumes of alcohol and water. 11. On the Diffusion of very Dilute Solutions of Chlorine and Iodine. By A. P. LAUrie. TUESDAY, AUGUST 14. The following Papers and Report were read: 1. Investigations on Tautomerism. By Professor W. J. BRÜHL, 2. On Ortho-dinitroso Derivatives of the Aromatic Series. In a paper jointly published about two years ago by Messrs. Grandmougin, Michel, and myself, the fact has been mentioned that ortho-nitrodiazobenzeneimide, upon being heated with water or distilled with steam, evolved nitrogen, and that finally a new body fusing at 70°-71° was produced. A recent investigation of that body, undertaken in collaboration with Dr. Karl Kohn, has proved its composition to agree with the formula C ̧H ̧ÑO, and the determination of its molecular weight gave numbers corresponding to this simplest formula and not to that of a polymere. The action of reducing agents resulted in the formation of ortho-phenylenediamine. These facts render it probable that the new body is ortho-dinitroso-benzene, a para-isomere of which had already been made known by Nietzki. It was produced by the oxidation of para-quinone-dioxime. The analogous oxidation of orthonaphto-quinone-dioxime into the corresponding ortho-dinitroso-compound has, some time before Nietzki, been carried out by Ilinski. This oxidation is rendered evident by the following formula:— Now the investigation of Mr. Kohn and myself on the products of decomposition of the two ortho-nitronaphtalenediazoimides NO by heat, has proved the identity of the products thus obtained with the orthodinitroso compounds described by Ilinski. analogous to that of Bamberger's nitrosobenzene, is however also possible. No conclusive argument can be adduced in favour of either of them, but, beyond all doubt, our benzene derivatives must have the same constitution as Ilinski's naphthalene compounds. The reaction then proceeds according to the equation A convenient method of obtaining dinitroso derivatives from the corresponding ortho-diazoimides consists in heating the latter in a solution of glycerine at temperatures ranging from about 100° to 120°. When the evolution of nitrogen has ceased, water is to be added and the nitroso-compound thus precipitated is filtered off. By crystallisation from boiling alcohol it may be easily obtained in a state of perfect purity. Ortho-dinitrosobenzene crystallises from its aqueous solution in needle-shaped crystals, from alcohol in plates, fusing at 70°-71°. It is sparingly soluble in water, easily in alcohol, ether, &c. It sublimes readily, and may be volatilised in a current of steam. The odour of its vapour is slightly irritating, and resembles in a dilute state that of nitrobenzene. When heated, either with nitric acid of 65 per cent. on the water bath, or treated with concentrated nitric acid at 0° in sulphuric acid solution, it yields a mono-nitro derivative, crystallising in yellow needles; fusing-point 143°. This mono-nitro derivative has the constitution NO NO NO2 as, by reduction, it yields the known 1. 2. 3, triamidobenzene. We have also obtained an isomeric compound NO NO NO2 by decomposing the dinitrodiazobenzeneimide N N NO2 NO2 in the same manner. in which a methylic group occupies an ortho-position in respect of the N N group, have hitherto resisted every attempt of decomposition by heat, from some cause yet to be investigated. 3. On the Formation of Indazol Derivatives from Aromatic Diazo-compounds. By Professor E. NOELTING, Mulhouse, Alsace. Professor Witt, Dr. Grandmougin, and myself showed a few years ago that the diazo-derivative of nitro-orthotoluidine, fusing-point 107° NO N=N-CI when heated with water, not only yields the corresponding nitrocresol он CH, NO but in about equal quantity a new body, which we proved to be nitro-indazol, |