ANALIYTICAL CHEMISTRY.THE contributions to this branch of chemistry have been of amuch more varied character during the past year than has beenthe case during the preceding five years. America has continuedto show the greatest activity in this respect, whilst the conditionsin this anld other countries more seriously affected by the war havenot yet become normal.The main material difficulties in the way of analytical workhave been the lack of an adequate supply of suitable glassapparatus and the continued shortage of platinum. Furtherdevices have, therefore, been proposed t o obviate the use of thatmeta1.l For example, i t has been shown that for certain quan-titative estimations, such as the analysis of chrome iron ore,platinum basins may be replaced by lead basins,2 whilst iron basinsare suitabla for sulphide fusions.3 An efficient substitute forplatinum wire for flame tests may be prepared from a strip ofrolled filter paper,* or from the “ lead” OP a black-lead pencil.6Physical Methods.A new form of viscosimeter of the capillary type has beendevised. The oil or other liquid is forced up into the tube by thecompression of a bulb, the temperature brought to a definite pointby means of a thermostatic jacket, and the time required for theliquid to fall between given points is noted.The absolute vicosityis then calculated by means of the formula: q = K d t , where Rrepresents a constant of the apparatus, d the specific gravity ofthe liquid, and t the time in seconds.6The errors associated with the falling sphere type of viscosi-meter have been taken into consideration in a new instrument,Compare Ann.Reports, 1918, 15, 118 ; 1919, 16, 127.C. Hutter, Zeitsch. angew. Chem., 1919, 32, 380; A., ii, 189.H. Sertz, ibid., 1920, 33, i, 156.* A. Ehringhaus, Centr. Min., 1919, 192; A., ii, 263.li C. C. Kiplinger, J . I n d . Eng. Chem., 1920, 12, 500; A., ii, 381.G. Baume and H. Vigneron, Ann. China. anal., 1919, [ii], 1, 379 ; A., ii, 92.13ANALYTICAL CHEMISTRY. 131in which the steel ball is delivered into the centre of the tube,and corrections for the wall and end effects are applied, the resultbeing calculated into absolute viscosity by means of a modifica-tion of Stokes' equation.7Fischer's TJiscosimeter, which is particularly suitable for veryviscid liquids, consists essentially of an inner tube surrounded bya water jacket, resembling a vertical Liebig's condenser, throughwhich water a t a definite temperature is run.The viscosity ismeasured by the time required by a metal ball to fall betweentwo points on the inner tube.8A viscosimeter map be used for determining the density of aliquid of which only a few C.C. are available? the calculation beingmade by a combination of the formulae of Ostwald-Poiseuille andof Scarpa .loThe determination of the surface tension afforzds a means ofidentifying minute quantities of organic liquids. A capillary tubecontaining a short column of the liquid is turned1 until the lowermeniscus of the latter changes to a plane surface, and the angulardeviation of the tube from the vertical position is then read on aquadrant scale.The surface tension, T, is calculated by means ofthe usual formula: T=d x h x T x 98012 dynes per sq. cm.llFor estimating the acidity of liquids, the use of surface-activeindicators, as originally suggested by Traube and Somogyi,lZ isoften more convenient than the use of colour indicators.13 Byusing salts of alkaline nature, such as eucupine dihydrochloride, itis possible to titrate highly dissociated acids in the presence of aprimary phosphate,l4 whilst the relative alkalinity of the carbonatesand hydrogen carbonates of potassium, sodium, calcium, andmagnesium may be determined in an analogous manner.15A study of the soap-bubble method of determining the ignitiontemperature of gaseous mixtures 16 has shown that, even aftercareful standardisation of the coils, size of bubble, and otherdetails, the results cannot be depended on, even for comparativepurposes.17W.H. Gibson and (Miss) L. M. Jaaobs, T., 1920, 117, 473.Chern. Zeit., 1920, a, 622.W. Block, Zeitsch. angew. Chern., 1920, 33, 198 ; A., ii, 590.lo Qazzetta, 1910, 40, ii, 261 ; A., 1911, ii, 17.l1 C. C. Kiplinger, J. Amer. Chem. SOC., 1920, 42, 472; A., ii, 291.la Int. Zeitsch. phys.-chem. Biol., 1914, 1, 479 ; A., 1916, ii, 101.W. Windisch and W. Dietrich, Biochem. Zeitsch., 1919, 100, 130; A.,l4 Ibid., 136 ; A., ii, 706. l6 Tbid., 101, 82 ; A., ii, 707.16 J. W. McDavid, T., 1917, 111, 1003.l 7 A. G. White and T.W. Price, ibid., 1919, 115, 1248.ii, 48.F 132 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.The melting-point curves obtained with binary and ternarymixtures of nitronaphthalenes may be used in the analysis of theindustrial products of nitration. The simultaneous presence ofmono- and tri-nitronaphthalene is indicated by points of arrest inthe cooling curves.IS Even when the melting point of the puresubstance is not known, the freezing-point curve affords an indica-tion of the amount of impurity present.19A new physico-chemical method of examining double salts hasbeen based on the measurement of the temperature produced onmixing the solutions with a constant liquid, such as phenol. Whena double salt is in solution, there is a variable difference betweenthe observed and calculated results.20Turning to optical methods, it has been shown that the ion con-tent of an aqueous solution of salts may be calculated from therefractive indices of the liquid before and after precipitation ofone of the component sa1ts.nSmall amounts of lead in copper may be estimated spectroscopic-aIly,22 the time required for the bright lead line (405-8pp) underconstant conditions being noted.Another application of spectro-photometry is in the analysis of colourless organic compounds,which f o r this purpose are converted into colourd compounds.23Reference may also be made to a new form of nephelometer24and to a photographic t'urbidimeter, in which one beam of lightis passed through the column of suspended substance and a secondbeam of equal intensity through standardised glass disks.25 I nthe nephelometric estimation of chlorides, the intensity of theopalescence of the silver chloride suspension is increased andrendered more constant by heating the liquid to 40° after theprecipitation .26 The nephelometric values of cholesterol and thehigher fatty acids have been shown to be affected by hydrolysisand by the presence of other substances, which, by themselves, donot produce turbidity.27The comparison of the fluorescence produced by ultra-violet rays18 P.Pascal, Bull. SOC. chim., 1920, [iv], 27, 388 ; A., ii, 514.lS W. P. White, J . Physical Chem., 1920, 24, 393 ; A., ii, 529.a. R. Dubrisay, Compt. rend., 1920, 170, 1582 ; A., ii, 508.11 M.de Crinis, Zeitsch. physiol. Chem., 1920, 110, 254 ; A., ii, 700.Za C. W. Hill and G. P. Luckey, Trans. Amer. Electrochem. SOC., 1917, 32,23 W. E. Mathewson, J . Amer. Chem. SOC., 1920, 42, 1277; A., ii, 566.t4 C. Chheveau and R. Audubert, Compt. rend., 1920, 170, 728; A.,ZJ W. G. Bowers and J. Moyer, J . Biol. Chem., 1920, 42, 191 ; A., ii, 444.26 A. B. Lamb, P. W. Carleton, and W. B. Meldrum, J . Amer. Chem. SOC.,a7 F. A. Csonka, J . Biol. Chem., 1920, 41, 243 ; A., ii, 272.335 ; A., ii, 193.ii, 327.1920, 42, 251 ; A., ii, 383ANALYTICAL CHEMISTRY. 133on substances in the cylinders of a nephelometer affords a newmeans of quantitative analysis.28A physical method of identifying and determining the purity ofacids has been based on the observation of their dissociationconstant, which may be determined, for example, by comparing thechange of colour of the same indicator in the solution and in astandard solution of known hydrogen-ion concentration .2Q Asimple method of determining the ion concentration of ultra-filtrates and other solutions free from proteins is to add an ionforming a sparingly soluble salt with the ion in question, and tonote the limit of solubility, as indicated by the formation of amilky turbidity.I n the case of calcium, the method gives resultsaccurate within 2 t o 3 milligrams per litre.30A method of analysis by fractional distillation under a constantreduced pressure is useful for the separation of substances havingboiling points which are close together.The fractionation is con-tinued until the final fractions show, not only identical boilingpoints, but also agree in their other physical properties, such asdensity, refractive index, and viscosity.31Gas Analysis.Various instruments for the automatic analysis of gases or forthe detection of an individual constituent have been described.The katharometer, which is intended for the estimation of smallquantities of hydrogen in air, is based on the change in the electricalresistance of a platinum helix through the increase in temperaturecaused by surface combustion of the hydrogen.32 An analogousprinciple has been utilised in the estimation of carbon monoxidein air.33The thermal conductivity method can only be applied quanti-tatively when the probable identity and amounts of constituentslikely to be present in a gaseous mixture are known, but undersuch conditions it gives good results in many cases.3428 L.J . Desha, J . Amer. Chem. SOC., 1920, 42, 1350 ; A., ii, 552.2 9 I. M. Kolthoff, Pharm. Weekblad, 1920, 57, 514 ; A., ii, 628.30 R. Brinkman and (Miss) E. van Dam, Proc. K. Akad. Wetensck.3l C. Moureu, C. Dufraisse, and P. Robin, Bull. SOC. chim., 1920, [iv], 27,33 H. A. Dayncs and G. A. Shakespear, Proc. Roy. SOC., 1920, 97, [A],33 A. B. Lamb and A. T. Larson, J. Amer. Chem. SOC., 1919, 41, 1908;34 E. R. Weaver and P. E. Palmer, J . Ind. Eng. Chem., 1920, 12, 894 ;Amsterdam, 1920, 22, 762 ; A., ii, 510.523 ; A., ii, 562.273 ; A., ii, 503.A., ii, 126.A., ii, 701134 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Another instrument which may be used for the estimation ofmall amounts of hydrogen or helium in air is the interferometer,which measures the difference in the refractivity of two gases.35A weighing burette for use in gas analysis has been devised, theweight of the gas and alterations in volume caused by the absorp-tion of the constituents being determined by weighing the quantityof mercury which is removed from, or introduced into, theburette.36 Improved modifications of Orsat’s apparatus have alsobeen described.373 88A source of error in the analysis of gases by fractional combus-tion with copper oxide is the dissociation of the latter, with theliberation of oxygen.This may be obviated by subsequently pass-ing the nitrogen repeatedly over the oxide a t a moderate redheat .39The use of an ammoniacal copper solution has several advantagesover alkaline pyrogallol as an absorbent for the estimation ofoxygen. If the usual ammonium carbonate solution be replacedby saturated ammonium chloride solution, the gas will be free fromcarbon dioxide derived from the reagent.40 On the other hand,when freshly prepared, i t may impart traces of ammonia to thegas, and cannot be used in the case of gases containing carbonmonoxide or acetylene.41For the estimation of the latter, good results may be obtained,under certain conditions, by the use of ammoniacal cuprouschloride solution as absorption reagent, the solution being thenacidified with acetic acid, and the cuprous acetylide separated and,estimated.42 Another suitable reagent for the absorption ofacetylene is a solution of mercuric cyanide in sodium hyaroxidesolution.This effects its separation from ethylene and benzenevapour. For the absorption of ethylene in the presence ofbenzene, a solution of mercuric nitrate in dilute nitric acidsaturated with sodium nitrate may be used.43 Ilosvay’s reagent d485 J. C. McLennan and R. T. Elworthy, Trans. Roy. SOC. Canada, 1919,13, [iii], 19 ; A., ii, 508.su E. R. Weaver and P. G. Ledig, J . Amer. Chem. Soc., 1920, 42, 1177 ;A., ii, 602.87 T. B. Smith, Bas World, 1919, 71, 379 ; A., ii, 263.s8 G. W. Jones and F. R. Neumeister, Chem. and Met. Eng., 1919, 21,89 E. Ott, J .Gasbeleucht., 1919, 62, 89 ; A., ii, 52.40 W. Haehnel and M. Mugdan, Zeitsch. angew. Chem., 1920, 33, 35 ; A.,41 W. L. Badger, J. I n d . Eng. Chem., 1920, 12, 161 ; A., ii, 264.42 J. A. Muller, Bull. SOC. chim., 1920, [iv], 27, 69 ; A., ii, 198.734 ; A., ii, 119.ii, 191.W. D. Treadwell and F. A. Trtuber, HeZu. Chim. Acta, 1919, 2, 601 ;A., ii, 61. u Ber., 1899, 32, 2697 ; A., 1900, ii, 52ANALYTICAL CHEMISTRY. 135removes acetylene quantitatively from coal gas or air, and goodresults may be obtained by igniting the precipitate with nitricacid and weighing the copper oxide. Hydrogen sulphide must notbe present, and in the case of air containing not more than 0.04 percent. of acetylene, about 5 per cent. of carbon dioxide must beadded to prevent oxidation.45 Methods depending on the reactionof acetylene with silver nitrate, and titrattion of the liberated nitricacid, are inaccurate, owing to the impossibility of controlling theconditions of the reaction.Good, results may be obtained, how-ever, by causing the copper acetylide, separated with Ilosvay’sreagent, to react with a sulphuric acid solution of ferric sulphate,and titrating the resulting ferrous sulphate.46A satisfactory reagent for the absorption of carbon monoxidemay be prepared by reducing cupric chloride in acid solution bymeans of stannous chloride. The presence of a slight excess ofthe latter prevents oxidation by the air, and the reagent may berenewed by expelling the absorbed gas a t 60-70O.47A method of detecting anad estimating traces of Bp’-dichlorodi-ethyl sulphide (mustard gas) in air has been based on its reducingaction on a solution of selenious acid in sulphuric acid,, the amountof the orange-red suspension of selenium being estimated bynephelometric comparison with standard suspensions.Theselenious reagent also reacts in the same way with arsines and:other toxic gases.48A g r i d tural A naly sis.There has been a considerable amount of investigation as to themost suitable methods of estimating the acidity and the limerequirement of soils. Extraction of an acid soil with potassiumnitrate solution, as in the method of Hopkins and Pettit, has beenshown to give the same results as extraction with equivalent solu-tions of potassium chloride, sodium nitrate, sodium chloride, orcalcium chloride, whilst variations in the temperature between2 5 O and 90° do not affect the acidity of the extract.There islittle, if any, exchange of acid radicles during the extraction.4945 H. Arnold, E. Mollney, andF. Zimmermann, Bey., 1920, 53, [B], 1034 ;IS R. WillstBtter and E. Maschmann, ibid., 939 ; A., ii, 514.47 F. C. Krauskopf and L. H. Purdy, J . Ind. Eng. Chem., 1920, 12, 168 ;46 M. Yablik, G. St. J. Perrott, and N. H. Furman, J . Amer. Chem. SOC.,49 H. G. Knight, J . I n d . Eng. Chem., 1920, 12, 340; A., i, 468.A . , ii, 613.A., ii, 267.1920, 42, 266 ; A., ii, 272136 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.For estimating the lime requirement, the vacuum method 50 givestrustworthy results agreeing with those obtained by the use of ahydrogen electrode after a specified period of shaking, whilst inthe case of limed soils the reduction in the acidity is best ascer-tained by the hydrogen electrode method or the Hopkins method.51The method of estimating the acidity or alkalinity of a soil interms of P H 5 2 has been adapted for use in field tests, and a tablehas been constructed showing the reactions of the soil correspond-ing with the colour changes of a series of indicators, each of whichoverlaps the next in the scale.53For the estimation of the total calcium in soils, a method hasbeen devised in which the precipitation with ammonia, and con-sequent occlusion of calcium in the precipitate, are eliminated.The precipitation is effected by adding solid ammonium oxalate tothe boiling solution, which has been rendered first just alkalinewith ammonia, and then just acid with hydrochloric acid, andmanganese is subsequently separated from the precipitatedoxalate.54 I n an investigation of ten different methods of estim-ating calcium in calcite, the most accurate results were obtainedby precipitation as calcium oxalate and weighing as calcium oxide,by precipitation as oxalate from a slightly acid solution, anmd bythe residual titration method.55The small amounts of manganese in plant ashes and similar pro-ducts can only be estimated colorimetrically.Of the differentmethods suggested, only that of Marshall56 has been found to givetrustworthy results.57 In that method, the manganese is oxidised,to permanganate by means of potassium persulphate.Theinfluence of various factors on the estimation of chlorides in soilhas been studied, and i t has been shown that Volhard’s methodgives more accurate and concordant results than Mohr’s method.58For estimating carbonates in limestones, marl, and soil, themethod of Van Slyke 59 is the most suitable for substances poor in5o J. W. Ames and C. J. Schollenberger, J . Ind. Eng. Chem., 1916, 8, 243;51 H. G. Knight, ibid., 1920, 12, 457, 559 ; A., i, 587 ; ii, 557.52 E. T. Wherry, J . Washington Acad. Sci., 1919, 9, 305; A., 1919,63 Ibid., 1920, 10, 217 ; A., ii, 400.0. M. Shedd, Soil Sci., 1920, 10, 1 ; A., ii, 636.66 G. E. &we, Chern. News, 1920, 121, 53; A., ii, 557.6Q Ibid., 1901, 83, 76; A., 1901, ii, 350.67 D.H. Wester, Rec. true. chirn., 1920, 39, 414 ; A., ii, 451.5a C. T. Hirst and J. E. Greaves, Soil Sci., 1920, 9, 41 ; A., ii, 384.59 J . Biol. Chem., 1918, 36, 351; A., 1919, ii, 78; compare Ann. Reports,A., 1916, i, 459.i, 428.1919, 16, 142ANALYTICAL CHEMISTRY. 137magnesium carbonate, whilst in other cases Van Slyke's gasometricmethod 60 gives more accurate results.61A method for the approximate estimation of phytin in plantextracts has been based on its precipitation in the presence ofinorganic phosphates by an acetic acid solution of copper acetateof definite concentration, but the amounts of phytin precipitatedvary with the nature of the plant.62Organic Analysis.Qualitative .-Formaldehyde gives colorations, which are notalways distinctive of the aldehyde, with certain aromatic com-pounds, such as pyrogallol, P-naphthol, and salicylic acid, andthese chromatic reactions may sometimes be used as tests ofidentity.63 Another reagent which gives distinctive colorationswith polyhydroxyphenols and other compounds of a phenoliccharacter is sodium p-toluenedisulphochloroamide in neutral oralkaline solution, but the tests must be made under definite con-ditions as to the proportions of reacting substances and thetemperature.64Most of the tests for methyl alcohol are based on its conversioninto formaldehyde, which is then identified either by chromaticreactions or by the formation of crystalline derivatives.Thesemethods have been critically examined and their relative trust-worthiness determined.Of the direct tests, all of which require aconsiderable proportion of methyl alcohol, that of Vivario,65 inwhich the methyl alcohol is converted into potassium cyanide, givesvery good results.66A test to distinguish between methyl and ethyl alcohols hasbeen based on the solubility of qrystallised copper sulphate in theformer,67 but is untrustworthy in the case of mixtures of the diluteal.cohols.68 Lieben's reaction for iodoform has been modified so asto afford a sensitive test for traces of ethyl alcohol.69A specific reaction for acetoacetic acid and its esters has been8 o J . Biol. Chem., 1017, 30, 347; A., 1917, ii, 422.C. S. Robinson, Soil Sci., 1920, 10, 41 ; A., ii, 635.6a A.Rippel, Biochem. Zeitsch., 1920, 103, 163; A., ii, 518.63 A. Rossi, Boll. Chim. farm., 1919, 58, 265 ; A., ii, 63.1 3 ~ A. Berthelot and M. Michel, Bull. Sci. Phurmacol., 1919, 26, 401 ; A . ,65 J . Pharm. Chim., 1914, [vii], 10, 145; A., 1914, ii, 780.66 A. 0. Gettler, J. Biol. Chem., 1920, 42, 311; A., ii, 562.O 7 Pannwitz, Pharm. 2entr.-h., 1919, 60, 441 ; A., ii, 62.6 8 T. Sabalitschka, ibid., 1920, 61, 78 ; A., ii, 271.69 R. Kunz, Zeitsch. anal. Chem., 1920, 59, 302 ; A., ii, 711.ii, 336.I?138 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.based on its conldensation in hydrochloric acid solution t o formP-methylumbellif erone, which in alkaline solution shows an intenseblue fluorescence.70From a study of various reactions for the identification of tracesof hydrocyanic acid, the conclusion has been drawn that the ferro-cyanide test is the most characteristic.71 The Prussian-blue testgives a reaction a t a dilution of about 1 : 17 x 104 with hydrocyanicacid, whilst the Schonbein test, which is not distinctive, gives areaction a t 1 : 43 x 106 in the light and a t 1 : 55 x 106 in the dark.72A test paper impregnated with a solution of o-tolidine, copperacetate, and dilute acetic acid is capable of detecting one part ofhydrocyanic acid in two million parts of air.73In the colour reaction with manganese salts for ~ x a l a t e s , ~ ~sufficient oxidation is produced, by agitating the hot solution witha little alkali hydroxide.75 Oxalic acid! may be distinguished fromtartaric and lactic acids by giving a violet coloration when heatedwith sulphurk acid and resorcinol.76 If the test be applied as azone reaction, a blue ring is formed a t the junction of the liquids.77A biochemical method of detecting dextrose in the presence ofother sugars, such as mannose, lzevulose, or arabinose, has beenbased on its conversion, in 70 per cent.methyl alcohol solution,by emulsin into P-methylglucoside, which may be identified incrystalline form.78 Of various tests recommended for the detec-tion of dextrose in urine, Fehling's test is the most sensitive, beingcapable of detecting 0.00125 per cent .79 Traces of acetylmethyl-carbinol formed in the butylene-glycollic fermentation of sugarsmay be identified by oxidising the carbinol to diacetyl, which isthen distilled, and identified by precipitation as nickel dimethyl-glyoxime.80A very sensitive test for benzoic acid or for substances such ascocaine, which contain a benzoyl group, or which, like atropine,yield benzoic acid on oxidation, has been based1 on the diazotisa-tion of the aminobenzoic acids produced on nitration and reduc-7 0 V.Arreguine and E. D. Garcia, Ann. Chim. anal., 1920, [ii], 2, 36 ; A.,ii, 273.72 J. B. Ekeley and I. C. Macy, Proc. Colorado Sci. SOC., 1919, 11, 269;A., ii, 202.7l L. Chelle, ibid.. 21 ; A., ii, 202.J. Moir, J. S. African Assoc. Anal. Chem., 1920, 3, 16 ; A., ii, 715.74 V. Mwri, Boll. chim. farm., 1917, 56, 377 ; A., 1917, ii, 511.76 H.Cmon and D. Raquet, Ann. Chim. alzaZ., 1919, [ii], 1, 205; A., 1919,7 6 K. Brauer, Chem. Zeit., 1920, 441, 494; A., ii, 517.7 7 L. H. Chernoff, J. Amer. Chem. SOC., 1920, 42, 1784; A., ii, 712.78 E. Bourquelot and M. Bridel, Compt. rend., 1920, 170, 631 ; A., ii, 337.7 9 G. E. *we, Amer. J. Pharm., 1919, 91, 717 ; A., ii, 132.ii, 438.M. Lemoigne, Compt. rend., 1920, 170, 131 ; A., ii, 198ANALYTICAL CHEMISTRY. 139tion, and the formation of an orange-red precipitate on treatingthe diazo-compounds with an ammoniacal solution of &naphthol .S1The behaviour of guaiacol with an oxydase affords a means ofdistinguishing it from creosote, the former giving a yellow color-ation immediately, whilst the latter is colourless a t first and thenshows a slight violet tint.s2A new colour reaction for quinine, which takes place on theaddition of pyridine in the presence of chlorine water, distinguishesthe alkaloid from quinidine and euquinidine.The colour changesfrom yellow to rose, and finally t o purplish-red.83A sensitive and characteristic test for strychnine consists intreatment of the alkaloid salt solution with hydrochloric acid andzinc amalgam, and finally with potassium ferricyanide. I n thepresence of quantities down t o 0.001 milligram of strychnine, apink to rose-red coloration is obtained .83 Various colour reactionsof emetine have also been described.84Pyrrole reacts with p-dimethylaminobenzaldehyde in ananalogous manner to indole, and this must, be borne in mind whenapplying Ehrlich's test .85Quantitative.-A new form of absorption apparatus forelementary analysis has been 'devised, in which the carbon dioxideis absorbed by potassium hydroxide solution, and special precau-tions are used in drying the gas.86 Certain organic substances,such as methyl esters and chloro-compounds, may be quantitativelyoxidised by means of a mixture of sulphuric and chromic acids, thechlorine in the latter compounds being retained by an amalgamatedcopper spiral preceding the burette in which the carbon dioxide ismeasured .87 Another new method of estimating carbon andhydrogen in organic compounds is based on their combustion incontact with platinum and cerium dioxide.88Fusion with an alkaline mixture containing alkali nitrates, pre-cipitation of the carbonate as calcium carbonate from a solution ofthe fused mass, and titration of the washed precipitate with hydro-chloric acid, has been recommended for the estimation of carbon.89A simple and rapid method of estimating halogens in organiccompounds is to volatilise the substance with air through a quartzM.Guerbet, Compt. rend., 1920, 171, 40 ; A., ii, 517.D. Ganassini, Arch. ital. biol., 1919, 69, 7 3 : A., ii, 339.813 H. E. Buc, J . Assoc. Off. Agric. Chem., 1919, 3, 193; A., ii, 397.84 A. Lahille, Arch. m a . exp., 27, 336 ; A., ii, 134.E. Salkowski, Riochem. Zeitsch., 1920, 103, 185 ; A., ii, 566.e6 F. Friedrichs, Zeitsch. angew. Chem., 1919, 32, 388 ; A., ii, 192.87 J. Guyot and L. J. Simon, Cornpt. rend., 1920, 170, 734 ; A., ii, 332.8Q L.Lescceur, J . Pharm. Chim., 1920, [vii], 21, 267 ; A., ii, 332.I<. Sumikura, J . Tokyo Chem. Soc., 1919, 40, 593 ; A., ii, 126.F* 140 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.tube heated a t 900-1000°, to absorb the products of combustionin sodium hydroxide solution containing sodium sulphite, and,after oxidising the excess of the latter with potassium perman-ganate, t o estimate the halogen by Volhard's pr0cess.mIn using Kjeldahl's process for the estimation of nitrogen inaromatic nitro-compounds, it should be noted that accurate remltsmay be obtained with ortho-compounds, but that those given bymeta- and para-derivatives will be much too low.91Estimation of nitro-groups by reduction with stannous chlorideand titration of the excess of the reagent with iodine gives too lowresults with mononitrotoluenes, owing t o the formation of p-chloro-toluidines, whilst the use of titanous chloride gives resuTts about3 per cent.too low in the case of o-nitrotoluene.92 The low resultsgiven by a-nitronaphthalene and similar compounds may beobviated by substituting titanous sulphate for titanous chloride,and thus preventing chlorination.93 Diazo-compounds, such asbenzenediazonium chloride, may be titrated in hydrochloric acidsolution by means of titanous chloride, with suitable indicators.94A method for the estimation of methyl alcohol has been basedon its oxidation with sulphuric acid and potassium dichromate, andgravimetric estimation of the carbon dioxide produced or volu-metric estimation of the excess of dichromate.95 I n anothermethod, the methyl alcohol is oxidised by means of ammoniumpersulphate, and the resulting formaldehyde estimated colori-met r ically.96Por the estimation of ethylene glycol, oxidation with potassiumdichromate and! sulphuric acid, as in glycerol analysis, gives trust-worthy results, but the acetin method is not applicable to dilutesolutions of the gly00l.~~An apparatus for estimating the carbon dioxisde formed in thefermentation of sugars has been devised, and the method has beenshown to be applicable to the differentiation of yeasts and enzymesby fermentation with appropriate sugars.98W. A.van Winkle and G. McP. Smith, J. Arner. Chem. SOC., 1920, 42,333 ; A., ii, 328.91 B.M. Margosches and E. Vogel, Ber., 1919, 52, [B], 1992 ; A., ii, 50.ga D. Florentin and H. Vandenberghe, BUZZ. SOC. chim., 1920, [iv], 27, 158 ;93 T. Callan, J. A. R. Henderson, and N. Strafford, J. Soc. C'hem. Ind., 1920,94 E. Knecht and L. Thompson, J. sbc. Dyers and Col., 1920, 36, 215;95 A. Heiduschka and L. Wolff, Pharm. Zen&.-h., 1920, 61, 361 ; A . , ii, 515.96 S. B. Schryver and C. C. Wood, Analyst, 1920, 45, 164; A., ii, 393.97 €3. Miiller, Ghem. Zeit., 1920, 44, 513 ; A., ii, 515.98 A. Slator, J. SOC. Chem. I d . , 1920, 39, 1 4 9 ~ ; A., ii, 448.A., ii, 271.39, 861. ; A., ii, 331.A., ii, 647ANALYTICAL CHEMISTRY. 141Sucrose may be accurately estimated by a cryoscopic method.99For the estimation of sugars by inversion, chemical catalysts, suchas benzenesulphonic acid, have advantages over enzymes.1 Theoptical rotation of laevulose is destroyed by heating the sugar forseven hours with dilute hydrochloric acid, whereas that of dextroseis not affected.This has been made the basis of a method ofestimating these sugars.2 I n the case of fruit juices, however, themost trustworthy method is to calculate the proportions of therespective sugars from the cupric-reducing power and the iodinevalue, the latter being characteristic for each sugar.3 Anothermethod of estimating dextrose has been based on the action ofpotassium cyanide for two days a t 20°, and estimation of theexcess of the reagent, or measurement in the change in rotation ofthe d e ~ t r o s e .~A modification of the phenylhydrazine method for estimatingpentosans consists in distilling the substance with sulphuric acidand estimating the furfuraldehyde by precipitation as hydrazone,and determination of the excess of phenylhydrazine in the filtrate.5The coloration given by phenol when heated with Millon'sreagent and nitric acid affords a means for its colorimetric estim-ation in the presence of other phenols.6 Another method ofestimating phenol is based on its titration with a solution of adiazonium compound, hydroxyazo-compounds being formed in thereaction.7The conditions under which phenolphthalein combines quanti-tatively with iodine to form tetraiodophenolphthalein have beeninvestigated, and a gravimetric method based on them.8Silico- and phospho-tungstic acids are suitable reagents for thegravimetric and volumetric estimation of alkaloids under specifiedconditions, phosphotungstic acid being preferable in the case ofaconitine and n i ~ o t i n e .~ For the quantitative separation ofstrychnine from quinine, advantage has been taken of the fact thatthe former is only very slightly soluble in ethyl ether, and is leftH. H. Dixon and T. G. Mason, Sci. Proc. Roy. Dubl. SOC., 1920, 16, 1 ;A., ii, 395.E. Hildt, Ann. Chirn. anal., 1920, [ii], 2, 103 : A., ii, 395.F. Lucius, Zeitsch. Nahr. Benusmn., 1919, 38, 177 ; A., ii, 132.8 (Miss) H. M. Judd, Biochem. J . , 1920, 14, 255 ; A., ii, 395.J. Bougault and J. Perrier, C m p t . rend., 1920. 170, 1395 ; A., ii, 452.P.Menad and C. T. Dowell, J . I n d . Emg. Chem., 1919, 11, 1024; A.,ii, 200.6 R. M. Chapin, ibid., 1920, 12, 771 ; A., ii, 645.8 S. Palkin, ibid., 1920, 12, 766: A., ii, 643.R. M. Chapin, ibid., 568 ; A., ii, 563.A. Heiduschka and L. Wolff, Schweiz-Apoth. Zeit., 1930, 58, 213, 229 ;A., ii, 780142 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.in the aqueous solution on extraction with ether in the presenceof ammonia.1° Reference may also be made to a method of estim-ating codeine by treatment of the plant extract with calcium hydr-oxide, and extraction of the filtrate with toluene.11Tyrosine cannot be estimated quantitatively in the products ofthe hydrolytes of proteins by the method of Folin and Denis,12 sincethe blue coloration is also given by tryptophan and othercompounds in the mixture.13It has been shown that the growth of yeast cells under specifiedconditions depends on the amount of so-called vitamins present,and a gravimetric method of estimating vitamins has been basedon this fact.14 The method has been found applicable to theestimation of the vitamin in f0od.15 On the other hand, the resultsof other experiments have indicated that the test is influenced byso many factors as t o have little, if any, value.leInorganic Analysis.Qualitative .-Sodium alizarinsulphonate has been proposed asa sensitive indicator for the titration of mineral acids, and has thefurther advantage that i t can be used both with ammonia andphosphoric acid.17Several new schemes for the separation of cations into groupshave been published.These include a modification of Petersen'smethod18 of separating the cations of the third and fourthgroups,1g and a new method for the separation of various metalsin the presence of phosphoric acid.20 The use of thioacetic acidhas been shown to offer several advantages over hydrogen sulphideas a reagent for the precipitation of the cations of Group II.21Sensitive tests for manganic, ceric, cobaltic, and thallic com-pounds have been based on the fact that they give intense bluecolorations with benzidine and other diphenyl derivatives.22lo A. R. Bliss, J. Amer. Pharm. ASSOC., 1919, 8, 804; A., ii, 276.11 H. E. Annett and H. Sen, Analyst, 1920, 125, 321 ; A., ii, 644.12 J. BioZ.Chem., 1912, 12, 245; A., 1912, ii, 1012.13 R. A. Gortner andG. E. Holm, J . Amer. Chem. Soc., 1920, 42, 1678;l4 R. J. Williams, J . Bid. Chem., 1920, 42, 259 ; A., ii, 648.l6 W. H. Eddy and H. C. Stevenson, ibid., 1920, 43, 295 ; A., ii, 716.l6 G. de P. Souza and E. V. McCollum, ibid., 1920, 44, 113; A., i, 919.W. Mestrezat, J . Pharm. Chim., 1920, [vii], 21, 185 ; A., ii, 263.Zeitsch. anorg. Chem., 1910, 67, 253 ; A., 1920, ii, 654.P. de Pauw, Chem. Wee,kblad, 1920, 17, 191 ; A., ii, 451.N. Alvarez, BoE. minero POC. nac. min. Chile, 1919, 31, 181 ; A., ii, 381.A., ii, 643.2Q H. Remy, Zeitsch. anal. Chem., 1919, 58, 385 ; A., ii, 186.22 F. Feigl, Chem. Zeit., 1920, a, 689; A., ii, 710ANALYTICAL CHEMISTRY. 143Turning to the specific tests for individual substances, it hasbeen shown that iodic acid may be used as a distinctive micro-scopical reagent for the detection of ammonia, characteristiccrystals of ammonium iodate being formed.23The differences between the deposits formed by hydrogenarsenide, cacodylic acid, methylarsinic acid, and neosalvarsan inMarsh’s test have been studied, and i t has been found that theaddition of platinum chloride t o promote the evolution of hydrogenmay fix some of the arsenic in the flask.2* The best method ofdetecting arsenic in sulphur is by oxidation with bromine andnitric acid, and application of Gutzeit’s test to the product.25 Atest for salts of tin has been based on the insolubility of stannousor stannic iodide in sulphuric acid. The yellow precipitate maybe distinguished from the similar compound formed by arsenic bythe fact that it is soluble in dilute hydrochloric acid.The corre-sponding antimony salt is brick-red and flocculent.26 A distinctivetest for osmium tetroxide is afforded by the blue coloration whichit gives with pyrogallol, whilst the vapour may be identified bygiving a permanent stain, due to the reduced osmium, with aslightly greasy finger-print .27Xanthic acid may be used as a distinctive reagent for molyb-denum,2* but i t is essential that oxalates should not be present,whilst copper, cobalt, nickel, iron, and uranium interfere with thetest .29For the detection of traces of cobalt, the coloration given by0-nitroso-a-naphthol is much more sensitive than that obtainedwith a-nitroso-P-naphthol.30 Cobalt may be detected in thepresence of nickel by means of potassium xanthate.The xanthatesof both metals are precipitates, but the cobalt compound is in-soluble in ammonia, whereas the nickel compound dissolves,forming a blue solution.31Most cerous and ceric compounds give a bright blue colorationwhen moistened with a solution of benzidine in acetic acid, andthe reaction may be used for detecting cerium in the presence ofother metals of the ammonium sulphide group of rare earths, withthe exception of thalli~rn.3~t3 G. Denigbs, Compf. rend., 1920,171, 177 ; A., ii, 555.24 D. Ganassini, Boll. Chim. farm., 1919, 58, 395 ; A., ii, 51.25 H. S. Davis and M. D. Davis, J . Ind. Eng. Chem., 1920, 12, 479 ; A.,27 C.A. Mitchell, Analyst, 1920, 45, 125 ; A., ii, 335.28 S. L. Malowan, Zeitsch. anorg. Chem., 1919, 108, 73 ; A., ii, 59.29 J. Koppel, Clzem. Zed., 1919, 43, 777 ; A., ii. 58.30 I. Bellucci, Qazzetta, 1919, 49, ii, 294 ; A., ii, 194.31 L. Compin, Ann. Chim. anal., 1920, [ii], 2, 218 ; A., ii, 559.32 F. Feigl, Oesterr. Chern. Zeit., [ii], 22, 124 ; A., ii, 54.ii, 448. 26 A. Mazuir, Ann. Chim. anal., 1919, [ii], 2, 9 ; A., ii, 197144 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Iodic acid has been found to be a useful microchemical reagentfor calcium, strontium, or barium, the distinctive forms of thecrystalline iodates being observed.33 It may also be used to dis-tinguish between the salts of barium and radium.34For the detection of magnesium, advantage has been taken ofthe fact that it gives a bluish-violet coloration with an alcoholicsolution of alkannin, the similar colorations given by strontiumand manganese being distinguished by their behaviour afteracidification .35The applicability and sensitiveness of various reactions forpotassium have been determined, and the most sensitive reagenthas been found to be sodium-bismuth thiosulphate, which iscapable of detecting 1 part in 57,000.36The colour changes which take place in the diphenylamine reac-tion for nitrates have been studied, and a sensitive modificationof the reagent devised for microchemical tests.37 A blue colorationmay be produced in the diphenylamine reagent by the presence offerric salts in the sulphuric acid. To prevent this, such acidshould be boiled, and cooled to reduce the ferric salts.38A sensitive test for nitrites in the presence of nitrates is theproduction of a red coloration on treating the solution successivelywith acetic acid, potassium oxalate solution, manganous sulphate,and hydrogen peroxide.39 For the detection of phosphates in thepresence of barium, a modification of Denigk’ strychnine-molybdate reagent is sensitive, but must be freshly ~repared.~OQuantitative.-It is in this branch of the subject that thegreatest activity has been shown, and the contributions to differentjournals have been so numerous that it has been necessary to selectonly the more important for mention in this Report.A method of using potassium chlorate as an original standardfor the titration of alkali has been described.The chlorate isreduced by means of sulphur dioxide, the excess of the latterremoved, and the sulphuric acid formed in the reduction titratedwith alkali-KC10, + 350, + 3H,O = 3H,S04 + KCl.41 Anothersuitable standard is potassium hydrogen phthalate, which containsG. DenigBs, Compt. rend., 1920, 170, 996 ; A., ii, 388.34 G. DenigBs, ibid., 1920, 171, 633 ; A., ii, 706.s5 F. Eisenlohr, Ber., 1920, 53, [B], 1476 ; A., ii, 708.s6 0. Lutz, Zeitsch. anal. Chem., 1920, 59, 145 ; A., ii, 509.37 E. M. Harvey, J . Amer. Chem. Soc., 1920, 42, 1245 ; A., ii, 504.3i3 F. Ham, Zeitsch. Nahr. Genusmn., 1920, 39, 355 ; A., ii, 555.39 P. H. Hermans, Pharm. Weekblad, 1920, 57, 4 6 2 ; A., ii, 448.40 L.DBbourdeaux, Bull. Sci. pharmacol., 1920, 27, 70 ; A , , ii, 505.41 H. B. van Valkenburgh, J . Amer. Chem. SOC., 1920, 42, 757; A.,ii. 387ANALYTICAL CHEMISTRY, 145no water of crystallisation and is not hygro~copic.4~ I n preparingthe salt, i t is advisable to crystallise i t above 20°, to prevent theformation of a more acid salt.43 A useful indicator for colouredliquids has been found in ferrous sulphide, the formation of whichis prevented by a slight trace of acid. A crystal of pure ferrousammonium sulphate is added to the solution, which is then treatedwith hydrogen sulphide, and titrated with alkali until a permanentblack coloration is obtained.44Various experiments have been made with salts of magnesium,zinc, mercury, and aluminium to determine the degree of accuracyobtainable in the acidimetric titration of the salts of heavy metalswhich form insoluble hydroxides.45I n the volumetric estimation of sulphates by oxidation ofbenzidine sulphate with potassium ~ermanganate,~~ it is essentialthat the solution from which the benzidine sulphate is precipitatedshould be free from organic matter, iron, heavy metals, nitrates,and phosphates.47A new volumetric reduction method with arsenic trioxide hasbeen devised, in which an oxidising agent, such as a chlorate orchromate, is treated with excess of arsenic trioxide in the presenceof hydrochloric acid, and the solution subsequently titrated withpotassium bromate solution.4*The con-ditions for the iodometric estimation of aci'ds, more especially weakacids, have been investigated, and the applicability of variousmethods has been determined .49 The direct estimation of chloricacid in a strongly acid medium gives too high results, owing tooxidation of hydrogen iodide, but Rupp's method50 is trust-worthy.51 For the iodometric estimation of arsenic acid, the react-ing mixture should be heated to looo and contain specified propor-tions of potassium iodide and hydrogen chl0ride.5~Stannous tin may be more accurately estimated by a volumetricmethod in an acid solution. Titration with standard iodine solu-Several new iodometric methods have been published.42 W.8. Hendrixson, J . Amer. Chem. SOC., 1920, 42, 724; A., ii, 382.43 F. D. Doclee, ibid., 1655 ; A., ii, 628.44 J.Houben, Ber., 1919, [B], 1613; A., ii, 53.45 I. M. Kolthoff, Zeitsch. anorg. Chem., 1920, 112, 172; A . , ii, 709.46 G. W. Raiziss and H. Dubin, J . Biol. Chem., 1914, 18, 297 ; A., 1914,47 P. L. Hibbard, So31 Sci., 1919, 8, 61 ; A., ii, 191.48 F. de Bacho, Annali Chim. Appl., 1919, 12, 153 ; A., ii, 188.49 I. M. Kolthoff, Pharm. Weekblad, 1920, 57, 53 ; &4., ii, 121.50 E. Rupp, Zeitsch. anal. Chem., 1917, 56, 580 ; A., 1918, ii, 125.51 I. M. Kolthoff, Pharm. Weekblad, 1919, 56, 460; A., ii, 190.62 P. Fleury, J . Pharm. Chim., 1920, [vii], 21, 385; A., ii, 448.ii, 671146 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.tion in the presence of hydrochloric acid is satisfactory in the caseof many compounds.53 The iodometric estimation of chromiumgives too high results, owing to the influence of atmospheric oxygenon the reaction.54To obtain accurate results in the iodometric estimation ofpotassium f erricyanide, the acid concentration must be keptsufficiently high and the time of the reaction restricted.55A new volumetric method of estimating nickel has been basedon the titration of the acid which is liberated in the reaction withdimethylglyoxime .56Further work on the use of organic solvents f o r the quantitativeseparation of metals has shown that the method of progressiveprecipitation previously described 57 is suitable for the separationof magnesium from sodium and potassium, the salts of the lattermetals being precipitated first by means of alcohol and ether.58New applications of the use of cupferron (the ammonium saltof nitrosophenylhydroxylamine) have been published.The reagentgives good results in the estimation of iron, copper, titanium,zirconium, thorium, and vanadium in the absence of certaininterfering substances, but in many cases offers no advantageover previous methods.59 It is most serviceable for theseparation of copper from arsenic, and especially from antimony.Like ccnitroso-&naphthol, it is a useful reagent for separatingelements into groups.60 I n neutral solution it gives precipitateswith all metals except the alkali metals, and the precipitates maybe separated into two groups by their behaviour towards chloro-form and1 dilute acilds.61 I f corrections for the solubility of theprecipitate be applied, iron may be accurately separated frommanganese by cupf erron.62 A trustworthy method has also beenworked out for the use of the reagent in the separation of tin andantimony.63A simple method of estimating mercury is to cause it to be6a J.G. F. Druce, Chem. News, 1920, 121, 173; A., ii, 710.5* 0. Meindl, Zeitsch. anal. Chem., 1919, 58, 629 ; A., ii, 390.56 I. M. Kolthoff, Pharm. Weekblad, 1919, 56, 1618; A., ii, 67.6t3 J. Holluta, Monatsh., 1919, 4-43? 281 ; A., ii, 57.57 S. Palkin, J. Amer. Chem. SOC., 1916, 38, 2326; A., 1917, ii, 43.&* Ibd., 1920, 42, 1618 ; A., ii, 637.G. E. F. Lundell and H. B. Knowles, J. Ind. Eng. Chem., 1920, 12,344 ; A., ii, 390.6o I. Bellucci and A. Chiucini, Gaxzetta, 1919, 49, ii, 187 ; A., ii, 54.V.Auger, Compt. rend., 1920, 170, 995 ; A., ii, 391.E. H. Archibald and R. V. Fulton, Trans. Roy. SOC. Canada, 1919, 13,[iii], 243 ; A., ii, 512.6* A. Kling and A. Lassieur, Compt. rend., 1920, 170, 1112 ; A., ii, 452ANALYTICAL CHEMISTRY. 147deposited on a coil of copper gauze, and to estimate the amountof the deposit by the loss in weight after heating the coil in acurrent of hydrogen .64Arsenic may be separated from antimony, tin, copper, lead,mercury, and iron by a modification of the distillation method,whilst antimony may be separated from tin by volatilisingantimony chloride a t 155-165O from solutions to which phosphoricacid has been added to render the tin non-volatile65 A modifi-cation of the Marsh-Berzelius test has been described in which thehydrogen arsenide is conducted over red-hot copper, which retainsthe arsenic as arsenides.The results are about 1.4% too low,possibly owing to retention of arsenic in the flask in the form of astable complex.66I n order to obtain a precipitate of secondary zirconium phos-phate of constant composition in the precipitation of zirconiumby the phosphate method, the solution should contain from 2 to 20per cent. of sulphuric acid, and an excess of diammonium hydrogenphosphate ten to one hundred times in excess of the theoreticalquantity should' be added.67 A new method, appliaable to zirconiaores, is to precipitate the zirconium with selenious acid, and toignite the basic selenite, which leaves a residue of zirconia.68The difficulty of precipitating molybdenum quantitatively assulphide may be obviated by having a sufficient quantity of formicacid in the solution, and seeing that the whole of the molybdenumis present as molybdate.69From a study of various methods of estimating uranium, theconclusion has been drawn that precipitation with ammoniumsulphide or with ammonia gives the most trustworthy results.I neither case, the precipitate leaves uranoso-uranic oxide on igni-tion .70 Precipitation of uranium as uranyl ammonium phosphateis also a good method, but has the drawback that the ignited uranylpyrophosphate rapidly absorbs moisture .71 For the estimation ofminute quantities of uranium, a colorimetric method has beenbased on the red coloration given by uranyl salts with sodiumsalicylate .7264 H.B. Gordon, Analyst, 1920, 45, 41 ; A., ii, 194.66 W. Strecker and A. Riedemann, BPr., 1919, 52, [B], 1935 ; A., ii, 51.G6 B. S. Evans, Analyst, 1920, 45, 8 ; A., ii, 125.G7 G. E. F. Lundell and H. B. Knowles, J. Amer. Chem. SOC., 1919, 41,6s M. S. Smith and C. James, ibid., 1920, 42, 1764; A., ii, 710.69 J. StGrba-Bohm and J. Vostfebal, Zeitsch. anorg. Chem., 1920, 110, 82 ;'O R. Schwarz, HeZv. Chim. Acta, 1920, 3, 330 ; A., ii, 391.71 C. A. Pierre, J. Ind. Eng. Chem., 1920, 12, 60 ; A., ii, 197.73 Muller, Chem. Zeit,, 1919, 43, 739 ; A., ii, 60.1801 ; A., ii, 60.A., ii, 335148 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.The perchlorate method for the estimation of potassium givesaccurate results under specified conditions.73 The methodpreviously described 74 is rendered more trustworthy by extractingthe sodium perchlorate with alcohol containing perchloric acid,instead of with a saturated solution of potassium perchlorate.75By using aniline perchlorate in place of perchloric acid, the processis rendered more rapid than any other method of estimatingpotassium.76 The proportion of potassium in the mixed precipitateof potassium and sodium cobaltinitrite may be calculated from agravimetric or electrolytic estimation of the cobalt.77 Anothermethod of estimating the proportions of the two metals in mixturesof the salts is to convert them into nitrates and determine themelting points.78Chromic acid liberates bromine quantitatively from bromides a tthe ordinary temperature, and the bromine may be removed byaspiration, whereas chlorides under the same conditions yield onlytraces of chlorine.A method of estimating bromine in mineralwaters has been based on these f a ~ t s . 7 ~The conditions for effecting the quantitative estimation of phos-phoric acid by precipitating and weighing it as ammoniumphosphomolybdate have been investigated, and a trustworthymethod devised. For the nephelometric estimation of traces ofphosphoric acid, the use of a reagent prepared by the interactionof strychnine sulphate and sodium molybdate gives excellentresults .SOElect roc h e mica1 A nal ysis.There have been relatively few contributions to gravimetricelectrolytic methods of analysis, but much work has been done inconnexion with the investigation of methods of electrometrictitration and their extension t o further substanoes.It has been shown that in many cases the end-point of a titra-tion is sharply indicated by measurement of the terminal voltagebetween two electrodes, one of which is immersed in the solution73 R.L. Morris, Analyst, 1920, 45, 349 ; A., ii, 707.74 G. P. Baxter and M. Kobayashi, J. Amer. Chem. SOC., 1917, 39, 249 ;75 G. P. Baxter and M. Kobayashi, ibid., 1920, 42, 735 ; A., ii, 385.76 S. B. Kuzirian, Proc. Iowa. Acad. h+ci., 1917, 24, 547 ; A., ii, 450.77 P. Wenger and C. HBmen, Ann. Chim. anal., 1920, [ii], 2, 198; A.,A., 1917, ii, 270.ii, 556.A.Quartaroli, Gazzetta, 1920, 50, ii, 64 ; A., ii, 635.7s W. F. Baughman and W. W. Skinner, J. I n d . Eng. Chem., 1919, 11,954 ; A., ii, 48.H. Kleinmrtnn, Biochem. Zeitsch., 1919, 99, 150 ; A., ii, 634ANALYTICAL CHEMISTRY. 149under examination, and is capable of yielding the same ions to thesolution, whilst the second, or comparison electrode, is composedof the same metal as the first and of the precipitate or otherproduct of the reaction. Owing to the slow action of the electrode,the method does not give sharp results in the titration of hydrogen-ion concentration.8' A hydrogen eleotrode giving a sharp end-point in acidimetric titrations has been described, and shown tobe suitable for the estimation of strong acids in the presence ofweak acids.8" Other electrometric methods may be used for theestimation of weak acids, such as acetic acid, in the presence ofstrong aoids, such as hydrochloric acid.83 A simple metho'd ofacidimetric or alkalimetric titration is to connect the solution tobe titrated with another of known I', value, each being providedwith a hydrogen electrode, and to continue the titration until thesame hydrogen-ion concentration is shown by both solutions.84Carbonic acid and its salts in dilute solution can be titrated in thisway,85 whilst phosphoric acid in dilute solution behaves like amono- or di-basic aoid, the first end-point being sharp, whilst thesecond is less distinct.86A method for the electrometric estimation of arsenic in colouredsolution is t o titrate arsenic trioxide with iodine solution in thepresence of sodium hydrogen carbonate, and to titrate arsenicpentoxide in sulphuric acid solution a t 9 5 O with sodium iodidesolution.87 Lead and zinc salts may be accurately titrated withpotassium ferrcoyanide, but in the case of other metals the pre-cipitates are not constant in composition .88The fact that mercuric acetate forms a stable, complex coni-pound with ally1 alcohol has been utilised for the electrometrictitration of that alcohol by means of a standard solution of themercury salt.89The presence of soluble ferricyanides or ohlorides in moderateproportion does not interfere with the titration of ferrocyanideswith potassium pernianganate solution, but salts yielding pre-W. D.Treadmelland L.Weiss, HeEu. Chim. Acta, 1919, 2, 680; A.,W. D. Treadwell and L. Weiss, ibid., 1920, 3, 433 ; A . , ii, 553.8a I. M. Kolt.hoff, Rec. trav. chim., 1920, 39, 280 ; A., ii, 327.84 P. E. Klopsteg, Science, 1920, 52, 1 s ; A., ii, 700.85 I. M. Kolthoff, Zeitsch. anorg. Chem., 1920, 112, 155 ; A., ii, 705.86 I. M. Kolthoff, ibid., 165 ; A., ii, 705. *' C. S. Robinson and 0. B. Winter, J . I n d . Eng. Chem., 1920, 12, 7 7 5 ;ii, 119.A., ii, 635.Erich Muller, Zeitsch. angew. Chem., 1919, 32, 351 ; A;, ii, 54.E. Biilmann, Medd. K. Vetenskapsakad. Nobel-Inst., 1919, 5, 1 ; A . ,i, 131150 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.cipitates with either ferro- or ferri-oyanides must not bepresent.90The alcoholic strength of beer or wine may be rapidly estimatedby distillation with magnesium oxide (after removal of carbondioxide), treatment of the distillate with oxalic acid solution, anddetermination of the specific conductivity of the mixture and ofthe original oxalic acid solution.The required result is thenobtained by the use of a formula.91Amino-acids may be estimated alkalimetrically by means of thehydrogen electrode, and a simple method for the purpose has beende~ised.9~ The conductometric process has also been adapted tothe titration of phenols, cresols, and certain hydroxy-acids.93 Asimilar process has been worked out for the estimation of alkaloidsand their salts, the alkaloids themselves being titrated with acid,whilst alkaloid salts are titrated with alkali.94An indirect method for the electrolytic estimation of halogenshas been based on the fact that the solution can be electrolysed bymeans of a silver anode, with the formation of an adherent depositon the anode, and without precipitation in the solution, providedthat an anode potential of 0.59 volt is not exceeded.95When an ammoniacal solution of nickel and a salt of arsenicacid is electrolysed, the nickel is quantitatively precipitated with-out any arsenic, whereas cobalt under the same conditions carriesdown a certain proportion of arsenic.96Reference may also be made to methods in which rotatingreductors are used in the estimation of iron97 and of m0lybdenum.~8Water Analysis.The principal contributions to the analysis of drinking-waterhave been in connexion with the dissolved carbon dioxide.It hasbeen shown that the simple mass-law equation of the primaryionisation of carbonic acid enables the hydrogen-ion concentrationof natural waters to be approximately calculated .99I n titrating the free carbon dioxide and that present as hydrogenG. L. Kelley and R. T. Bohn, J . Amer. Chem. SOC., 1919, 41, 1776 ; A.,ii, 134.91 I. M . Kolthoff, Rec. trav. chim., 1920, 39, 126 ; A., ii, 198.ga E. L. Taube, J . Amer. Chem. SOC., 1920,42, 174; A., ii, 396.98 I. M. Kolthoff, Zeitsch. anorg. Chem., 1920, 112, 187 ; A., ii, 711.94 I. M. Kolthoff, ibid., 196; A., ii, 781.96 J. H. Reedy, J . Amer. Chem. Soc., 1919, 41, 1898; A., ii, 122.96 N. H. Furman, ibid., 1920, 42, 1789 ; A., ii, 710.g7 W. Scott, J. Ind. Eng. Chem., 1919, 11, 1135; A., ii, 128.B8 W. Scott, ibid., 1920, 12, 578; A., ii, 578.g s R. E. Greenfield and G. C. Baker, ibid., 1920, 12. 989; A., ii, 771ANALYTICAL CHEMISTRY. 151carbonate in moorland waters, errors are caused1 by the presenceof weak organic aaids which are simultaneously titrated. To deter-mine the solvent action of such waters on limestone, the water isshaken a t intervals with pow'dered marble in a closed flask, and aportion then titrated with N / 10-acid. The difference between theresult and that obtained without the addition of marble gives theamount of calcium carbonate dissolved.1It has been shown by Tillmanns and Heublein2 that the freecarbon dioxide in soft water has a much greater solvent action oncalcium carbonate than the same amount of carbon dioxide in hardwater. In the oase of waters containing iron hydrogen carbonate,the semi-combined and the combined carbon dioxide in the ironcompound are titrated simultaneously with the free carbondioxide, and, as a correction, 1.1 milligram of carbon dioxide mustbe deducted for each milligram of ferric oxide present.3In using Escaich's colour test for nitrites4 in water, the resultsare uncertain in the presence of ohlorides, which must therefore beremoved by means of silver nitrate before applying the test.5C. AINSWORTH MITCHELL.1 V. Rodt, Chem. Zeit., 1920, 44, 469 ; A., ii, 507.2 Gemndheits-Ing., 1912, 35, 669 ; A., 1913, ii, 51.3 H. Noll, Zeitsch. angew. Chem., 1920, 33, i, 182 ; A., ii, 555.J . Pharm. Chim., 1918, [vii], 17, 395; A., 1918, ii, 273.A. Escnich, ibid., 1920, [vii], 22, 138 ; A., ii, 644