ANALYTICAL CHENISTRY. ii. 33 An a1 y ti eal Chemi s t r y. Bimetallic I. Systems WILLARD and Electrode Systems in Electrometric Analysis. comprising Two Dissimilar Metals. H. H. FLORENCE FENWICK ( J . Amer. Chem. SOC.. 1922. 44 250&2515).-The constant half -;ell of the usual electrometrid titration apparatus may be replaced by a metal or certain alloys of metals of the platinum group other than pure platinum or pure palladium and by tungsten. The end-point obtained with such a bimetallic system differs in character from &hat given by an uni- metallic electrode but is situated in exactly the same position. Such a bimetallic electrode system has the advantage of being essentially simpler than the usual system and the end-point is much more distinct than with the ordinary electrode.J. F. S. Bimetallic Electrode Systems in Electrometric Analysis. 11. Theory of Bimetallic Systems Systems comprising Two Similar Metals. H. H. WILLARD and FLORENCE FEN- WICK ( J . Amer. Chem. SOC. 1922 44 251&2529).-1n aqueous solutions of multivalent elements in which the concentration of either state of oxidation approaches the limiting value the potential unattackable electrodelsolution is determined by the magnitude of the solution pressure of gas absorbed by the electrode and the end-point obtained in oxidimetric titrations with systems com- prising two unattackable electrodes is given only by virtue of a difference in the solvent power of the two metals for gas. When t'he concentration of oxygen-ion or hydrogen-ion becomes vanish- ingly small the osmotic pressure is too low for a saturation of the electrodes with the corresponding gas.The two elements receive the charge in accordance with the distribution law and a difference in potential develops which is at its maximum with minimum concentration of the corresponding ion and falls rapidly as this ion-concentration rises to a value in excess of that required ,for thp saturation of both electrodes. The bimetallic systems described (cf. preceding abstract) provide a type of electrode system essentially different from those previously used. A com- parison of the relative values of the usual unimetallic and bi- metallic systems may be based on the difference in the mechanism of the end-point in the two cases. The change in voleage with the former which is a true oxidation-potential is continuous khroughout and rises to a maximum at the end-point.With the latter practically all change is confined to within less than 0.5 C.C. of the completion of the titration. Although the actual mag- nitude of the break is normally less than with unimetallic com- binations it is relative to the preceding rise much greater hence the sharpness of the end-point is correspondingly increased insuring greater speed and accuracy. With the use of polarised bimetallic systems the break may be increased so much as to leave no possible 9 VOL. cxxrv. ii. Yii. 34 ABSTRACTS OF CHEMICAL PAPERS. comparison favourable to the unimetallic system. It thus becomes possible t o titrate solutions which offer too small potential differences between two possible states of oxidation for a good end-point with the usual apparatus.The polarising circuit may be made an integral part of the usual apparatus which requires rro attention and presents no difficulties. The disadvantages of the new system lie in the greater localisation of the total potential change which makes it more difficult to anticipate the end-point. Enough warning is given however to make the danger of over- titration negligible after very little practice. Estimation of Chlorine in Benzaldehyde. J. VOIGT (2 angew. Chem. 1922 35 654-655).-An apparatus is described by means of which traces of chlorine in volatile organic com- pounds particularly in synthetic benzaldehyde can be accurately estimated. From 10-20 g. of benzaldehyde are weighed into a wickless lamp arranged in an air-bath which can be heated by bunsen burners.The exit of the lamp is connected with a pipette- shaped combustion tube through which oxygen is passed. Hydro- gen is passed over the warmed benzaldehyde and mixing with the vapours it carries them forward through a jet situated in the wide part of the combustion tube. The issuing gas is ignited in the oxygen by heating the tube externally and it continues to burn quietly until all the benzaldehyde has been volatilised. The products of combustion pass over 2-3 g. of granulated anhydrous sodium carbonate packed in the narrow stem of the combustion tube and heated by a row of burners. The water is collected in a calcium chloride tower. About 7 g. of benzaldehyde can be burnt in an hour and a t the end of the combustion the soda is washed out of the tube by means o€ dilute nitric acid and titrated with standard silver nitrate.As little as O . O O l ~ o of chlorine can be estimated accurately by this method and it is therefore necessary to apply a correction for the chlorine content of the hydrogen employed as although undetectable by ordinary methods the amount which is always present may make an appreciable difference in the present case. Hydrogen in cylinders contains on the average 0.003 g. of chlorine per ni3. Detection of Chlorides and Bromides in the Presence of Thiocyanates. G. SPACU (BuZ. SOC. hytiinje CZuj 1922 i 302- 305 ; from Chem. Zentr. 1922 iv 733).-Neutral solutions con- taining chlorides and bromides in the presence of thiocyanates are treated with a little pyridine and excess of 10% copper sulphate solution.The thiocyanate ion is quantitatively precipitated as copper pyridine thiocyanmte [CuPy,](CNS) and the filtrate may be tested in the ordinary way for chlorides and bromides. J. F. s. G. I?. M. G. W. R. A Gasometric Method of Estimating the Halogen in Organic Compounds. ALEXANDER KILLEN MACBETH (Chem. News 1922 125 305-306).-Some organic halogen compounds in which the halogen possesses an induced electropositive natureANALYTICAL CHEMISTRY. ii. 35 (cf. Henderson and Macbeth T. 1922 121 892) are reduced quantitat,ively by hydrazine the volume of nitrogen evolved depending on the amount of halogen or halogen compound dealt with. A list of substances to which the method has been success- fully applied is given.With some halogen-substituted malonic esters the reaction is not quantitative and in the case of certain di-subst,ituted compounds only one halogen atom reacts the other being unattacked in the cold. Estimation of Fluorides. N. K. SMIT (Chem. Trade J. 1922 71 325).-A simplification of Greeff's method (A. 1913 ii 975). To 5 or 10 C.C. of the neutral fluoride solution are added 5 C.C. of 10% ammonium thiocyanate solution and 25 C.C. of alcohol and the mixture titrated directly with alcoholic ferric chloride stand- ardised against potassium fluoride. The reaction between ferric chloride and fluorides can also be applied to the detection of the latter. A. A. E. GEORGES CHAUDRON and LOUIS BLANC (Compt. rend. 1922 175 885-887).-In estim- ating oxygen in steel by heating the metal in a current of hydrogen and measuring the quantity of water formed the addition of other metals to enable the estimation to be carried out at a lower tem- perature was not found to affect the result obtained.When dealing with molten mixtures of iron and oxide of manganese or of iron and silicon dioxide the method is inaccurate owing to incomplete reduction of the oxides and to the relatively small amount of water obtained which is of the same order as the correc- tions necessitated by the method. Estimation of Sulphur in Iron and Steel. FERDINAND NIKOLAI (Chem. Ztg. 1922,4.6,1025-1026).-The iron or steel turn- ings or powder are treated with hydrobromic acid (d 1*48=47.4y0) in an apparatus consisting of a small flask ground on to a tube the first 12 cm.of which act as a reflux air condenser and is then bent round and carried vertically downwards into a cylinder containing 2.5% sodium hydroxide solution. The mixture of iron and hydro- bromic acid is gradually raised to its boiling point and the hydrogen sulphide liberated is absorbed in the sodium hydroxide solution which after adding 10 C.C. of potassium iodide solution and some starch is titrated with ,V/150-iodine solution. The whole estim- ation requires twenty to twenty-five minutes and very concordant results are obtainedl which agree well with those obtained using Fresenius's method. The results obtained by the above method are higher than those obtained using either dilute or concentrated hydrochloric acid because the aqueous solution of hydrobrolnic acid which distils unchanged contains 47.4% HBr whereas the corresponding hydrochloric acid solution only contains 20 yo HC1.H. J. E. The Estimation of Oxygen in Steel. H. J. E. H. C. R. The Rapid Estimation of Sulphur Dioxide. R. MARCXLLE (Ann. Fdsif. 1922,15 395401).-For the rapid and approximate 2-2ii. 36 ABSTRACTS OF CHEMICAL PAPERS. estimation of sulphur dioxide in wines 10 C.C. of the wine are heated with LO C.C. of a solution of 60 g. of sodium hydroxide per litre in a wide-necked conical flask until the volume of the contents is reduced to one-half. The solution is cooled diluted with 100-150 C.C. of water and enough dilute sulphuric acid added to neutralise the sodium hydroxide. Starch solution is added and the solution is titrated with iodine solution containing 4 g.or 8 g. per litre (1 c.c.=l mg. or 2 mg. of sulphur dioxide). If the alcohol and aldehydes are not driven off by heating the results may be 200 mg. per litre too low. The presence of more than 5 g. of dextrose per litre in the wine affects the results obtained and the sulphur dioxide must then be separated by distilling 10 C.C. of the wine acidified with 5 or 6 drops of syrupy phosphoric acid the distillate being led into 10 C.C. of the sodium hydroxide solution which is afterwards treated as in the direct method described above. The results obtained by this method agree well with those given by Haas's method and the degree of accuracy is sufficient for all practical purposes. H. C. R. New Process for the Volumetric Estimation of Ammonia and Carbamide Nitrogen by the Mypobromite Method.J. TTUMANS and A. KRUGER (Z. angew. Chem. 1922 35 686- 687).-A simple glass apparatus is described for the estimation of ammonia or carbamide by the hypobromite method. It consists essentially of an upper cylindrical portion provided with a ground- glass stopper and a draw-off cock drawn out to a h e opening. The upper vessel fits with a ground-gla,ss joint into a lower vessel and communication between them is established through a vertical tube extending nearly to the top of the upper chamber which is charged with brine to a level of 3 4 cm. below the top of the vertical communication tube. When the draw-off cock is opened and the apparatus is otherwise closed brine runs out until equilib- rium is attained between the inside and outside pressures.The lower vessel being previously charged with hypobromite solution the reaction is then started by rotating a boat containing the carbamide or ammonium salt on its axis so that its contents fall into the solution and the nitrogen evolved causes the displace- ment of an equal volume of brine from the upper vessel through the draw-off cock. The boat is fixed to a horizontal axis which is ground to pass through and fit into a tubulure in the side of the lower vessel and the boat can be reversed by turning the pro- jecting portion of the axis through 180". The apparatus can be utilised for urine investigations by placing the urine in the lower vessel and running in the hypobromite by means of a tap funnel through the vertical communication tube.The Estimation of Aliphatic Nitrates in the Presence of certain Nitro-aromatic Compounds. WILBERT J. HUFF and RICHARD D. LEITCH (J. -4mer. Chem. SOC. 1922,4-4,2643-2645).- Aliphatic nitrates may be estimated in the presence of certain aromatic nitro-compounds by reduction with ferrous sulphate in G. El. M.ANALYTICAL CHEMISTRY. ii. 37 excess the unoxidised ferrous salt being titrated with standard permanganate solution. A Kjeldahl flask is fitted with a two- holed rubber cork carrying a pear-shaped dropping funnel and a tube bent twice a t right angles and having its external orifice closed by a mercury seal. The air is expelled from the flask by boiling in it 25 C.C. of distilled water and when the flame is removed the nitrate ester mixed with the aromatic compound and dis- solved in glacial acetic acid is slowly run into the flask.Then a known volume of a standard solution of ferrous sulphate in dilute sulphuric acid is slowly added and this is followed by concentrated hydrochloric acid in volume equal to that of the liquid already in the flask. The reaction mixture is then evaporated to 10-15 C.C. and nearly neutralised with sodium carbonate and when cool is diluted to 600 C.C. and after the addition of 2-3 g. of manganous sulphate is titrated with permanganate solution. Satisfactory results were obtained for mannitol hexanitrate but those for glyceryl trinitrate were slightly low probably owing to the partial vaporisation of the nitrate before it was completely hydrolysed.W. G. New Reagent for the Detection of Nitrites in Water. GEORGES RODILLON ( J . Phurm. Chim. 1922 [vii] 26 376-379).- To the water to be examined contained in a test-tube 3 4 C.C. of a reagent composed of a 6% solution of resorcinol in pure sulphuric acid are added so as to form two layers. In presence of nitrites a rose-coloured ring is formed a t the junction of the two layers. An approximate estimation of nitrous acid can also be arrived a t if desired by comparing the density of colour and the appearances of the ring and the aqueous and sulphuric acid layers with those obtained with standard sodium nitrite solutions. Thus the coloration of the ring with solutions containing in 1 litre I g. of sodium nitrate is brownish-black fading away above and below to red ; 0.1 g.brownish-red similarly fading to red at the edges ; 0.01 g. carmine and 0.001 g. a very pale rose. The sulphuric acid layer ranges in colour with these concentrations from an intense amethyst- violet to the unchanged yellow of the original reagent. G. F. M. Volumetric Estimation of Phosphate in Solution. FRANK W. BURY ( J . Xoc. Chem. Ind. 1922 41 352~).-In the volumetric estimation of phosphate in solution by Rosin’s method (A.? 1911 ii 768) the solution is kept neutral by means of zinc oxide. The neutralisation is very tedious and copper carbonate was tried but showed no advantage. The best results were obtained by neutralising with borax the procedure being as follows. A measured quantity of phosphate solution is added to a measured excess of N/lO-silver nitrate and N/lO-borax solution is run in until the liquid is neutral to litmus.The solution is then filtered and the excess of silver nitrate estimated either by Volhard’s method or by adding a measured excess of N/10-potassium chloride and titrating back the excess with N/lO-silver nitrate using potass- ium chromate as indicator. G. F. M.ii. 35 ABSTRACTS OF CHEMICAL PAPERS. The Titration of Boric Acid in Presence of Phosphoric Acid. I. M. KOLTHOFF (Chem. Weekblad 1922 19 545-546; cf. A. 1922 ii 867).-An answer t o the criticisms of Deerns (A. 1922 ii 867) on the citrate method put forward by the author. s. I. L. Estimation of the Radium Content of Low-grade Radium- Barium Salts. VICTOR F. HESS and ELIZABETH E. DAMON (Physical Rev.1922 20 59-64 ; cf. Trans. Amer. Electrochem. SOC. 1922 41).-The y-ray method for the estimation of radium is inapplicable to salts containing much less than lo4 g. of radium per g. if the customary apparatus is used and the eman- ation method involves errors due to dilution and other causes if the sample contains more than lo-' g. of radium per g. For the examination of salts of intermediate radium content the y-ray method has been modified by the adoption of a shallow container with two curved sides each concentric with the cylindrical string electrometer and only a small fraction of the radius of curvature apart. Readings are taken with the container in position full of the salt t o be measured first alone and then with a small radium tube of known strength placed first immediately in front and then immediately behind the container.The ratio of the first reading to the mean of the other two gives that of the radium content of the salt under examination to the sum of the same and the standard to within 1%. Bismuth Sodium Thiosulphate ; its Preparation and Use in the Estimation of Potassium. V. CUISINIER (Bull. SOC. chim. 1922 [iv] 31 1064-1088).-0bservations on the thio- sulphates of bismuth and alkali metals have been made by Carnot (A. 1876 ii 426) Hauser (A. 1003 ii 487) Sanchez (A. 1912 ii 562) and by Vanino and Mussgnug (A. 1920 ii 44) but the substances have not been fully described and their use in the estimation of potassium salts has given inconsistent results. A salt of the approximate composition Na,Bi( S203) can be prepared crystallising in yellow prismatic plates rapidly turning brown on exposure to air.On attempting to carry out estimations of potass- ium by precipitation as potassium bismuth thiosulphate and sub- sequent treatment of the precipitate with iodine it was found that the results obtained were too high. Further the amount of precipitate obtained appears to be influenced by the concentration of the reacting solutions; the iodine value is not proportional t o the potassium content of the solution investigated and the volume of the alcohol used as precipitant also affects the result obtained. [Cf. J.S.C.I. 1922 981A.l A Photochemical Test for Silver in Thin Sections of Ores. G. SILBERSTEM and E. WIESS (2. anorg. Chem. 1922 124 355- 356).-On placing the section in a saturated solution of an iodide for five minutes the silver is covered with a thin film of silver iodide which under the microscope showed the characteristic reaction to light.The authors propose investigating the practical A. A. E. H. J. E.ANALYTICAL CHEMISrnS. ii. 39 liinits of this test and also the light reaction of silver sulphide with the object of applying it to test for sulphur in metals by shaking them with a solution of a silver salt. W. T. Contradictions and Errors in Analytical Chemistry. 111. The Separation of Zinc from Magnesium Calcium and Aluminium by Precipitation of the Phosphates. IV. The Conversion of Alkali Sulphates into Chlorides. FRIEDRICH L. HAHN [with (111) J. DORNAUF and (IV) R. OTTO] (Ber. 1922 55 [R] 3434-3436; cf.A. 1922 ii 873).-111. The separation of magnesium calcium and aluminium from zinc can be effected according to Voigt (A. 1910 ii 74) by precipitation with phosphate in ammonical solution containing ammonium salts whereby only the zinc remains dissolved. This method of separation has great advantages when only small quantities of foreign metals are present but the subsequent operation of precipitating the zinc as the zinc ammonium phosphate cannot be effected quantitatively by simply boiling the solution until the excess of ammonia is expelled. Accurate results are obtained if the filtrate containing the zinc is treated with hydrochloric acid until just acid to methyl-red and subsequently with an excess of phosphate. If the solution is rich in ammonium salts it is preferable to expel the bulk of the ammonia before addition of the acid.IV. Potassium cannot be separated from sodium as the per- chlorate or chloroplatinate if the metals are present as sulphatles. The usual procedure of precipitating the latter with barium chloride removal of excess of barium with ammonium carbonate and sub- sequent volatilisation of the excess of ammonium salts is inaccurate since barium sulphate always carries down alkali sulphate. This difficulty can be overcome by precipitating the barium sulphate in extreme dilution. For this purpose moderately dilute solutions of sulphate and barium chloride are added simultaneously and a t approximately equivalent rates to a small quantity of boiling hydrochloric acid. Pure barium sulphate is thereby precipitated in well-formed coarse crystals which are readily filtered and washed.The presence of varying amounts of alkali chloride and even of nitrate in the sulphate solution is without influence. If ferric iron is present ammoniacal barium chloride solution must be used and the solution be acidified after the precipitation. H. W. I. M. KOLTHOFI and J. C. VAN DIJK (Pharm. ‘CVeelcbZud 1922,59 1351- 1360).-The fact that zinc sulphide is almost always precipitated to some extent with copper sulphide from solutions of the two metals is not due to adsorption since zinc sulphide is only with difficulty soluble in acids of low concentration. The precipitation of zinc from acid solutions by hydrogen sulphide depends on time and temperature as well as on the concentration of zinc- hydrogen- and sulphide-ions and is accelerated by presence of copper In the quantitative separation of copper and zinc by hydrogen The Carrying-down of Zinc by Copper Sulphide.‘ sulphide .ii. 40 ABSTRACTS OF CHEMICAL PAPERS. sulphide the time must be as short as possible. If sulphuric or hydrochloric acid is used the acidity must be at least 0.5N; with the former room temperature is best; with the latter the gas should be led in a t the boiling point. s. I. L. Some Applications of Sodium Peroxide in Analytical Chemistry. W. M. STERNBERG (Xcience 1920 52 162) .-The decomposition of lead and zinc ores by fusion in an iron crucible with a t least six to eight parts of sodium peroxide is rapid and complete. I n the case of lead ores the aqueous solution of the fused mass after treatment with 1.4 parts of oxalic acid to reduce the lead dioxide is acidified with sulphuric acid boiled cooled and the lead sulphate collected and washed with 5% sulphuric acid.It is then dissolved in a solution of ammonium chloride or of ammonium chloride and sodium acetate and titrated with ammonium molybdate. I n the case of zinc ores the fused mass is dissolved in ammoniacal ammonium chloride solution the solution boiled filtered and washed with hot ammonium chloride solution acidified with hydrochloric acid boiled and titrated with potassium ferrocyanide after the addition of hydrogen sulphide. It is necessary that the solutions should be standardised against a standard ore treated by the same method. A. A. E. Gravimetric Analysis.XXVI. Estimation of Lead. L. W. WINKLER (2. angezu. Chem. 1922 35 662-663).-For the gravimetric estimation of lead in neutral solution 100 C.C. of lead nitrate or lead chloride solution containing from 0.60 t o 0.01 g. of lead is acidified with 1 C.C. of normal nitric acid solution and boiled. Ten C.C. of 10% ammonium sulphate solution are slowly run in and the whole is allowed to remain then filtered. The precipitate is washed with 50 C.C. of a cold saturated solution of lead sulphate and the edge of the filter washed with 1-2 C.C. of dilute alcohol. The precipitate is dried for two hours a t 130". Modifications of the procedure necessitated by the presence of nitric or hydrochloric acid and salts of potassium or other metals are discussed and details are given of the method of estimation of lead in crude lead minium lead chromate and lead salts of organic acids.J. S. G. T. A New Sensitive Reaction for Copper Thiocyanate and Pyridine. G. SPACU (Bull. Xoc. Lytiinfe Cluj 1922 1 2 8 6 2 9 1 ; from Chem. Zentr. 1922 iv 737).-The reaction consists in the formation of a light green precipitate when a neutral solution of a copper salt is treated with a neutral thiocyanate and a few drops of pyridine. The precipitate has the composition [CuPy,]( CNS) and is almost completely insoluble in water. The reagent must be neutral and excess of pyridine must be avoided. The pyridine should be added before the thiocyanate. One part in 300,000- 800,000 of copper can be detected according to the conditions of experiment. For pyridine the sensitiveness is one part in 2000 and for thiocyanate one in 50,000.G. W. R.ANALYTICAL CHEMISTRY. ii. 41 The Hindering of the Precipitation of Copper Sulphide by the Presence of Sodium Chloride. WILHELM BILTZ (Ber. 1922 55 [B] 3393).-In a recent communication Dede and Bonin (A. 1922 ii 766) have shown that lead and cadmium sulphides are only incompletely precipitated in hydrochloric acid solution in the presence of relatively large quantities of the chlorides of the alkali or alkaline-earth metals. Similar observations have been recorded in the case of copper (cf. Biltz and Marcus A. 1909 ii 101 1). Precipitation of copper sulphide is however quantitative in a 5% solution of sodium chloride. Magnesium chloride or carnallite does not appear to have an inhibiting action.H. W. A New Microchemical Reaction for the Estimation of Copper. G. SPACU (Bull. Xoc. Stiinfe Cluj 1922 1 296-301; from Chem. Zentr. 1922 ivy 737-73S).-After precipitation of copper as copper pyridine thiocyanate CuPy,(CNS) (see this vol. ii 40) the precipitate is separated in an appropriate apparatus (Rothe’s shaking funnel) by successive extraction with chloroform. The precipitate is then weighed after removal of the chloroform in a vacuum desiccator. Alternatively the precipitate may be ignited and the copper estimated as cupric sulphide or cupric oxide. G. W. R. Volumetric Estimation of Copper and its application to the Estimation of Reducing Sugars. ED. LASAUSSE ( J . Pharm. Chim. 1922 [vii] 26 401406).-The precipitated cuprous oxide formed by the reducing action of the sugar on a cupro-tartrate solution is collected on a Gooch crucible and dissolved by means of 2-3 C.C. of nitric acid.The filter is finally washed with water the total volume of filtrate and washings being kept below 25- 30 C.C. The nitrous acid present is then eliminated by adding 2% permanganate solution drop by drop to the boiling liquid until the pink coloration just persists and the latter is reduced by the addition of a few drops of alcohol. After boiling for a minute longer the precipitated oxides of manganese are filtered off the filtrate is made up to 100 c.c. and the copper is estimated iodo- metrically in the usual way after the addition of an excess of sodium acetate to displace free mineral acid. In order to obtain the best results in the estimation of reducing sugars the procedure of Quisumbing and Thomas should be followed (A.1922 ii 92) as a granular cuprous oxide is thereby obtained which can readily be collected and washed. The amount of dextrose is given by the formula 0~474x+O~Oool 15x2 and of invert-sugar by 0504x+ 0~000087x2 where x is the weight of copper reduced. Iodometric Estimation of Copper in Bronze and Brass. G. BATTA and G. LATHIERS (Bull. SOC. chim. Belg. 1922 31 297- 302).-The application of the iodometric method of copper estim- ation to bronze and brass gives results which are of sufficient accuracy for many purposes. It is rapid and useful when electro- lytic methods cannot be employed. G. P. M. H. J. E.f i b 42 ABSTRACTS OF CHEMICAL PAPERS. Rapid Electrolytic Estimation of Mercury.A. DE M E E ~ S (Bull. Soc. Aim. Belg. 1922 31 302-323; cf. Bottger A. 1921 ii 351).-A study of the factors which are said to cause inaccuracies in the electrolytic estimation of mercury. Good results were ob%&twd,wi$4 a platinum cathode but gold was found to be more suitable forbthe purpose as the mercury deposit is more adherent a& so pepnits of the use of alcohol and ether for washing ; in the ease of other metals this should be done with water only as the more rapid method occasions loss. Experiments carried out with .mercuric salts in presence of substances which form complexes 'give similar results with gold and platinum cathodes but the method is more satisfactory when simple acid electrolytes are used as solvents. The method is applicable to mercurous salts which are oxidhed before electrolysis. Constancy of the Titer of Permanganate Solutions and Different Methods of Standardisation.0. HACKL (Chem. Ztg. 1922 46 1065).-N/lO-solutions of potassium permanganate made from the freshly prepared pure salt remained permanent if stored in a dark place for one year whereas solutions made from a ten-year old preparation altered considerably in the course of a few months. Tests showed that the most satisfactory com- pound for use in standardisation of permanganate solutions is sodium oxalate. The composition of oxalic acid crystals and ferrous ammonium sulphate was found to vary within narrow limits but these are sufficient to lead to discordant results in the titration.H. J. E. A. R. P. The Electrometric Estimabion of Iron and Vanadium when present together. ERICH MULLER and HANS JUST (2. anorg. Chem. 1922 125 155-166) -Ferrous salts and vanadium salts can be accurately estimated electrometrically with perman- ganate. Ferrous salts are more vigorous reducing agents than the salts of quadrivalent vanadium and the authors find that they can be estimated \+hen together in the same way. To the mixture a decinormal solution of potassium permanganate is gradually added until there is a sudden change in the E.M.F. This indicates the end-point for the ferrous salt; the titration is then continued until a second sudden change occurs which repre- sents the end-point for the vanadium salt. The latter titration is best carried out a t about 70" as the reaction is very slow a t the ordinary temperature.W. T. W. S. HENDRIXSON and L. M. VERBECK ( J . Amer. Chem. Xoc. 1922 44 2382-2386).-A slightly acid solution of titanous sulphate may be used to standardise solutions of potassium permanganate or potassium dichromate. The method consists in adding the titanous ;solution to the permanganate or dichromate solution and measuring the E.M.F. of the solution toward a platinum electrode against a calomel electrode. The E.M.F. changes very rapidly at the end-point which is thelhefore made very precise; in the case of Electrometric Standardising of Titanous Solutions.ANALYTfCAL CHEMISTRY. ii. 43 dichromate the change in E.M.P. on adding a slight excess of titanous ion to a slight excess of dichromate is about 0.6 volt whilst in the case of permanganate the change in E.M.F.is about 0.9 volt. This difference makes it possible to estimate two sub- stances of quite different oxidising power in the same solution. The authors have investigated the estimation of titanous and ferrous ions by means of permanganate or dichromate of f e h - i r o n and dichromate or permanganate by means of a titanous solution and have found that the method is trustworthy. New Method €or the Gravimetric Estimation of German- ium. JOHN HUGHES M~~I;LER (J. Amer. Chem. Xoc. 1922 44 2493-2498).-A new method of estimating germanium is described which is based on the precipitation of this metal as magnesium orthogermanate and weighing the dried and ignited precipitate. Magnesium orthogermanate is a snow-white infusible substance which is prepared by the action of magnesium sulphate ammonium sulphate and ammonium hydroxide on a solution of germanic acid.No precipitate is formed until the ammonium hydroxide has been added and the reagents are added in the order named. Magnesium orthogermanate is amorphous and is very soluble in dilute solutions of acids but after ignition it is only slowly dissolved by concentrated acids. It dissolves in water to the extent of 0.000016 g. per C.C. at 26". In a mixture of two volumes of ammonia (0.880) and three volumes of water the solubility is 0.00002 g. per c.c. whilst in an ammoniacal solution of 10% ammonium sulphate the solubility is 0-00013 g. per C.C. a t 26". The estimation of germanium is effected by adding an excess of 2N-magnesium sulphate to the cold solution of germanic acid ; this is followed by at least an equal volume of 2N-ammonium sulphate and then ammonium hydroxide (0.880) is added in amounts varying from 15 C.C.to 20 C.C. for each 100 C.C. of solution and the mixture stirred vigorously raised to the boiling point for a few minutes and kept for ten to twelve hours before filtering. The precipitate is filtered washed with a mixture of 90 C.C. of water and 10 C.C. of ammonia dried and ignited with free access of air. Analysis of Mixtures of Hydrogen with Paraffin Hydro- carbons. J. G. KING (Fuel 1922 1 103-106).-The estimation of hydrogen in admixture with paraffin hydrocarbons is effected by fractional combustion over copper oxide a t 280" in an apparatus which is a modification of that of Jager ( J .Gasbeleuchtung 1898 41 764). Carbon monoxide and hydrogen are thereby consumed whereas the paraffins remain unchanged. Identification of Methyl Alcohol in Liquids containing Ethyl Alcohol. I. M. KOLTHOFF (Pharm. Weekblad 1922 59 1268-1274).-The method of Denig&s (A. 1910 ii 461) namely oxidation with permanganate in presence of acid removal of excess with oxalic acid and testing for formaldehyde with Schiff's reagent will detect 0.05% of methyl alcohol in ethyl alcohol. The reagent J. F. S. The results of the analyses are excellent. J. F. S. CHEMICAL ABSTRACTS.ii. 44 ABSTR.ACTS OF CHEMICAL PAPERS. is best prepared as suggested by Elvove (A 1917 ii 341) and if kept in stoppered bottles is stable for long periods.The usc of qhosphoric acid in place of sulphuric acid as recoiiiiiiended by Chapin (A 1921 ii 50&) is found advantageous. The method can be applied to the examination of tinctures and inedicinal preparations containing alcohol without preliminary distillation. The reagent will also detect fornialdehyde in milk. 8 I. L. Colour Reaction for Phenols based on the Use of Selenious Acid. VICTOR E. LEVINE (Bcience 1920 52 207).-A reaction of great sensitivity and wide applicability consists in the faci; that phenols in contact with a solution of selenium dioxide (O*5(y0) or sodiuin selenite (0-75 yo) in concentratfed sulphuric acid give rise to characteristic green or blue colorations. On heating or on the addition of water the colour changes to brown or xed.It is considered that the phenol causcs the liberation of seleiziuiii which then dissolvcs with a green colour in concentrated sulphuric acid t o forin selenosulphur trioxide. A list is given of phenolic substances to which the test liar; been applied. The Action of Proteins on the Phenol Reagent of Folin and Denis. VICTOR E. LEVIKE (Science 1920 52 612-G13).- The phosphotungstic-phosphoinolybdic reagent of Polin and Denis (*4. 1913 ii 1011) is not specific for the phenolic group. The colour reaction is given by proteins and in faet by a large number of iiiorgahic and organic substances. The reagent appears t o be uflccted by substances possessing iiiore or less reducing properties. VICTOR E. LEVINE (Science 1920 52 301).-0ne to 2 C.C. of urinc m e heated for several minutes .with 5 C.C. of i5 2yo solutioh cf sodiuiii teilurito in 10% sodiulil carbonatc. Itcduction to telluriuitl takes piace in the presence of carbohydmtes possessing a frce carbonyl group. With small aiiiounts of sugar the telluriuni foriiis a colloidal solution; which has a characteristic browii colour i:i transmitted light and is reyish- blnclr in reflected light ; largcr amounts yield a greyis%-bhcic precipitate. A. A. E. The Partition Coefficiebts and the. Estimation by Extrac- tion of Organic Acids. JOX. PINKOW (2. Unters. iVnJu. Genussvz. 1932 44 204-209).-Bcetic and formic acids forin cloublc inoleculcs in ethereal solution and the irrcgularities observcd in the partition cocfficicnt of the latter acid- may be ascribed t o this cause. Assuming that no doublc molecules occur in the aqueous solutions tlie dissociation coefficients of the doublc mole- culcs 1iavc the following vafues a t L 5 O acetic acid 1.776; formic acid 4-77 The partition cocfkieiits (watcr/ether) of tlic siiiiple molccules at 15' are acetic acid 2 21; formic acid 2.38. At 26.3" formic acid 2.79. Saga hais no cffect oil the cxtractiozi of succiiiic acid frroni its aqueous solution. A. A. E. A. A. E. Test for Sugar in Urine. H. C. R.