Of the chlorides to distil over into B. The temperature was raised to 250°, and the distillation was continued for about one bour. The spark spectrum of the material that condensed in B showed the presence of a preponderating amount of gallium, and the lines of indium were only faintly visible. The material remaining in the tube A showed strong lines of gallium, indium, and zinc. Purified gallium chloride from the tube B (Fig. 1) was again distilled in chlorine, the side-arm for the appiratus in this case being fused to a tube of Jena glass that had two depressions, BC (Fig. 2). Gallium chloride was Dry chlorine A 257 These distillations were carried out in a tube of Pyrex glass of about 20 mm. external diameter, bent into a series of nine depressions (Fig. 3). The tube was constricted between the depressions. 5'3978 grms. of the original alloy was placed in the first bend A, and dried chlorine was conducted through a glass tube, the end of which dipped below the surface of the alloy. The reaction between the chlorine and alloy was started by beating the latter with a Bunsen flame. The action was at first violent, the union of the chlorine with the metals being accompanied by the emission of a bluish light and a play of incandescent 8 To hood particles throughout the tube in the neighbourhood of the sample. When a layer of the molten chloride had formed above the metal the chlorination proceeded quietly. After the metals in the alloy had been completely converted to chlorides, the portion of the tube marked a was surrounded by the asbestos oven mentioned above and was heated from 220-250°. A portion of the chlorides distilled over into depression 1. The chloride was sucessively distilled in this manner from one section to the next until it finally was collected in the last section, No. 8, of the tube. The distillation of the chloride from one depression to the next was effected by slipping over the glass tube the barrel of a porous cup from which the bottom had been cut off, and which had been wrapped with nichrome ribbon so that it could be electrically heated. This heating tube was about 7 cm. in diameter and 15 cm. long. In distilling the chloride from section I to section 2 the heating tube was placed around depression 1, and the current was adjusted to raise the temperature to 254°. In the succeeding distillations the heating tube was moved along until it covered the section containing the chloride and the current was adjusted to give a temperature ranging from 235-245°. A slow current of chlorine was passed through the tube during the distillations. The spark spectrum of the material which condensed in section 8 of the tube, and which consisted chiefly of gallium chloride, still showed faintly the lines of indium and zinc. The indium spectrum was stronger than from the material obtained in the preceding experiment. The failure to obtain pure gallium chloride by these eight distillations may have been due either to (1) insufficient vertical rise between the different depressions in the distilling tube, which might allow the chlorides of zinc and indium to be carried along mechanically; or to (2) over-heating during the chlorination of the alloy, which might cause either the distillation of some zinc chloride, or the distillation of indium monochloride or indium dichloride before the conversion of the indium to the trichloride was complete; these lower chlorides of indium would first be formed, and they are more volatile than the trichloride. To remove the first of these possible difficulties an apparatus of different form was constructed (Fig. 4). This was blown entirely of Jena glass, and the tubes A and D were fitted at the top with widened inlet tubes that were ground into the necks of the two distilling tubes. The gallium chloride from the two preceding experiments, which was nearly free from zinc and indium, was melted Glass-wool that had preand poured into the tube A. viously been washed with hydrochloric acid and carefully dried was inserted in A and D around the inlet tubes. The of a and D. A This was stopped in the next distillation by sealing these joints with water-glass. A second portion of gallium weighing 9.2982 grms. was chlorinated and distilled, and then a third sample weighing 8.4075 grms. was later chlorinated and distilled. The residue left in a after the first two samples had been distilled over was tested for the presence of zinc with the microscope, using solid potassium mercuric thiocyanate (Behrens, Zeit. Anal. Chem., 1891, xxx., 141), and zinc was found to be present. The residue remaining in c where the chlorination was carried on and the chloride that condensed in D gave no This method, however, has test for zinc by this method. been found to be not wholly reliable for the detection of minute amounts of zinc in the presence of large amounts "The Mercuric Sulphoof gallium (Ruth E. Chipman, cyanates as Reagents in Microscopic Qualitative Analysis," a thesis presented to the Faculty of the Graduate School of Cornell University, June, 1917). In an attempt to increase the delicacy of this microscopical test for zinc the gallium chloride was first volatilised at 250° and the test was then applied to the residue, but the results were not successful. The most satisfactory method for determining the purity of the gallium chloride prepared by the above method was found to be the photographing of the arc spectrum of the material (see ante). This redistilled gallium chloride was found to be spectroscopically free from zinc and indium, and by the method above described it was shown that it could not contain more than o'005 per cent of either of these two elements. Two successive distillations of gallium chloride in the apparatus shown in Fig. 3 therefore yield a product of high purity. (To be continued). PRESENCE OF TUNGSTEN.* By G. WATSON GRAY and JAMES SMITH (Liverpool). chlorides were distilled from A into D in a current of dry THE ESTIMATION OF PHOSPHORUS IN THE chlorine by heating a with a Snowdon furnace from 230255°. The material that condensed in D showed the lines of neither zinc nor indium. This chloride was removed from D, dissolved in water, and its solution was electrolysed with a rotating rod of platinum as cathode until a small amount of metal was deposited. The spark spectrum of the solution of this deposit in hydrochloric acid showed the lines of neither zinc nor indium. Because of the success of this method of distillation in freeing gallium chloride from indium and zinc it was decided to convert the large amount of metallic gallium that had been prepared by electrolysis (see ante) to the chloride, and to subject this chloride to all distillations. This gallium was nearly free from indium and zinc. The chlorination of the metallic gallium was carried out in the bent tube c (Fig. 5) of Jena glass that was fused to the inlet tube в of the flask A (Fig. 4). Metallic gallium was placed at the depression c and the chlorination was made as described above, в not being connected with A during the procedure. This avoided the possibility of the distillation of zinc chloride and lower chlorides of indium into A and D during the chlorination of the metal. After chlorination was complete the tube B was inserted into the neck of the flask A, and c was warmed to such a temperature as caused the The current of chlorides to melt and flow down into A. chlorine was continued through the apparatus, and A was heated in a Snowdon furnace while the chloride distilled When all of the through the glass-wool from a into D chlorides had condensed in D the heating of A was discontinued, and the furnace was brought up around the tube D and the chloride was again distilled through the glass-wool into the tube E, where it condensed. In the first chlorination and double distillation 5'009 grms. of the purified metallic gallium was employed. A slight leakage of the vapour of gallium chloride into the outer air took place through the ground joints at the top In view of the large amount of tungsten used in the manufacture of high-speed steel, the question of the estimation of phosphorus in the presence of tungsten is an important one, but, although important, does not appear to have had the attention it deserves. Judging from numerous analyses which have come under the authors' notice, many chemists seem to find less phosphorus than is actually present. On investigating the cause of some of the low results they were found due to faulty methods. Taking ferro-tungsten as an example, the authors found that the method adopted by many chemists was to decompose the ferro-tungsten by fusion or other methods, separate the tungstic acid by evaporating to dryness with hydrochloric acid, and estimate the phosphorus in the filtrate. This method is altogether wrong, as a large amount of the phosphorus present is precipitated with the tungstic acid as phospho-tungstic acid, and the phosphorus so precipitated is lost. Most ferro-tungstens contain only o'015 to 0.030 per cent phosphorus, and with these the error is perhaps not quite so glaring, but odd samples made from phosphoric tungsten ores contain up to o 100 per cent, and sometimes even more. It is in these high phosphoric samples that the error stands out prominently. The authors have seen samples containing O'100 per cent phosphorus returned as containing o'020 per cent. To illustrate this point a table of results obtained by using the faulty method described above is given, together with the figures obtained by further examining the residue of tungstic acid for phosphorus. These results are only a few selected from a large number of samples examined. A paper read before the Iron and Steel Institute, May 8, 1919. Phosphorus in fil trate · .... Phosphorus in tung Inter-allied Conference of Chemistry at Paris. 0023 0017 0 022 0014 0.017 0.013 stic acid residue.. 0057 0071 0·080 0.055 0.032 0.020 Total phosphorus.. 0080 0.088 0102 0.069 0.049 0.033 From the above table it will be seen that the larger proportion of the phosphorus contained in ferro-tungsten is precipitated with the tungstic acid, and any method of analysis which is based on separating the tungstic acid before estimating the phosphorus must give low results for phosphorus. It occurred to the authors that there might possibly be some definite ratio between the amount of phosphorus precipitated and the amount not precipitated, but after making numerous analyses of samples of ferro-tungsten it was concluded that there was no ratio, neither have they been able to determine the conditions necessary either to precipitate all the phosphorus with the tungstic acid or to precipitate the tungstic acid free from phosphorus. These low phosphorus results not only occur with ferro-tungsten, but also with tungsten ores. To quote one example out of many, a sample of wolfram which contained o 200 per cent phosphorus was returned by one chemist as containing o'030 per cent. The same state of things also obtains to some extent with tungsten steels-in fact, some tungsten steels when analysed by the usual method show only a trace or no phosphorus at all, even when it is known that phosphorus is present. Indeed, it has been suggested that the phosphorus is slagged out in making tungsten steel because the amount of phosphorus in the ingredients used was so much higher than that found in the finished steel. It appears to the authors that the discrepancy is due not to slagging out the phos. phorus, but to the whole of the phosphorus in the finished steel not being obtained in the analysis owing to the cause stated above. The method devised by the authors for the estimation of phosphorus in ferro-tungsten, tungsten powder, and tungsten ores, which has proved to be an accurate one after some years of service, is as follows: 259 dissolve in hydrochloric acid, boil, add a few cubic centimetres nitric acid, and precipitate with ammonia. Boil, filter, wash, dissolve in nitric acid, and precipitate the phosphoric acid with ammonium melybdate as usual. The method may also be used for the estimation of phosphorus in alloy steel, but as steel cannot be fused directly with nitre mixture it should be dissolved in nitric acid in a dish or wide beaker, evaporated to dryness, and baked on a hot plate until the nitrate of iron is decomposed to oxide. The resulting oxide of iron, &c., can then be easily and cleanly removed from the dish or beaker with the aid of a flexible spatula and wiping with a little damp filter paper. It is then fused with nitre mixture and the analysis proceeded with as directed for ferro-tungsten. The method is especially useful for vanadium steels, as by its means the vanadium (which interferes very considerably with the estimation of phosphorus by the molybdate method) is completely got rid of before precipitating with molybdate. The authors have submitted the method to chemists in several large steelworks, who have found it satisfactory. CHLORINATION OF BENZENE. AT a meeting of the Birmingham and Midland Section of This method is similar in principle to that employed by H. G. Colman for the determination of toluene in commercial toluol, and consists in distilling from an Engler flask 100 cc. of the sample at the rate of 7 cc. per minute and interruping the distillation at 122° C. (corr.) and 142° C. (corr.). From the amounts distilling below 122° and above 142° the percentages of benzene and chlorobenzene are obtained from a graph, and the dichlorobenzene may be found by difference. Experiments with mixtures of known composition have shown that the percentages of benzene and chlorobenzene deduced from the graph are not affected by the other products which are likely to be formed during the chlorination of benzene. The method is an expeditious one-a distillation being complete in half to three-quarters of an hour, and it proves eminently suitable for investigating the products obtained from the chlorination of benzene either on the experimental or industrial scale. Fuse 2 grms. of the finely powdered sample with 10 grms. nitre mixture (Na2CO3+KNO3 in molecular proportions) in a large covered platinum crucible, dissolve the melt in the least possible quantity of water, filter, and wash free from tungstates with boiling water containing a little ammonium nitrate (six washings are sufficient). Ignite the residue in the original crucible, transfer to a 400 cc. beaker, dissolve in hydrochloric acid, and evaporate to dryness. To the filtrate add 20 cc. hydrochloric acid, and about 2 cc. bromine (bromine, not bromine water), stir well until the liquid is distinctly coloured by bromine, add ammonia until the precipitated tungstic acid is dissolved, and then a further quantity of strong ammonia INTER ALLIED CONFERENCE OF CHEMISTRY equal to one-fourth of the original bulk of the liquid. The volume of the liquid should now be about 250 cc. and add 3 cc. magnesia mixture, stir well, and allow to stand for six hours, or preferably over night. It is generally supposed that ammonium magnesium phosphate is much more soluble than it really is. The authors found that under the conditions described, the phosphorus is completely precipitated by magnesia mixture. Cool Filter through double papers and wash six times with ammonia water (10 per cent). Dissolve the precipitate in bydrochloric acid, allowing the liquid to run into the beaker containing the iron, evaporate to dryness, take up with 10 cc. hydrochloric acid, dilute, saturate with sulphuretted hydrogen to remove arsenic, tin, &c.; filter, wash, boil off sulphuretted hydrogen, add 50 cc. ferric chloride (1 grm. pure iron dissolved in hydrochloric acid, oxidised with nitric acid, and made up to 1 litre), cool completely, add ammonia until a distinct dirty green precipitate is obtained, then acetic acid until the precipitated ferrous hydrate is dissolved, boil well, and filter off the basic ferric acetate which contains all the phosphorus, AT PARIS. ON the initiative of the Société de Chimie Industrielle an which took place in private, and at which the statutes of an international confederation, which is to aim at bringing about an intimate alliance between France, England, Belgium, the United States, and Italy, were fixed. An inter-allied council, which is to meet in London on July 15-18, was appointed, each nation appointing two members as follows: Belgium-MM. Chavanne and Crismer. United States-Dr. Cottrell and Lieut.-Col. Zanetti. At the public meetings, which took place on April 14 and 15, communications of the greatest interest were presented. Prof. Louis described modern methods of enriching iron minerals by magnetic separation, and Dr. Cottrell gave an account of the preparation of helium for balloons and dirigibles. A communication on the potash industry in the United States by Mr. MacDowell was also read at the meeting on April 14, and on the following day an exposition of the America patent laws was given by Mr. John Pennie. A banquet was held in the evening, and on the 16th a visit was paid by the members of the Conference to the devastated region of Chauny. EXPORT OF FOODSTUFFS TO GERMANY. (a) The quantitative limitations on the import of food- it has been decided to authorise the resumption, by firms All arrangements for finance should be made by, and at the risk of, the private traders and the neutral or allied firm or Government concerned. Foodstuffs on List C of prohibited exports may be exported to Northern Neutrals and Switzerland, as well as to allied countries without licence; but foodstuffs on Lists A and B can be so exported only under licence from the Export Licence Department, 4, Central Buildings, S.W. 1. The term "Foodstuffs" means food, beverages, spices, edible oils, or other articles intended solely for the manufacture of human food. BOARD OF TRADE ANNOUNCEMENTS. IMPORT RESTRICTIONS THE President of the Board of Trade, after duly considering the recommendations of the Consultative Council on Imports, has given the following further directions in regard to the prohibitions of imports : The restrictions on the importation of the following articles are to be removed. 201. Printing inks. 202. Oil-lamp burners. 203. Gas-burners. 204. Metal parts and accessories of pedal cycles except 205. The following painters' colours and pigments except 207. Fancy goods (articles de Paris). 209. Paints and enamels (from July 1). PROCEEDINGS OF SOCIETIES. ROYAL SOCIETY. Sir J. J. THOMSON, O.M., President, in the Chair. "On the Area of Surfaces." By Prof. W. H. YOUNG, F.R.S. Many attempts by well-known writers have been made to frame a theory of the area of surfaces. These efforts have been attended with so little success that even the most recent text-books define the area of a curved surface by means of the formula known to hold in the case of a surface of revolution. Not even in the matter of the definition itself has anything which can be regarded as final been achieved, still less has it been found feasible to proceed from the definitions which have been given to the formula required. In the present communication the author attacks the question from an entirely new point of view. The defini tion given is based on what is itself a new concept, namely, that of the area of a closed skew curve. It is characterised further by the use to which is put the idea that the surface is, like a curve, an ordered manifold, the ultra-order being double instead of single. marine blue, white lead, satin white, lamp black. 227. Metal fittings and frames for bags and trunks. 228. Raw spirits for industrial purposes. 229. Reclaimed rubber. In accordance with the above, general licences have been issued for the articles mentioned in items 202 to 206. In the case of printing inks and aerated mineral and table waters general licences are already in operation, and a general licence will be issued in due course for paints and enamels. Applications for special licences should be made as usual to the Department of Import Restrictions, 22, Carlisle Place, London, S.W. 1. The surface is accordingly supposed defined by equations of the formx = x (u, v) y=y (u, v) tat (u, v) and divided up by the curvesu = const. v = const. On the fact that the sum of the areas of the boundaries of the portions of surface thus obtained bas a unique limit, the definition of the area of a surface is based. The curve boundaries have in fact an area whenever they possess a length. Moreover, the unique limit obtained for their sum is shown under very general conditions to have precisely the value given by the well-known formula. CHEMICAL NEWS. } Metrology in the Industries It may be added, though this is not proved in the paper, that the theory developed may be utilised to indicate the limitations of the older method, based on triangulation, and at the same time to notably exterd such results as have been obtained by earlier writers. "On Change of the Independent Variables in a Multiple Integral." By Prof. W. H. YOUNG, F.R.S. "Researches on the Chemistry of Coal. Part I. The Action of Pyridine upon the Coal Substance." By Prof. W. A. BONE, F.R.S., and R. J. SARJANT. The paper records the result of an experimental investigation of the so-called solvent action of pyridine and homologues upon the coal substance, with the double object of clearing up certain discrepancies in the work of previous investigators, and of determining the real nature of the action in question. It is shown that the presence of oxygen has an important retarding action upon the extraction process (the extent of which varies considerably with the nature of the coal), and that in order to obtain consistent results in any such process, it is necessary not only to employ an anhydrous solvent, but also to exclude oxygen. A suitable apparatus and method are described for carrying out the process under standard conditions, such as will give consistent results of comparative value with various coals. The application of the method to two typical isomeric bituminous coals is fully described. It is shown that when such extraction is carried out at ordinary pressures, with exclusion of oxygen, a practical limit is finally attained. In the case of the two coals in question, this limit considerably exceeded the amount of "volatiles" yielded by them on carbonisation at 950°. At higher pressures this first limit was considerably passed, and when conducted in sealed tubes between 130° and 150° as much as twothirds of the coal substance was rendered soluble. The results as a whole indicate that neither pyridine alone, nor even pyridine in conjunction with chloroform, is capable of extracting in a pure condition the resinic constituents of the coal substance, and that in addition to any ordinary solvent action which pyridine may have upon such constituents it also at the same time slowly attacks and resolves into simpler molecular aggregates the complex structure of the coal substance as a whole, for which it has a marked affinity. "A New Method of Weighing Colloidal Particles." By Prof. E. F. BURTON. When fine colloidal particles are dragged up and down for equal periods in a liquid by the application of a vertical electrical field a net settling of the particles is noted. It ia thought that, although for small forces such as gravity alone the Brownian movement prevents the attainment of any limiting velocity, yet when the particles are dragged by a much larger force the comparatively insignificant gravitational force is added to the electrical for downward motion and subtracted for upward motion, thus becoming effective in producing a net settling of the particles. Application of Stokes's Law to this net settling gives a value for the size of the particles very closely agreeing with that obtained by the counting method (e.g. 2.2 X 10-5 cm. and 1'7 x 10-5 cm.) even though values to band are taken from old observations made when this net settling was not appreciated and not closely observed. PHYSICAL SOCIETY. Ordinary Meeting, March 28, 1919. Prof. C. H. LEES, President, in the Chair. 261 A DISCUSSION on "Metrology in the Industries” was held. Sir RICHARD GLAZEBROOK, C.B., F.R.S., in opening the discussion, traced briefly the early history of metrology. While in some form or other it must have been employed in the very earliest times-if what antiquarians told us of the Pyramids was correct-the first application of really accurate measurement to mechanical engineering was chiefly due to Sir Joseph Whitworth, who taught people to make their length measurements with great accuracy and introduced reference gauges. The next step was the use of limit gauges. This greatly simplified the gauging of repetition work. At the time of the Boer War the supplies of ammunition, especially breech plugs of guns, were not interchangeable as obtained from different shops. This led to the formation of the Engineering Standards Committee on Gauges, which tackled the problem of producing accurate gauges with defined limits and tolerance, and by 1914 a certain number of firms had introduced the use of limit gauges. In 1915 the cry for munitions on a great scale brought home the great importance of interchangeability and the need for strict standardisation of gauges. When Screw gauges were first tested at the National Physical Laboratory the rejections total'ed 75 to 80 per cent; but after two years this was reduced to about 20 per cent. Now, if we are to maintain our position in peace, the maintenance of interchangeability in the engineering manufacture is equally necessary, so that we may manufacture in quantity. Much has yet to be done if we are to keep ahead, and the co-ordination of research with routine testings is vital to the progress of the science. Mr. TAYLOR (of Messrs. Taylor, Taylor, and Hobson) agreed that the war had produced a great advancement in the standardisation of manufactures; but we were apt to overlook much that had been done in this way before the war. Interchangeabilty of manufacture had existed, for instance, in the watch trade in Switzerland for centuries. In America, interchangeable manufacture had been introduced by Col. Colt in the Civil War, and was widely practised in America before it was generally understood in this country, But something, nevertheless, had been done in this country. The Linotype Company had done much in this way, while bicycle making had taught us a great deal of interchangeable manufacture. Our schools clusion of the science of the workshop. were at fault in teaching the science of design to the ex Sir HENRY FOWLER (Chief Mechanical Engineering sure the Department, Midland Railway) said he was country did not realise the debt it owed to the National Physical Laboratory in connection with gauges. While, as Mr. Taylor bad said, a certain amount of interchangeable work was done in this country before the war, the country, as a whole, did not appreciate the advantage of To produce economically we must produce in quantity. limit gauges. A great deal of education is still wanted. For this it is essential to have interchangeability, and we can only have this if the science of metrology goes ahead. Dr. P. E. SHAW (University College, Nottingham) said he thought the extent to which metrology had now permeated the workshops of the country made decentrali Experiments are now in hand to determine this settling very exactly. By the application of ultra-microscopic illumination of single particles in a slowly alternating field, it is hoped that the setting will be observed over an extended interval and that a very exact method of weighsation desirable, and he thought the universities were the ing these particles will be developed. In addition, the method should prove useful in determining the effect of the addition of various electrolytes on the size of colloidal particles and thereby shed much light on incipient processes of coagulation. "The Value of the Rydberg Constant for Spectral Series." By W. E. CURTIS. proper sites for the subject. He had instituted a department of metrology in University College, Nottingham, in 1916. The instruction would be mainly practical, but addition to the instruction work testing should be done, preliminary lectures would be necessary as well. In the standards used being verified periodically at the National Physical Laboratory. Such work is essential to the vigorous growth of the subject. |