REFERENCES. 1 Etard, Compt. rend., 1877, LXXXIV., 127; ibid., 1878, LXXXVII., 989; Ann. Chim. Phys., 1881, [5], XXII., 218. Henderson and Campbell, Journ. Chem. Soc., 1890, LVII., T. 253. 'Henderson, Munro and Campbell, ibid., 1907, XCI., T., 1871. 4 5 ་ Henderson and Cameron, ibid., 1909, XCV., T., 969. Henderson, Brown and Robertson, ibid., 1922, CXXI., T., 2717; Robertson, ibid., 1924, CXXV., T., 765. Bredt and Jagelki, Ann., 1900, CCCX., 112. Henderson and Heilbron, Journ. Chem. Henderson and Gray, ibid., 1903, Henderson and Heilbron, ibid., 1908, • Henderson and Chisholm, ibid., 1924., CXXV., T., 107. 10 Wallach, Ann., 1908, CCCLIX., 266. "Rhode and Miller, Ber., 1890, XXIII., 1070; ibid., XXIV., 1856. "Henderson and Gray, Journ. Chem. Soc., 1904, LXXXV., T., 1041. BRITISH ASSOCIATION-SECTION B. DISCUSSION ON THE ALTERNATING EFFECTS IN CARBON CHAINS. September 1st, 1925. (Summary of the Positions taken up by the Different Participants in the Discussion.) PROFESSOR A. LAPWORTH. The term "polarity theory" is a wide one, covering many views and may not arbitrarily be restricted to those which are expressed in terms of electronic valency conceptions. His own views were based on and regulated by his hypothesis of 1901, which postulated alternate effects, recognised by chemists even before that date, and was in fact originally devised to explain alternate effects. The hypothesis assumed that many organic chemical reactions proceed by formation and decomposition of complex ions, broadly on the lines of the theory of Abegg and Bodländer. The types considered in 1901 were the simple ones the superimposed signs denoting electrical changes, localised or diffuse, as the case may be. If the new complexes to be produced belong to a type which is "stable"-that is, associated with a low free energy or high entropy-then the reactions are likely ceteris paribus, to be the more readily formed. The usefulness of the hypothesis depends largely on a previous knowledge of what are 66 99 stable forms in the sense indicated, but many types are quite well known and easily recognised. For example, many anions of the general form R-0-0 typified by the anions of water, alcohols, phenols and carboxylic acids, are well known to be water-stable, the stability varying with the composition and structure of R. As an illustration of this idea it was suggested (1901) that acetanecy anohydrin is formed as follows CN which leads at one step to a form of recognisably "stable" type, and it was afterwards proved that such ions actually exist in aqueous solutions of ketone- and aldehydecyanohydrins, some being so stable that the crystalline pottassium salts can be obtained from aqueous solutions. The corresponding form of the complex ion from, say, crotonaldehyde, would arise as follows: The application of the same principle to the case of the union of an aromatic amine at the para position with a positively charged ion, such as a diazonium ion, had been referred to by Professor Lowry; it was one which the speaker had entirely overlooked until Professor Robinson entered the field. The extension of the complex ion hypothesis to the addition reactions of ethylenic and aromatic hydrocarbons is much more speculative, but the possibility at least suggests new lines of experimental enquiry. Formulæ containing free polar partial valencies or free polar affinities are doubtless speculative, but as electrovalence is the only kind of externally directed attractive force between molecules which has been recognised, formulæ with free non-polar affinities are more speculative still. The speaker had used them with the polar sense only in so far as they seemed to run parallel with the requirements of otherwise sound chemical theory; but he was beginning to believe that they often had a mere mnemonical significance. The question, "From where is the necessary energy required to produce polar activated forms supposed to come?" had various answers. In cases where activation is due to induction from, or association with, pre-existing polar molecules (as assumed in some cases by Thomson), then the answer was not quite the same as when activation is what may be termed "automatic," or better, perhaps, " purely thermal." The speaker had found that this latter case, really the simplest of all, was still widely misapprehended, and he therefore thought it worth while to paraphrase very briefly a fundamental wellknown principle. The most stable homogenous systems are not those which contain all their molecules in the most stable forms ;for example, if a gas molecule is capable of existing in two forms then at complete thermal equilibrium, the gas, in the absence of any absorption of radiant energy, will in general contain both types of molecules in the ratio K (the familiar Mass-law constant) and K may be calculated from the relation : F= RT log K where F is the amount of work required to convert a gram molecule of the more stable form of the gas into the less stable form without change in volume or temperature. The energy, if any, required for the con version of the one form into the equilibrium quantity of the other form, comes from the same source as that required for any endothermic reaction, namely, from the heat of the surroundings if the reaction is adiabatic. These are points which broadly do not admit of dispute except in slight academic details. It may be contended that F for activation is so large that K is negligably small, but that is an objection which can be tested in any given case when the value or even the order of magnitude of F has been discovered. PROFESSOR R. ROBINSON. Dr. It has been stated by Flürscheim that it is not possible to construct a hydrogen molecule with shared electrons without conflicting with the Rutherford-BohrSommerfeld theory of the atom. The mathematical difficulties may be great indeed, but it is unnecessary to regard them as insuperable, especially in view of the fact that Bohr's theory already includes postulates which are foreign to classical dynamics. When Copernicus constructed his system he might well have been met with the criticism, criticism, analogous to Flürscheim's present argument, that all the mathematical problems of the solar system had not yet been solved. The theory of shared electrons supplies a convincingly simple explanation of a large number of numerical coincidences, and it is unquestionably a useful working hypothesis. In the speaker's view, the charges, small or large, of ions or polarised complexes, merely function in facilitating a reaction by bringing the parts into contact. The process of chemical change which then occurs will include the reduction of the external fields, but will by no means be confined to this. In agreement with Lapworth, an alternating polar effect, represented + − + − + occasionally by symbols such as A B C D E, is not taken to mean the simultaneous exhibition of all the charges represented. These symbols are mnemonics combining in one expression all the possible polarised forms of the molecule. In view of the development of the underlying electronic theory the use of such symbols may well give place to that of arrows showing the direction of displacement of electrons due to the general electrical effect or to conjugation. In nitrobenzene, for example, we have : The process (B) inhibits o-p substitution and (C) inhibits p-substitution and assists o- and m- substitution. The signs in (A) seen in the light of the particular explanation (B) could well be omitted, although it is not necessary to do so since + and are symbols having a significance relative to some standard only. Much has been said about " point addition," and this has been contrasted with addition to a conjugated system. The speaker considered that this contrast is misleading and unnecessary. In the many cases where point addition of an ion occurs an electrical charge appears at some other point. Thus in the case of dimethylaniline, there is an electron flow towards the p-position and the reaction with a diazonium salt consists in the addition of the diazonium ion to the p-position and of a positive charge to the nitrogen atom. Neutralisation of the latter synchronises with loosening of the p-hydrogen atom and the completion of the reaction may be due to oscillations of the electrons in the system, each of which produces a very small charge on the nitrogen atom. Every surge of electrons from the nitrogen atom enables the carbon atom in the p-position to tighten its grip on the diazonium ion and every return to the normal similarly allows the proton in the p-position to escape from the molecule. (To be Continued.) H Mez H D3 NMe 2 ON THE TERNARY DIAGRAM OF THE (Science Papers of Tohoku Imperial INTRODUCTION. The investigation of the effect of copper on the physical properties of steel is a matter of great importance for the metallurgy of iron in our country; because in this country iron ores almost always contain a small amount of copper. According to the previous results, when the copper content exceeds some 5 per cent., the steel is red short and difficult to forge; if, however, copper content is lower than this amount the steel can be either rolled or forged. As regards the properties, at ordinary temperatures of steels containing copper, it is known that hardness and strength as well as their resistance to chemicals are promoted by the addition of this element. This advantage effect of copper on steel is similar to that produced by nickel or chromium and for this reason the investigation of copper steel has been carried out by many metallurgists such as Stead, Brenil, Shirotani, and others. SUMMARY. 1. Theoretical: The equilibrium diagrams for the systems cementite-copper and ironcementite-copper have been considered; especially in the latter case, the following equilbria have been dealt with : (i) Peritectic equilibrium of pseudo- (ii) Invariant equilibrium consisting of 2. Experimental: The portion of the ternary diagram representing the state of the alloys containing less than 30 per cent. of copper has been determined. Microscopic: The binary and the ternary eutectoid structures were determined. NATURAL IONIZATION IN GASES Measurements of the pressure variation of the natural ionization in air, nitrogen, oxygen and carbon dioxide contained in a steel sphere of one foot inside diameter, made by H. F. Fruth for pressures from 1 to 75 atmospheres, were repeated with Results obtained slight modifications. 66 corresponded with those of Fruth with the following exceptions: (1) although the pressure rate of variation of the natural ionisation at high pressures was very small (0.4 to 1.0 ion/cc sec. atm.) no abrupt change in the rate of variation and no true saturation " value of the ionization were "anomalous "behaviour observed; (2) no was noted in the case of commencial nitrogen. Measurements to detect the variation with size of chamber were carried out with spherical iron ionization chambers of 7, 9, and 12in. respective internal diameters. The rate of production of ions per cc at high pressures did not vary unidirectionally with the size of the ionization chambers, but was least in the 9in. sphere. To study the effect of the lining the experiments were repeated with oxygen in each of the spheres lined with aluminium of 1/16in. thickness, and in the 12in. sphere lined with copper of 1/16in. thickness. The aluminium lining increased the ionization at low pressures, and decreased its pressure rate of variation at high pressures. The effect of water vapour on the natural ionization in aged, dust-free air, or upon its variation with pressure was found to be negligible. Both the absolute and the fractional variation of the natural ionization with time the were found to depend upon pressure, the average value of the former during a 24 hour period being about twice and that of the latter being about 1/5 as great in oxygen at 5.4 atm. pressure as in air at atmospheric pressure. Very slight evidence of a singly periodic diurnal variation in air observed. at atmospheric pressure was Attempts to determine the effects of slow chemical reactions upon the natural ionization in air and oxygen at atmospheric pressure led to negative results. waste) in 1925, totalled 258,017 tons, or 53,268 tons in excess of the quantity of the quantity exported in 1924. Shipments to Great Britain amounted to 8,709 tons, to United States 209,879 tons, and to Germany 11,120 tons. Exports of asbestos to Australia, Belgium, France, Italy and the Netherlands increased materially. In the asbestos mining industry in Canada during 1925 the amount of capital employed by the fourteen producing firms was $38,138,046. Employment was furnished 117 salaried employees, 1,315 mine workers and 1,150 mill workers and their combined earnings amounted to $,997,107. hour day was usual throughout most of the field. Fuel and electricity cost $923,239, including $546,175 for 60,508,285 kilowatt hours of electricity. Plant equipment included 406 electric motors having a combined rating of 28,621 h.p. A ten Six plants in Canada manufacture asbestos products, including the following commodities: asbestos paper and mill board; asbestos roofing of all kinds ;asbestos rigil shingles; asbestos building materials; asbestos cellular and sponge-felted pipe insulation; insulating sheets and blocks; asbestos brake linings and clutch facings (woven on special looms); and asbestos packings for steam, oil and hydraulic operations. PROCEEDINGS AND NOTICES OF SOCIETIES. THE ROYAL SOCIETY. Papers read on Thursday, June 24, 1926. On the Structure of the Arc Spectrum of Gold. By J. C. MCLENNAN, F.R.S., AND A. B. McLAY. We have been able to classify by our term scheme nearly all the wave-lengths known to belong to the gold arc-spectrum. The most important wave-lengths not yet classified are six, given in a table, that were found to be absorbed by the vapour in the under-water spark of gold. These wave-lengths undoubtedly involve the metastable term 1D, and will probably be found to lead to the indentification of certain of the terms of the group F2D2P2 with the limit D1 in the table, 2b, or to certain of those of the groups in the table, 2c. But we have not any further evidence in support of this view. Zeeman-effect experiments will probably furnish the best means of definitely settling this question and any other features of the arc-spectrum of gold not yet clear. On the Series Spectra of Palladium. By J. C. MCLENNAN, F.R.S., AND H. G. SMITH. The regular arc-spectrum of palladium, involving disturbances of a single electron outside a core of nine 4, electrons, includes very nearly all the strong lines of the arc, and most of the faint lines of wave-length longer than 3400 A. In the region of shorter wave-lengths there are also a large number of faint lines, apparently belonging to the arc-spectrum, for which no place can be found in the regular series system, since all predicted for prominent electronic configurations have been found, and any terms due to additional possible configurations would gives lines in infra-red or extreme ultraviolet. These faint lines can probably be ascribed to a secondary-series systems, with terms arising from configurations of type (43) 5, nk. The analysis of the spark-spectrum has shown that (4,) 5, and (4,)" are of almost equal stability, and consequently terms of this type should occur with considerable prominence in the arc-spectrum, but might not combine readily with the regular terms. An attempt to find a clue to this part of the spectrum by means of the inter-combinations has been unsuccessful, and no serious effort has been made to analyse it separately. Similar configurations are also likely to be prominent in the sparkspectrum, accounting for the very large number of lines, only a small fraction of which have been included in the present analysis. The Mobility of Ions in Air. By A. M. TYNDALL AND L., R. Communicated by Prof. A. P. F.R.S. Part III. PHILLIPS. Chattock, Measurements of the mobility of ions have been made in air containing organic vapours over a range of concentration extending up to saturation. In every case a reduction in mobility is produced by the addition of vapour, though the amount depends upon the constitution of the The gradient of the mobility vapour general steep at low concentrations, but gous series of normal aliphatic alcohols the steepness increases as one ascends the series, the effect of the higher alcohol being very striking. Similar effects have been observed for the positive ion, but the initial drop in mobility is much less marked. The factors entering into the cluster theory of gaseous ionic structure are |