ANALYTICAL CHEMISTRY.THE number of papers with a bearing on analysis which haveappeared during the period under review has not diminished and thetask of selection is no easier. In surveying the literature, however,it has again been noticed that certain papers appear to be devoted tothe rediscovery of well-established facts or of methods which are byno means new. It is true that such contributions make a Reporter’stask somewhat easier, but with the literature in its over-burdenedcondition they are surely superfluous.In this Report a separate section has been devoted to Colorimetry,a, branch of analysis which, in recent years, appears t o be coming intoits own, with a wider future before it. Hence the time seems ripefor a separate treatment of the subject in its quantitative aspect,with especial reference to the instrumental side.L.S. T.MACNETO-OPTIC METHOD OF CHEMICAL ANALYSIS.has grown round this much-debatedmethod of analysis since it was first launched by Allison in 1930.2Subsequently, the method has been the subject of much contro-versy. Some investigators have claimed success with it, whilstothers have reported a complete failure, and not only has the inter-pretation of the results been in dispute, but also the very existenceOF the minima has been called in question.Recently, F. G. Slack has reported a critical investigation of themethod, and although minima were observed he regards them as atype of N-ray phenonienon, and as such, subject to physiologicaland psychological effects. Further, with J.A. Peoples, jun.,4 hereports that attempts to reproduce time-lag measurements and1 Bibliographies which cover the field will be found in the articles bySlack, ref. (3), and by Cooper, ref. (9). Tho method has been mentioned in theseReports from time to t h e , see Ann. Reports, 1931, 28, 181; 1932, 29, 300;1933, 30, 349 ; 1934, 31, 372 ; 1935,32, 142, a detailed description being givenin Ann. Reports, 1930, 27, 203.2 F. Allison and E. J. Murphy, J . Apner. Chern. SOC., 1930, 52, 3796; thefirst papers on the time-lag in the Faraday effect by Allison appeared in 1927;see J. W. Beams and-F. Allison, Physical Rev., 1927, 29, 161; Phil. Mag.,1927, [vii], 3, 1199.3 J . Franklin In&., 1934, 218, 445; cf. also H. W. Farwell and J.B.Hawkes, Physical Rev., 1935, [ii], 47, 78.4 Ibid., 1934, [ii], 45, 126.An extensive literaturTHEOBALD : INORGANIC ANALYSIS. 433chemical analyses have failed. Working independently, H. G.MacPherson finds himself in substantial agreement with Slackand Farwell and Hawkes, and is unable to confirm sharp minimacharacteristic of the substance under investigation. More evidenceof a similar nature comes from M. A. Jepperson and R. M.On the other hand, G. Hughes and R. Goslin seem to have madea pertinent reply to many of these doubts and misgivings by demon-strating the reality and reproducibiIity of the minima photographic-ally, and they claim that these objective tests are characteristicof the inorganic acids and salts which they have examined.Further-more, G. M. Wissink and J. W. Woodrow * have such confidence inthe method that they use it to detect vitamin-A, to the presence ofwhich they attribute the characteristic minimum given by manyanimal and vegetable products.Finally, we have a recent series of articles by S. S. Cooper andT. R. Ball9 who, after a somewhat enthusiastic introduction, inwhich they describe the magneto-optic apparatus as “ perhaps themost important tool for chemical research developed in the pastdecade ” and extol the advantages of this method of analysis, pro-ceed in the first paper to deal with the history and present status ofthe method. From their review they believe conclusively that theminima do exist and that they are characteristic of the substance orsubstances under investigation.The second paper discusses insome detail the construction, arrangement, and adjustment of theapparatus, and describes a successful test of the method in solving aseries of unknowns, whilst the last paper gives the actual procedureused in locating the minima for a specific example, vix., salicylic acid.The authors’ concluding remarks are illuminating and make it clearthat the technique is difficult, whilst a long and very specialisedtraining is necessary before success in the use of the method canhope to be achieved. It would seem, therefore, that for the presentat least the practising analyst will still have to rely on the morepurely chemical methods which are at his disposal for the identifica-tion and determination of small concentrations of unknowns.L.s. T.INORGANIC ANALYSIS.Quantitative.Standards for Volumetric Analysis.-In last year’s Report (p. 452)The stabilisation of5 Physical Rew., 1935, [ii], 47, 310.6 Ibid., p. 546.8 Ibid., 1934, Lii], 45, 126.9 J. Chew,. Educ., 1936,18, 210, 278, 3%; cf. also T. €3. Ball, Physical Bev.,this subject was dealt with in some detail.Ibid., p. 317.1936, Lii], 47, 548434 ANALYTICAL CHEMISTRY.Q.1N-sodium thiosulphate solutions by the addition of borax wasmentioned. Now it is pointed out lo that such solutions give, inneutral solution, iodine values which are too low even when onlyone-twelfth of the usual amount of crystalline borax is added.Experimental work indicating the suitability of borax as an acidi-metric standard appears from time to time, but the methods adoptedfor drying the hydrated salt have probably prevented its more ex-tended use.Drying by alcohol and ether is now shown to be aneasy and satisfactory procedure which eliminates this disadvantage.llFurther, loss of water by exposure to air is not a serious source oferror over periods of less than a week or ten days.I. M. Kolthoff and J. J. Lingane l2 have shown that potassiumthiocyanate is a suitable standard for work of ordinary accuracy(& O.lyo), and when stored in the dark the pure, melted salt is stableindefinitely ; deliquescence, moreover, is harmless a t a relative humid-ity of less than 45%. When protected from light, aqueous solutionsgave no detectable change in titre after eleven months’ st0rage.1~In spite of their high precision or reproducibility, potentiornetrictitrations of silver with thiocyanate are not suited to work of ahighly exact nature, owing to side reactions which have been foundto take place, and the relatively good results obtainable with theVolhard method are due, as is often the case in an analytical process,to a compensation of errors.12G. I?.Smith, V. R. Sullivan, and G. Frank l4 have proposed thesalt (NH4),Ce(N03)6, which, incidentally, is indicated to be the com-plex ammonium hexanitratocerate, as a reference substance inceriometry. They have worked out an easy method of preparationin a degree of purity sufficient for this purpose and have shown thatthe stability of the salt in hot, dilute sulphuric acid‘is satisfactory.Many desirable properties are claimed for the new standard, and itsuse may well prove to be a definite advance in volumetric work.I.ndicators.-There is now available a wide range of indicatorssuitable for the titration of acids and bases, but in those of the neweroxidation-reduction type there is still room for improvement and awidening in scope.As time goes on this extension is graduallybeing made, and one of the latest additions offering promise of wideuse is phenylanthranilic acid which has now been recommended 18for many of the usual oxidation processes of volumetric analysis,10 P. Horkheimer, Pharm. Z t g . , 1935, 80, 1330.11 I?. Hurley, jun., I n d . Eng. Chem.(Anal.), 1936, 8, 220.12 J . Amer. Chem. SOC., 1935, 57, 2126.13 Cf., however, E. N. Taran, J . Ben. Chem. Russia, 1935, 5, 602.1 4 I n d . Eng. Chem. (Anal.), 1936, 8, 449.1 5 A. Kirssanov and V. Tscherkassov, BUZZ. SOC. chi?n., 1936, [v], 3, 817;W. S. Syrokomsky and V. V. Stiepin, J . Amer. Chem. SOC., 1936, 58, 928THEOBALD : INORGANIC ANALYSIS. 435especially for titratioiis with ceric sulphate. It gives, it is claimed,a sharp, reversible colour change and is more stable in the presenceof excesB of oxidant than diphenylamine and its derivatives; theoxidation potential is + 1.08 volts and the indicator error is negli-gible. In iron titrations the addition of phosphoric acid is nolonger necessary, and the ease with which it can be preparedgives it a decided advantage over the costly o-phenanthrolinecompound.A 1 yo solution of brucine in 3N-sulphuric acid is another indicatorof this type which presents possibilities, and has been described asa suitable internal indicator for iron in dichromate titrations.16The colour change from green to red is said to be more distinct thanthat of diphenylamine and to be unaffected by ferric ions in largeexcess, by mercury, and by stannic ions; here again, phosphoricacid is unnecessary, and permanganate can be used as the titrant withhydrochloric acid present, This indicator has been employed withadvantage in the analysis of chromiurn-iron alloys, or of ores wherethe concentration of ferric iron is high.Sodium diphenylbenzidinemonosulphonate and the diphenylaminederivatives, NHPh*C,H,*SO,Na and NHPh*C,H,Me*SO,Na, havealso been described l7 as oxidation-reduction indicators for dichro-mate titrations.The details for the successful and accurate use of diphenylcarb-azide as an internal indicator in the volumetric determination ofiron have been fully worked out,18 but the close attention to pro-cedure which is demanded, the indicator correction which has to befound, and the careful control which is necessary in the titration itselfmay restrict the popularity of the method.The same substance has been satisfactorily applied l9 also todetermine the end-point in titrating chloride ion with mercuricnitrate solution.Benzopurpurin-B and -4B have been put forward as indicatorsfor the bromometric titration of tin and antimony.20The advantages to be gained from the recently-introduced fluores-cence indicators are discussed by &I. DBrib6r4, who recommendsumbelliferone, p-methylumbelliferone, and uraiiyl salts (in theabsence of halogens) as suitable for strong acids and bases, @-naphtholor eosin-BN for weak acids, and axxulin or fluorescein for weak16 D. S.Narayanmurthi and T. R. Seshadri, Proc. Indian Acad. Sci., 1936,17 S. Cohen [with R. E. Oesper], Ind. Eng. Chem. (And.), 1936, 8, 364.18 H. E. Crossley, Analyst, 1936, 61, 164.1s 1. Roberts, Ind. Eng. Chem. ( A w l . ) , 1936, 8, 366.20 Z . Raichinschtein, J . Appl. Chern. Rueeia, 1935, 8, 1470.3, A, 38; $3. Miyagi, J. SOC. Chern. Znd. Japan, 1933, 36, 146436 ANALYTICAL CHEMISTRY.bases.21 l-Naphthol-4-sulphonic acid (Nevile and Winther’s acid)shows a sharp change from no fluorescence to an intense blue atp , 6-66,22 and naphthionic acid and Schaeffer’s salt give, re-spectively, changes in fluorescence from pH 3 to 12 and 5 to > 11, bywhich pH can be measured to 0.5 unit.23In connexion with adsorption indicators, A.J. Berry 24 describesthe conditions under which phenosafranine, tartrazine, and rose-Bengal are best adapted to systems such as silver-halogen25 andthallous-thallic halides, and by titrating nitric or acetic acid withsodium hydroxide in the presence of lead nitrate and fluorescein oreosin, S. N. Roy 26 extends their application to acidimetry. Otherindicators worthy of note are a universal indicator for the pH range1-2-12.7,27 and 4-nitrocatechol,28 and oximinothiocamphor, pHrange 8.6-9.0,29 for acidimetry.The theories of Bjerrum and Brmsted have been applied to thetitration of weak acids and bases in water-ethyl alcohol mixtures,and it is found that the titration of weak bases is less practicable thanin water, but the salts of weak bases, e.g., alkaloid hydrochlorides,may be titrated with much enhanced accuracy in concentratedaqueous-alcoholic solutions.30Finally, I.M. Kolthoff 31 has discussed possibilities for the furtherdevelopment of acid-base indicators for the measurement of hydro-gen-ion activity and concentration, as well of adsorption, oxidation-reduction, and specific indicators for volumetric purposes .32Reagents.-No new reagent * comparable with 8-hydroxyquinoline21 Ann.Chim. analyt., 1936, [iii], 18, 37.22 Idem, ibid., p. 120.24 Analyst, 1936, 61, 315.26 Cf. also R. Ripan-Tilici, 2. anal. Chem., 1936, 104, 16, for axgentometrictitration of halide, thiocyanate, selenocyanate, and cyanate with fluoresceinas adsorption indicator.23 Idem, ibid., p. 173.26 J . Indicm Chem. Xoc., 1936, 13, 486; cf. idem, ibid., 1935, 12, 584.27 F. &ha and K. KBmen, Chem. his&, 1936,30, 22, 129.t 8 S. R. Cooper and V. J. Tulane, I n d . Eq. Chem. (Anal.), 1936,8, 210.29 D. C. Sen, J . Indian Ghem. Xoc., 1935,12, 751.30 H. Baggesgaard-Rasmussen, 2. and. Chem., 1936,105, 269.31 I n d . Eng. Chem. (Anal.), 1936, 8, 237.32 For a useful general article see “ Universal and Other Indicators,” byT.G. Pearson in Thorp’s “ Dictionary of Applied Chemistry,” 1935, Supple-ment, Vol. 2, p. 617.* Styryl dyes (P. Krumholz and E. Krumholz, Mikrochem., 1935, 19, 47),8-hydroxy-5-methylquIuinoline (C. E. Giete and A. SB, Anal. Asoc. Qzcim.Argentina, 1935, 23, 45), a mercaptan-like substance named “ thiocarbin ”(A. Steigmann, Phot. Ind., 1936, 34, 499), and what is probably ethyl 5-keto-2-thionhexah~drop~rimidine-4-carbox~late (S. E. Sheppard and H. R. Brigham,J . Amer. Chem. Soc., 1936, 58, 1046) are newly-described reagents for certainmetalsTHEOBALD : INORGANIC ANALYSIS. 437or thionalide has recently come to the fore, but a review which hasjust been published 33 of the use of organic reagents in both qualita-tive and quantitative analysis and their increasing significance willbe of much interest to all engaged in analytical work.Gravimetric and Volumetric Methods for the Determination, of theElements.-In a report of this nature it is impossible to do more thanmention a comparatively small number of the many papers whichhave been concerned with the determination of the elements duringthe period under review, and although all new analytical methodsshould be approached with an open mind tempered by a criticaloutlook, any criticisms of the present work would be presumptiveand premature, for the true value of a method can be assessed onlyin relation to the purpose for which it is designed, and after it hasbeen put to the test of practical experience.Group I .The accuracy of the potentiometric iodide-silver titra-tion as distinct from its precision or reproducibility has been recentlyin~estigated.~~ In the slow titration of silver with iodide, errors areproduced by the adsorption of iodide ions by the silver iodide formed,but the magnitude of the error can be reduced to a small value(0.017% or 0.028%) by titrating a t go", or by digesting the precipi-tate a t 90" in presence of a slight excess of silver prior to completionof the titration a t the ordinary temperature. Silver can also besuccessfully titrated with potassium iodide, even in the presence ofcupric or ferric ions, by using ceric ammonium sulphate and starch asinternal indicators, since oxidation of the iodide ion to iodine byCe"" is not permanent until the end-point is reached.35 Havingfound that, using the iodine monochloride end-point, the titrationof thaEZous salts with either potassium permanganate or ceric sulphateis unreliable, E.H. Swift and C. S. Garner 36 recommend titrationwith potassium iodate instead.The method previously described 37 for the removal of tungstenfrom tin by means of 8-hydroxyquinoline in an oxalate medium giveshigh results for tungsten owing to retention of tin, and a modificationnow put forward 38 is designed to overcome this.Group I I . Further details 39 have now been supplied 40 concerning33 F. Feigl, I n d . Eng. Chem. (Anal.), 1936, 8, 401.34 I, M. Kolthoff and J. J. Lingane, J . Amer. Chem. SOC., 1936, 58, 1524;See idem, ibid., 1935,57, 2126, 2377, for similar studies cf.idem, ibid., p. 1528.of the systems silver-thiocyanate and mercury-thiocyanate.2.5 A. Bloom and W. M. McNabb, I n d . Eng. Chem. (Anal.), 1936,8, 167.36 J . Amer. Chem. Soc., 1936, 58, 113.3 7 A. Jilek and A. RyBhek, Coll. Czech. Chem. Comm., 1933, 5, 136.313 Idem, ibid., 1936, 8, 246.39 Cf. Ann. Reports, 1935, 32, 457.40 G. Spacu and M. Kurag, 2. and. Chem., 1936,104, 88438 ANALYTICAL CHEMISTRY.the gravimetric determination of lead, thallium, bismuth, and gold,which are precipitated by thiolbenzthiazole as C7H,NS2PbOH,C7H,NS,T1, (C7H4NS2)3Bi, and (C7H,NS2)3Au, respectively ; andlead has been separated as carbonate, from copper, or cobalt andnickel, by the passage of carbon dioxide into a solution of the nitratesin presence of ~ y r i d i n e .~ ~ After a critical review of the determinationof lead as sulphate, and of antimony as sulphide by the method ofVortmann and Metz1,42 modifications of procedure have beenrecommended for these metals and their alloys.43 The chromatemethod for lead has receiyed further in~estigation,~~ and the B.P.(1932) method has been criticised as inaccurate owing to the non-quantitative liberation of oxalic acid from the lead oxalate precipi-tate; more accurate results are obtained, it is stated, by determin-ation of the excess of oxalic acid in the filtrateeq5In the gravimetric determination of mercury, mercuric sulphide isremoved from the weighed sulphide-sulphur precipitate by dis-solution in cold, concentrated hydriodic acid and the residual sulphuris weighed.The procedure can conveniently be applied to the rapidevaluation of technical grades of the ~ulphide.*~ An indirectvolumetric method, based on a critical study of the dichrornate-pyridine method of G. Spacu and J. Diekt7 consists of titrating thedichromate ion in the precipitated fHg py,]Cr,O, by one of the con-ventional methods, whilst substitution of acetone, in which thecomplex is less soluble than in alcohol, simplifies and improvesthe washing technique originally advocated by Spacu andDick.A critical comparison of a number of methods by an independentworker often serves a useful purpose in assessing the true value ofproposed methods and helps to clarify the position for those seekingan alternative to the older methods available. Several such reviewshave been published during the year.Various separations ofbismuth from lead, vix. , bromide-bromate hydrolysis,48 the pyro-gallol preci~itation,4~ and the cupferron methodY5O have now been4 1 A. Jilek, J. Kot’a, and J. Vfegt’al, Chem. Liaty, 1935, 29, 299.42 F. P. Treadwell and W. T. Hall, “ Analytical Chemistry,” 1935, Vol. 11,43 H. Vdoviszevski, 2. anal. Chem., 1936, 104, 94.44 Idem, ibid.; L. Guzelj, ibid., p. 107; Z. Karaoglanov and M. Jblichov,ibid., 1935, 103, 113.45 S. Wetherell, Quart. J . Pham., 1935, 8, 453.48 E. R. Caley and M. G. Burford, Ind. Eng. Chem. (Anal.), 1936,8,43.4 7 2. anal. Chem., 1929, 76, 273.O 8 L. Moser and W. Maxymowicz, 2.anal. Chem., 1925, 67, 248.49 F. Feigl and H. Ordelt, ibid., 65, 448.60 A, Pinkus and J. Dernies, Bull. SOC. cham. Belg., 1928, 37, 267.p. 220THEOBALD : INORQBNIC ANALYSIS. 439checked,51 but the hydrolysis in formic acid solution 52 leads toirregular results. Bismuth can also be separated from copper by amodification of the cyanide method of Fresenius and Haidlen,53but for both the bismuth-lead and the bismuth-copper separationsthe bromide-bromate procedure of Mosdr and Maxymovicz is pre-ferred. L. ICielt and G. C. Chandlee 64 find that precipitation withgallic acid at 70" serves to separate bismuth from lead, copper,cadmium, and many other metals, but not from mercury, antimony,tin, or silver. The phosphate method has also been re-examined 55and found to be inaccurate when lead is present.F. Hecht and R.Reissner 56 report that the micro-determination of bismuth asoxyiodide 57 is inexact, and that the macro-determination asC,H,N*OH,HBiI, 58 is unsatisfactory owing, inter alia, to the partialdecomposition of the precipitate during washing.59 Good results onboth the macro- and the micro-scale can be obtained, however, underthe exact conditions prescribed, by precipitation with 8-hydroxy-quinoline.The well-known Haen-Low volumetric method for copper isaccurate but long, and a shortening of the time required is much tobe desired. B. Park's procedure 6o which aims at effecting this hasnow been modified ; 61 potassium hydrogen phthalate has no materialeffect on the pH of the solution and can be omitted, but in order toensure complete oxidation of an ore containing sulphide, iron, andarsenic, a double treatment with nitric and hydrochloric acids, or asingle treatment with the two acids followed by one with saturatedbromine-water, is found to be essential.H. W. Foote and J. E.Vance 62 apply their modified iodometric method,63 with controlledpE, to determine copper in the presence of AsV and not more thanapproximately 20 rng. of antimony, interference by iron being pre-vented by the now common device of adding sodium fluoride.*61 E. A. Ostroumov, 2. (Mzal. Chm., 1936,106, 36.58 A. L. Benkert and E. F. Smith, J . Amer. Chem. Soc., 1898, 18, 1065.59 See Treadwell and Hall, op. d., p. 206.54 Ind. Eng.Chem. (AnaZ.), 1936, 8, 392.55 W. C. Blasdale and W. C. Parle, ibid., p. 352.56 Mikmchem., 1935, 18, 283.67 R. Strebinger and W. Zins, Mikrochem., 1927, 5, 166; 2. awl. Chem.,58 R. Berg and 0. Wurm, Ber., 1927, 80, 1664.59 F. Hecht and R. Reissner, 2. anal. Chem., 1935,103, 261 ; see also, idem,60 Ind. Eng. Chem. (Anal.), 1931, 3, 7 7 .6 1 W. R. Crowell, T. E. Hillis, S. C. Rittenberg, and R. F. Evenson, ibid.,68 Ibid., p. 119. * See K. Heller and F. Machek (Mikrochm., 1936,19, 147), for a review of1927,72, 417.ibid., pp. 88, 186.1936, 8, 9.63 Ann. Reports, 1935,32, 459, ref. (73).the literature on the determination and detection of cadmium440 ANALYTICAL CHEMISTRY.The complete precipitation of molybdenum by hydrogen sulphideis generally a matter of difficulty in an analysis, and it is noteworthythat the trisulphide can be precipitated quantitatively in the presenceof formic acid by initial reaction with a solution of hydrogen sulphidesaturated a t 0°,63a a finding which supports previous claims of asimilar nature ; 64 separation from tungsten is accomplished byadjusting the p , to 2.9 by means of a suitable buffer.There is littleto be gained, in general, by conducting the precipitation of the molyb-denum under pressure-anot her point which has been much disputedin the past, According to H. G o ~ o , ~ ~ molybdenum is completelyprecipitated by 8-hydroxyquinoline in the range pH 3.3-7.6, andvanadium a t pH 2.7-6.1 : and the conditions under which the formercan be determined volumetrically by oxidation of MoV to MoVI with0-1N-ammonium vanadate have been investigated by R.Lang andS. Gottlieb; G6 iron, vanadium, and large amounts of copper renderthe method impracticable. N. H. Furman and W. M. Murray,j ~ n . , ~ 7 reduce MoVI quantitatively to MoV by shaking with mercuryin a solution which is 2-3.5N with respect to hydrochloric acid,and titrate the quinquivalent molybdenum with ceric sulphate atthe ordinary temperature using the o-phenanthroline-ferrous com-plex indicator. They state that the presence of considerablequantities of phosphate, arsenate, or of ammonium salts is withouteffect on the accuracy of the molybdenum determination.A fractional distillation method has been worked out at theBureau of Standards for the separation of arsenic, antimony, and tinfrom one another and from elements having non-volatile chlorides ;germanium and rhenium, but not bismuth, interfere.68 The op-timum conditions of acidity for the volumetric determination ofantimony and arsenic by Andrews's iodine monochloride method 69have been fully in~estigated,'~ and a critical examination of variousmethods for arsenic in iron, steels, and iron ores has shown that thebest results are obtained by dissolution in nitric acid or bromine-water (bromine and hydrochloric acid for ores) followed by distilla-tion of the arsenic as chloride.71 Published methods for the Gutzeitreaction carried out on paper strips have also been critically re-viewed.7863a H. Yrtgoda and H. A.Fales, J . Amer. Chem. SOC., 1936,58, 1494.64 Cf. I. Koppel, Chem.-Ztg., 1924, 48, 801; J. fit6rba-Bohm and J. Vos-6 5 J . Chern. SOC. Japan, 1935, 56, 314. O6 8. awl. Chem., 1936, 104, 1.67 J . Amer. Chem. SOC., 1936, 58, 1689.6s J. A. Schemer, J . Res. Nat. Bur. Stand., 1936, 16, 253.69 J . Amer. Chem. Soc., 1903, 25, 756.70 A. Mutschin, 8. anal. Chem., 1936, 106, 1.71 A. Stadeler, Arch. Eisenhuttenw., 1935-36, 9, 423.72 W. Mddsteph, Z. anal. Chem., 1936, 104, 333.tfebal, 8. anorg. Chern., 1920, 110, 81THEOBALD : INORGANIC ANALYSIS. 441Dry distillation in oxygen serves to separate O ~ l - O - O O 1 ~ o ofselenium from sulphur and sulphur-containing materials.73 Arsenicand tellurium present in an amount equivalent to the selenium donot interfere.74Volumetric methods for tin have long been a source of tribulationt o the analyst, and two papers which have recently been publishedmay help to clear the situation.According to F. L. Okell and J.L ~ m s d e n , ~ ~ low results in tin titrations are due to oxygen dissolvedin the iodine solution and not necessarily to incomplete exclusion ofair in the flask, and the interfering action of titanium experienced inthe analysis of tin ores is eliminated when oxygen-free iodine is usedfor titration. Reduction with aluminium turnings is recommendedin ore analysis.76 For alloys of tin and lead with less than 2% ofantimony, the tin and lead are best dissolved directly in concentratedhydrochloric acid in absence of oxygen, and the antimony removedby filtration before titration with potassium iodate.Removal ofantimony is essential since cold stannic chloride reacts with freshly-precipitated antimony too rapidly to permit accurate titration ofSn" in its presence.77 The titration of Sn" has also been carriedout with ceric sulphate, diphenylamine being the indicator.78When heated a t 400-500" with excess of ammonium iodide, tindioxide is quantitatively volatilised as stannic iodide and advantagecan be taken of this to ascertain the purity of the ignited metastannicacid obtained in the usual course of an analysis.79R. Gilchrist and E. Wichers have made an important contributionto the analysis of the phtinum metals in which a new procedurefor the separation of osmium, ruthenium, platinum, palladium,rhodium, and iridium from one another, and their gravimetricdetermination, have been described ; no specialised equipment orreagents are necessary and an accuracy comparable with that of thebest analytical procedures for the common metals is claimed fortheir methods.Goup 111. For the determination of small amounts of iron,aluminium, and titanium in admixture, 8-hydroxyquinoline and its5 : 7-dibromo-derivative, which precipitates titanium in acid andaluminium in alkaline solution, have been utilised.*l J.Dewar and73 See G. G. Marvin with W. C. Schumb, I n d . Eng. Chem. (Anal,), 1936, 8,109, for determination of selenium in 18 : 8 stainless steels.74 Idem, ibid., 1935, '7, 423.7 5 Analyst, 1935, 60, 803.76 See also L. Deutsch, Ann.Chim. analyt., 1936, [iii], 18, 10.7 7 H. F. Hourigan, Analyst, 1936, 61, 328.7 8 N. A. Rudnev, Trans. Butlerov Inst. (.'hem. Tech. Kazan, 1934, No. 2, 51,79 E. R. Caley and M. G. Burford, I n d . E n g . Chem. (Anal.), 1936, 8, 114.80 J . Arner. Chem. Soc., 1935, 67, 2565.81 A. M. Zanko and A. J. Bursuk, J . Appl. Ckem. Russia, 1936, 9, 895442 ANALYTICAL CHEMISTRY.P. A. Gardiner 83 find that, contrary to the adverse criticism of L.Moser and M. Nie~sner,*~ a slight modification of H. Britton'smethod 8* (hydrolysis of the alkali beryllate) furnishes an accurateseparation of aluminium and beryllium when the former is not greatlyin excess of the latter. For the case of aluminium and zinc, F. H.Fish and J. M. Smith, jun., adopt the aluminate method,85 in whichaluminium is weighed as 2Li,O ,5A1,0,, zinc being determined in thefiltrate as the pyrophosphate, whilst T.K8zu 86 separates aluminiumfrom manganese, cobalt, and nickel, but not zinc, by precipitationwith a saturated, aqueous solution of aniline, the aluminium beingweighed as the oxide.This use of organic bases in preference to ammonia is e~tending,~'and 20% pyridine is now employed to precipitate iron, chromium,and aluminium as hydroxides from dilute nitric or hydrochloric acidsolution in the presence of the corresponding ammonium salts. Asingle precipitation is said to give a separation, which is practicallycomplete, from manganese, cobalt, and nickel, but not zinc, and themethod has been applied to pyroluaite and to cobalt ores.88 Uran-ium, also, is precipitated, as HzUzO,, from solutions of uranyl salts,thus providing a quantitative separation from the alkaline earths.89For the removal of gullium from beryllium, titanium, zirconium,and thorium, S.Ato 90 has recourse to the ether-extraction methodfrom hydrochloric acid solution, and finally precipitates the galliumwith sodium camphorate ; and for the determination of this metal inaluminium, J. A. Scherrer 91 adopts the same device but precipitatesthe gallium with cupferron and then ignites it to the oxide, afterremoval of any iron, tin, etc., which has accompanied the galliuminto the ether extract. An alternative method, obviating an extrac-tion and based on precipitation with cupferron in sulphuric acidsolution, is also given.In view of the use which is made in some laboratories of perchloricacid as an oxidising agent for chromium, it is of interest to learn thatthe incomplete oxidation which results when this acid alone is usedis due to the production of it small amount of hydrogen peroxide;82 Analyst, 1936, 61, 536.84 Analyst, 1921, 46, 361, 437 ; 1922, 47, 50.85 Ind.Eng. Chem. (Anal.), 1936, 8, 349; see J. T. Dobbins and J. P.86 J . Chem. SOC. Japan, 1935, 56, 562, 683.88 E. A. Ostroumov, 2. anal. Chem., 1936, 106, 170.89 Idem, ibid., p. 244.90 Sci. Papers I m t . Phys. Chem. Res. Tokyo, 1936,29, 71 ; cf. 1:ba'd., 1931, 24,91 J . Rerr. Nat. BUT. Stand., 1935, 15, 585.83 ~?lonatsh., 1927, 48, 113.Sanders, J . Amer.Chem. SOC., 1932, 54, 178.See Ann. Reprt8, 1935, 32, 462, ref. (10); 460, ref. (86).270; 1931, IS, 289THEOBALD : INORGANIC ANALYSIS. 443quantitative results, however, are obtained with a mixture of per-chloric and sulphuric acids in certain proportion^.^^Analytical methods for the determination of zirconium and thecomplete analysis of zirconium minerals have been re~iewed.9~W. R. Schoeller and his collaborators have now completed theirinvestigations into the analytical chemistry of tantalum, niobium,and their mineral associates : the four papers recently published com-plete the series. I n the first of theseYg4 the fate of beryllium in themore important, operations advocated for the separation of the vari-ous earths and the analysis of minerals containing them is dealt with ;the secondg5 is concerned with the determination of tungsten inearth-acid minerals,g6 and the third with the separation of phosphorusand vanadiuimg7 The last paper 98 contains a general summarywhich is intended as a key to facilitate the study of the whole series.Not the least striking feature of these remarkable investigations isthe simplicity of the apparatus and reagents whereby the resultshave been achieved; and, as is pointed out in the final paper, “ thesimple classic processes of mineral analysis have proved adequatefor the solution of some of its most difficult problems.” The mono-graph based on these researches, which is to be published under theaegis of the Society of Public Analysts, will be awaited with greatinterest by allwhosework takes them into the field of mineral analysis.Anthranilic acid has been confirmed as a suitable, althoughrestricted, reagent for the determination of zinc, cadmium, cobalt,nickel, and copper; the conditions specified by 13. Funk and M.Dittg9 are satisfactory except in the case of nickel, and are notimproved by the addition of ammonium or sodium acetate, orsodium tartrate.9ga This acid, in the form of its sodium salt, hasalso been employed for the micro-determination of zinc.1The post -precipitation of zinc sulphide with mercuric sulphidehas been studied in its theoretical and practical aspects by I.M.Kolthoff and R. Moltzau ; and J. R. Caldwell and H. V. Moyer3 find92 G. F. Smith, L. D. McVickers, and V.R. Sullivan, J . SOC. Chern. Ind.,1935, 54, 3 6 9 ~ .93 G. A. Ampt, J . Proc. Austral. Chem. I m t . , 1935, 2, 321.94 W. R. Schoeller and H. W. Webb, Analyst, 1936, 61, 235.9 5 W. R. Schoeller and E. F. Waterhouse, ibicl., p. 449.96 Cf. Ann. Reports, 1935, 32, 462, ref. (21).97 W. R. Schoeller and H. W. Webb, ilna,lyst, 1936, 61, 585.98 W. R. Schoeller, ibid., p. 806.99 Z . anal. Chem., 1933, 93, 241.9% R. J. Shennan, J. H. F. Smith, and A. M. Ward, AnaZyst, 1936,61, 395.1 C. Cimerman and P. Wenger, Arch. Sci. phys. na,t., 1935, [v], 17, Suppl.,2 J . Physical Chem., 1936, 40, 779.9 J . Amer. Chem. SOC., 1935, 57, 2372.94-98444 ANALYTICAL CHEMISTRY.that the addition of small amounts of gelatin or agar producesimmediate and complete flocculation of zinc sulphide suspensionsand permits filtration 15 minutes after precipitation.Satisfactoryseparations from nickel, manganese, aluminium, or chromiumcan $hus be ~btained,~ and when a small amount of acraldehydeis also added, a good separation from cobalt can be effected witha single pre~ipitation.~Precipitation is generally preferable to extractionfrom solids in quantitative work, and one such method has beenworked out 6 for the nitrates of the alkaline-earth metals. Strontiumnitrate is completely precipitated in a dense, crystalline form andseparated from numerous other metals by the addition of lOOyonitric acid until the resultant solution contains not less than 79% ofnitric acid. Barium and lead can be obtained free from othermetals in a similar way.The solubility of calcium nitrate decreasesrapidly with increasing acid concentration, and caZcium and stron-tium nitrates are separated from each other when the content ofnitric acid lies between 79 and 81% : the method seems preferableto that of S. G. Rawson.7 Attempts to separate these two metalsby precipitation of strontium nitrate with nitric acid in variousorganic media were not encouraging.Rosolic acid has been used as indicator for the direct titration ofbarium with potassium chromate by the method of K. Jellinek andJ. Czerwinski,g and this furnishes an indirect method for the deter-mination of sulphate and of sulphur in pyrites and slags.9Group V . The determination of potassium on the micro-scaleforms the subject of a detailed examination by P.Wenger, C.Cimerman, and C. J. Rzymowska,lo who find that Emich's chloro-platinate method gives good results for this element alone, but not inthe presence of more than four times as much sodium. By an adapt-ation of the macro-method of F. G. Smith and J. L. Gring,ll in whichpotassium is converted into its perchlorate before treatment withchloroplatinic acid, they have been able to determine gravimetricallyamounts of potassium of the order of 0-5 mg. in the presence of tenGroup I V .4 Idem, J . Amer. Chem. Soc., 1935,57, 2372.5 Idem, ibid., p. 2375.6 13. H. Willard and E. W. Goodspeed, I n d . Eng. Chem. (Anal.), 1936, 8,7 J. SOC. Chem. Ind., 1897,16, 113; cf. W. Noll, 2. alzorg. C'hem., 1931,199,414.193.2.anorg. Chern., 1923, 130, 253.9 A. V. Vinogradov, Ann. Chirn. analyt., 1935, [iii], 17, 285.10 MiErochem., 1936, 20, 1 ; Arch. Sci. phys. nat., 1935, [v], 17, Suppl.,11 J . Amer. Chem. SOC., 1933, 55, 3957.p. 89THEOBALD : INORGANIC ANALYSIS. 445times as much sodium. A volumetric ending, which is particularlysuitable in the determination of potassium in biological media, hasalso been worked out by these authors, and depends on the conversionof the K,PtCl, into K,PtI, by treatment with potassium iodide, andtitration of the iodoplatinate with thiosulphate.12C. H. Greene l3 has examined the sensitivity of the magnesiumuranyl acetate reagent to sodium and potassium, and his data showthat addition of alcohol increases the sensitivity of the reagent to-wards sodium more than towards potassium-a desirable result.The distinction between these two metals also improves as the ratioof the volume of the reagent to that of the test solution is increased.Errors arising from the micro-determination of sodium by means ofan aqueous-alcoholic solution of this reagent have also been discussedby A.Krassilchik,14 who puts forward an improved technique.Disturbing discrepancies have often come to light in the valuesobtained for the alkalis by independent analysts using the LawrenceSmith method on the same material, and incomplete attack in thefusion may often be the explanation of the disagreement. This andother sources of error are discussed by M. 0. Lamar, W.M. Hazel,and W. J. O’Leary,15 who suggest remedies and improvements parti-cularly for substances known to be refractory. For aluminiumrefractories, other workers l6 fuse with ammonium fluoride as amethod of attack, and then, after expulsion of hydrogen fluoride,precipitate the sodium with zinc uranyl acetate, reduce the uraniumin the precipitate by a coil of aluminium, and titrate back withpermanganate. For amounts of sodium ( ? Na,O) from 0.6 to 4.574,test results show goodagreement with those obtained by the LawrenceSmith method.Anions. There remain to be mentioned in this section some in-vestigations concerned with the determination of the anions.According to G. L. Jenkins and C. F. Bruening,17 the official(National Formulary) methods are suitable for ferric but notfor ammonium, calcium, sodium, potassium, and manganesehypophosphites,18 and improvements are suggested.H. Terlet andl2 For a “ Semmelreferat ” of methods for the micro-determination ofpotassium, see C. Cimerman and C. J. Rzymowska, Mikrochem., 1936, 20,129.13 Ind. Eng. Chem. (AnaE.), 1936, 8, 399.14 Compt. rend., 1936, 203, 78.15 Ind. Eng. Chem. (AnaE.), 1935, 7, 429.16 H. V. Churchill, R. W. Bridges, arid A. L. Miller, ibid., 1936, 8, 348.17 J. Amer. Pharrn. Aasoc., 1936, 25, 19.18 For these last five salts the methods are based on oxidation t o phosphatewith nitric acid, precipitation of phosphate with excess of silver nitrate,m d a back-titration with ammonium thiocyanate446 ANALYTICAL CHEMISTRY.A.Briau l9 report that for phosphoric acid Scheffer's method,awhich consists in the titration of ammonium phosphomolybdatewith sodium hydroxide in the presence of formaldehyde, is trust-worthy and accurate when steps are taken to remove co-precipitatedinolybdic acid. They outline procedures for the examination ofnatural and artificial phosphates of various kinds.21The conditions necessary for the precipitation of benzidine sul-phate (pH 2-3),22 the titration of sulphnte with lead nitrate with eosinas incticat~r,~~ and indirect methods with sodium rhodizonate 24and rosolic acid as indicators 25 have also been communicated.Much of the work recently done on the determination of $wineis concerned with H. I€. Willard and 0. B. Winter's distillationmethod 26 and suggested modifications of it,27 and a150 with Sanchis'Owing to the marked adsorption properties of lanthanumfluoride, determination of fluorine by precipitation in this form 29is found to be impra~t~icable.~~ Precipitation as PbFBr 31 or asK2SiF6 32 forms the basis of gravimetric or volumetric methods ; andfor the determination of fluorine in minerals, calcium fluoride pro-tected as a colloid with gelatin forms the basis of a nephelometricmethod.33 A review of various procedures for silica in the presenceof fluorine has also been made.34When iodine monochloride solution is treated with saturatedpotassium bromide, iodine is a t first liberated, but on addition ofmore bromide this disappears, iodine bromide being re-formed.19 Ann.Faiaif., 1935, 28, 546.20 J. OfficieZ, 1934, Aug. 30th.21 See H. Trapp, J. pr. Chem., 1935, [iif, 144, 93, for calcium phosphates.23 E. C. Owen, Biochem. J., 1936, 30, 352.23 J. E. Ricci, Ind. Eng. Chem. (Anal.), 1936, 8, 130.24 R. Strebinger and L. von Zombory, with L. PollBk, Z. a d . C'hem., 1936,25 A. V. Vinogradov, loc. cit., ref. (Q), p. 444.26 Ind. Eng. Chern. (Anal.), 1933, 5, 7.27 D. Dahle and H. J. Wichmann, J. Assoc. 03. Agric. C'hem., 1936,19, 313,320; D. S. Reynolds, J. B. Kershaw, and K. D. Jacob, ibid., p. 156; W. K.Gilkey, H. L. Rohs, and H. V. Hansen, Ind. Eng. Chenz. (Anal.), 1936, 8, 150;W. D. Armstrong, ibid., p. 384 (micro-method).28 E. H. Ducloux, Anal. Aaoc. Quim. Argentina, 1935,23,63; J . M. Muiioz,Rev.Xoc. Argentin. Biol., 1934, 10, 395; A. H. de Carvalho, Rev. C'him. puraappl., 1936, [iii], 11, 99.105, 346.20 Cf. R. J. Meyer and W. Schulz, Z. angew. Chem., 1925,38,203.30 J. Fischer, with E. Muller and H. Knothe, 2. and. Chem., 1936,104, 344.31 A. A. Vesiliev, J . Appl. Chem. Ru88&3, 1936, 9, 747; see also, ideln, ibid.,3% A. A. Vasiliev, with N. N. Martianov, 2. anal. Chem., 1935, 103, 107.33 R. E. Stevens, Ind. Eng. Chem. (Arttrl.), 1936,8, 248.34 S. S. Korol and V. M. Ka,lushskajja, J. Appl. Chem. Russia, 1936, 9, 148.p. 943, for the determination of fluorine in the presence of berylliumTHEOBALD : INOR(3ANIC ANALYSIS. 447To apply the iodine bromide process tjo the determination of iodineor iodide, the sample is treated with a large excess of potassium bro-mide, and concentrated hydrochloric acid.After dilution and addi-tion of carbon tetrachloride, the solution is titrated with potassiumiodate-other oxidising agents such as ceric sulphate, potassiumperiodate, permanganate , or dichromate can be used-until thecarbon tetrachloride is decolorised. Antimony in the presence ofa hydrochloric acid concentration too great for the iodine mono-chloride process to be satisfactory can also be determined by this~ i i e a n t i . ~ ~Iodides can be titrated with ceric sulphate to a visual end-point inthe presence of acetone and sulphuric acid, o-phenanthroiine-ferrousion being used as indicator : moderate amounts of chloride have noadverse effect, and interference due t,o bromide can be largely elimi-nated by appropriate d i l u t i o i ~ .~ ~ Other methods for bromide andchloride,37 and bromine and iodine in the presence of each 0ther,~8have also been given.The use of mercury 39 or amalgams seems to be increasing in popu-larity in analytical work, and advantage has been taken40 of thereduction of chlorates, brornates, and iodates by zinc amalgam or byWood’s alloy for their determination.Vanadous sulphate serves for the quantitative reduction of chlor-ates, nitrates, and persulphates in an inert atmosphere, -the excessVS04 being titrated with potassium ~ermanganate.~~ M. B. Donaldreports that the optimum conditions for the reduction of nitratesare very different from those originally specified by Devarda, muchless sodium hydroxide being required.42A routine method, based on the ‘‘ partition ” of boric acid betwcenwater and ether in the presence of hydrochloric acid and ethylalcohol, has been developed for the determination of boron in gla~s.4~Boric oxide contents varying from 0.7 to 16% can be rapidly deter-mined with no material sacrifice of accuracy; zinc interferes seri-ously, but barium, fluorine, and abnormal amounts of iron onlyslightly.Phosphoric acid (40%) has been recommended for the expulsiona5 R.Lang, 2. unal. Chem., 1936, 100, 12.36 D. Lewis, Ind. Eng. Chern. (Anal.), 1936, 8, 199.37 G. G. Longihescu and E. I. Prundeanu, Bull. Acad. Sci. Rotmaiiw, 1935,38 L. Spitzer, Ind. Eng. Chern. (Anal.), 1936, 8, 466.39 Cf. N. H. Purman and W.M. Murray, jun., Zoc. cit., ref. (67), p. 440.40 P. G. Popov, Ukrain. Chem. J . , 1936, 10, 413.41 P. C. Banerjee, J . lndkn Chem. Soc., 1936, 13, 301.42 A w l y e t , 1936, 61, 240.43 F. W. Glaze and A. N. Finn, J . Rm. Nat. BUT. Stand., 1936,16,421.17, 47448 ANALYTICAL CHEMISTRY.of carbon dioxide from carbonate^.^^ This is by no means new,for the use of this acid was advocated over 30 years ago by G. T.Morgan,45 and since that time many hundreds of determinationsof carbon dioxide in dolomite and rocks, for which it is particularlysuitable, have been carried out by this means in the laboratory towhich the Reporter is attached. It is surprising that the method isnot more widely adopted, for it has many advantages, chief amongwhich are the elimination of the condenser and absorption tubenecessitated by the volatility of hydrochloric acid, and the verysmall “ blank ” which it affords.The preparation of a solution of manganic sul-phate as a reagent for volumetric work has been recently described.46When protected from light the solution showed no change in titreover a period of nine days, and it rapidly oxidises nitrites, oxalates,ferrous iron in presence of chloride, and VII to V V .The reactionsare stoicheiometric, and the end-points well-defined, thus affordingresults superior to titrations with permanganate.There are many signs in the literature that more attention is beingpaid to the statistical evaluation of the possible errors to which ananalysis is subject. In the past, this has generally been taken intoaccount in the determination of atomic weights, but it has beensomewhat neglected in ordinary analytical work, and this interestwhich is being displayed in the theoretical aspect of the reliabilityof an analytical measurement is all to the good.It is also desirablethat there should be a clear recognition of the meaning of the termserror, precision, and accuracy, which have often been too vaguelyused. An interesting chapter on this subject is to be found in I. M.Kolthoff and E. B. Sandell’s recently published book on analysis:’and an article which should be of much interest to analysts has beenwritten by A. A. Benedetti-Pichler 48 who discusses the statisticalaspects of chemical measurements applicable to analytical data.MiscelZaneous.Qualitative.Methods for the -Detection of Anions and Cations.-During the periodunder review several new schemes for the systematic separation ofanions have been put as well as methods for the commoner44 F. Vojif, Chem.Listy, 1935, 29, 185.4 5 J . , 1904, 85, 1004.46 A. R. J. P. Ubbelohde, J., 1935, 1605.4 7 “ Text-book of Quantitative Inorganic AnalyEiis,” Macmillan, New Yo&,48 Ind. Eng. Chem. (Anal.), 1936, 8, 373.49 E. Umblia, Keem. Teated, 1935, 2, 79; J. T. Dobbins and H. A. Ljung,J . Chem. Educ., 1935, 12, 586; E. W. Flosdorf and C. Henry, ibid., 1936, 13,274; F. Pozna and E. Migray, Ann. Chim. appl., 1936, 26, 81.1936, Chapter XV, p. 250THEOBALD : 1NORG.ANIC ANALYSIS. 449cations which dispense with the use of hydrogen ~ulphide.~O Asimplified method for Group I1 51 has also been recommended.E. R.Caley and M. G. Burford 52 find that concentrated hydriodicacid is a valuable reagent for the detection and separation of com-pounds such as lead sulphate, stannic oxide, silver halides, calciumfluoride and fluorspar, and certain chromium compounds, whichhelp to form the " insolubles " of qualitative analysis. Reactionsare often distinctive and frequently more rapid and convenient than afusion, and some of the separations are quantitative. L. C. Hurd 53shows that rhenium concentrates with arsenic in the Prescott-Johnson system 54 of analysis. He points out that sublimationmethods for the detection of rhenium in minerals may fail, and re-commends opening up by a fusion when the mineral is insoluble inhydrochloric or nitric acid.He adds that rhenium is probably bestdetected 55 by the thiocyanate reaction after molybdenum has beenremoved as a xanthic acid complex soluble in chloroform.Antipyrine is chosen from a number of organic bases as the bestreagent for antimony, with which it yields a yellowish-orange pre-cipitate in the presence of potassium iodide; tin gives a white pre-cipitate, but here the reaction is less sensitive. The reaction isapplied after digestion of the arsenic, antimony, and tin sulphides ofGroup IIb with hydrochloric acid56For the detection of platinum in small amount in minerals, alloys,and the like, the alkaline solution is treated with potassium iodideand acetic acid and a reddish-brown or rose colour appears if Pt'**'is present.When precipitated from sodium tellurite by sulphurdioxide, tellurium separates platinum, gold, selenium, molybdenum,and mercury from other metals and so concentrates the platinum.57The formation of red compounds with 4-methyl-1 : 2- and 4-chloro-1 : 2-dithiolbenzene (" dithiol ") is used as a test for tin by R. E. D.Clark.58 These reagents, it is said, can be employed in the presenceof all other metals when the colour of the mercaptides which they mayform is not intense enough to mask the red colour due to tin, but50 A. B. Levin, 2. anal. Chern., 1936,105,328; M. B. Schtschigol and N. M.Doubinski, Ann. Chim. analyt., 1936, [iii], 18, 257; V. J. Petraschenj, 2.anal.Chem., 1936, 106, 330.61 E. ChirnoagB, ibid., 1936, 104, 356.62 I n d . Eng. Chem. (Anal.), 1936, 8, 63; E. R. Caley, J. Arner. Chem. SOC.,53 I n d . E q . Chem. (Anal.), 1936, 8, 11.54 R. K. McAlpine and B. A. Soule, '' Qualitative Chemical Analysis," 1933.66 See L. C. Hurd and B. J. Babler, fnd. Eng. Chem. (Anal.), 1936,8, 112, for66 J. A. Gautier, J . Pham. Chim., 1936, [viii], 23, 283.67 S. K. Hagen, Mikrochem., 1936, 20, 180.68 A d y 8 t , 1936, 61, 242.1932, 54, 4112.determination of rhenium.REP.-VOL. XXXIII. 450 ANALYTIOAL CHEMISTRY.the only metals likely to interfere me copper, bismuth, and nickel.Udortunately, like many reagents recommended for colour reactions,these thiolbenzenes are unstable and must be freshly prepred orstored in hydrogen.The test will undoubtedly be useful, but itseems that the ideal reagent for tin has yet to be f0und.~9The white precipitate which Sn"" in hydrochloric or sulphuricacid forms with nitrophenylarsinic acid on boiling provides anotherselective test for this element which can be applied to its detectionin alloys since the majority of metals likely to be present in such acase do not interfere. The material under investigation is firsttreated with concentrated nitric acid and the test made on the hydro-chloric acid solution of the metastannic acid thus produced.60In the absence of mercury, silver, and thallium, copper can bedetected (and determined) in the presence of relatively large amountsof bismuth, cadmium, lead, and zinc by means of the yellow ammineCu[Cr(CNS),(NH,),], which is precipjtated by the addition ofReinecke's salt to Cu" ions reduced to Cu' by K,SnCI4,2H2O in hydro-chloric acid solution,61 and since cupric ions give no precipitate,the same reagent serves for the sensitive detection (and determina-tion) of mercury, as Hg[Cr(CNS),(NH,),],, in the presence of manyother metals.62 After removal of copper by means of potassiumthiocyanate, cadmium can be detected by 2 : 7-diaminofluorene,which is preferred to hydrogen sulphide for this purpose.@In view of the utility of ammonium mercuric thiocyanate as aconfirmatory test for zinc, determinations of the solubilities of zincmercuric thiocyanate in alkali-salt solutipns are of interest, as is theconclusion that zinc should be in the form of nitrate when this testis applied.64According to H.Ditz and R. Helleb~and,~~ the sensitivity of theammonium thiocyanate-acetone reaction for cobalt is much reducedif accompanying iron is removed either by sodium carbonate or bythe formation of a complex fluoride. Removal by calcium carbon-ate, however, leaves the sensitivity unchanged and then 1.5 mg. ofcobalt per litre can be detected in the presence of no less than 15 g. ofiron. F. P. Dwyer 66 has also examined this reaction and finds that69 Ann. Repork?, 1935, 32,459; cf. also this Report, p. 463.60 B. Tougarinoff, Bull. SOC. chim. Be@, 1936, 45, 542.61 C. Mahr, 2. anorg. Chem., 1935, 225, 386.62 Idem, 2. a d . Chem., 1936,104, 241.63 E.L. NSO and F. Calvert, A&. ma. Q U ~ P ~ . , 1934,353,698; cf. &ISO A. w.Scott and E. G. Adam, J . Amer. Chm. SOC., 1935,57,2541.Ellso B. V. J. Cuvelier, 2. anal. Chem., 1935,102, 16.64 B. V. J. Cuvelier and F. Bosch, Natuwwetenach. Tijds., 1936, 18, 9; see6s 2. m r g . Chm., 1936, 225, 73; 888 dso idem, &id., 1934,2U, 97.e6 J . Proc. Austral. Cht?rn. Inat., 1936, 3, 239TREOBBLD : INORQAIYIU ANALYSIS. 451the addition of ammonium acetate and tartaric acid to preventinterference from iron also leads to a serious loss of sensitivity,and he prefers to add potassium ammonium fluoride or, better still,sodium ammonium hydrogen phosphate, for this purpose.I n view of the peculiar behaviour of precipitated nickel and cobaltsulphides towards mineral acids, which is utilised in so many schemesof analysis and for which a satisfactory explanation has yet to beestablished, it is of interest to note that, according to A.M. Middletonand A. M. Ward,67 the precipitates which are usually obtained inqualitative analysis are oxygenated and not the normal sulphides.A mixture of potassium ferrocyanide and Cu(NH3),S04,H,0 isstated to give a sky-blue precipitate with calcium ions,@ whilst newand more sensitive reagents described for magnesium 69 are p-nitro-benzenediazoamino-4-nitronaphthalene, p-nitrobenzenediazoamino-benzene, and 4-nitro-4'-amino-l: 1 '-azonaphthalene. It is claimedthat no other metal hydroxide, even those of beryllium or therare earths, gives a blue colour with the last reagent.Several papers have been concerned with the removal of phosphateions in qualitative analysis, and of these S.Ishimaru's contributions 70are by far the most comprehensive. He finds that (i) POL" can becompletely removed from a solution just acid to methyl-orange byaddition of ferric nitrate, and this is more convenient than the leadmethod which can be applied satisfactorily only after removal ofGroup IIIa and manganese, (ii) precipitation with BY' leads to lessocclusion and adsorption of other ions by the precipitate than theiron method, but is less satisfactory in the presence of a high ironconcentration, (iii) the zirconium method equals that of the ironin merit, (iv) precipitation with tin leads to loss of iron and chromium,but can be adopted after removal of the aluminium group and man-ganese, Reynoso's procedure 71 being the best, and (v) T.B. Smith'sformate method 72 is the most suitable of all the methods advancedfor elimination of the effects of phosphate ions based on the additionof excess of phosphate or oxalate. It is finally concluded that, in sofar as accuracy is concerned, the bismuth method, except in the caseof a high iron content, is the best of the many methods which havebeen examined in considerable detail in this series of investigations.L. J. Curtman and T. B. Greenslade 73 also find that with the tinand the ferric chloride method loss of cations is serious. Both these67 J., 1935, 1459.68 8. A. Celsi, Anal. Fam. Bwquh., 1934, 5, 85.69 F.P. Dwyer, J . Proc. Awrtmt. C h m . Imt., 1936, 3, 184, 224.70 Sci. Rep. TbhoXru, 1935,24,426, 439, 448, 461, 473.7 1 An%. Chim. Phy8., 1862, 84, 320.72 J., 1933, 253.73 J . Chm. Educ., 1936,18,238452 ANALYTICAL CHEMISTRY.and the zirconyl chloride method are all efficient in removing phos-phate, but they consider that the last is the most rapid, effective, andconvenient. C. N. Potschinok 74 eliminates this ion as aluminiumphosphate, and S. Augusti 75 uses lead acetate in acetic acid solutionto remove oxalate, fluoride, silicate, and silicofluoride ions as well.V. J. Petraschenj 76 has outlined a scheme for cations of the thirdand the fourth analytical group in the presence of phosphate, andW. Fischer, W. Dietz, K. Brunger, and H. Grieneisen ‘7 have in-vestigated the same subject in considerable detail, putting forwardnew schemes for these groups.The sensitivity of the phospho-molybdate reaction is said to be enhanced by the addition of a suit-ably-prepared glycerol-gelatin solution.78In order to detect very small percentages of non-metallic impuritiesin metals, the sample of metal is made the cathode in the electrolysisof dilute sulphuric acid or dilute sodium hydroxide plus potassiumcyanide. Phosphorus, arsenic, antimony, sulphur, selenium, andtellurium, combined or in solid solution, are reduced to their hydrides,which are identified by filter-paper impregnated with suitable re-agents ; 0.001 yo of phosphorus, for example, can thus be dete~ted.7~Small amounts of bromide in sodium chloride can be identified by amodification of the colour reaction with fuchsin,sO and bromates aredetected in the presence of potassium chlorate and bromide by agreenish-yellow colour which develops with fluorescein.In 4N-hydrochloric acid solution, bromates rapidly decolorise methyl-orange, and this forms the basis of a test in the presence of otheroxidising agents such as chlorates, iodates, nitrates, persulphates,dichromates, ferricyanides, and nitrites. The same reaction canalso be used to detect small amounts of bromate in a large excess ofchloride or bromide. 82Methods for dealing with insoluble ferricyanides and the detectionof the ferricyanide ion with leuco-malachite green or benzidine aredescribed by L. K ~ h l b e r g .~ ~ According to J. Plank,m freshly-prepared ceric sulphate plus potassium carbonate will detect 1 part ofhydrogen, peroxide in 160,000 parts of solution.74 J . Appl. Chem. Rumia, 1936, 9, 140.v 5 Ann. Chim. appl., 1935, 25, 448.76 Z. anal. Chem., 1936, 106, 241.A. Steigmann, Chem.-Ztg., 1936, 60, 129.79 K. W. Frohlich, Angew. Chem., 1935, 48, 624.8o R. C. Lbpez, Farm. moderna, 1935,. 46, 55; see also F. Feigl, “ Qualit-81 F. L. Hahn, Mikrochem., 1936, 20, 236.a2 I. M. Korenman, 2. anal. Chem., 1935, 103, 269.83 Ibid., 1936, 106, 30.84 Magyar Chem. Pol., 1934, 40, 105.7 7 Angew. Chem., 1936, 49, 719.ative Analyse mit Hilfe von Tiipfelreaktionen,” 1935, p. 278THEOBALD : INORGANIC ANALYSIS. 453Drop Remtions.-The output this year of papers dealing with thesetests, conveniently but unfortunately spoken of as “ spot ” tests, issomewhat less than in previous years, but the importance and inter-est of the subject make it desirable again 85 to sumrnarise the workwhich has been done.The tendency too readily to regard thesetests as specific still exists in some quarters, although the series ofcritical investigations which are being undertaken by certain workersmay help to correct this erroneous view.* For example, havingshown that the cacothelin test for tin is by no means specific,s6J. B. Ficklen, I. L. Newell, and N. R. Pike87 have turned theirattention to the cinchonine-potassium iodide reagent for bismuth,which likewise is not truly specific, and to p-nitrobenzeneazoresor-cinol for magnesium, which should be used only when ions of GroupsI, 11, and 111, and ammonium have been removed.88The use of drop reactions in the identification of substances solublewith difficulty in acids, such as the silver halides, insoluble sulphatesand fluorides, ignited oxides, silica, etc., is described by P.Feigl,*9and A. A. Benedetti-Pichler and W. F. Spikes have presented a schemefor the separation, identification, and estimation of mixtures ofthallium, Zead, and siZver using 0.5--P mg. of solid material.gO Thisis the first of a series of papers on qualitative separations on themicro-scale, and those who, like the Reporter, hold the view thatgroup separations are still essential for the analysis of any but thesimplest materials, and that the future of qualitative analysis liesin the judicious combination of these separations and drop reactions,will look forward with considerable interest to the contributions85 Cf.Ann. Reports, 1935, 32, 471.86 Cf. ibid., p. 472, ref: (36).87 8. anal. Chem., 1936, 104, 30.8 8 I. L. Newell, N. R. Pike, and J. B. Ficlrlen, 2. worg. Chem., 1935, 225,89 Nikrochem., 1936, 20, 198.90 Ibid., 1936,19, 239.* Part of the misapprehension which has arisen as t o the true selectivity oforganic reagents now employed in drop reactions may be due t o a confusionof terms. Feigl, who has done so much to advance this branch of analysis-and, presumably, German-speaking authors have followed him-uses theterm “ specific ” in a wider sense than is customary with English-speakingpeople, but he is careful both in his book ( o p .cit., p. 10) and in his latestarticle [ I n d . Eng. Chem. (Anal.), 1936, 8, 4011 t o differentiate between‘6 specific ” reagents and those he calls ‘‘ special ” reagents, few of which, as hepoints out, are known at the present time. ‘‘ Special ” reactions or “ Sonder-reaktionen” are, in the ideal case, limited solely to one substance and arenot subject to interference by the presence of any other, and it is these(‘ special ” reactions and reagents which we regard as “ specific,” since withus this term implies a property characteristic of and peculiar t o one substancealone.281454 ANALYTICAL CHEMISTRY.which are to follow, based, aa they are to be, on the thorough andexhaustive investigations of Noyes and fiis school.Q1A new test for siEver with what is probably ethyl 5-keto-2-thion-hexahydropyrimidine-4-carboxylate is said Q2 to be more selectivethan, and about as sensitive as, the rhodaqine base used by F.Feigl.93Lead can be detected in the presence of a, large excess of barium,copper, iron, manganese, nickel, etc., by applying the triple nitritetest to the ammonium acetate solution of the precipitated sulphste,"and tervalent thallium can be recogniaed by the intense green colourwhich it gives with leuco-o-nitrodiamant green.95 The action ofcupric salts on benzidine has been discussed, and a reagent consistingof o-tolidine (which gives tolidine-blue) and ammonium tbiocyanatein acetone is preferred Q6 as a more sensitive test.After conversionof the iron into [FeF6]***, it is claimed that O~OOOIS~o of copper iniron salts can be detected. F. Feigl and R. Uzel Q7 utilise as a dropreaction for copper the yellow to red c0lour,9~ due to tervalent copper,which is formed with potassium tellurate or periodate in alkalinesolution and in the presence of an oxidising agent such as potassiumpersulphate, A reversal of the test serves to detect either tellurium,for which drop reactions are all too few, or periodic acid which canthen be sought in the presence of other oxidising anions: Periodicacid, and tellurium in the presence of a very large excess of selenium,can both be identifiedQ' by their inhibiting action on the catalyticeffect of copper in the oxidation of manganese to manganate ionsby sodium hypobromite.Drop methods for bismuth have also beenadapted fkom well-known macro-reacti0ns.lInduced precipitation is well known as a device for collectingtraces of an element from a very dilute solution, and it is nowemployed for the detection of small quantities of titanium and zir-conium.2 Zirconium is added to the suspected titanium solution andprecipitated by means of arsenic acid ; coprecipitated titanium isconfirmed in the usual way with hydrogen peroxide. In effect, thesensitivity of the hydrogen peroxide test thus becomes muchincreased. The procedure detailed is particularly useful whenmuch iron, vanadium, etc., which interfere in the ordinary way91 See A. A. Noyee and W.C. Bray, " A System of Qualitative Analysis92 S. E. Sheppard and H. R. Brigham, J . Amer. Chem. SOC., 1936,68, 1046.Ss Z. w a l . Chem., 1928, 74, 380.94 I. M. Korenman and S . S . Messonshnik, Mikrochm., 1936,20, 189.95 L. Kuhlberg, ibid., 1936, 19, 183.95 Idem, aid., 1936, 20, 153.9* B. Brauner and B. Kuzma, Ber., 1907,40, 3362.for the Rare Elements," 1927.O7 Ibid., 1930, 19, 132.2 N. A. Tsnanrtev and A. V. Tananaeva, J . AppZ. Chm. Ruesia, 1936, 8,S F. Feigl and E. Rajmann, lkZikro&em., 1936,19, 60.1457THEOBdLD : INORQANIO ANALYSIS. 455with this test, are present. By a reversal of the process, smallamounts of zirconium can be gathered with titanium amenate andconfirmed with azoarsinic acid.In addition to providing a serviceable test for aluminium, morincan also be used for gallium, but unlike the case of alumhiurn, thefluorescence which it gives is not suppressed when sodium fluorideis added.3 In daylight the fluorescence of these two elements withmorin appears to be specific, but in ultra-violet light other elementsalso fluoresce.This fluorescence with morin has served to detectgallium in minerals such as zinc blende and arsenopyrite.8 Othercolour reactions for gallium, vix., a bright red lake with alizarin inpresence of ammonia and ammonium chloride, and a reddish-browncoloration or precipitate with potassium ferrocyanide, manganesechloride and potassium bromate, applicable when aluminium andindium are present, have also been described and applied to minerals.*Drop methods for cerium with leuco-malachite-green,5 berylliumwith alkannin and naphthazarinY6 cobalt with Na,[Fe(CN),NO] andpiperidine in acetic acid s~lution,~ and ammonium with Nessler's re-agent 8 have been worked out.The sensitivity of the benzidine-bluetest for dichrornate is increased by the addition of hydrogen peroxide,according to L. Kuhlberg: who adds that greater sensitivity isattained when o-tolidine replaces benzidine. The formation ofoctahedral crystals of a triple nitrite with praseodymium nitrate andsodium nitrite provides a sensitive test for cmium which is notaffected by the presence of potassium or rubidium.lOThe ferrous sulphate test for hydroxylamine has been put on amicro-basis by F. Feigl and R. Uze1,l1 the ammonia evolved beingidentified by its action on silver nitrate and manganese su1phate;Pi0.1 y of hydroxylamine, it is claimed, can thus be detected, and 0-5 yin the presence of a 3000-fold excess of hydrazine.A selection of the Behrens tests l3 considered to be the most satis-factory in the actual practice of determinative mineralogy," is de-8 G.Beck, Mikrochem., 1936, 20, 194.4 N. S. Poluektov, ibid., 1936, 19, 248.5 L. Kuhlberg, J . Appl. Chem. Russb, 1935, 8, 1452.6 J. Dubskf and E. Krametz, Mikrochern., 1936, 20, 57.7 F. Feigl and R. Uzel, ibid., 1936, 10, 132.8 N. A. Tananam and A. A. Budkevitsch, J . Appl. Chem. RzWreia, 1936,0,362.Mikrochem., 1936, 20,244.10 H. C. Goswami and P. B. Sarhr, J . Indian Chem. Soc., 1935,12, 608.11 Mikrochem., 1936,19, 132.18 F.Feigl, op. cit., p. 271. 18 See Ann. Reports, 1936,32,474. * For the dehtion in rocks and minerals of molybdenum, lead, and cobaltby means of their colour reactions with calcium xanthate, benzidine, andaJkali thiocyanates, respectively, 8ee H. Leitmeier and F. Feigl, Tach. Nin.Par. ilia., 1936,47, 313456 ANALYTICAL CECEMISTRY.scribed and illustrated byL. W. Staples,l* who has also given details 15of a microchemical test for siZicon.16 This depends on the formationof sodium silicofluoride from silicon tetrafluoride and its recognitionunder the microscope, and is claimed to be better than the meta-phosphate bead, the rubidium silicornolybdate, or the benzidinetest.Electro-capillary methods of drop analysis have also been dis-cussed by various authors l7 during the past year.In the drop reactions mentioned in this section, the limitingamount of an element or ion which can be detected is of the order of10-6 g.or less, but attention must always be paid to the possiblepresence of other constituents which frequently necessitate a modifiedprocedure with a resultant loss of sensitivity.I;. S. T.QUANTITATIVE COLORIMETRIC ANALYSIS.During the past few years publications relating to colorimetricmethods of analysis have been so numerous that for the presentReport it has been possible to refer only to a selection which con-veniently illustrates the progress that has been made.The essentials of colorimetric analysis are : (1) preparation of asolution of a suitable coloured derivative, (2) evaluation of thissolution by measurement of its light absorptive power.It is desir-able to consider (2) in some detail, since it is the factor common to allcolorimetric analyses.‘ ‘ Colour ” Measwement or Comparison.-Measurements areessentially relative, ultimately in terms of similar solutions of thesame substance in known concentration. For convenience, artificialstandards are sometimes used, e.g., Lovibond glasses, liquids such asferric chloride solutions,l aqueous picric acid, and aqueous potassiumchromate? The disadvantage of this method is that the spectralabsorption curves of the test solutions and the sub-standard glassesor solutions may be far from identi~al.~ Errors due to this cause are14 Amer.Min., 1936, 21, 613.16 Ibid., p. 379.16 See W. R. Schoeller and E. F. Waterhouse, Analyst, 1936, 61, 454, con-17 S. I. Dijatschkovski, J. Gen. Chem. Russia, 1935, 5, 728; A. F. Orlenko,1 M. G. Mellon and C. T. Kasline, Ind. Eng. Chem. (Anal.), 1935, 7, 187;2 R. Strohecker, R. Vaubel, and K. Breitwieser, 2. anal. Chern., 1935,3 Cf. J. P. Mehlig and M. G. Mellon, J . Physical Chem., 1931, 35, 3397;cerning the inadequacy of certain tests for tungsten.{bid., 1936, 5, 1091 ; H. Fritz, Mikrochem., 1935, 19, 6.cf. also ibid., 1936, 8, 463.103, 1.A. L. Bacharach and E. Lester Smith, AnaZyst, 1934,59, 70STRAFFORD : QUANTITATNE COLORIMETRIC ANALYSIS. 457greatly magnified in the case of observers suffering from partialcolour-blindne~s.~ Better results may be anticipated in the rarecases where it is possible to prepare a sub-standard which hasabsorption curves practically identical with those of the test solu-tion.5Instruments.In the simplest form of measurement using comparison tubes of theNessler-cylinder type, the probable error is rarely less than & 3%,and in some cases may be as high as & 8%. The accuracy obtain-able with colorimebers of the Duboscq type is not much greater.6It appears to be now generally recognised, however, that in aninstrument using white light, Beer’s law cannot apply rigidly, withthe result that empirical correction curves are necessary.’ More-over, owing to the diluting effect of the white light, the sensitivityis less than that of spectrophotometric methods.The use of suitablecoloured filters appreciably increases the accuracy of the Duboscqtype of instrument .8Measurement of Absorption Density .-In modern instruments, therelative absorption density of the coloured solution, for light of wave-length approximating to that for which the solution shows a maxi-mum selective absorption, is measured. The obvious advantagesare (1) maximum sensitivity, and (2) a rectilinear (logarithmic)calibration curve, in accordance with Beer’s law. A further impor-tant advantage is that, once a calibration curve has been constructed,there is no necessity to prepare the colour standards each time ananalysis is made.Undoubtedly the best method of measuring the absorption densityis by means of the spectrophotometer, whereby measurements maybe made over a very narrow range (ca. 50 A.) a t any desired wave-length in the visible spectrum.By measurement at two suitablewave-lengths, it is often possible to determine two coloured substancesin admixture.* By use of the quartz spectrograph, measurement canbe extended to the ultra-violet, opening up an important field, e.g.,the determination of vitamin A by measurement of the absorptiondensity at 3280A.Possibly from considerations of cost, references to the employmentF. Twyman and G. F. Lothian, Proc. Physical SOC., 1933, 45, 643.H. W. Swank and M. G. Mellon, Ind. Eng. Chem. (Anal.), 1934, 6, 348;A. Thiel, Ber., 1935, 68, 1015; 2. anal. Chem., 1936, 106, 281.W. D. McFarlane, ibid., 1936, 8, 124.7 J. H.Yoe, “ Photometric Chemical Analysis,” Vol. 1, Colorimetry; cf.A. P. Mussakin, 2. anal. G‘lzem., 1936, 105, 351.8 W. D. Armstrong, Ind. Eng. C’hem. (Anal.), 1933,5,300; A. Thiel, loc. cit.,ref. (6); R. J. Robinson and H. E. Wirth, Ind. Eng. Chem. (Anal.), 1935, 7,147458 ANALYTICAL CHEXISTRY.of the spectrophotometer for colorimetric chemical analysis axerelatively few, but it has undoubtedly proved of value in initialresearch on individual colorimetric methods for the determinationof the full spectral absorption curve of the coloured solution. Thisis exemplified in the studies of H. W. Swank and M. G. Mellon 5 oncolorimetric standards for silica, and in the determination of17ibmin-D.~In the Pulfrich step photometer, colour filters are employed;consequently, the absorption densities determined are for light ofcomparatively broad wave-length range, and have relative ratherthan absolute significance. As with the spectrophotometer, theinstrument enables a permanent calibration curve to be constructed,As an alternative to the mechanical (variable shutter) photometricdevice of the Pulfrich, use is made of a neutral grey solution of vari-able thickness for reducing the intensity of the direct beam in the6c absolute ” colorimeter.l*Photo-electric Instruments.-These should be described as spectro-photometers or absorptiometers, rather than colorimeters. Thephoto-electric cells, which may be of the photo-emission or the semi-conducting type, can be used (1) as a null-point instrument to replacethe eye in a spectrophotometer with polarising photometer 11 orDuboscq colorimeter,12 (2) to afford a direct measure of light in-tensity and hence of absorption density.Instruments in thesecond class may be divided into two groups depending upon thenumber of photo-cells employed.The solution to be measured is interposedbetween the cell and the source of light, and the absorption of lightby the solution is measured directly by determining the currentoutput of the photo-electric cell in relation to the value obtainedwith the pure s01vent.l~ In instruments of this type it is of para-mount importance to use a light source of constant intensity; thismay be realised by incorporating a Barretter (current-regulating)lamp in a circuit buffered ” by an acc~mulator.~~ In the case of9 H.BrockmRnn and Y. H. Chen, 2. p h y h l . ChMn., 1936,241,129.10 A. Thiel, 2. amZ. Chm., 1933, 94, 170; A. Thiel and W. Thiel, Chern.p&., 1932, 409; A. Thiel, ibid., 1934, 7, 383.11 M. G. Mellon and C. T. Kasline, Zoc. &t., ref. (1); A. G. Winn, Tmns.Fcaraday Soc., 1933, 29, 689.1% G. Bernheim and G. Revillon, Ann. Faleiif., 1936, 29, 6; cf. &O A.G o u d d t and W. H. Summerson, J . Biol. Chem., 1935,111, 421; E. W. H.Selwyn, J . Sci. Inetr., 1933, 10,116.I* J. H. Yoe and T. B. Crumpler, Ind. Eng. Chem. (AmZ.), 1935, 7, 281;N. StraiTord, Analyst, 1936, 61, 170; R. S. W. Thorne and L. R. Bishop,J . In&. Brew., 1936, 42, 15; L. E. Howlett, Canadian J. Res., 1936, 14, A ,38; R.A. O h m , J . A880C. Off. AQ&. Chm., 1934,17, 136.(a) One-cell type.14 N. Stra$ord, Zoc. cit., ref. (13)STRAFFORD : QUANTITATIVE OOLORIMETRIC ANALYSIS. 459cells showing a fatigue ” effect it is necessary to allow the photo-cellto attain its equilibrium current after each change of light intensity.15(b) Two-cell type. Two photo-electric cells illuminated by thesame source of light are balanced against each other through a,galvanometer. The test solution is placed before one cell, the puresolvent before the other, and the current output difference measureddirectly.ls Variations due to small fluctuations of the light sourceare automatically cancelled in an instrument of this type.Photo-emission cells with different characteristic response curvesare available, permitting the choice of one having maximum responseto the coloured light under measurement.17&loured light in most instruments is obtained by atering whitelight by colour filters, as in the Pulfrich photometer.More nearlymonochromatic light is obtained by filtering the light from metal-vapour discharge lamps, e.g., mercury or sodium.lsWhen a monochromator of the spectroscope type is used, the lightis of such relatively low intensity that valve amplification of thephoto-electric current is necessary.Since most colour filters are comparatively transparent to infra-redrays, to which the photo-electric cells are responsive, an additionalfilter must be used (“ minus-infra-red ”) in order to obtain maximumabsorption-density readings.lQ Weston Electrical Instrument Co.provide a “ Viscor ” filter which fulfils this purpose.20Accuracy of Photometric Colorimetric Analysis.-Even with thebest instruments, careful choice of operating conditions is necessaryto ensure the highest accuracy.According to I?. Twyman andG. F. Lothian,21 the percentage error is a t a minimum at an absorp-tion density of between 1.5 and 2.0 in the case of a visual spectro-photometer, and of 0.43 in objective (photo-electric) instruments.Theseauthors also consider that, although theoretically photo-electricmethods give much greater sensitivity of discrimination than the eye,yet visual methods are more trustworthy as far as absolute measure-ments are concerned.In the Reporter’s opinion, a simple and16 J. H. Yoe and T. B. Crumpler, loc. cit., ref. (13); N. Strafford, Zoo. cit.,ref. (13).16 C. Zinzadze, Id. Eng. Chem. (Awl.), 1935, 7 , 280; R. B. Withrow,C. L. Shrewsbury, and H. R. Kraybill, ibid., 1936, 8, 214; B. Lange, Chem.Fabr., 1934, 7 , 45.17 R. B. Withrow, C. L. Shrewsbury, and H. R. Kraybill, loc. cit., ref. (16).18 R. Sewig and F. Miiller, Chem. Fabr., 1934, 7, 26; H. Alterthum andM. Reger, ibid., 1933, 6,283.19 E. R. Bolton and K. A. Williams, Analyst, 1936, 60,447; N. Strafford,too. cit., ref. (13); cf. also R. Fonteyne and P, de Smet, Mikrochern., 1933,13,289.a0 J . Sci. Imtr., 1936, 13, 338.21 LOC. cit., ref. (4)460 ANALYTICAL CHEMISTRY.relatively inexpensive photo-electric instrument,22 properly used, isas accurate as a spectrophotometer for the relative measurementsrequired by colorimetric analysis.Nephelometry .There is no sharp division between colorimetry and nephelo-metry.Some of the organometallic derivatives, such as copperdiethyldithiocarbamate in aqueous media, and sulphides such aslead sulphide may be described as being in colloidal “ solution’’in colorimetric tests. Particular care in preparation of the standardand test solutions is necessary in order to obtain reproducible resultsand compliance with Beer’s law. The use of protective colloids isof value in this connection.23Photo-electric measurements have shown that transmission oflight by two similarly prepared colloidal lead sulphide ‘‘ solutions ”apparently equal in shade to the eye may be appreciably different.=For true nephelometric determinations, e.g., of zinc as ferrocyanide,the photo-electric absorptiometer affords an appreciably greateraccuracy than visual c~mparison.~~ K.W. Franke, R. Burris,and R. S. Hutton 26 describe a novel procedure by which colouredprecipitates of colloidal fineness, e.g., selenium, are filtered on toa mat of barium sulphate. Permanent colour standards are thusprepared.Colorimetric Determination of the Elements.Although so far almost exclusively used for determination ofminor amounts of an element, in the Reporter’s opinion there areadequate reasons (mainly saving of time) why much wider useshould be made of colorimetric methods for determination ofelements present as major constituents.* At the same time, theopinion is recorded that, for micro-analysis, colorimetric methodsare at least as accurate as, and usually simpler and more convenientthan, alternative methods ; in many cases, moreover, alternativemethods of adequate sensitivity do not exist.Cf. N.Strafford, Zoc. cit., ref. (13).83 L. de Brouckbre and S. Solowiejczyk, Bull. SOC. chim. Belg., 1934, 43,597; C. Zinzadze, Ind. Eng. Chem. (Anal.), 1935, 7 , 227.e4 Second Report of the Sub-Committee on the Determination of Arsenic,Lead and other Poisonous Metals in Food Colouring Materials to the AnalyticalMethods Committee of the Society of Public Analysts : 11, The Determinationof Lead; Analyst, 1935, 60, 541.26 N. Strafford, loc, cit., ref. (13).26 I d .Eng. Chern. (Anal.), 1936, 8, 435.* As an example, J. P. Mehlig [Id. Eng. Chem. (Anal.), 1935,7,387] claimsthat the copper content of ores (up to 22%) may be determined with a photo-electric spectrophotometer to an accuracy of * 0.1%STRAFFORD : QUANTITATIVE COLORIMETRIC ANALYSIS. 461The first task of the analyst is to devise conditions ensuring speci-ficity, and in spite of the availability of numerous organic reagents,cases in technical analysis where a preliminary separation is notnecessary are the exception rather than the rule. The most impor-tant advance in recent years is in the application of organic reagentsin organic solvents €or separation of interfering elements, as well asfor the formation of a coloured derivative.The following will serveas examples of modern technique in the inorganic field.Lead.-A general method for the determination of lead in the pre-sence of other metals involves isolation of the lead, first as sulphideand then as sulphate, followed by the application of the colloidalsulphide method, the limitations of which are disc~ssed.~4The use of dithizone is rccommended both for the separation of leadfrom most other metalsY2' and from bismuth,28 a.nd also for theactual colorimetric measurement. The latter may be applied tothe red lead dithizone or to the regenerated greendithizone .30P. A. Clifford and H. J. Wichmann 31 show, from considerationsof distribution between aqueous and organic media, that both thesemethods €ail to give complete recovery of the lead, and they recom-mend a " mixed colour " photometric method in which completeextraction of lead is ensured by use of an excess of dithizone appliedto a solution of optimum pE value (9.0).They also criticise themethods €or removal of bismuth referred to above, and suggestmethods €or preventing interference from tin.It is of particular importance to employ a purified reagent freefrom oxidation products,30 and to avoid the presence of oxidisingreagents .32In certain cases, the extraction of lead from aqueous solutionscontaining insoluble matter by means of a solution of dithizone is notcomplete owing to occlusion of lead by the insoluble matter.24Dithizone is, however, used for extracting insoluble lead compoundsfrom spray residues.3327 N.L. Allport and G. H. Skrirnshire, Analyst, 1932, 57, 440; P h a m . J.,1932,129, 248; G. Rocho Lynch, R. H. Slater, and T. G. Osler, Analyst, 1934,59, 787.28 S. L. Tompsett, ibid., 1936, 61, 591; C. E. Willoughby, E. S. Wilkins,jun., and E. 0. Kraemer, Ind. Eng. Chem. (Anal.), 1935, 7, 285.29 J. R. Ross and C. C. Lucas, J . Biol. Chem., 1935,111,285; 0. B. Winter,H. M. Robinson, F. W. Lamb, and E. J. Miller, Ind. Eng. Chem. (Anal.), 1935,7, 265.30 H. Fischer and G. Leopoldi, Angew. Chem., 1934, 47, 90; F. Morton,Analyst, 1936, 61, 465.31 J . A880c. 08. Agric. Chem., 1936,19, 130.32 S. L. Tompaett, Zoc. cit., ref. (28).33 W. E. White, Ind. Eng. Chem. (And.), 1936, 8, 231462 AXALYTICAL CHEMIS!L%Y.Hercury.-For the determination of mercury, N.Strafford andP. F. Wyatt 34 recommend the reaction with p-dimethylamino-benzylidenerhodanine, which is less subject to interference than thatwith diphenylcarbazide or diphenylcarbazone. Preliminary separa-tion of the mercury from most other metals is essential. H. Fischerand G. Leopoldi 35 and W. 0. Winbler 36 separate and determinemercury as the dithizone complex.Tin.-Since previous methods for tin depend merely on the reduc-ing action of stannous salts, the appearance of a reagent which formsa coloured derivative with tin is welc0me.3~ It must be noted,however, from the author’s own findings, that the reaction is notspecific for tin (see p. 449), a number of other metals, particularlycopper, bismuth, and nickel, giving coloured derivatives.Zinc.-W. Deckert 38 determines zinc, after removal of copperas sulphide, as the coloured complex obtained with dithizonein alkaline solution.Dithizone is also recommended by H. E’ischerand G. Le~poldi.~~ A novel method for zinc consists in precipitatingthe metal, again after a preliminary separation and removal ofcopper as sulphide, as its complex with 8-hydroxyquinoline.The absorption density of the 8-hydroxyquinoline is then de-termined, after decomposition of the complex, in ultra-violetlight .40Caper.-Numerous papers continue to appear, most of themdescribing variations of well-known methods. C. A. Goethals 41discusses the sensitivity of several colour reactions. General opinionseems to favour the dithiocarbamate methodF2 although it is recog-nised that preliminary separation of other metals is often necessary.&G. Bertrand and L.de Saint Rat 44 recommend urobilin as a selectivereagent. H. Fischer and G. Leopoldi45 determine copper as thedithizone complex in an organic solvent. E. Stolze46 points outa4 Analyst, 1936, 61, 628.a6 J . Aesoc, Off. AgTic. Chena., 1935, 18, 638.87 R. E. D. Clark, Am?y8t, 1936,61,242.38 2. anal. Chem., 1935, 100, 305.39 Ibid., 1936, 107, 241.40 J. D9browski and L. Marchlewski, Biochem. Z., 1935,282,387.4 1 2. anal. Citem., 1936,104,170.** T. Callan and J. A. R. Henderson, Andgat, 1929,54,650.2. anal. Chem., 1935,103,241.L. W. Cow, A. H. Johnson, H. A. Trebler, and V. Karpenko, Ind.Eng.Chem. (Anal.), 1935, 7, 15; N. D. Sylvester and L. H. Lampitt, Analyst* 1935,60, 376; E. Lasausse and L. Frocain, J . PJmm. China., 1936, [viii], 23, 77;B. Eisler, K. G. Rosdahl, and H. Theorell, Biochem. Z., 1936, 285, 76.44 Compt. rend., 1936, 2U3, 140.4O Alzgew. Chem., 1934,47, 90.40 Bodenk. Pflamenernlihr., 1936,l, 115STBAFFORD : Q U B N T I T A ~ COLORIBIETRIC ANALYSIS. 463that it is necessary to reduce any ferric mlts before applying thismethod. N, Straf€ord4‘ suggests the use of salicylaldoxime as asensitive and specific nephelometric reagent.Iron.-Attention continues to be paid to improvement of thethiocyanate 48 and the sulphosalicylic acid 49 method. [Incident-ally, calcium (as oxalate) may be determined by its bleaching actionon the iron-sulphosalicylic acid c01our.l~ ad-Dipyridyl is recom-mended as a colorimetric reagent.61Aluminium.-Calcium salts and other interfering substancescause fictitiously high values in determinations with ali~arin.5~Interference from iron is prevented by its removal as thiocyanate inamylGermanium.-N.S. Poluektof describes the determinationof small quantities of germanium, the method depending onthe blue colour formed by the reduction of germanomolybdicacid.Titanizm.-Titanium is determined by hydrogen peroxide 55 or bysalicylic acid.56AZkaZi Metals.-Existing methods for the colorimetric determin-ation of potassium, involving the precipitation as cobaltinitrite, arereviewed, and a modified procedure is described by R. J. Robinsonand G.L. P~tnarn.~’Sodium is determined as an aqueous yellow solution of its tripleacetate with uranium and magnesium.58PhoslpF,orus and XiZicon.-Considerable attention has been given47 ‘ ‘ m e Detection and Determination of Small Amounts of InorganicSubstances by Colorimetric Methods,” Institute of Chemistry Publication,1933, p. 27; cf. also F. Alten, B. Wandrowsky, and E. Knippenberg, MiEm-chem., 1936, 20, 77.48 K. Steinhauser and H. Ginsberg, 2. anal. Chem., 1936, 104, 385; G. E.Farrer, jun., J . Biol. Ohm., 1935,110,685.40 F. Alten, H. Weiland, and E. )Iille, 2. w r g . Chem., 1933, 215, 81; A.!Chiel and 0. Peter, 2. anat!. Chm., 1935, 103, 161.60 L. Jendrassik and F. Takhs, Bwchem. Z., 1934, 274, 200.61 G. Bode, Woch. Brau., 1933,50, 321; L.L. Engel, J . Dent. Rea., 1934,14,273 : W. D. McFarlane, Zoc. cit., ref. ( 5 ) ; F. B. Shorland and E. N. Wall,Biochem. J., 1936,30,1049; H. I. Coombs, ibid., p. 1688.62 D. F. Eveleth and V. V. Myers, J . Biol. Chm., 1936, 113, 449.68 A. P. Mussakin, J . A@. Chem. R u s k , 1936,9, 1340.54 Z . m l . Chem., 1936,105,23.55 H. Ginsberg, 2. amrg. Chern., 1933, 211, 401; 1935, 226, 67; G. P.Lutschinski and A. I. Lichrttscheva, 2. a d . Chm., 1936, 103, 196; H. A.Liebhafsky, ibid., 105,113; L. Maillard and J. Ettori, C m p t . rend., 1936, $302,594.56 M. Schenk, Hdv. Chh. A&, 1936,19,1127.57 Ind. Eng. Chm. (Anal.), 1936,8,211.68 A. EUas, A d . Aaoc. Qub. A r g d h z , 1936,23,1464 ANALYTICAL ClIEMISTRY.to the important determinations of phosphorus 59 and of silicon 60by the molybdenum-blue reaction.Determinution of pH Value.According to E.A. Guggenheim and T. D. Schindler,61 resultsof colorimetric pa determinations by means of a Gillespie com-parator are reproducible within 1%. M. Kilpatrick, E. F. Chase,and L. C. Riesch 62 maintain that the experimental error of thecolorimeter method is about 5%, and of the electrometric methodabout 2.5%. A. G. de Almeida 63 plots dominant hue, expressedas a wave-length, against pH for a series of indicators, and claimsthat it is possible to determine the pH of an unknown solution moreaccurately than by electrometric methods.Textile assistants such as sulphonated oils, sulphated and sul-phonated fatty alcohols, sodium alkylnaphthalenesulphonates, etc.,exert a specific effect on most indicators which is frequently a sourceof considerable error, even up to 1.0 unit, in the colorimetric deter-mination of pH.64 This is discussed elsewhere in these Reports(p. 110).Photometric Titrations.Photometric methods may be employed for the accurate objectivedetermination of the end-point in a colorimetric or turbidimetrictitration. E.T. Bartholomew and E. C. Raby 65 determine alka'licyanides by titration with silver nitrate to the incipient developmentof turbidity, detected by a system of balanced photo-electric cells.S. Hirano 66 titrates mercuric chloride or nitrate with 0.01 or 0-001N-sodium sulphide in presence of gum-arabic as protective colloid, theend-point being indicated when the turbidity reaches a maximumvalue.In a similar manner he determines silver by titration withsodium chloride, in presence of starch.67 He also describes methodsfor determining soluble sulphides,68 gold,69 and halides.'O69 K. C. Scheel, 2. mal. Chem., 1936,105,256; R. Amrnon and K. Hinsberg,Z. phy8si0l. Chem., 1936,239,207; C . Zinzadze, Ind. E n g . Chem. (Anal.), 1935,7, 227; K. Boratyiiski, 2. awl. Chem., 1935, 102,421; H. Etienne, Bull. SOC.chirn. Belg., 1936,45,516; H . L. Brose and E. B. Jones, Nature, 1936,138,644.60 F. De Eds and C. W. Eddy, J . Biol. Chem., 1936, 114, 667; R. Stro-hecker, R. Vaubel, and R. Breitwieser, 2. anal. Chern., 1935, 103, 1.6r J . Physical Chem., 1934, 38, 543.64 J. E. Smith and H.L. Jones, J . Physical Chern., 1934, 38, 243; H. L.66 I n d . Emg. Chem. (Anal.), 1935, 7, 63.66 J . SOC. Chem. I n d . Japan, 1935, 38, 646.67 Ibid., 1934, 37, 75413.gg Ibid., 1935, 38, 5 9 8 ~ .6D Ibid., 1934, 37, 1 7 8 ~ , 5 6 1 ~ .J . Arner. Chem. SOC., 1934, 56, 2051. 63 Diss., Lisbon, 1935.Jones and J. E. Smith, Amer. Dyleatu, Rep., 1934, 23, 423.'O Ibid., p. 1 7 7 ~ ; 1935, 38, 175WEST : ORGANIC ANALYSIS. 465F. Muller reviews the use of photo-electric cells in automatictitrati~ns.'~CoZorirnetric Oxidation-reduction Reactions.-W. P. Lundgren 72describes a modified Duboscq colorimeter used for following fadingcurves in oxidation-reduction reactions.N. S.ORGANIC ANALYSIS.Elements.-Difficulties experienced in the determination ofsulphur, oxygen, and nitrogen by the ter Meulen method have nowbeen overcome,2 and it is claimed that satisfactory results for theestimation of nitrogen in compounds of high halogen content areobtained if soda-lime is placed both in front of and behind the cata-l y ~ t .~ Ultimate analysis of an organic substance may be based onpressure measurements after combustion in a bomb calorimeter .*Sulphur may also be satisfactorily determined by the bomb methodY5and chlorine may be determined accurately by the rapid lampmethod.6 Convenient wet oxidation methods have been describedfor the determination of (a) sulphur and copper,' and (b) carbon andnitrogen. 8Qualitative.-When treated with nascent chlorine, i.e., potassiumpermanganate in dilute hydrochloric acid, ketoximes give a bluish-green coloration, whilst aldoximes yield a solution which gives ared colour with ferric chloride : certain aromatic aldoximes do notgive this reaction. The reaction with hydroxylamine hydro-chloride has been the basis of a somewhat elaborate scheme for therecognition of aldehydes, ketones, acids, acid chlorides, anhydrides,amides, esters, and lactones.1°The following reagents have been investigated for the identific-ation of aldehydes and ketones : p-bromo-,ll m-chloro-,12 and m-7 1 2.Elektmchem., 1934, 40, 46.73 Science, 1934, 80, 209.1 5. Gauthier, Bull. Soc. chim., 1935, [v], 2, 506.3 If. ter Meulen and H. J. Ravenswaay, Chem. Weekblad, 1936, 33, 248.4 P. J. Merkus and A. H. White, Proc. Amer.Gas ASSOC., 1934, 991.5 H. C, Chiang and C. L. Tseng, Sci. Kep. Nat. Univ. Peking, 1936,1, 19.6 W. N. Malisoff, Ind. Eng. Chem. (Anal.), 1935, 7 , 428.7 N. N. Melnikov, J. Gen. Chem. Russia, 1935, 5, 839.8 C. N. Acharya, Biochem. J., 1936,30, 241.9 E. Graf, Anal. $'is. Quh., 1936, 34, 91.10 Idem, ibid., p. 95.11 8. M. Wang, Cheng-Heng Kao, Chung-Hsi Kao, and P. P. T. Sah, Sci.19 P. P. T. Sah and C. S. Wu, ibid., 1936, 3, 443.H. ter Meulen, ibid., p. 1692.Rep. Nat. Tsing Huct Univ., 1935, A , 3, 279466 ANALYTICAL CHEMISTRY.bromo- benzhydrazide,lS m- tolylhydrazine,14 which in conjunctionwith the o- and p-isomerides is suggested as a reagent for the ident-ification of sugars, l5 p-naphthoylhydrazide,l6 and m-tolylsemi-carbazide.l' It has been shown that the melting points of manyof the recorded 2 : 4-dinitrophenylhydrazones are incorrect ; the useof isopropyl instead of ethyl alcohol in Brady's method gives purercompounds.lS3-Nitrobenzazid<lg p-bromobenzazide,20 and m-nitrobenzoylthio-carbimide 21 have been investigated as reagents for the identificationof amines , and benzylamine and a-phenylethylaminc salts 22 andl-chloro-2 : 4-dinitrobenzene 23 are recommended for use in theidentification of acids and phenols respectively.Japanese acid clay is a powerful adsorbent for diamino-acidsobtained in the hydrolysis of proteins, arginine being adsorbed to theextent of 86*8% compared with 8.45% for g l y ~ i n e .~ ~ A process isdescribed for the removal of the adsorbed amino-acids and for thcregeneration of the clay.25It has been found that most ethers form peroxides on storage andtend t o explode on dry distillation, and the use of certain reagentsas peroxide destroyers and oxidation inhibitors has been investi-gated.26 Phenolic ethers may be identified through their picrates,although some of the compounds are unstable in air.27A simplified copper solution for use in sugar analysis is made up of100 C.C.of %.970 sodium hydroxide solution, 1 g. of copper sulphate,and 10 C.C. of water.a8Quulztitative.-Accurai;e determination of .the hydroxyl groupsmay be rapidly effected by heating the substance with pyridine-13 Chung-Hsi Kao, T. Tao, Cheng-Heng Kao, and P. P. T. Sah, J . Chinese14 P. P. T. Sah and C.Z. Tseu, SCi. Rep. Nat. Tsing Hua Univ., 1936, A,X 5 Idem, &bid., p. 409.16 H. Chen and P. P. T. Sah, J . Chinese Chem. SOC., 1936,4,62.1' P. P. T. Sah, S. M. Wang, and C. H. Kao, ibid., p. 187.19 K. Meng and P. P. T . Sah, J . Chinese Chem. Soc., 1936, 4, 75.2* P. P. T. Sah, C. H. Kao, and S . M. Wang, ibid., p. 193.81 W. L. Tung, Cheng-Heng Kao, Chung-Hsi Kao, and P. P. T. Sah, Sci.c'hem. SOC., 1936, 4, 69.3,403.N. R. Campbell, Analyst, 1936, 61, 391.Rep. Nat. T8iny H m Univ., 1935, A, 3, 285.2181.C. A. Buehler, L. Carson, and R. Edds, J . Amer. Chem. SOC., 1935, 57,2s R. W. Bost and F. Nicholson, ibid., p. 2368.z4 M. Mashino and N. Shikazono, J . SOC. C'hern. I n d . Japan, 1936, 39, 5413,25 Idem, ibid., p. 1 3 6 ~ . 26 E. C.Williams, C h m . and I&., 1936,580.27 0. L. Baril and G. A, Megrdichian, J . Amer. Chem. SOC., 1930, 68, 1416.88B.E. J. Mueller, J . Pham. Cltim., 1936, [V;ii], 24, 18WEST : ORGANIC ANALYSIS. 487acetic anhydride, decomposing the product with water, and titratingthe excess acetic The method is not valid with tert.-alcoholsand gives low results with sugar alcohols. Small quantities of ethylalcohol may be determined by an oxidation method.30Reducing sugars may be determined by direct potassium ferri-cyanide titrationF1 or by back titration with ceric sulphate.32Quantitative differentiation of fructose and mannose in glucose-fructose-mannose mixtures is possible by the use of micro-organ-isrnsg3 and sucrose inversions may be followed by means of the glassbut not the quinhydrone ele~trode.~* Accuracy is claimed formethods of determining the acetyl 35 and the nitrate 86 group incarbohydrate derivatives.Errors in the titration of the amino-groups in amino-acids, etc., areminimised if glacial acetic acid is used as the solvent.37Convenient methods are described for the determination of :toluidines in aqueous solution:* xylidine isomerides,3Q camphor bythe titration of the hydrogen chloride liberated from hydroxylaminehydrochloride after oxime formationto phenol and cresols in thepresence of each other,4l and pyridine in the presence of nicotine.42The methods available for the titration of alkaloids in anhydrousmedia, have been critically reviewed.&Colorimetry .-Carbamide, urethanes, and carbazides give a cerisecoloration on heating with vanillin in concentrated hydrochloric acidafter pre-heating with phenylhydrazine hydrochloride.44 Colourreactions are described for certain o-nitro-compounds,45 cyclopentadi-ene,46 and tartaric, citric, and aconitic acidsP7a9 M.Freed and A. M. Wynne, Ind. Ertg. Chem. ( A d . ) , 1936,8, 278.30 T. E. Friedemann and R. Klass, J . BioZ. Chem., 1936,115, 47.31 G. I. Solomos, BuW. SOC. Chim. bioZ., 1935,17, 1465.32 W. Z. Hassid, Id. Eng. Chm. (And.), 1936,8, 138.33 T. F. Nicholson, Biochem. J., 1936,30,1804.34 H. P. Cady and J. D. Ingle, J . Phy8icaZ Chm., 1936, 40, 837.35 A. Friedrich and H. Sternberg, Biochm. Z., 1936, 286, 20.36 J. Dewar and G. W. Brough, J. SOC. Chem. Ind., 1936,55,207.37 L. J. Harris, Biochem. J., 1935,29,2820; G. F. Nadeau and L. E. Bran-38 D. Cismaru, BuZ. SOC. Chim. Rorrz&nia, 1934, 16, A , 37.39 B. P. Fedorov and A. A. Spriskov, 2. anal. Chem., 1936,105,412.40 R. Vandoni and G. Desseigne, Bull. SOC. chim., 1935, [v], 2, 1686.41 V. Mikhschevskaja, Chim. Tverd. Tog., 1934,5, 553.42 R. L. Fratkin, L. P. Juravleva, and A. G. Blankachtein, Sborn. Robot43 G. N. Thomis, J . Pham. Chim., 1936, [vs], 24, 162.4 1 J. A. Sanchez, Ann. Chim. curacclyt., 1936, [c], 18, 66.45 P. K. Bose and S. Ram, J. Indian Chm. Soc., 1935,12,687.46 B. N. Afanasiev, Ind. Eng. Chem. ( A d . ) , 1936, 8, 16.47 R. Casares, A w l . Pis. Qdm., 1936,34,694; 0. Fiirth and H. Herrmann,chen, J . Amer. Chem. Soc., 1935,57,1363.Chim. OtdeEa, 1935, 88.Bwchem. Z., 1936,280,448468 ANALYTICAL CHEMISTRY.The colours formed in the bromide-resorcinol reaction are tracedto the intermediate formation of glyoxylic acid, and a solution of thisacid is recommended as a reagent for the characterisation of phenols,phenolic acids, and the chief opium alkaloids.48Spot tests are described for the detection of phenol,4s oxalicacidYS0 and of benzidine and tolidine in the presence of each other.51fMicro-metho~~.-Methods for the determination of various organicgroups have been reviewed,52 and the technique detailed for thedetermination of carbon, hydrogen, nitrogen, halogen, and methoxylas applied to samples of 0-001-0.01 g.53The necessary data are given for the characterisation, by theexamination of crystals under the microscope, of : alkaloids aspic rate^,^* d- and Z-cocaine after treatment with potassium perman-ganate,55 and amino-acids as their phosphotungstates, phospho-molybdates, picrates, and Aa~ianates.~~ Berberine gives character-istic micro-crystalline precipitates with a number of compounds ofpharmacological importance. 57Accurate methods are described for the determination of : acetylgroups,58 acetone, 59 oxalic acid, the permanganate titration in hotsolution bcing shown to be untrustworthy,m glucose by the ferri-cyanide-ceric sulphate metl.lod,61 and protein nitrogen in thepresence of ammonium salts.62Apparatus.-Improved apparatus is described for the micro-hydrogenation of organic comp0unds,~3 and for avoiding loss byspurting during ashing with nitric acid.64An electrically-heated melting-point apparatus gives an accuracyof 0.5°,65 and a method for the determination of setting pointsenables metastable forms fusing within 1" of each other to be de-4 8 M. Pesez, Bull. SOC. chim., 1936, [v], 3, 676.40 Y. Kondo, Mikrochem., 1936, 19, 214.F. Feigl and 0. Frehden, ibid., 1935, 18, 272.s1 L. Kulberg, J . Gen. Chem. Russia, 1935, 5, 1754.s2 A. Lacourt, Bull. Soc. china. Belg., 1936, 45, 313.63 C. Weygand and H. Hennig, Chem. Pabr., 1936,9, 8.64 A. Ionesco-Matiu and E. Iliesco, J . Pham. Chim., 1936, [viii], 23, 117.s5 R. Ceeconi, Ann. Chim. appl., 1936, 26, 218.66 B. Bullock and P. L. Kirk, Mikrochem., 1935,18, 129 ; B. L. Crosby and67 C. van Zijp, Pharm. Weekblad, 1936, 73, 764.E8 A. Elek and It. A. Harte, Ind. Ena. Cherq. (Anal.), 1936, 8, 267.56 A. Lindenberg, Compt. r e d . SOC. Biol., 1936, 122, 317.6 o J. Reneudin, J. Pham. Chim., 1936, [viii], 23, 447.61 R. Vanossi and R. Ferramola, Anal. Asoc. Quim. Argentina, 1935, 23, 162.62 A. Roche and F. Marquet, Bull. SOC. Chim. biol., 1935,17, 1630.63 H. Jackson and R. N. Jones, J., 1936, 895.64 H. Kaunitz, Mikrochem., 1936, 20, 104.B G E. Dowzard and M. J. Russo, Ind. Eng. Chcm. (AnaZ.), 1936, 8, 74.P. L. Kirk, ibid., p. 137WEST : ORGANIC ANALYSIS. 469tected.66 A differential mercury manometer 67 is used for measuringboiling points with only a few drops of liquid.Modern resistance glass enables simple all-glass water stills to beconstructed, which compare favourably with metal stills for effici-ency.6* R. W. W.N. STRAFFORD.L. S. THEOBALD.R. W. WEST.66 F. Francis and F. J. E. Collins, J., 1936, 137.6 7 R. Dolique, Bull. SOC. chim., 1935, [v], 2, 1832.68 B. Siede, Chem.-Ztg., 1935, 59, 925