with the observed facts. Fluoran, though not itself tautomeric, might give tautomeric fluorescent oxonium salts; these have been isolated. The non-fluorescence of the nitro- derivatives of fluorescein is readily explained; the nitro- groups enter into the ortho- positions to the hydroxyl groups, and since compounds of the type —C(NO,) = C(OH)-yield sodium salts which are, in all probability, of the general formula-C(NO,Na)-CO-the fluorescence which depends on a doubly symmetrical tautomerism is necessarily inhibited. Whilst in the greater number of cases the theory propounded agrees with the observed facts, exceptions such as the following must not be overlooked. 1. Substances having the necessary constitution, but not exhibiting fluorescence. A secondary tautomerism might inhibit the vibration between two extreme similar configurations; this case has been considered in dealing with the nitroderivatives of fluoresceïn. In some cases it is possible that fluorescence has not been detected owing to the emitted radiant energy corresponding to an invisible part of the spectrum. Another cause which may preclude the necessary vibration is that the symmetrical intermediate configuration may correspond to more molecular free energy than the extreme unsymmetrical configurations. The molecule would then have no tendency to vibrate regularly, the case being analogous to that of an inverted pendulum. 2. Substances which fluoresce but cannot possess doubly symmetrical tautomeric formulæ. The author does not think the occurrence of such substances can be taken as a serious argument against his theory, which may be formulated as follows: If the molecules of a tautomeric substance possess such a structure that the passage from the configuration of least free energy to the less stable configuration may be effected by equal and opposite atomic displacements, the molecules will vibrate between the extreme positions and the substance exhibit the phenomenon of fluorescence. That fluorescence may be due to other causes is not negatived by this assertion. 3. Preliminary Note on some Electric Furnace Reactions under High Gaseous Pressures. By J. E. PETAVEL and R. S. HUTTON. The paper gives an account of some work carried out in an inclosed electric furnace constructed to work with gaseous pressures up to 200 atmospheres. The power employed has been usually about 15 kilowatts per hour, the furnace containing a charge of about 20 lb. of material and 1,000 to 2,000 litres of gas. A second furnace of about one-tenth the capacity was used for gas reactions with high-tension current. The reactions at present under investigation include the direct reduction of alumina by carbon, the conditions of formation of calcium carbide, particularly as modified by the change of gaseous atmosphere, and the formation of graphite. With regard to gaseous reactions a study of the production of nitric acid and cyanogen compounds has already been commenced. The preliminary experiments have shown that under pressure alumina is reduced to the metallic condition, but in all cases accompanied by a large amount of aluminium carbide. This reaction is most unfavourably influenced if the carbon monoxide which is formed be retained, whereas it is favoured by the rapid removal of the gaseous products of reaction. So far as calcium carbide is concerned, contrary to expectation, the yield is in no way diminished by the presence of carbon monoxide gas even at high pressures. An important difference in the methods of working is necessary in those cases where it is desired to effect purely gaseous reactions. Here a high-tension current is required. For instance, the formation of nitric acid, even at pressures of 100 atmospheres, is only accomplished in appreciable amount where the electromotive force used is of several thousand volts. The account includes a general description of the plant employed for preparing and compressing the various pure gases required in quantity for this work. 4. The Atomic Latent Heats of Fusion of the Metals considered from the Kinetic Standpoint. By HOLLAND CROMPTON. X According to the kinetic theory, if a fluid is composed of monatomic molecules, the kinetic energy given to the particles of the fluid on heating, if no external work is done, is equivalent to 296 T calories for the gram-molecule. If it is assumed that in the solidification of such a fluid the process consists mainly in bringing the molecules (in this case atoms) to rest, or in largely restricting their motion, kinetic energy approximately equivalent in amount to 2.96 x T calories should be lost by the gram-molecule. Hence in the above case the molecular (atomic) latent heat of fusion Ar should approximate to 2.96 × T calories, or Ar T = 2.96. As a matter of fact, the values of Ar/T in the case of fourteen different metals, presumably of monatomic structure, vary between 1.82 for potassium and 3.05 for tin, the average value being 24. Up to the present only two cases have been met with among the metals in which the determined values of r do not bring Ar/T within these limits, these being gallium and bismuth, for which the values of Ar/T are 4.67 and 4.82 respectively. 5. The Influence of Small Quantities of Water in bringing about Chemical Reaction between Salts. By EDGAR PHILIP PERMAN, D.Sc. Many experimenters have investigated the influence of traces of moisture in reactions between gases, but so far as I am aware no one has hitherto made similar experiments with solids. The substances chosen for experiment were salts of lead or mercury, and salts of potassium, usually the iodide, which would show the occurrence and progress of a reaction by a colour change. a. Experiments with Lead Chloride and Potassium Iodide.-Equivalent quantities of the two salts were dried over strong sulphuric acid in a suitable apparatus, and then mixed; it was found that after forty-eight hours' drying no visible change took place on mixing the salts, but on keeping the mixture for a week (in a sealed flask) a faint yellow colour appeared, which gradually deepened, until after some months it became a bright yellow. Attempts were made to discover how much water was necessary in order to make the reaction immediately visible; the results were not very concordant, but indicated about 5 mg. as the amount necessary in the conditions of the experiment, viz. two grams of potassium iodide and an equivalent quantity of lead chloride were mixed in a glass flask of about 100 c.c. capacity. b. Experiments with other Lead Salts and Potassium Iodide.-Lead formate and lead nitrate were found to act in a similar way to the chloride. Lead sulphate reacts much more slowly, although exposed to the air, while the carbonate and the oxide react very slowly indeed. c. Experiments with Mercuric Chloride and Potassium Iodide.-Mercuric chloride and potassium iodide treated in exactly the same way as already described gave a strong red coloration on mixing; the same result was obtained when commercial phosphoric anhydride was used as a drying agent. By drying with specially prepared phosphoric anhydride, however, the mixture obtained has been kept for some months without change. d. Other Experiments with Mercuric Salts.-Mercuric cyanide showed no reaction with potassium iodide. Mercuric chloride and potassium chromate reacted very slowly, although exposed to the air. Discussion of Results.-There is no reason for thinking that these reactions take place in any way essentially different from similar reactions in solution, and I believe that the only difference is the extreme slowness of the reaction. It is noteworthy that the velocity of the reaction between mercuric chloride and potassium iodide is enormously greater than that between lead chloride and potassium iodide when dried in the same way. The factors to which this difference may be referred are (1) solubility, (2) volatility, (3) degree of ionisation, (4) specific reaction velocity. We will consider these in order. (1) Mercuric chloride is about ten times as soluble as lead chloride in cold water, but this alone would not account for the difference; e.g. mercuric cyanide is still more soluble, but does not react at all. (2) Judging from the boiling-points, mercuric chloride would appear to be more volatile than lead chloride, the boiling-point of the former being 300° C. and that of the latter about 900° C.; but I find that on aspirating air over each salt and then over potassium iodide, the vapour from the lead chloride affects the potassium iodide much the sooner. The difference in the speed of the two reactions (a and c) cannot therefore be caused by the difference in volatility. (3) The degree of ionisation cannot be the cause of the difference noted, for mercuric chloride is known to be very slightly ionised in solution, while lead chloride may be taken as completely ionised. (4) The specific reaction velocity appears to be the real determining factor, and the reaction is probably of the form AB + CD = AD+ BC. If it is only free ions that react (which seems to me improbable), then the velocity of ionisation in the case of mercuric chloride must be extremely great. There may also be other factors not yet understood. 6. Report of the Committee on the relation between the Absorption Spectra and Chemical Constitution of Organic Substances.-See Reports, p. 126. TUESDAY, SEPTEMBER 15. The following Papers and Reports were read :— 1. Freezing-point Curves for Binary Systems. When a liquid mixture of two components is cooled, a point is reached at which separation of solid takes place. For complete interpretation of the phenomena, it is necessary to know not only (1) this temperature of initial freezing, but also (2) the composition of the separating solid, and (3) that of the liquid from which it separates. The varying character of the relation between (2) and (3) is best demonstrated graphically by plotting the one against the other in a square diagram. It is then found that the cases experimentally known fall into one or other of two classes, according as the composition of the solid varies continuously with that of the liquid, or is constant for certain ranges of concentration, and to that extent independent of the composition of the liquid. To the former class belong systems of two components that form mixed crystals; the components of systems in the latter class do not form mixed crystals, and the definite solids that separate out, each within its own range, are either the pure components or compounds of these. If consideration is contined to the latter class of cases, it is found that on the freezing-point curves (ie. the curves obtained by plotting the temperature of initial freezing against the composition of the liquid) there is a branch corresponding with each range of concentration over which the separating solid is definite and constant. With the intersection of two branches on the freezing-point curve, there corresponds a vertical line on the square diagram, and where the freezing-point curve has an intermediate branch with a summit, there is on the square diagram a horizontal line cutting the diagonal. Further, the composition at a summit point on the freezing-point curve is exactly that of the definite solid separating out over that branch of the curve. Experimental examples of these relationships are supplied, for example, by Roozeboom's work on the hydrates of ferric chloride, Stortenbeker's work on iodine and chlorine, Heycock and Neville's work on gold and aluminium, and the author's work on the freezingpoint curves for mixtures of organic substances. Two cases are specially referred to: (a) where the freezing-point curve exhibits a branch that does not reach a summit; (b) where the two components form a compound that is dimorphous, and the corresponding freezing-point curve exhibits two intermediate branches, the one enveloping the other. Examples of case (a) are furnished by the freezingpoint curve for gold and aluminium, and, to a certain extent, by that for phenol and urea. Examples of case (b) are found in the freezing-point curve for iodine and chlorine, and in that for phenol and p-toluidine.3 2. A Contribution to the Constitution of Disaccharides. It is shown in a recent communication to the Chemical Society that tetramethyl a-methyl glucoside, obtained by the action of methyl iodide and silver oxide on a-methyl glucoside, yields on hydrolysis a well-defined crystalline tetramethyl glucose possessing the ordinary properties of an aldose. The reactions of this substance prove that its unmethylated hydroxyl group is in the y position, and direct evidence is thus obtained of the correctness of Fischer's formula for the parent methyl glucoside. The authors have extended their experiments to the hydrolysable sugars, and the present paper deals with the results obtained in the methylation of cane-sugar and maltose. Methylation of Cane-sugar. An aqueous solution of cane-sugar was mixed with methyl alcohol and alkylated as usual by the addition of silver oxide and methyl iodide. The product, which was readily soluble in alcohol, was then treated as in the alkylation of methyl glucoside, two further alkylations in alcoholic solution and one in methyl iodide being necessary to complete the reaction. The alkylated cane-sugar is a viscid neutral syrup readily soluble in ether, alcohol, or methyl iodide, and showing no action on Fehling's solution. The compound has not yet been obtained in a state of purity, but combustions of the substance dried at 100° in a vacuum and methoxyl determinations by Zeisel's method showed that the alkylation was practically complete. On hydrolysis, which was effected by boiling with dilute hydrochloric acid, the initial dextro-rotation though not inverted was much reduced, and the cane-sugar ether was resolved into a mixture of methylated glucose and fructose. The former proved to be identical with the tetramethyl glucose (m.p. 81°-84°) above referred to; in one experiment the compound crystallised spontaneously from the oily product of the hydrolysis, whilst in other cases it was obtained only after fractional distillation of this product, being then found in the higher boiling distillate. The more volatile fractions, judging from their lower dextro-rotatory power, contained the methylated fructose, but the substance did not crystallise, and it was found impossible to effect a complete separation by vacuum distillation. At our suggestion, Mr. D. M. Paul undertook the preparation of tetramethyl fructose from methyl fructoside, in the hope of establishing the identity of the substance with the methylated fructose produced in the above hydrolysis. The preparation yielded tetramethyl methyl fructoside as a colourless mobile oil boiling at 132°-136° under 10 mm. pressure and having no action on Fehling's solution. 1 Journal of the Chemical Society, 1903, 83, 814. 2 Loc. cit. 3 Loc. cit. The hydrolysis of this compound gave a colourless liquid (b.p. 153°-156° under 13 mm. pressure), which readily reduced Fehling's solution and showed in approximately 5 per cent. alcoholic solution a specific rotation of -21.7°. The substance was evidently comparatively pure tetramethyl fructose. No crystalline derivative of the compound being obtained, we have thus so far no direct evidence of its production in the hydrolysis of alkylated cane-sugar. The production from cane-sugar of the identical tetramethyl glucose previously obtained from methyl glucoside shows that the linking of the glucose residue in the former is the same as in the latter compound, and consequently Fischer's formula,1 so far as the glucose half of the molecule is concerned, is proved to be correct. Alkylation of Maltose. The method adopted was similar to that already described for cane-sugar, save that no solvent water was required. The process was at first attended with an appreciable amount of oxidation, and after the first treatment the syrup obtained was acid in reaction. The final product was a thick neutral syrup without action on Fehling until hydrolysed. The assumption is that the free aldehyde group had been oxidised, and subsequently methylated. The substance was hydrolysed by boiling for an hour with 1 per cent. hydrochloric acid. The solution was neutralised exactly with barium hydrate, evaporated in a vacuum at 60°, and the residue extracted with boiling alcohol. A mixture of the methylated glucose with the barium salt of the oxidation acid was thus obtained, from which the alkylated sugar was extracted with ether. From this extract tetramethyl glucose was obtained on distillation in a vacuum, and the substance, after removal of a trace of organic acid and some incompletely alkylated glucose, was finally obtained crystalline. After recrystallisation from ligroin the compound melted sharply at 84°, and its identity was further confirmed by analysis. The acid produced by the oxidation was recovered from the barium salt. It distilled in a vacuum apparently without decomposition, and the figures obtained from combustions and methoxyl determinations agreed approximately with those required for tetramethyl gluconic acid. The above results show that the linkage of the glucose residues in maltose is not of the acetal, but of the glucosidic type, and are in agreement with the formula CHO-(CHOI), 'CH,-0–CHI•(CHOH),•CH CHOHCH,OH suggested by Fischer.2 2 The alkylation of polysaccharides and glucosides, and the identification of the alkylated products obtained by hydrolysis, seem to furnish a general method for elucidating the constitution of these compounds, and the authors are continuing their experiments in this direction. 1 Ber. (1893), 26, 2405. 2 Loc. cit. |