ANALYTICAL CHEMISTRY. An a1 y t i c 8 1 C h 8 mist r y. ii. 55 Application of the Quinhydrone Electrode to Electrometric Acid-base Titrations in the Presence of Air and the Factors Limiting its Use in Alkaline Solution. V. K. LA MER and T. R. PARSONS ( J . Biol. Chem. 1923 57 613-631).-Measurements of the pE of solutions of hydrochloric acetic and boric acids and of acid potassium phosphate have been made with the quinhydrone electrode and the results have been compared with those obtained with the hydrogen electrode. In buffered solutions more acid than pH 8.0 the quinhydrone electrode gives trustworthy results pro- vided that no rapid oxidising or reducing agents are present. With more alkaline solutions the results are untrustworthy. This is mainly due to the autoxidation of quinol and t'o its weak acidic properties.E. S. New Methods of Determining Chlorine Bromine and Iodine in Organic Compounds. I. Hydrogenation Method. H. TER MEULEN and J. HESLINGA (Rec. trav. chim. 1923 42 1093-1096) .-The halogen derivative to be analysed (20-50 mg.) is placed in a small platinum boat near the left-hand end of a quartz tube 80 cm. long the other end of the tube being plugged loosely with cotton wool. A mixture of ammonia and hydrogen is passed through t'he tube from left to right and the tube to the right of the boat heated to redness over a distance of about 20 cm. Heating is then commenced to the left of the boat which during the courseii. 56 ABSTRACTS OF CHEMICAL PAPERS. of about twenty minutes is progressively heated until no further decomposition occurs The whole of the halogen element present even in the case of iodine is converted into ammonium halide which condenses in the right-hand end of the tube.The latter is washed out with water the solution acidified with acetic acid any hydrogen cyanide removed by boiling and the halogen estimated by Volhard’s method. Explosive substances are mixed with six times their weight of oxalic acid before being analysecl whilst volatile liquids are intro- duced in small tubes such as are used in ordinarv combustions. With a tube partly packed with platinised quartz ;cry low results are obtained. E. E. T. Determination of Dissolved Oxygen in the Presence of Iron Salts. A. M. BUSWELL and W. -U. GALLAHER (Ind. Eny. Chem. 1923,15 118G-l188).-The Winkler method for the deter- mination of dissolved oxygen in water depends on the oxidation of bivalent manganese in alkaline solution and the liberation of iodine from potassium iodide in acid solution by the oxidised manga- nese.If iron is present however it will be in the ferric state in consequence of the preliminary use of potassium permanganate to oxidise nitrites etc. and will itself liberate iodine from potassium iodide. Hence too high a value for the dissolved oxygen present is obtained in this case and as the action of ferric iron on the iodide solution is incomplete and depends on the concentration temper- ature and other conditions no correction can be applied. An investigation of the effect of the presence of iron salts on the Levy- Mohr method in which the dissolved oxygen is allowed to oxidise ferrous iron in alkaline solution showed that in this case no error was introduced.c. I. Determination of Sulphite and Thiosulphate by Oxidation with Nascent Bromine. C. MAYR and J. PEYFUSS (2. anorg. Chem. 1923 127 123-136).-Nasceiit bromine liberated from an acid bromide-bromate mixture oxidises sulphurous acid according t o the equation SO + Br + H,O =S03+2HBr; and thiosul- pliuric acid according to the equations S,O2+€I,O+Br,=2HBr+ S03+S and S+4H,0+3Br,=6HBr+H2S0,. Thus Na2SO,=.2Br and Na,S20,=8Br. By using an excess of standard bromide- bromate solution sulphites may be estimated in the absence of air and excess of bromine used to liberate iodine which may be deter- mined with standard thiosulphate. Advantage may be taken of the fact that iodine osidises thio- sulphates t o dithionates whilst bromine oxidises them to sulphates to determine sulphites and thiosulphates in one and the same solu- tion using the equivalents Na2S0,_2I and Na,S,O,=H together with those given above and calculating the amounts of sulphite and thiosulphate from the simultaneous equations so obtained.BROTHERTON & Co. LTD. (Chemistry and Industry l923,42,1131).-1n an iodometric method proposed for the determination of sodium hyposulphite the reaction H. H. Determination of Hyposulphite.ANALYTICAL CEEMISTRY. ii. 57 proceeds according to the equation 3Na,S,0,+4KI03+2KI= 31,+3Na,SO,~3K,SO,. Six g . of potassium iodate and 10 g. of potassium iodide are dissolved in about 500 C.C. of water in a litre flask 300 C.C.of N/lO-sodium thiosulphate solution is added the mixture is diluted to about 950 c.c. and 3 g . of the sample of hyposulphite is added from a weighing bot,tle. The whole mixture is then diluted to 1,000 c.c. shaken until the hyposulphite is dis- solved and the excess of thiosulphate titrated in an aliquot portion (100 c.c.) with N/lO-iodine solution. The method cannot be applied to sodium hyposulphite which contains decomposition products or sodium carbonate. w. P. s. Determination of Ammoniacal Nitrogen in certain Nitro- genous Materials and particularly in Proteins and Products of Proteolysis. J. FEOIDEVAUX (Compt. rend. 1923 177 1043- 1046).-The method previously described (A. 1922 ii 454) has been improved. The material is distilled in presence of aqueous lithium carbonate keeping the volume of liquid in the distilling flask con- stant by continued addition of water.The distillate is collected in a series of graduated flasks containing known quantities of standard acid. The distilling flask is heated in a calcium chloride bath kept at 110". The results are calculated as before and me satisfactory with proteins uric acid amino-acids and carbamide ; special procedure is necessary in the case of some compounds (diphenylamine oxamide etc .). Determination of Nitrate Nitrogen in the Presence of Calcium Cyanamide and some of its Derivatives. K. D. JACOB (Ind. Eng. Chem. 1923,15,1175-1177).-The determination of nitrogen present as nitrate in soil extracts by reduction with Devarda's alloy is interfered with by the presence of cyanamide derivatives. These however can be eliminated :-urea by decom- position with urease and calcium cyanamide dicyanodiamide and guanylurea by precipitation with silver aulphate when an accurate J .MTNICFI (Biochenz. Z. 1923 142 26&273).-The residual nitrogen of various types of mammalian blood has been determined by Pregl's method after removing the proteins by precipitation with trichloroacetic phosphomolybdic phosphotungstic and meta- phosphoric acids and with uranium acetate. The most consistent results were obtained with the first-mentioned acid whilst rneta- phosphoric acid gave the least trustworthy figures and the other reagents were intermediate between these extremes. No propor- tionality was observed between the quantity of precipitant used and the residual nitrogen found.Variations in the latter are not solely due t o differences in degree of precipitation. Detection of Small Quantities of Nitric Acid in Poisoning Cases. C. GHIGLIOTTO (Ann. Chim. Annlyt. 1923 5 325).- Tissues that have been in contact with nitric acid are stained blue by treatment with a sulphuric acid solution of diphenylamine. E. E. T. estimation becomes possible. [Cf. B. Jan.] c. I. Determination of the Residual Nitrogen in the Blood. J. P.2. 58 ABSTRACTS OF CHEMICAL PAPERS. The lining of the stomach responds to the test if it has been in contact with acid of a greater strength than 0.1 yo. Determination of Organic Phosphorus. E. J. BAUMANN (Proc. SOC. Exp. Biol. Med. 1922 20 171-173).-The loss by volatilisation of phosphorus experienced with Bloor’s or Bell and Doisy’s method is avoided by using 0-2 C.C.of 30% hydrogen peroxide and eight drops of sulphuric acid solution. After removal of most of the water by evaporation on a water-bath Bell and Doisy’s colorimetric method is used for the determination. Determination of Phosphorus in Organic Materials. ( ~MLLE) J. GAROLA (Ann. Chim. Analyt. 1923 5 326-328).-The loss of phosphorus during the incineration of organic materials containing it may be avoided by mixing 5 g. of the substance with 0.2 g. of light calcined magnesia and a little water drying the mixture on a sand-bath and heating it to dull redness until the residue is free from carbon. It is then dissolved in hydrochloric acid the solution filtered and the phosphorus precipitated as magnesium ammonium phosphate by the addition of an ammoniacal solution of ammonium citrate.This process is considerably quicker than the usual Kjeldahl method. A R. P. Determination of Phosphoric Acid by Means of Uranyl Acetate Solution. G. JANDER and K. REEH (2. anorg. Chem. 1823 129 302-305) .-The phosphate is precipitated as uranyl ammonium salt by the addition of a slight excess of uranyl acetate. The precipitate is filtered washed and redissolved in sulphuric acid (about 25%). The solution is heated to incipient ebullition and then reduced with aluminium as previously described (this vol. ii 65). The quadrivalentl uranium is then reoxidised to the sexavalent condition by means of potassium permanganate of known strength every atom of available oxygen used being equiva- lent to 1 mol.of uranyl ammonium phosphate. Modification of the Bell-Doisy-Briggs Method for Colori- metric Determination of Phosphoric Acid. B. SJOLLEMA and H. GIETELING (Chem. Weekblad 1923 20 658-659).-The method (cf. A. 1920 ii 769; 1922 ii 718) is now applied to cases such as foodstuffs urine etc. in which a preliminary destruction with sulphuric acid is necessary. Comparison is made with a standard phosphoric acid solution to which the reagents used in the destruc- tion (sulphuric acid potassium and copper sulphates) are added in the same proportions. Since the depth of colour is not proportional to the phosphoric acid present a correction graph is employed. A. R. P. CHEMICAL ABSTRACTS. H. H. s. I. L. Separation of Phosphates from Fluorides.E. M~~LLER and W. WAGNER (2. anorg. C‘hem. 1923 129 306-308).-The usual method is to precipitate the neutral phosphate as silver salt remove excess of silver with sodium chloride and then precipitate calcium fluoride. The authors criticise trhe method of neutralising theANALYTICIAL OHEMISTRY. ii. 59 solution before adding the silver nitrate. With Methyl-orange as indicator neutrality is obtained at NaH,PO ; with phenolphtha- lein at Na,HP04. As the silver salt is Ag,PO free hydrogen-ions are liberated during the precipitation causing solution of the silver phosphate to the extent of 37%. If the solution is neutralised to Methyl-orange after addition of the silver solution some silver oxide is included in the precipitate.The authors therefore recom- mend neutralisation with alkali hydroxide followed by acidificat,ion with a very small quantity of nitric acid. Determination of Arsenic in Organic Compounds. A. KIRCHER and F. VON RUPPERT (Ber. Deut. pharm. Ges. 1923 33 185-186).-1t has been suggested by StolIB and Fechtig (A. 1923 ii 335) that the authors' method for determining arsenic in neosalvarsan (A. 1921 ii 130) is liable to errors due to the passage of iodine-fixing organic substances into the absorption tube and the loss of arsenic on boiling to remove sulphur dioxide. The con- cordance of results obtained by the authors' method and that of StollB and Pechtig proves the above contention to be groundless. W. T. K. B. A. PIUTTI (Rend. Accud. Sci. Pis. Mat. Napoli 1922 [iii] 28 91-96).-The author has investigated the ability to absorb carbon monoxide exhibited by a number of different mixtures the best results being obtained with cuprous chloride mixed with fine charcoal granules of pumice and powdered soda-lime.This mixture is stable in the air and absorbs per 100 g. 3,960 C.C. of carbon monoxide. A. KARL and S. LOMBARD (Compt. rend. 1923 177 1036-1037).- The difliculty of obtaining a solution containing the constituents of minerals of the titano-niobate type has led to the following method The ground mineral is treated with hydrofluoric acid the product being evaporated to dryness and fused with potassium hydrogen sulphate until sulphuric anhydride vapours cease to be evolved. The cooled fused mass is mixed with sodium sulphate or even with lithium sulphate as well to depress the melting point to the minimum obtainable the whole is raised to a red heat cooled broken into pieces and so transferred to a Pyrex glass tube.By heating at 500-600" the product becomes completely liquid and the radium content may be determined in the usual way the emanation being displaced using a current of air and freed before measurement from acid fumes by being passed over soda-lime heated at 150". Gravimetric and Volumetric Determination of Potassium. W. STREKHER and A. JUNGCK (2. anal. Chern. 1923 63 161- 180).-Comparison is made of a number of methods of determining potassium. Precipitation as potassium cobaltinitrite does not lead to concordant results. The precipitate produced by the addition of sodium cobaltinitrite to potassium salts contains varying pro- portions of sodium presumably as the salt K,NaCo(NO,),. It is H.H. Mixtures Absorbing Carbon Monoxide. T. H. P. Determination of Radium in Natural Titano-niobates. E. E. T.ii. 60 ABSTRACTS OF CHEMICAL PAPERS. shown however that potassium can be precipitated completely from solution by this means and can be determined in the pre- cipitate by the perchlorate method. Satisfactory results are obtained by reduction of the perchlorate by (1) titanous sulphate (2) fusion with potassium nitrate and sodium hydroxide and (3) fusion with hydrazine sulphate and sodium hydroxide and titra- tion of the chloride produced with silver nitrate. Sodium and potassium can be determined together by determining the mixed chlorides volumetrically and subsequently precipitating the potassium as perchlorate which is reduced to chloride and titrated with silver nitrate.Concordant results can be obtained by pre- cipitation of potassium as the hydrogen tartrate and titration with sodium hydroxide. The picric acid method is not of sufficient accuracy owing to the solubility of potassium picrate in the solutions used. A. G . P. MARJAN G~RSKI (Prxemysl Cheznicxny 1922 6 311-312 ; from Chem. Zentr. 1923 iv 488).-Directions are given for the determination of potassium in potassium salk by the perchloric acid method. Determination of Potassium in Potassium Salts. G. W. R. A New Qualitative Test for Sodium. I. M. KOLTHOFF (Pharm. WeeEBlnd 1923 60 1251-1255).-The pyroantimonate test for sodium is not very sensitive and is trustworthy only in the absence of heavy metals and salts of ammonium and the alkaline earths. Magnesium uranyl acetate which has long been employed for the micro-chemical detection of sodium may be employed in ordinary qualitative analysis and in 50% alcohol solution will detect as little as 50 mg.of sodium per litre. The reagent is very suitable when potassium salts are present and will indicate 0.5% of sodium in a potassium salt ; salts of ammonium lead zinc aluminium magnesium and the alkaline earths do not interfere and the reagent may be used directly to detect sodium in them. s. I. L. Determination of Caesium as Perchlorate. E. MURMANN (Oesterr. Chem. Ztg. 1923 26 164).-About 0-5 g. of c3esium nitrate is dissolved in 50 C.C.of water an escess of 10% yerchloric acid is added and the mixture is evaporated. The residue of cEsium perchlorate when cold is treated with a small quantity of water containing perchloric acid collected on a weighed filter washed successively with 50 yo alcohol (containing a drop of perchloric acid) 95% alcohol and ether then dried a t loo" and weighed. Sulphates and potassium rubidium barium and ammonium salts New Method of Titrating Silver and Halogen-Ions with Organic Dyes as Indicators. I<. [FAJANS and 0. HASSEL (2. Elektrochem. $923 29 495-500).-Very dilute solutions of the alkali salts of fluorescein and its halogen substitution products show characteristic colour changes in the presence of ,silver halide sols when the solutions contain an excess of silver-ions.The must not be present. w. P. s.ANALYTICAL OHEMISTRY. ii. 61 colour changc is attributed to the association of the dye anion with the excess silver-ions which are adsorbed by the silver halide. By this association the electron sheaths of the dye anions undergo a similar change to that brought about in the formation of their silver and other salts. The colour change brought about in the presence of an cxcess of silver-ions can in some cases be reversed by an excess of halogen-ions. The halogen-ions in such cases displace the dye anions from the silver halido surface. The ease with which this occurs depends on the one hand on the adsorb- ability of the halogen-ion and on the other on the adsorbability of the dye anion. In the case of the halogen-ions the adsorbahility increases in the order I’ Br‘ Cl’ and in the case of the dye anions in the order fluorescein dibromofluorescein eosin di-iodofluorescein and erythrosin.On account of this behaviour these substances can be used as indicators in the titration of halogen-ions with silver-ions. In the case of chlorides fluorescein alone may be used the concentration of indicator should be about 1 /3OOOOOAV and the titration carried to a red colour. In the case of bromine and iodine dibromofluorescein and eosin may be used the concentration of the indicator (eosin) being 1/3OOOOM the titration being carried to il; violet colour. I n a mixture of chloride and iodide the iodide alone can be titratecl with eosin as indicator and the sum of the two halogen-ions with fluorescein as indicator.Determination of Calcium by the Qpacimetric [Nephelo- metric] Method. C. CHBNEVEAU and R. Boussu (Compt. rend. 1923 177 1296-1298).-A continuation of previous work (A. 1920 ii 327). It is shown that suspended calcium oxalate pre- cipitates obey within certain limits the general optical laws for turbid liquids and that the determination of calcium by the nephelometric method is always possible but necessitates careful preliminary experiments. E. E. T. Lead. VIII. The Microchemical Detection of Lead. L. T. FAIRHALL ( J . Biot. Chern. 1923 57 456461).-Behrens and Kley’s ‘‘ triple nitrite test,” which depends on the formation and recognition of the crystalline hexanitrite of potassium copper and lead [K,CuPb(NO,),] is employed. The conditions for carry- ing out this test on the ash from biological material containing traces of lead have been standardised.Amounts as small as l p g . may be detected. Electro-analytical Determination of Thallium by Means of the Anodic Deposition of Thallic Oxide. W. DIETERLE (Z. Elektrochem. 1923 29 493405) .-Using the method previously described by Gutbier and Dieterle (A. 1923 ii 880) the author has investigated the determination of thallium by anodic deposition of thallic oxide. The followiiig method is recommended a solution containing 0 - 1 4 - 5 g. of thallous nitrate and 0.1 g . of free nitric acid is placed in a roughened platinum dish diluted to 100 C.C. with water and either 10 C.C. of absolute alcohol or 5-10 C.C. of rectified acetone added. The dish is made the anode in the J.F. S. E. S.if. 62 ABSTRACTS OF CHEMICAL PAPERS. electrolysis and the cathode is a disk of platinum of 12 cmm2 which is rotated at 300 revolutions per minute. The electrolysis is carried out at 60-62" directly from a single lead accumulator. After ten hours the voltage is raised to 2-5-3-0 and a current of 0.05 ampere passed for a short time. The deposit which is firm is washed with water and dried at 160-170" preferably in an electrically heated oven but in any case reducing gases must not come into contact with the deposit. When dry the dish is cooled in a desiccator and weighed rapidly since the deposit is somewhat hygroscopic. The method is good the results being associated with an error of &0.2%. J. F. S. The Quantitative Separation from a Complex Mixture of Very Small Quantities of Copper Zinc Nickel and Cobalt.G. BERTRAND and M. MOKRAGNATZ (Bull. Xoc. chim. 1923 [iv] 33 1539-1547).-A method of separating traces of copper zinc nickel and cobalt from mixtures is described in detail. It is based on the precipitation of the hydroxides of the metals by ammonia in the presence of calcium hydroxide. In the case of zinc a crystal- line compound of the formula (OH*Zn*O),Ca,n,H,O is obtained but the precipitate given by each of the other metals may be a mixture of the hydroxide of that metal with a calcium compound. Experimental results for comparison of the method with standard procedure are given; these &ow good agreement. [Cf. B. Jan.] H. J. E. The Solubility of Mercuric Sulphide in Ammonia and Its Influence on the Tests for Mercury and Arsenic.C. GHIGLIOTTO (Ann. Chim. Anulyt. 1923 5 326).-In toxicological analysis the organic matter is destroyed by heating with sulphuric acid ; this solution after treatment with hydrogen sulphide for the precipitation of arsenic and mercury yields a mixture of sulphur and organic sulphur compounds. If the precipitate is extracted with ammonia to dissolve any arsenic all the mercury will also dissolve owing to the presence of the sulphur and organic sulphides and on subsequent treatment for arsenic by the Marsh test no mirror will be obtained if mercury is present. Thus neither arsenic nor mercury may be detected. The solution should be again evaporated with sulphuric acid and the heating continued for some time; subsequent treatment according to the usual methods will then give satisfactory separations.Determination of Aluminium and its Separation from other Metals. G. JANDER and B. WEBER (2. ungew. Chem. 1923 36 586-590).-The method of determining aluminium in alloys which is based on the sublimation of the chloride in a current of dry hydrogen chloride gas (Jander and Wendehorst A. 1922 ii 529) suffers from the disadvantage that traces of -magnesium and manganese sublime together with the aluminium chloride. A complete separation of these metals is effected by a series of re- sublimations carried out under carefully controlled conditions in a A. R. P.ANALYTICAL CHEMISTRY. ii. 63 slightly modified apparatus a diagram of which is given. The slight trace of manganese subliming in the first sublimation is easily separated but the somewhat large quantity of magnesium is only separated after four or five resublimations which are carried out by driving the sublimate on to successive portions of a glass tube lying inside the sublimation tube proper on which the aluminium chloride is ultimately allowed to deposit.The subliming operation takes about three hours. The improved method affords an accurate means of determining aluminium directly in alloys and of separ- ating it from the other metals present in much smaller proportions so that the determination of these metals can be effected with much greater accuracy. The method also allows of the determin- ation of oxides such as silica and alumina which occur in alloys and cause important variations in their properties. The aluminium chloride can eventually be obtained with less than 0-007% of manganese and 0.025% of magnesium.Details are given of the behaviour of the other metals present on heating in a current of dry hydrogen chloride gas. Any elementary silicon present sublimes as hydrides or silico-chloroform and is subsequently separated from the aluminium chloride by evaporation with acids and filtration. The silicon in “ Silumin ” can be accurately deter- mined by dissolving the alloy in dilute hydrochloric acid filtering off the silicon and weighing. Only 0.05y0 of silicon was found to escape in this process in the form of volatile compounds. Speci- men analyses of various alloys are given. Titrimetric Determination of Aluminium in Aluminium Salts in particular in ” Liquor Aluminii Acetici.” A. WBHLK (Ber. Deut. p h r m . Ges. 1923 33 195-204).-0wing to various objections to the titrimetric method proposed by Valentin (Apoth.- Ztg. 1912 590) notably the uncertainty of the composition of the precipitated basic aluminium arsenate and of the accuracy of iodimetric titrations of arsenic acid in strongly acid solution the following method which gives accura,te results is preferred. To a known quantity (about 10 c.c.) of “ Liquor aluminii acetici” is added 10 C.C. of a solution (12 g. of sulphuric acid and 6-5 g. of potassium sulphate made up to 100 C.C. with water) and the potassium alum formed is completely precipitated by addition of 90 yo ethyl alcohol (100 c .c .). The precipitate is collected repeatedly washed with 20 C.C.of 70% alcohol and finally dissolved in boiling water. The subsequent procedure depends on the reactions AlJS0,),+3BaC1,=3BaS0,+2AlC13 and AlC13+3NaOH=A1(OH) +3NaC1. To the alum solution is added 20 C.C. of a barium chloride solution (containing 4 g.) and five drops of 1% phenol- phthalein and N / l -sodium hydroxide is run in without removing the barium sulphate until a red colour is obtained. The solution is diluted to 200 c.c. and the hydrolysis completed by heating to boiling. The titration with sodium hydroxide is then carried to the true end-point. The application of this method to aluminium salts in the presence of ammonium and iron salts is being investi- gated. W. T. K. B. H. C. R.ii. 64 ABSTRACTS OF CHEMICAL PAPERS.The Simultaneous Electrometric Determination cf Iron and Manganese. E. MULLER and 0. WAHLE (2. anorg. Chem. 1923 130 63-68).-When the titration of a ferrous salt with potassium permanganate is continued after the normal end-point has been reached in presence of hydrofluoric acid a second reaction takes place namely the oxidation of the manganous salt according to the equation MnO,' +4Mn"+ SH'= 5Mn"'+4H20. The second reac- tion requires one-quarter as much permanganate as the first. Both end-points are sharply marked on the electrometric titration curve but if the first reaction takes place in presence of hydrofluoric acid the first end-point is reached too soon because ferrous salts in presencc of fluoride-ions are very susceptible to atmospheric oxidation.Obviously the method of titration can be applied to the determin- ation of manganese already present with the ferrous salt. It is recommended to carry out the &st titration colorimetrically in presence of sulphuric acid in a platinum dish at 18" then to add 7 g. of potassium fluoride and 5 C.C. of 2N-sulphuric acid per 100 C.C. of solution aiid titrate the manganese electrometrically at 80". E. H. E. Interference of Cobalt in the Bismuthate Method for Manganese. G. E. F. LUNDELL (J. Amer. Chem. Soc. 1923 45 2600-2603) .-The proper conditions for the determination of manganese by Blum's bismuthate method (A. 1912 ii 1214) have been investigated and it is shown that in addition to the conditions laid down by Blum the following must also be observed. The solution should not contain more than 0.05 g.of manganese; moderate amounts of sulphuric acid are not harmful as for example 5 C.C. of sulphuric acid and 10 C.C. of nitric acid in 50 C.C. of solution ; ehlorides must be absent; the temperature of the solution may be varied from 5" to 25" ; half a minute is a sufficient time for agitating the solution but a longer agitation will do no harm ; the bismuthate should always be tested for its oxidising power before use the asbestos used in the filtration must be very carefully purified all hydrochloric acid and chlorides must be removed from it; the ferrous sulphate solution should not be kept more than ten minutes before titration and the permanganate solution which may be 0.03N or O-lON is best standardbed against sodium oxalate as described by McBride (A.1912 ii 494). When thcse precau- tions are taken the accuracy of the method for amounts up to 0.06 g. is within 1 part in 500 parts. The process is singularly free from interference by other elements the only elements which have been recognised as troublesome being chromium and cerium. The interference caused by these and other elements is discussed. Cobalt is oxidised by the bismuthate and the oxidised compound reacts with both ferrous sulphate and permanganate. The latter reaction does not commence until some permanganate has been decomposed and then proceeds so rapidly that it causes low results for manganese. In the presence of cobalt Bord's method is to be preferred for accurate analyses and the persulphate arsenite method for routine work.J. F. S.ANALYTICAL CHEMISTRY. ii. 65 Determination of Small Quantities cf Mo3ybdenum. Application to Ammonium Phosphomolybdate for the Indirect Titration of Phosphorus. A. VILA (Compt. rend. 1923 177 1219-1321) .-If ammonium molybdate or phosphomolybdafe is reduced in hydrogen a t 700" (silica tube) it is quantitatively reduced to molybdenum. A solution of molybdic acid (3 g.) in 60% phosphoric acid (10 c.c.) is decolorised with permanganate and diluted to 50 C.C. The molybdenum is dissolved in this solution. After diluting and boiling molybdenum-blue Mo0,,4Mu03 is formed quantitatively and may be titrated against permanganate (cf. Font& and Thivolle A. 1923 ii 264). The method is used for determining small quantities of phosphorus.Determination of Rare Metals and their Separation from Other Metals. 111. The Separation of Uranium from Titanium Iron and Aluminium. L. MOSER (Nonutsh. 1923 44 91-96).-The separation by hydrogen sulphide from an ammoniacal sulphosalicylic acid solution occurs readily but the uranium sulphide is precipitated in a colloidal condition so that filtration is extremely diflicult) and the precipitate carries down other salts from the solution. The separation by means of the soluble complex carbonate (NH4)4'cTog(cj0& formed by addition of ammonium carbonate which throws down the common metals is quantitative only in presence of titanium if very small amounts of uranium compounds are present. Titanium may be separated as the hydroxide Ti(OH),; the two oxides are precipitated together by ammonia and after ignition and weighing are dissolved by fusion with potassium sodium carbonate and treatment with hydrochloric acid..Addition of potassium bromate and sulphate precipitates the titanium hydr- oxide which is filtered ofl dried and weighed; uranium may be separated from the bromate-free filtrate by precipitation with ammonia but if aluminium is present ammonium carbonate in excess is added to the neutralised solution to remove this as hydr- oxide ; the precipitation must be repeated to get a good separation. Iron is also separated by the complex carbonate method the E. E. T. precipitation being repeated for quantitative work. s. I. L. Determination of Uranyl Compounds. G. JANDER and K. REEH (2. anorg. Chem. 1923,129,293-301).-Sexavalent uranium is reduced to the quadrivalent state in sulphuric acid solution and may then be reoxidised by means of standard permanganate according to the equation 2MnO,'+5U""+2H,O=2Mh"+ 5U02"+4H'.In view of the fact that statements are to be found that acid reduction of F proceeds to a point below UIV tests were carried out using zinc zinc amalgam and aluminium for the reduction of a standard uranyl solution. Both zinc and zinc amalgam were found to be most unsuitable but aluminium in sulphuric acid reduces UVI quantitatively to F. A correction must be applied for any iron present in the aluminium nsed. H. H. The end-point is quite sharp. VOL. CXXVI. ii. 3ii. 66 ABSTRACTS OF CHEMICAL PAPERS Determination of Titanium by Reduction with Zinc and Titration with Permanganate.G. E. F. LUNDELL and H. B. KNOWLES ( J . Amer. Chem. SOC.. 1923 45 2620-2623).-Titanium may be determined volumetrically as follows. A solution of titanic sulphate in 5 vol.% sulphuric acid is run through a Jones's reductor 19 mm. bore containing a column of zinc 43 cm. in length into a bottle containing three times as much ferric sulphate solution as is necessary to oxidise the tervalent titanium back to the titanic state of oxidation. The solution is then titrated with a 0-1N solution of permanganate. The results are extremely accurate in the absence of substances which interfere with the reaction. All substances with such interfering action except niobium can be readily removed by well-known methods. Substances which interfere are some organic compounds tin arsenic antimony molybdenum (ous) iron (ow) chromium (ous) vanadium (ous) tungsten uranium and niobium. The above method for the determination of titanium is superior to the method in which the reduced material is run into standard permanganate and to methods in which the titanous solution is collected and prepared in atmospheres of hydrogen or nitrogen.Titanous solutions required for standard solutions can be prepared by means of a Jones's reductor which passes through the cork of a bottle through which a current of hydrogen is being circulated. 5. F. S. Electrometric Determination of Vanadium and Uranium Separately in the Presence of One Another and in the Presence of Iron. E. XULLER and A. FLATH (2. Elektrochem. 1923 29 500-508) .-The authors have repeated the electrometric titrations of Gustavson and Knudson (A.1923 ii 185) of vanadium uranium and iron separately and in the presence of one another. It is shown that vanadium when titrated electrometrically with potassium permanganate at 80" shows three sharply marked breaks in the titration curve. The volume of permanganate used between the first and second break corresponds with the oxidation from tervalent vanadium to quadrivalent vanadium and that between the second and third break to the oxidation from quadrivalent vanadium to quinquevalent vanadium. These breaks are very sharp very much sharper than those obtained in a titration at 18". In the presence of iron at SO" the three vanadium breaks are found in the curve but the break for the oxidation FeU+FeIII is not clear; at the ordinary temperature the iron break is sharp and the VI=+VIv disappears.Iron and vanadium can be estimated in mixtures by reducing in a Jones's reductor then titrating at 80" to the second break which leaves the vanadium in the VIV stage then in cold solution to the third break where the iron is all oxidised and finally at 80" to the fourth break where the vanadium is in the stage VV. The amount of iron can then be calculated from that of permanganate used for the cold titration between the second and third breaks and the vanadium from the amount used between the first and second or the third and fourth breaks. Uranium solutions reduced by zinc to the stage UIV are oxidisedANALYTICAL CHEMISTRY. ii. 67 by permanganate in the presence of sulphuric acid to the stage UVI and at this point a definite break is shown in the titration curve for SO".In the case of mixtures of uranium and iron the titration with permanganate after reduction is carried to the first break in the titration curve at SO" which indicates the completion of the oxidation Urn to UVI and then to the second break at 18" which indicates the completion of the oxidation FeII +Fern. Mixtures of vanadium and uranium are titrated a t 80" with permanganate the permanganate used between the first and second breaks effects the oxidations VIII+VIV and UIV+UVI and that between the second and third breaks Vm-+vV. Mixtures of uranium vanadium and iron are reduced with zinc and titrated at 80" to the second break in the titration curve with permanganate ; at this point the oxidations F1+V and Um-+VI are com- plete then the titration is continued a t 18" to the third break where the oxidation FeII-Fem is complete and then a t 80" to the fourth break where the oxidation Vm+-VP is complete.J. F. S. Volumetric Determination of Vanadium in Steel. A. T. ETHERIDGE (Analyst 1923 48 588-590).-!L'wo g. of steel is dissolved in 5 C.C. of dilute sulphuric acid in a 500 C.C. flask. Two to three C.C. of nitric acid (d. 1.42) is added and any residual carbide dissipated if necessary by digestion with a little more nitric acid. The solution is evaporated until fumes of sulphur trioxide appear. Nitric acid destroys the diphenylcarbazide subsequently used as an indicator and must be removed. The sulphatjes are dissolved in 400 C.C.of water and the solution is boiled porous platme being added to prevent bumping after which a saturated solution of potassium perrnanganate is added drop by drop until a permanent precipitate is formed and boiling is continued for fifteen minutes. A large excess of manganese dioxide must be avoided or chromium and vanadium may be associated with the precipitate. The solutioii is cooled filtered through asbestos and the precipitate washed with cold water the filtrate being received in a large Erlenmeyer flask. Fifty C.C. of syrupy phosphoric acid is added and after cooling a slight excess of ferrous ammonium sulphate to reduce chromate and vanadate. This excess is titrated with 0-IN permanganate to a permanent pink. Fifty C.C.of concentrated sulphuric acid is now added and the solution cooled after which sufficient standard ferrous ammonium sulphate solution is added to reduce all the vanadium and also to leave a small excess the approximate amount of vana- dium present being found by preliminary tests either by this process or colorimetrically with hydrogen peroxide. After thorough shaking standard 0-1N dichromate solution is added until within a few C.C. of the end this end-point being approximately known from the preliminary test. The indicator is prepared as follows 0.05 g. of diphenylcarbazide is dissolved in a few C.C. of glacial acetic acid and diluted to 500 C.C. Five C.C. of this solution is acidified with three drops of dilute sulphuric (1 3) and one drop of 0.1N dichromate solution added. About half of this solution is added to the liquid 3-2ii.68 ABSTRACTS OF CHEMICAL PAPERS. in the flask and dichromate solution continually added unt,il the purple colour has faded away. The remainder of the indicator is now added and the titration continued until the purple colour is discharged and the solution becomes apple-green. The ferrous ammonium sulphate is standardised under the same conditions of acidity and volume as in the test. The difference between the two tit'rations represents vanadium in terms of 0-1N dichromate solution. The process was tested on electrolytic iron to which known amounts of vanadium and chromium had been added. Up to 2% of vanadium results correct to wit'hin 0-020/ are obtained but beyond 2% of chromium the green chromic sulphate obscures the end-point and for 4% chromium steel it is recommended that work be carried out on a 1 g.sample. The presence of cobalt interferes with the permanganate end-point. Colorimetric Determination of Small Quantities of Bis- muth. c. E. LAPORTE ( J . Phurm. Chim. 1923 [vii] 28,304-305 ; cf. A. 1923 ii 792).-The method is based on the precipitation of b'smuth as quinine iodobismuthate followed by solution in acetone. clear yellow to orange-yellow proportionally to its bismuth content ; 0.0001 g. of bismuth can be determined with an error not exceeding 2-3%. To determine bismuth in urine the ash is taken up in 10% nitric acid in sufficient excess to prevent the formation of the basic salt and is treated with Leger's reagent as modified by Aubry (A.1922 ii 165) containing 1% quinine sulphate and 27k potassium iodide. It is better to make up the solutions of quinine sulphate and potassium iodide separately and mix the two as required. The precipitation of bismuth is quantitative and the orange-yellow precipitate that is formed is dissolved in acetone and its intensity compared with a standard solution prepared similarly. An Improved Method €or the Separation of Gas Mixtures. M. SHEPHERD and F. PORTER (Ind. Eng. Chem. 1923 15 1143- 1146) .-The apparatus described was designed for the separation of the hydrocarbons in natural gas by fractional distillation for the purpose of analysis and for the preparation of constituents in a state of purity with greater expedition than was possible with older types.The whole apparatus including distillation bulbs pump reservoirs and measuring burette is constructed in one unit. Three distillation bulbs are provided permitting of a rapid distillation from one and condensation of the vapour in another (by varying the pressure on the liquid air jackets surrounding them). I n the case of natural gas the nitrogen and helium which remain uncondensed at - 190" are a t once removed and measured. The methane fraction is twice distilled and condensed in successive distillation bulbs and is then removed and measured and the higher homologues are similarly separated one by one the temperatures of the bulbs being appropriately controlled throughout. Tests with an artificially prepared gas mixture have demonstrated the accuracy of this procedure.For the preparation of pure gases suitable for the determination of physical constants repeated distillations are H. C. R. T In e intensity of the coloration of the acetone solution varies from D. R. N.ANBLYTICAL CHEMISTRY. ii. 69 necessary. Analyses performed by the above method differ from those obtained by the combustion method but this is attributable to the impossibility of obtaining a correct percentage composition of a mixture of more than two hydrocarbons by the latter method and to the daculty of obtaining complete combustion. Determination of Methyl Alcohol. A. €3. LYOKS ( J . Amer. Pharm. Amoc. 1922 11 682-686; cf. ibid. 12).-Dried egg- albumin is used in preference to milk or peptone and Chapin's method of oxidation using permanganate is employed.A very dilute solution of methyl alcohol containing about 1% of ethyl alcohol is oxidised and 1 C.C. is mixed with 1 C.C. of water containing 0~005-0.010 g. of egg-albumin and 2 C.C. of sulphuric acid contain- ing a trace of ferric chloride. The purple colour which reaches a maximum intensity in ten minutes is compared with a standard. CHEMICAL ABSTRACTS. C. I. Detection of Methyl Alcohol in Ethyl Alcohol. C. H. LA W~LL (Amer. J. Phurrn. 1923 95 812-820).-The literature of the subject is reviewed. It is considered that the test depending on the oxidation of the alcohols with permanganate and the testing of the solution of aldehydes for formaldehyde with the magenta- sulphuric acid test solution described in the U.S. Pharmacopoeia tenth revision is satisfactory.The cooling of the solution after the addition of the sulphuric acid can be obviated by adding acid previously diluted (1 3) and cooled. The test is sensitive to 1 part of methyl alcohol in 500 parts of ethyl alcohol and if smaller pro- portions are present or suspected a preliminary fractioiiation may be carried out and the test applied to the 1 C.C. fraction coming over first from a 10 C.C. sample. In this way positive reactions were obtained with 1 part of methyl alcohol in 10,000 parts of ethyl alcohol. H. C. R. Ether Anssthesia. I. The Estimation of Ethyl Ether in Air and in Blood and its Distribution Ratio between Blood and Air. P. A. SHAFFER and E. RONZONI ( J . Biol. Chem. 1923,57 741-760).-A modification of the Nicloux method for the estimation of ether in air water or blood is described.The solution is grated and the air passed through a series of three tubes containing 500/ sulphuric acid and dichromate in excess of that required for the oxidation. When absorption is complete the solutions are heated to boiling to complete the oxidation. After cooling t>he contents of the tubes are washed into a flask diluted wit'h water potassium iodide is added and the iodine liberated titrated with thiosulphate. The difference between this result and that obtained with a blank experiment gives the amount of dichromate used in oxidising the ether to acetic acid. No carbon dioxide is produced if the tempera- ture and concentration of acid are properly controlled. If desired the ether may be absorbed in one tube of concentrated acid.This is then diluted to 40 or 50% and a measured quantity of dichromate added; the remainder of the procedure is as above. Using this method the distribution of ether between air and water and airii. 70 ABS!L'RACTS OF CHEMICAL PAPERS. and blood has been determined for a series of temperatures. The ratios obtained below 35" are somewhat higher than those obtained by Haggard (A. 1923 ii 343) ; between 35" and 40° the results agee. Determination of @-Naphthol in a-Naphthol and a-Naphthol in p-Naphthol. J. PROCHAZKA (Ind. Eng. Chem. 1923 15 944-945).-To determine @-naphthol in a-naphthol use is made of the fact that the latter couples in acid solution with p-nitro- diazobenzene whilst the former does not or only very slowly. a-Naphthol (0.36 g.) is dissolved in 3040 C.C.of alcohol cooled to ti" and titrated with p-nitrodiazobenzene solution (100 C.C. = 1 Q . of nitrite) containing about 30-50~0 excess of hydrochloric acid above the theoretical amount for diazotisation. Rather less than the calculated quantity of diazo-solution is run in fairly quickly and the solution left for about thirty minutes. The p-nitrobenzene- azo-a-naphthol separates and the disapperaance of the diazo- compound is determined by spotting on filter-paper with dilute sodium hydroxide solution. Addition of diazo-solution is then continued. When all the a-naphthol has combined the spot test after further addition of diazo-solution shows the red colour of p-nitrobenzeneazo- p-naphthol. The volume of diazo-solution used is a measure of the purity of the a-naphthol; the @-naphthol is obtained by difference. The method of determining a-naphthol in P-naphthol consists in adding to an alcoholic solution of the naphthol rather more than sufficient p-nitrodiazobenzene to com- bine with arll the a-naphthol present.After half an hour the solution is diluted with boiling water the precipitate of impure p-nitrobenzeneazo-wnaphthol collected on a filter washed with hot water and then boiled with 0.5% sodium hydroxide solution. The blue solution is filtered and compared colorimetrically with standard solutions of pure p-nitrobenzeneazo-cx-naphthol or since the colour is fugitive with a secondary standard for instance the colour from tetrazotised benzidine coupled with H-acid (Diamine Blue 2B).E. M. R. The Detection of Pentose Formaldehyde and Methyl Alcohol. J. B. SUMNER ( J . Arner. Chem. Soc. 1923 45 2378- 2380).-Bial's reagent for pentoses will keep if the six g. of orcinol and forty drops of 10% ferric chloride solution are dissolved together in 200 C.C. of ethyl alcohol without the addition of hydrochloric acid. Fifteen drops of this solution 5 C.C. of the sugar solution and an equal volume of fuming hydrochloric acid are heated at 100". A clear blue colour develops if 1 mg. of arabinose or xylose is present; with less pentose the colour is green with more a preci- pitate is formed. With formaldehyde a white precipitate appears but only on heating if the concentration is small. With a very small quantity of formaldehyde the precipitate is not formed but a yellow coloration is developed.When the material is heated at 100" for fifteen minutes the precipitate turns brown; it now dissolves in an excess of alkali to give a pink solution or with much precipitate pink E. s.ANALYTICAL CHEMISTRY. ii. 71 flocks are produced. With such a small quantity of formaldehyde that there i8 no precipitate after heating for fifteen to twenty minutes the addition of sodium hydroxide produces a pink or salmon-coloured solution with a green fluorescence which is so intense that this test will readily show one part of formaldehyde in ten million parts of water. One part of formaldehyde in one million parts of water may be detected by the use of an alcoholic solution of orcinol and two drops of ZOyo sodium hydroxide solution.The solution becomes pink owing to oxidation by atmospheric oxygen. The test is only of use in the absence of certain interfering substances Acetaldehyde gives a similar precipitate with orcinol ferric chloride and hydrochloric acid but only in concentrated solution; this precipitate does not darken on heating and dissolves in alkali to give a yellow solution which however does not show any fluorescence. The formation of the white precipitate may be used as a test for methyl alcohol after its oxidation to formaldehyde. The most suitable oxidising agent is potassium dichromate and sulphuric acid since this forms very little formaldehyde from ethyl alcohol. One C.C. of the suspected alcohol e.g. from the distillation of an alcoholic beverage 2 C.C. of 6.7 yo potassium dichromate solution and 2 C.C.of 62% sulphuric acid are allowed to remain. If the reduction to blue chromic sulphate takes much longer than forty seconds too much water is present. Fifteen C.C. of water is added and the mixture heated a t 100" for ten minutes. Five mg. of orcinol in one C.C. of water is added and the heating continued. If the alcohol contains 5% or more of methyl alcohol a precipitate is formed after about five minutes. With only lo/ fifteen minutes are required. The test will show methyl alcohol down to o*5y0 but it may tlhen be necessary to heat for thirty minutes and then cool. Even smaller quantities may be detected by removing the chromium by heating with sodium hydroxide. The filtered solution has a green fluorescence if traces of methyl alcohol are present.Formic acid amyl alcohol acetone and furfuraldehyde do not interfere with the reaction. Glycerol is eliminated during the distillation of the alcohol but gives a positive test if even as little as 5 mg. is added to the distillate w. s. N. Determination of Pentoses and Pentosans. 11. Determin- ation of Furfuraldehyde. N. C. PERVIER and R. A. GORTNER ( I d . Eng. Chem,. 1923 15 1255-1262).-Several new volumetric methods for determining furfuraldehyde in dilute aqueous solution were test'ed. Iodine in alkaline solution did not give dependable results and acid permanganate was unsuccessful owing to the catalytic reduction of the permanganate by furfuraldehyde. The use of potassium bromate in acidified furfuraldehyde solutions containing potassium bromide was eminentjly successful. Specific directions are given for obtaining theoretical yields of furfuraldehyde from pentose materials and for the volumetric determination of this substance in the resulting distillates.Representative results of determinations on pure pentoses and pure furfuraldehyde areii. 72 ABSTRACTS OF CHEMICAL PAPERS. recorded and the factors to be used for the conversion of potassiuin bromate used to furfuraldehyde pentose pentosan or the corre- sponding methyl derivatives are given. The acidity of the solution to be titrated should not exceed 4-5y0 as further oxidation of the primary product of the interaction of bromine and furfuraldehyde results in the presence of high concentrations of acid. The velocity of this secondary reaction appears to be so small as to be without appreciable effect on the titration if the conditions proposed are adhered to.Either hydrochloric or sulphuric acid may be used and a potassium bromide solut>ion of approximately 1 :/o concentra- tion is satisfactory. The potassium bromate solution must be added slowly and the presence of any considerable excess a t all times avoided. The amount of furfuraldehyde present does not influence the accuracy of the method. In titrating dilute solutions with potassium bromate the end-point can be readily located by the use of a simplified electrometric apparatus consisting of two platinum wires a galvanometer and a key. The time factor of the reaction involved is made the basis of the foregoing method for finding the end-point.Hydroxymethylfurfuraldehyde a product of the acid distillation of hexose materials apparently interferes slightly with the use of the proposed method but the effect is small enough to be disregarded. Lzevulic acid a further decomposition product of hexoses is without any effect whatever. Methyl pentoses or pentosans will interfere in that methylfurfuraldehyde is formed and this will react with the bromate solution. A bibliography of the subject is appended. J. B. SUMXER and V. A. GRAHAM (Proc. SOC. Exp. Biol. &led. 1922 20 96; cf. A. 1921 ii 564).-Blood (1 c.c.) is laked with 2 C.C. of water 4 C.C. of a 2.94% solution of neutral sodium dinitrosalicylate and 2 C.C. of 0-4N-sulphuric acid are added and the mixture is filtered ; 3 C.C. of the filtrate is heated for three minutes in a Folin sugar tube in a boiling water-bath 1 C.C.of 3% sodium hydroxide solution is added and the heating is continued for ten minutes. The solution is then cooled and diluted to a volume suitable for colorimetric comparison with a standard prepared by heating 2 C.C. of a o.01570 solution of dextrose with 1 C.C. of a 1.780,; solution of neutral sodium dinitrosalicylate for three minutes adding 1 C.C. of the alkali and heating for ten minutes. Determination of Lactose by the Polarimetric and the Gravimetric Methods. A. L. BACHARACH (Analyst 1923 48 521-528).-Careful redetermination of the polarimetric constants of lactose gave the following results [a]k=+62.40+(t-20) x 0.072 ; [a]~e54s=+61.94+(t-20) ~ 0 . 0 5 5 . The latter formula is applicable when the green band of the mercury vapour arc is the source of illumination.For the gravimetric determination of lactose the modified Fehling method of Quisumbing and Thomas (A. 1922 ii 92) was found satisfactory in the respects claimed by them. The copper was however determined after ignition by H. C. R. Dinitrosalicylic Acid as a Reagent for Blood-sugar. CHEMICAL ABSTRACTS.ANALYTICAL CHEMISTRY. ii. 73 weighing as cupric oxide. The ratio of cupric oxide to lactose was constant and equal to 1.565. A New Titrimetric Method for Determining Formic Acid. 0. RIESSER (Biochem. Z. 1923,142,280-281).-The author points out that the method described by Utkin-Ljubovzov (A. 1923 ii 588) for the determination of formic acid had already been published by him (2.physiol. Chem. 1916,97,355) and used by other workers. It is unnecessary to filter the precipitated mercurous chloride and the determination may be carried out in neutral solutions. Application of the Method of Diffusion to the Detection and Separation of Fatty Acids. A. HEIDUSCHKA and J. RIPPER (2. EZektrochem. 1923 29 552-555).-Details are given of the application of the method of diffusion to the separation of the so-called heptadecoic acid into its constituent fatty acids palmitic acid and stearic acid. An alcoholic solution of heptadecoic acid is allowed to diffuse through a collodion membrane and it is shown that the respective rates of diffusion of the constituents decrease as the molecular weights increase. The rate of diffusion is not always proportional to the concentration.A mixture containing an excess of palmitic acid may yield a diff usate containing an excess of stearic acid. The results were found to be applicable to a mixture of lauric and stearic acids. Determination of Tartaric Acid by Oxidation with Potassium Iodate and Sulphuric Acid. R. STREBINGER and J. WOLFRAM (Oesterr. Chem. Ztg. 1923 26 156-157).-The oxidation of the tartaric acid proceeds according to the equation C,H +2KIO + H,SO = K,SO + I + 4C0 + 4H,O. A weighed quantity of about 0.3 g. of the tartaric acid is added to a flask containing 1 g. of potassium iodate a few drops of water and 30 C.C. of concentrated sulphuric acid and the mixture is heated on a sand-bath for thirty minutes until the greater part of the liberated iodine has been expelled.After cooling the mixture is diluted with water the solution boiled to expel remaining traces of iodine again cooled and diluted to a definite volume. The excess of iodate is then determined by treating an aliquoit portion with potassium iodide and titrating the iodine with thiosulphate solution. If desired the tartaric acid may be precipitated as lead tartrate by means of lead acetate in acetic acid solution ; the lead tartrate is collected washed with hot water and then oxidised as described. Titration of Amino- and Carboxyl-groups in Amino-acids etc. 1-111. In Aqueous Solution. L. 3. HARRIS (Proc. Roy. Soc. 1923 [B] 95 440484).-From a consideration of the %itration curves of amino-acids it is deduced that the amino- and carboxyl-groups can be estimated by titration.At a pH of 0.7-0.1 the mono-amino-mono-carboxylic acids are completely dissociated as acid salts of the amino-group so that the amino-group can be estimated by titration with acid to this p,. Similarly their carboxyl groups are completely dissociated a t a pa of 11.75 and can be determined by titration with alkali to this degree of alkalinity. H. C. R. J. P. J. S. G. T. w. P. s.ii. 74 ABSTRACTS OF CHEMICAL PAPERS. I n practice the titration is carried to the mid-point i.e. the point zt which 50% of the amino- (or carboxyl-) group is neutralised. This point is determined from the titration curve of the acid under investigation. Mixtures of amino-acids may be determined by means of a series of titrations to different pH values.The original paper should be consulted for the theoretical basis of the work and the necessary corrections involved in the special cases of the various amino-acids. C. R. H. Action of Carbonyl Chloride on Ricinolein. A. PIUTTO and A. CURZIO (Rend. Accad. Sci. Fis. Bat. Napoli 1921 [iii] 27 40- 47).-The action of carbonyl chloride on castor oil or on ricinolein results in the formation of a compound which contains the per- centage of chlorine corresponding with glyceryl A-chlorocarbonato- ricinoleate (CH,*[ CH,],*CH[ OCOCl]=CH,*CH:CH*[CH,] ,*CO,),C,H ; with ricinoleic acid the chlorocarbonato-acid itself appears to be formed. On the other hand olive almond arachis and sesame oils take up only 0.08-0.18% of chlorine when treated with carbonyl chloride. Adulteration of castor oil with other oils is therefore detectable in this way.The name chlorine index is proposed for the percentage of chlorine thus absorbed the value for the samples of castor oil examined being 9.03. Melting Point and Iodine Value of Refined Natural &Camphor. MAURICE s. SALAMON (Analyst 1923 48,536-539). -The camphor was purified by sublimation only the middle fraction being used. No difficulty was experienced in obtaining samples by this means which had a melting point of 179" the melting point remaining constant after repeated sublimation. Iodine value determination on such samples by Wijs's method gave values not exceeding 0.1. It is extremely diflicult to remove the last traces of oil and it is probable that this small residual iodine value is due to a minute trace of oil still present in the camphor.To determine the iodine value of camphor oil samples of oil were expressed from both Chinese and Japanese crude camphor and the dissolved camphor was removed as completely as possible by repeated freezing. The iodine value of the oil finally obtained varied between 86.0 and 91.7. It is considered that this oil contained not more than 10% of camphor and in calculating the percentage of camphor in samples from the iodine value it is recommended t o take the iodine value of the oil as 100. L. KOFLER and 0. DAFERT (Ber. Deut. pharm. Ges. 1923 33 215-229).-Previous methods of determining the saponin content of the root (e.g. with baryta. magnesia lead acetate) are shown either to give too low results or to involve alteration of the substance.I n order to extract the saponin quantitatively and unchanged the powdered root is extracted exhaust'ively with boiling water the ext'ract; evaporated to a syrup and the saponin fractionally precipitated first with alcohol and then with ether. Each of the two precipitates represents about 10% of the starting material with an ash conteiif T. H. I?. H. C. R. Saponin from Gypsophila puniculatu.ANALYTICAL CHEMISTRY. ii. 75 of 10-19;& and 4-6:4 and a hamolytic index of 1 100,000 and 1 500,000 for the alcohol and ether fractions respectively. By means of Pauli's electrodialysis apparatus the ash in the saponin is reduced to o.0770. Elementary analysis of the alcohol and ether fractions gives respectively C = 47-88% H = 7-60% and C=48*40~0 H=7.21%.The saponin is considered to have a mo1. wt. of 3000 or more. Microchemical Detection of Hydrocyanic Acid. F. VON NEUREITER (Deut. 2. yes. gerichtl. Med. 1923 2 313-317; from Chem. Zentr. 1923 iv 230-231).-The material in which the presence of cyanides is suspected is mixed with concentrated oxalic acid solution in a glass vessel which is covered by a slide carrying a hanging drop of 1% silver nitrate solution coloured blue with Methylene-blue. In the presence of hydrocyanic acid a turbidity is produced which is seen under the microscope to consist of fine blue needles of silver cyanide the identity of which can ba established by ordinary reactions. The Colorimetric Determination of Hydrocyanic Acid as Thiocyanate. I. M. KOLTROFF (2.anal. Chem. 1923,63,188-190). -To 5-10 C.C. of a cyanide solution are added 1 C.C. of 1% sodium tetrrtthionate solution and 5 drops of 10% ammonia solution. The mixture is warmed on a water-bath for five minutes at 60-56" cooled and treated with 2 C.C. of 4Y-nitric acid and 3 drops of S-ferric chloride solution. The colour is matched against a standard. Working with 10 C.C. of test solution 1 mg. of hydrocyanic acid per litre is easily detected. The sensitiveness of the reaction is 0.3 mg. per litre. Determination of Uric Acid in Blood-serum and Tissue Extracts. K. HARPUDER (2. ges. Exp. Med. 1923 32 378-386; from Chern. Zentr. 1933 iv 443444).-The method for which detailed directions arz given consists in the case of sera of removal of proteins followed by precipitation of the uric acid as zinc mate.The precipitate after centrifuging is taken up with sodium chloride solution and the uric acid determined colorimetrically by means of phosphotungstic acid. In the case of tissues a modified method is described in which the material is first extracted with 3% sulphuric acid in a refiux apparatus. Determination of Uric Acid and Purines in Blood-serum and Urine by the Copper Urate Method. H. CHABANIER M. LEBERT and C. LOBO-ONELL (Bull. Xoc. Chim. biol. 1923 5 731-738).-The purines are precipitated in the form of their insoluble copper compounds and the nitrogen in the precipitate is determined by Kjeldahl's method. By this method uric acid added to urine or to serum can be determined to within about 6%. A Quantitative Colour Reaction given by Adrenaline and Urine.H. FRIEND ( J . Biol. Chem. 1923 57 497-505).- Adrenaline niay be determined coloriinetrically by means of the [Cf. B. Jan.] W. T. K. B. G. W. R. The above conditions must be closely adhered to. A. G. P. G. W. R. C. R. H.ii. 76 ABSTRACTS OF CHEMICAL PAPERS. red coloration which is produced when it is treated successively with sulphanilic acid nitrous acid and ammonia. For the deter- mination of adrenaline in urine the latter is precipitated with lead acetate and the excess of lead removed by means of ammonium sulphate. The determination is then carried out on two portions of the filtrate one of which is first treated with ferric chloride at 50" to destroy the adrenaline. The difference between the results gives the adrenaline content of the urine.According to this method normal urine contains 0.2-0.4 mg. of adrenaline Fer 100 c.c.; larger variations occur in pathological urines. Determination of Bilirubin in1 Urine. K. HOESCH (blunch. Xed. Woch. 1923,70 534; from Chem. Zenir. 1923 iv 444).-The method depends on coupling with diazoacetophenone. After treating the urine with a solution of diazoacetophenone potassium hydroxide and ethyl alcohol are added. The green precipitate is dissolved in hydrochloric acid and the solution extracted with chloroform. The blue chloroform solution is repeatedly extracted with water until the colour changes to a clear red. It is then evaporated. The residue consists of azobilirubin which can be determined colorimetrically by comparison with a standard solution of azo- bilirubin. G. W. R. E. S. The Refractometric Determination of Haemoglobin. J. L. STODDARD and G. S. ADAIX ( J . Biol. Chenz. 1923 57 437454).- When determined on hzemoglobin (from human blood) which has been dried a t 110" the refractive constant a of hzemoglobin has a value of 0.001942. The difference between this value and that obtained by Howard (A.' 1920 i 451) k probably clue to water of crystallisation in the latter's specimen. Two refractometric methods for the estimation of hzemoglobin in blood have been elaborated. In the first the red corpuscles from a measured volume of blocd are washed hzmolysed with water and saponin treated with salt to precipitate the stromata made up t o volume centrifuged and the refractive index of the solution determined before and after coagul- ation of the hzmoglobin by heat. The hzemoglobin content of the blood can be calculated from these readings by means of the constant a. Equal volumes of salt solution (0.8%) and salt solution containing sufficient saponin to haemolyse the blood are added respectively to separate portions of equal volume of the blood. After centrifuging the clear solutions are read in the refractometer and the hzemoglobin content is calculated as above. A correction must be made for the refraction due to the saponin. Determinations by these methods combined with determinations of the oxygen capacity of the blood indicate that the latter is determined more accurately by T'an Slyke's than The second method is shorter. by Haldane's method. E. s.