ANALYTICAL CHERilISTRY.THE fact that the function of the writer of this Report must, fromthe nature of the subject, be that of a judicious reporter ratherthan of a reviewer has been insisted upon in previous AnnualReports, and the difficulty of treating the subject of the progress inAnalytical Chemistry in the form of a connected narrative has beenpointed out. The field to be covered is a very large and ever-increas-ing one, and from some inquiries which he has made, the author isled to believe that the arrangement of the subject-matter, whichhe has adopted in past years, is perhaps the clearest and mostconvenient one for the reader. The work of the year will thereforebe dealt with as before, under the following headings :(1) Inorganic Chemistry, including electrochemical methods.(2) Organic Analysis.(3) Analysis of Foods and Drugs.(4) Toxicological Analysis.(5) Apparatus.The above subdivision of the subject is, of course, to some extentan arbitrary one, but it appears to permit of a more methodicaltreatment than any other with which the writer is acquainted.Inorganic Chemistry.I n connexion with the qualitative section of this branch, there arebut few communications which merit special reference.The detection of ferrocyanides, ferricyanides, and thiocyanates inthe presence of one another is not always easy, and an apparentlyuseful method which appears to be characterised by certainty anddelicacy has been proposed for this purpose by Browning andPalmer.1 This depends on the fact that the ferrocyanides can beseparated from the ferricyanides and from the thiocyanates byprecipitation with a salt of thorium, whilst the ferricyanides can inturn be separated from the thiocyanates by precipitation with a saltAbstr., 1907, ii, 721200 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.of cadmium. The main difficulty is in connexion with filtration,but the method has given good results in the writer’s hands.I n mineral analyses, it is frequently necessary t o test silica precipi-tates for the presence of small quantities of oxide of titanium, andEnecht2 has described a method based on the reduction of thetitanium oxide and the decolorising effect of the resulting solutionwhen added to a solution of Rochelle salt, coloured slightly withindigotin, or, better, with methylene-blue.The detection of sulphites in the presence of thiosulphates andthionates presents some difficulty, and a method proposed byVotoEek3 seems likely to be useful.It is based on the fact that,whilst normal sulphites destroy the colours of solutions of certaintriphenylmethane dyes, thiosulphates and di-, tri-, and tetra-thionates are without effect. The most suitable reagent consists ofa mixture of a solution of magenta with one of malachitegreen, andthe author states that as little as 0*00006 gram of sulphurous acid(as normal sulphite) can be detected.For the detection of traces of moisture in gases or liquids, W.Biltz 4 makes use of potassium lead iodide, an almost colourless salt.,which is readily decomposed by traces of water with separation ofyellow lead iodide.Test papers may be easily prepared with this re-agent, and constitute a convenient method of applying the reaction.Ehrenfeld 5 shows that the red precipitate sometimes obtainedon acidifying the ammonium carbonate solution, which hasbeen used for the separation of arsenic and tin in the ordinaryprocess of analysis, consists of arsenic disulphide, the reduction ofthe higher sulphides being readily effected by stannous chloride inacid solution.Pozzi-Escot 6 recommends for the detection of traces of nickel theuse of ammonium molybdate, which gives a green, crystalline pre-cipitate even in the presence of a considerable excess of cobalt.Tschugaeff? however, points out that this test is far less sensitivethan that based on the use of dimethylglyoxime, which is saidto be capable of detecting as little as one part of nickel in twomillion parts of water.The ionisation theory, which was, I believe, first put forward byOstwald to explain the colour changes manifested by indicators, hasbeen very generally accepted as correct, notwithstanding an impor-tant paper published some years ago by Eltieglitz,* who suggestedwhat he termed the ‘‘ chromophoric ” theory, pointing out that itwas more probable, for example, that phenolphthalein, in its colour-Abstr., 1907, ii, 654.a Ibid., 195. Ibid., 574.Ibid., 949. IbicZ., 818. Ibid., 989.Ibid.,1904, ii, 17ANALYTICAL CHEMISTRY. 201less solution, has the constitution of a lactone, whilst its salts arederivatives of a carboxylic acid containing the chromophoricquinonoid complex :C,H,:O.Following up his work on thecolorimetric measurement of affinity, Salm, in conjunction withFriedenthal, has recently published a paper 9 on this subject, inwhich he adduces some evidence for the beIief that the colourchanges of indicators are, in many cases, due rather to intramolecularchange than to ionisation. In this connexion, it may be mentionedthat Knowles 10 recommends alizarin-red I.W.S. (alizarinmonosul-phonic acid) for use in alkalimetry instead of methyl-orange, overwhich it appears to possess some advantages, and that 0. Stark11states that excellent results may be obtained with 3-amino-Z-methyl-quinoline as an indicator, especially in substitution f o r methyl-orange in the titration of ammonia.As a standard substance for usein alkalimetry and acidimetry, Phelps and Hubbard 12 recommendsuccinic acid, obtained by the hydrolysis of ethyl succinate.RonchBse 13 describes an ingenious method for the estimation ofammonia, based on the reaction of formaldehyde with ammoniumsalts, whereby hexamethylenetetramine is formed, and, if theformaldehyde is in sufficient excess, the whole of the acid of theammonium salt is liberated, and may be titrated with standardalkali.I n former reports, reference has been made to the estimationof sulphuric acid by Raschig’s benzidine method, and, as this methodis capable of giving good results and may occasionally be of use, itwould seem desirable to direct attention to a paper by Friedheimand Nydegger 14 on this subject.The authors have fully investi-gated the most favourable conditions for precipitation, and give fullparticulars in regard to the solubility correction.The ‘‘ nitron ” method for the gravimetric estimation of nitricacid has been submitted to a detailed and critical study by S. W.Col1ins,l5 who finds that good results can be obtained both withsimple and niixed pure nitrates and in commercial products, suchas natural waters and nitratecontaining fertilisers. In view of theexpense of the “nitron,” it is satisfactory to note that the authorhas succeeded in devising a process for the recovery of the base.Last year, Jannasch and Heimann 16 succeeded in quantitativeIyvolatilising phosphoric acid from its salts by making an intimatemixture of the phosphate with carbon, and distilling off the, phos-phoric acid in a stream of chlorine.This method, although itgave good results with ammonium hydrogen phosphate and withAbstr., 1907, ii, 389. lo B i d . , 389. l1 fiaid., i, 974.l2 &d., ii, 297. l3 Ibid., 651. l4 Ihid., 196.j 5 Ibid., 907. l6 Bid., 1006, ii, 745202 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.ammonium magnesium phosphate, was somewhat clumsy and diffi-cult to carry out. Jannasch and 3ilke17 have now improved onthis, in that the phosphoric acid is distilled off by heating stronglyin a stream of chlorine mixed with the vapour of carbon tetra-chloride. The results given for ammonium phosphate and for mag-nesium phosphate are very good, and the method is one which might,in cerbain cases, be usefully employed in mineral analysis.Thetime required for the carrying out of the process (between four andfive hours for the vaporisation of an amount of phosphoric acid equi-valent to about 0.3 gram of magnesium pyrophosphate) would, how-ever, constitute a drawback from the point of view of practicallaboratory procedure.The quantitative separation of the halogens constitutes a veryimportant analytical problem, and one t o the solution of which agreat many chemists have a t various times devoted themselves. Theprinciple underlying almost all existing methods is that of differen-tial oxidation, and, as the ‘( oxidation potential ” of all the oxidisingagents recommended is, with one exception, higher than that of anaqueous solution of chlorine, it follows that good results can onlybe obtained by accurately interrupting the process when the wholeof the bromine has been distilled over, and by paying minute atten-tion t o the experimental conditions, especially to the degree ofacidity and concentration.Iodic acid constitutes the single excep-tion referred to above, and was originally proposed for the separa-tion of bromine and chlorine by Bugarszky.18 This substance hasan “ oxidation potential ” which lies between that of chlorine, andthat of bromine, and, consequently, it is well suited for the separa-tion of these two halogens. The method, which does not appear t obe so widely known as it should be, has been submitted to it furtherstudy by Andrews,lg who has suggested certain improvements in theprocedure, and has shown, not only that accurate results can beobtained in the separation of bromides and chlorides, but that themethod is applicable to the estimation of chlorine in crude bromine.For the estimation of cyanogen in slightly dissociated salts, suchas mercuric cyanide, mercuric cyanonitrate, and mercuric cyano-chloride, Borelli20 has proposed a method in which the cyanogen isobtained in the form of sodium cyanide by treatment with alumin-ium powder in a solution rendered strongly alkaline with sodiumhydroxide.The results appear t o be good, and the process, whichis also applicable to the determination of cyanogen in complexiron cyanides, is more rapid of execution than that of Rose.Agood many methods have a t various times been proposed for the17 dbslr., 1907, ii, 564.ly Ibid., 1007, ii, 503.Ibid., 1896, ii, 216.2o ]bid., 825ANALYTICAL CHEMISTRY. 203analysis of mixtures of thiocyanates and chlorides, or bromides, andRosanoff and Hill 21 have studied and criticised some of these, andhave devised a method which appears to be both simple and rapidof execution. This depends on the oxidation of the thiocyanate tohydrocyanic acid by means of nitric acid, and is similar to oneproposed years ago by Volhard, but differs from that in that thesolution is not rendered alkaline, a treatment which, according toRosanoff and Hill, causes the re-formation of some thiocyanate.The separation of tellurium from certain heavy metals and itsaccurate estimation has occupied the attention of Brauner andKuzma,22 who show that, when tellurium in the presence of copper,bismuth, and antimony is precipitated by means of sulphur dioxide,appreciable quantities of the associated metals are carried down,and that this is especially the case with copper.I n order t o over-come this difficulty, the authors recommend the oxidation of thetellurium to telluric acid by means of ammonium persulphate, andthe subsequent precipitation of the heavy metals with hydrogensulphide. The tellurium can then be estimated by reduction withhydrochloric acid and sulphur dioxide. Incidentally, attention iscalled t o the interesting fact that, when solutions containing bothcopper and tellurium are oxidised with the persulphate, an intensered coloration is observed, which the authors attribute t o the forma-tion of a, derivative of cupric acid.Hinrichsen23 calls attention to the great difficulty there is incompletely removing the hydrofluoric acid when that substance hasbeen used in conjunction with sulphuric acid for the decompositionof silicates.He points out that, when ammonia is subsequentlyadded for the purpose of precipitating the iron and aluminium, adouble fluoride of ammonium and aluminium is formed, which issoluble in hot water, and he states that this behaviour may be thecause of introducing a very serious error into the aluminium esti-mation.This paper deserves the attention of all who are interestedin mineral analysis. The estimation of iron, aluminium, and titan-ium, when present together, is a problem of very frequent occur-rence in mineral analysis, and two papers dealing with this subjectwill be read with interest. One by Gallo24 deals especially with thevolumetric estimation of the titanium by titration with a solutionof ferric alum, whilst in the other, by Magri and Ercolini,25 amethod for the electrolytic separation of the iron is described.I n this connexion, attention may be directed t o a paper by Goochand Newton,26 who show that iron may be determined by titrationwith permanganate in the presence of titanium if some bismuth24 Ibid., 402. 25 Ibid., 400.3(i Ibid., 507.Abstr., 1907, ii, 984. 22 Ibid., 716. 23 lbid., 506204 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.oxide is added to the reduced solution. This appears to have theeffect of oxidising the titanium without affecting the ferrous salt, sothat the latter may be directly titrated after filtration. Inasmuchas the authors specially recommend the use of amalgamated zincfor the reduction of the iron, the writer of this Report. may perhapsbe permitted once again to direct the attention of analysts to thegreat advantage of using charged palladium for that purpose. I fthe metal is fully charged, reduction is readily brought about, andthere is obviously nothing to interfere with the sharpness of the endreaction with the indicator. The quantitative separation of zincand cadmium by means of hydrogen sulphide is, as is well known,somewhat troublesome, inasmuch as the acidity limits within whichsuch separation can be effected are very narrow, and in any caseseveral precipitations are usually necessary. Box 27 has shown thattrichloroacetic acid may be advantageously substituted for themineral acids usually employed in this separation, and that, unlessthe amount of zinc present is large in relation to that of.thecadmium, a single precipitation is sufficient. A critical paper byFunk,28 dealing with a kindred subject, namely, the separation ofzinc from nickel, cobalt, iron, and manganese by means of hydrogensulphide, is worthy of study, in view of the well-known difficultiesattaching to the problem.The author recommends either precipi-tation from a solution containing formate or Treadwell’s process,in which the precipitation is effected in the presence of an excessof an alkaline chloride or sulphate.Glixelli 29 contributes a very interesting communication dealingwith the theory of the action of hydrogen snlphide on zinc salts, andshows that, contrary to the generally accepted view, the reactionZnS0,+H2S=ZnS+H2S0, is not a reversible one, but that falseequilibria occur in acid solutions which are sometimes very per-sistent, and which can be upset, and precipitation hastened, by thepresence of zinc, cadmium, or copper sulphides. The author is ofopinion that the separation of zinc from nickel and cobalt in acidsolutions by means of hydrogen sulphide does not depend on thedifference in the equilibrium conditions, but on the inductionperiods.There is still considerable room for investigation in connexionwith the methods for estimating and separating many of the rareelements, and the results of such investigations are always welcome.This is the case with a paper by Paal and Arnberger,30 who havedevoted a good deal of attention to the analytical chemistry ofosmium.Although this rare element is of some technical import-27 Trans., r907, 91, 964.lbid., 868.28 Ahstr., 1907, ii, 398.3o ]bid., 404ANALYTICAL CHEMISTRY. 205ance, the methods recommended for its estimation do not appear tohave been well worked out, and leave much to be desired in pointof definiteness and accuracy.Of these, perhaps the most widelyknown is that in which the osmium is separated as Fr6my’s osmyldi-ammine chloride from alkaline osmate solutions. The abovemen-tioned authors have now shown that this method is very unsatis-factory, and have suggested several processes which will doubtlessbe carefully tested by those analysts who are actively interested inthe analytical chemistry of the platinum metals. In this connexion,it may be noted that Makowka31 has shown that osmium is com-pletely separated from acid solutions by acetylene, and he also statesthat palladium in acid solution may be precipitated and separatedfrom platinum and iridium by means of the same reagent. Thoseinterested in this method may also be referred t o two papers on thesame subject by Erdmann and Makowka.32Dittrich and Freund 32a have studied the separation of thorium,titanium, and zirconium from iron, and have devised a new methodfor separating titanium from zirconium.This consists in adding adilute and faintly acid solution containing these elements to a boil-ing solution of ammonium salicylate. Zirconium salicylate is pre-cipitated, the titanium remaining in solution. Small quantities ofthe latter element may be carried down with the zirconium, butcomplete separation may in such cases be effected by a second pre-cipitation.As an interesting example of the manner in which organic com-pounds are being pressed into the service of analytical chemistry,mention may be made of the use of dicyanodiamide sulphate as aprecipitant for nickel.This substance, the use of which had pre-viously been suggested for the detection of nickel, has now beenapplied by Grossmann and Schiick 33 to its estimation. The dicyano-diamide precipitate may be converted into nickel sulphate, orapparently more accurate results are obtained by drying a t 1 1 5 O andweighing, the precipitate then having the compositionNi( C,H,ON,),.Cobalt in the tervalent condition and zinc are not precipitated, andthe method may be used for the separation of nickel from iron andfrom aluminium. Similarly, Brunck 34 has employed dimethyl-glyoxime, which had previously been suggested as a test for nickel,for the estimation of that metal, and its separation from certainother metals, such as zinc, manganese, iron, aluminium, andchromium. The results appear to be good, and the method is rapid,31 dbstr., 1907, ii, 403.32a Ibid., 1908, ii, 134.32 Ibid., 399, 403.33 l b i d ., 1907, ii, 582, 819.34 Ibid., 582, 989206 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.but the costliness of the reagent must tend considerably t o restrictits useAlthough there is little that is new in a paper by Kolb and Feld-hofen,35 these authors have done good service in calling attention tothe fact that hydrogen peroxide may, under suitable conditions, beused for the purpose of reducing mercuric salts t o the mercurousstate. The best conditions for obtaining the mercurous chlorideprior to its estimation by Hempel’s well-known iodine method aregiven.The detection of very small quantities of mercury in explosiveshas recently become a matter of some practical importance, owingto the marked manner in which minute quantities of mercuricchloride affect the Abel heat test.A satisfactory spectroscopicalmethod for this purpose was devised by the late Dr. Dupr6, and, ina recent paper, Hargreaves and Rowe 36 have described an electro-lytic process, the mercury being deposited on a gold foil cathode,from which it can be obtained by sublimation in the form of globulesfor microscopical identification.The ‘‘ comparison of mirrors ” method for the estimation of minutetraces of arsenic, which was first suggested by Sanger, has beenextended by that author and Gibson37 to the estimation of verysmall amounts of antimony.It has been found that, when quanti-ties less than 0.1 milligram of that element are added to the reduc-tion flask under the conditions laid down by the authors, the wholeis evolved as hydride. The Gutzeit,although distinctly inferior for most purposes to the Marsh-Berzeliusmethod, has its uses, and Sanger and Black38 have dealt a t con-siderable length with the former process, and have made a detailedstudy of the conditions which must be observed if the greatestdegree of sensitiveness and accuracy is to be obtained. Robertsonand Nbpper 39 have dovised a method for the estimation of smallquantities of nitrogen peroxide which is specially applicable to theexamination of the gaseous products of the decomposition of gun-cotton and other explosives.This is based on a comparison ofthe absorption spectrum of the gas under observation with that ofa standard gaseous mixture, the spectrum of nitrogen peroxidebeing, as is well known, very characteristic, and altering markedlywith the concentration.The test results are good.ElectTolytic Methods.A good deal of very useful work has been done during the year inthe study of electrochemical methods of analysis.35 Abstr., 1908, ii, 69.37 Abstr., 1907, ii, 654.a6 J. SOC. Chem. Ind., 1907, 26, 813.3y Trcu~., 1907, 91, 761. 38 Ibid., 1908, ii, 65ANALYTICAL CHEMISTRY. 207Dorniaar 4O has investigated the cause of the high results whichare frequently obtained in the electrolytic estimation of antimony,and finds that, whilst they are partly due, as has usually been sup-posed, to the inclusion of sulphur compounds in the precipitate,they are chiefly due to oxidation changes inherent in the electrolyticprocess itself.The same conclusion has been arrived at by Foersterand Wolf,41 who have identified antimony oxide as well as sulphurin the precipitated antimony. Foerster, Blankenberg, Brunner,Lee, and Rommler have made a detailed study of the influence oftemperature on the electrolytic deposition of various metals, andsome of their results are of importance. Thus it has been observedthat, whilst a t the ordinary temperature the decomposition poten-tials of zinc and nickel are so near to one another that electrolyticseparation is impossible, the difference is so greatly increased a thigher temperatures that a satisfactory separation can be effected.Other advantages appear to result in special cases from working a televated temperatures.The advantages of employing rotating electrodes are now wellrecognised.With these, it is possible to work with much highercurrent densities, whilst the time required for the deposition of themetal to be determined is very much shortened, and the deposit is,moreover, frequently obtained in a better form for weighing. Avery important paper, and one embodying a considerable amount ofvery useful work, has been published by H. J. S. Sand,42 who de-scribes the results he has obtained working with rotating electrodesand employing graded potential.The test results are, as a rule, verygood, and the electrolytic separation of no fewer than seven metals inone solution must surely constitute a record. Miss Langness 43 alsodetails the results she has obtained in a number of estimations andseparations when working with a rotating anode, and the paper isone which may be read with advantage. From the point of viewof analytical chemistry, perhaps the most important, and certainlythe most useful, electrochemical investigations are those undertakenwith the object of ascertaining the precise working conditions neces-sary for the accurate separation of commonly associated metals.The above-mentioned papers afford instances of this class of com-munication, and, as another example, attention may be drawn toa paper by Miss Kollock and E.F. who have determinedthe time necessary for the complete deposition of a number of metalswhen varying amounts of sulphuric acid were present in the elec-trolyte, a ‘mercury cathode and a rotating anode being employed.Results are also given showing the best experimental conditions for41 Ibid., 508. 40 ABslr., 1907, ii, 200. 41 Tyans., 1907, 91, 373,43 Absty., 1907, ii, 585. 44 ]bid., 719208 ANNUAL lCEPORTS ON THE PROGRESS OF CHEMISTRY.effecting certain separations. Another interesting communicationis that of Hildebrand,45 who shows that the negative radicles in suchcompounds as the carbonates, thiocyanates, and ferrocyanides of thealkali and alkaline earth metals may be estimated by employinga cell with a mercury cathode and a rotating silvered platinumanode.This work has been extended by McCutcheon, jun.,lG byLukens and E. F. Smith,47 and by McCutcheon, jun., and E. F.Smith.48 These authors have observed that, when solutions ofmetallic chlorides are electrolysed according to Hildebrand's method,the metals are capable of being divided into two classes, accordingto the behaviour of the amalgams formed. Thus the amalgams oflithium, sodium, potassium, calcium (see below), strontium, andbarium decompose in the outer, or cathode, compartment withformation of the corresponding hydroxides, whilst the amalgams ofmagnesium, aluminium, and the heavy metals are decomposed withformation of the corresponding hydroxides in the inner, or anode,compartment.It was therefore possible in this way to carry outelectrolytically a number of interesting separations. It is note-worthy that in the electrolysis of solutions containing both mag-nesium and calcium chlorides, none of the latter metal passes intothe outer compartment, as is the case when calcium chloride aloneis used. It would appear therefore that on this behaviour an in-teresting method might be based for the electrolytic separation ofcalcium from barium and strontium. In order to obviate the neces-sity for using a motor for the purpose of rotating the electrodes,Frary49 has adopted the device of causing the electrolyte itselfto rotate by placing it within a solenoid, through which the currentused for the electrolysis passes.The conditions affecting the accu-racy of the results obtained in the electrolytic estimation of leadhave been somewhat exhaustively studied by Vortmann,6O who showsthat good results are not so easily obtained as is often imagined,and that there are a good many substances which interfere, and inthe presence of which special treatment becomes necessary. It iswell known that the ease with which many oxidation and reductionchanges can be effected-depends very greatly on the nature of themetal of which the electro,de concerned is constructed. This pheno-menon, usually referred to as " supertension " or " over-voltage," hasbeen ascribed by Tafel, Caspari, and others to the pressure a t whichthe hydrogen and oxygen respectively are set free at the surfaces.Kauflertl however, contends that it is largely, if not entirely, dueto local heating of the electrodes, and haa shown that a low " poten-45 Abstr., 1907, ii, 574.48 Ibid., 988.@ Ibid., 988.49 lbbid., 649.61 Ibid., 924.47 lbid.) 988.6o Ibid., 302ANALYTICAL CHEMISTRY.209tial " metal, such as platinum, is capable of bringing about changeswhich can usually be effected only by high (' potential " metals ifsuitable means are taken t o heat the electrode sufficiently.Organic Analysis.Reactions (chiefly dependent on colour changes) of more or lesscomplicated organic substances, even when characteristic, are, as arule, of very little general interest, but reference may perhaps bemade to communications dealing with some analytical reactions ofveronal (diethylmalonylcarbamide) by Lemaire,52 and of maretineby the same author,53 of adrenaline by Krullt4 of vesipirin (phenyl-acetosalicylate) by Zernik,55 and of antipyrine by Steensma,"6 asthese substances are of therapeutic importance.A simple reactionof adrenaline, which is said to be characteristic, is also described byGunn and Harrison.57 Molinari 58 has investigated the use of ozoneas a reagent for the study of the unsaturatedness of organic com-pounds. The author finds that unsaturated compounds, whetheraliphatic o r aromatic, containing double bonds combine readily withozone, whilst aliphatic compounds containing triple bonds, althoughuniting directly with a large proportion of iodine, do not absorbozone.F o r the detection of small quantities of hydrocyanic acid,Thigry 59 recommends the use of test papers, moistened first with adilute solution of copper sulphate, and then with an alkaline solu-tion of phthalophenone. I n the presence of hydrocyanic acid, afairly permanent rose-red coloration is produced, and the test issaid to be capable of detecting the presence of as little as one partof hydrocyanic acid in two million parts. The separation of dye-stuffs in mixtures is often a matter of considerable difficulty, andthe suggestion of E. LehmannGO that advantage may sometimes betaken of the different rates of diffusion into jellies is noteworthy.With a mixture of eosin and tartrazin, f o r instance, sufficientseparation could be effected to permit of certain identification.Avery comprehensive paper dealing with the identification of dye-stuffs on vegetable fibres by Green, Yeoman, Jones, Stephens, andHaley6l cannot fail t o be of great assistance, not only to colourchemists, but also to those whose incursions into this difficult fieldof analysis are of less frequent occurrence.W. E. Marshall 62 confirms the value of the p-dimethylaminobenz-Rep. Phawn., 1907, 19, 104. 63 Zbid., 49.55 Apoth. Zeit., 1907, 22, 152.57 Ibid., 591.@2 Abstr., 1907, ii, 996.61 dbatr., 1907, ii, 316.56 Abstr., 1907, ii, 995.58 IEid., i, 1039.REP.-VOL. IV. P59 lbid., ii, 408. Go lbbid., 234.J. Xoc. Dyers, 1907, 23, 252210 ANNUAL REPORTS ON THE PROGRESS OF CIIEXlSTltP.aldehyde test for indole, and shows that i t is capable of distinguisli-ing that substance from others giving very similar general reactions,and that it can, moreover, be employed quantitatively as a chlori-metric method.As I have already referred in previous reports63 t o Denn-stedt’s method of elementary analysis, it is well that attentionshould be called t o a paper by Baumert,64 who describes certainmodifications which he has introduced into this combustion process.In this connexion, reference may also be made to two papers dealingwith the relative advantages of platinum and palladium as contactsubstances in organic analysis, the one by A.Jacobsen and Lan-de~en,~5 and the other by Dennstedt.66 The latter author is ofopinion that platinum is for several reasons to be preferred to palla-dium, and points out that, although both these substances give betterresults with methane than oxide of copper, neither is capable ofbringing about the complete oxidation of that gas.Quite a numberof new methods for the estimation of halogens in organic compoundshas been suggested during recent years, some of which are bothrapid and exact, and the analyst is no longer compelled t o choosebetween the somewhat dangerous method of Carius and the trouble-some Iime-combustion process. Chablay, for instanceF7 describesa convenient and apparently accurate method, in which advantageis taken of the fact that halogen-containing organic compoundsreact with sodammonium with quantitative formation of the sodiumhalide, F o r the estimation of phosphorus and sulphur in organiccompounds by the fusion method, Stutaer 68 recommends the use ofbasic calcium nitrate, since this substance not only supplies thealkali necessary to prevent any loss of the elements in question, butalso melts below looo, and does not exhibit any tendency to detonateor spirt.For the estimation of hydrogen in a very large numberof organic and inorganic compounds, Lidoff recommends a volu-metric method involving the ignition of the compound with mag-nesium or aluminium powder. The results appear t o be good, andsulphur, nitrogen, and the halogens may be determined in theignited residue by the usual methods. It will be remembered thata few years ago, Hibbert and Sudborough 7^ showed that, by observ-ing certain precautions, Grignard’s magnesium alkyl halide method,which had been proposed by Tschugaeff for the detection of hydroxylgroups, might be made quantitative, and they published it numberof results in support of this contention.This method has now been63 Ann. Report, 1905, 195 ; 1906, 211.65 Ibid., 718. Ibid., 909.67 Ibid., 195. G8 Ibid., 906.69 Ibbid., 574. 70 Trans., 1904, 85, 933.Abstr., 1907, ii, 909ANALYTICAL CHEMISTRY. 211further examined a t Tschugaeff’s request by Zerewitinoff,7l who hasobtained very satisfactory results with a considerable number oforganic compounds, differing widely in their chemical characters,and there can be no ‘doubt that the method is a very useful one.The writer of this Report, in conjunction with Whitteridge,72 hasdevised a method for the estimation of tartaric acid which may beusefully employed in many cases in which the ordinary potassiumhydrogen tartrate method is, for various reasons, inapplicable.Thetartaric acid is precipitated as a basic bismuth tartrate, which isdissolved in acid, and determined volumetrically by titration withpermanganate. Lasserre 73 calls attention to a convenient methodof separating formic and acetic acids, on the one hand, from butyricand valeric acids, on the other, based on the fact that the two last-mentioned acids can be extracted from their aqueous solutions bymeans of benzene or toluene. Propionic acid cannot be separated inthis way, as it is equally soluble in benzene, toluene, and water.I f , asappears t o be the case, this separation is sharp, the metho,d is likelyto be much used. I n previous Reports,74 attention was called tothe employment by Knecht, and subsequently by obhers, of titaniumtrichloride as a reducing agent for use in chemical analysis. It hasan even more powerful reducing action than stannous chloride, andmay be used for the estimation, not only of such substances asferric iron, but also for the determination of organic nitro-com-pounds, such as trinitrocresol, dinitrobenzene, and dinitronaphtha-lene, as well as of azo-dyes, such as benzopurpurin. In a recentpaper, Knecht and Miss Hibbert75 describe the results of furtherwork in this direction, and show that in many cases where directtitration with the titanium trichloride gives inaccurate results, goodresults can be obtained by adding an excess and titrating back witha standard ferric solution.The authors also recommend the useof this reagent for the evaluation of commercial hyposulphites andfor the estimation of dissolved oxygen in water. One advantage ofthe titanium over the permanganate method in the latter estima-tion is that i t is not affected by any organic matters which may bepresent. This paper is one which will well repay study, especiallyby those chemists who are interested in the analysis of organicdyestuffs. I n my previous Report,76 I called attention to a paperby Valenta, in which he recommended the use of methyl sulphatefor the purpose of distinguishing between aromatic and paraffinhydrocarbons and separating them.This method has since beenstudied by Graefke,77 who has shown that aliphatic hydrocarbons are71 dbstr., 1907, ii, 509.74 Ann. Report, 1905, 195 ; 1906, 207.T2 Zbid., 513. 73 Ibid., 991.75 Abstr., 1907, ii, 907.Ann. Xeport, 1906, 214. 77 Chem. rev. E%tt.-€€arz.-Ind., 1907, 14, 112.P 212 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.not by any means completely insoluble in methyl sulphate, and thatin separations their solubility is oftten still further increased by thepresence of the dissolved coal-tar hydrocarbons. Notwithstandingthis, the method is, in certain cases, capable of being usefully em-ployed.Of all the problems in organic analytical chemistry, perhaps oneof the most difficult consists in the quantitative analysis of certaincomplex artificial carbohydrate mixtures.Indirect methods havealmost invariably to be employed, and errors in the estimation ofsome of the substances present, although not serious in themselves,are sometimes apt to become additive, and to fall with very seriouseffect upon the estimation of some one constituent. During recentyears, there has been an increasing tendency on the part of analyststo avail themselves of biological methods, and the application ofpure cultures of yeasts and other lowly organisms has rendered itpossible successfully t o attack analytical problems which had pre-viously been regarded as incapable of solution. The method isclearly one which must be applied with very great caution, anddemands a certain amount of biological training on the part of theoperator.Among others, Lindner has done a great deal of workin this direction, and atten€ion may be called to an interesting paperby Konig and €€ormann,78 in which some very useful confirmatorywork is recorded, and which contains the results of experimentsmade for the purpose of ascertaining to what extent the dextrinsobtained by the limited action on starch of acids and of diastaserespectively may be separated from the sugars. It is pointed outinter alia that the dextrinous body present in honey is of an entirelydifferent character from the dextrins prepared as above mentioned.This paper will well repay careful perusal by all who are activelyinterested in carbohydrate analysis.It is interesting to note thattwo species of torulz are capable of effecting a separation of dextroseand fructose from sucrose, and it seems very probable that a furtherstudy of the torulze (as opposed t o the true yeasts) might beproductive of useful results. Neuberg79 has shown that, by theaction of emulsin, raffinose is hydrolysed into a mixture of saccharoseand &galactose, and, in conjunction with Marx,gO has suggested theuse of emulsin as a test f o r the presence of raffinose in raw canesugars. I n view of the fact that the melting points of the osazonesare frequently made use of for purposes of identification, a paperby Tutinsl is of considerable importance. The author shows thatwhen d-phenylglucosazone is purified by recrystallisation from amixture of pyridine and alcohol, the compound melts at 2 1 7 O .TheA b s t ~ . , 1907, ii, 202.IDitl. ii, 408.7'3 ]bid., i, 388.a PYOC., 1907, 23, 250ANALYTICAL CHEMISTRY. 213melting point of 205O, which mas originally given by E. Fischer andthe correctness of which has always been assumed, must thereforeapparently be abandoned. It may be recalled that the meltingpoint of a-acrosazone has already been found t o be 2 1 7 O , instead of205O as originally stated, and it will be interesting to ascertainwhether similar differences will be observed in the case of the osazonesof other sugars. The nature of the influence exerted by basic leadacetate on the rotation of sucrose in aqueous solution has been in-vestigated by Bates and Blake,82 who show that von Lippmann’sstatement that basic acet’ate is without effect is incorrect.Theauthors find that on the progressive addition of the acetate thereis first a lowering of the rotation, and then a gradual rise, whichthey ascribe to the formation of soluble lead sucrates having rota-tions differing from that of sucrose itself. Lintner83 describes apolarimetric method for the estimation of starch in cereals, whichhas been further examined by Canet and D~rieux,84 who haverecorded a number of results with cereals and other starch-containingproducts. The method certainly appears to be capable of givinggood results, and will doubtless be useful as a substitute for themore lengthy and laborious diastase conversion method when rapidresults are required.The analysis of indigo has, during recent years, been made thesubject of a considerable amount of experimental work, and hasfurnished a good deal of controversy.The attention of all chemists who are interested in this matteris directed to an important communication by Orchardson, Wood,and W.P. BloxamJs5 constituting the second part of their paper on“ The Analysis of Indigo.” This contains an account of experimentsmade for the purpose of isolating the impurities present in cakeindigo, and of ascertaining how these substances behave with theanalytical reagents employed, and t o what extent they interfere.A new method is described for the estimation of indigotin obtain-able from the leaf, and the authors appear t o be confirmed in theiropinion that the separation of the indigotin in the form of a purifiedcompound (potassium indigotintetrasulphonate) must constitute thefirst step in the analytical process.The method of Bergtheil andBriggs, which was referred t o in my previous Report,*6 is adverselycriticised by the authors, who state that it gives irregular rmults.The attention of those interested in this matter may be also directedt o two other communications, one by Bergtheil and Briggs?7 andthe other by Gaunt, Thomas, and BIoxam,8* both dealing with the82 Abstr,, 1907, ii, 406,84 BuEL SOC. Chim. BeZg., 1907, 21, 329.@q dwz, Report, 1906, 216. 11’ Abstr. 1907, ii, 416. a8 &d., 1908, ii, 76.Ibid., 823.85 Abstr., 1907, ii, 203214 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRYestimation of indigotin in the plant The writer of this Report,in common with several other chemipts, has had an opportunityof witnessing a demonstration of the tetrasulphonate process, andin his opinion there can be no doubt that the indigotin is quanti-tatively separated from the impurities with which it is associatedin the crude product, and that, t o say the least of it, the processmarks a great advance.The estimation of choline and other bases in vegetable and otherextracts is a matter of considerable importance t o the physiologicalchemist, and any improvement in the methods for separating andestimating these substances is welcomed by a large and increasingbody of workers.Stankk’s periodide method for the separation ofcholine and betaine, to which reference was made in the lastReportFg appeared t o constitute a very useful addition to existingmethods, but Kiesel90 has shown that it cannot be applied to theestimation of choline in such materials as plant extracts, since thesecontain many other bases which seriously interfere with the accuracyof the results.During the past few years, rubber experts have shown a growingtendency to attach importance t o the results of chemical analysiswhen forming an opinion as t o the value of commercial rubbers.A t the present time, the most important constituent, caoutchouc,is almost invariably estimated by difference, and the desirability ofobtaining a direct method for its estimation is obvious.Severalsuch methods have been suggested, but by far the most promisingare those in which the rubber is submitted to the action of nitrousfumes, and the resulting so-called “ nitrosate ’’ weighed. I n a pre-vious Report,91 it was mentioned that Harries and Alexander hadfound that the nitrogen compound originally obtained by Weberwas not a t all definite. I n a recent communication, Alexander92has recorded the results of further work in this direction, and hassuggested a method which, if carefully carried out, would appearto be capable of giving fair results. Much work remains t o be donebefore i t can be said that a satisfactory process for the direct estima-tion of caoutchouc has been devised, but Alexander’s method appearst o offer a solid foundation for further building, and his results,moreover, may throw a good deal of light on the constitution ofcaoutchouc and on the nature of vulcanisation.Another promising direct method is that of Budde, which is basedon the readiness with which caoutchouc unites with bromine to forman insoluble tetrabromide.A carbon tetrachloride solution, con-taining both bromine and iodine, is added t o a solution of the89 Ann. Report, 1906, 220.Dl Ann, Report, 1005, 200,00 Abstr., 1907, ii, 994.AQatr., 1007, i, 433ANALYTICAL CHEMISTRY. 215rubber, also in carbon tetrachloride, and the tetrabromide, which isprecipitated by the addition of alcohol, is collected, washed, dried,and weighed. Budde has shown that the compounds formed bythe action of bromine on certain unsaturated rubber resins aresoluble in the above mixture of carbon tetrachloride and alcohol,and, consequently, do n o t interfere with the results.Two papers,the one by Budde,g3 and the other by Aselrod,94 have recently beenpublished, from which it would appear that this method, with slightmodifications, is applicable t o the direct determination of caoutchoucin certain vulcanised rubbers. The further development of thesetwo direct methods will be watched with very considerable interestby all who are concerned in the analysis and commercial evaluationof rubber.Analysis of Foods and Drugs.Comparatively little progress has been made during the year inconnexion with the analysis of milk and milk products. Analystsare well aware that the accurate estimation of fat in certain milkpreparations is not always an easy matter, and that differentmethods are required in dealing with various materials.Hals andKlykkeng5 have recorded the results they have obtained in theestimation of f a t in sweetened and unsweetened condensed milks bya number of standard prpcesses, and are of opinion that the Gottlieband the Gerber methods are the most accurate for the purpose.Cocliran g6 has investigated the behaviour of Wiley’s acid mercuricnitrate solution, and has shown that it has no effect on the polarisa-tion of lactose, and that a t temperatures below 1 5 O its invertingaction on sucrose is very slight indeed. I n these circumstances,it can be used with good results for the estimation of lactoseand sucrose in sweetened condensed milk. Although a good dealof attention ha$, been devoted during the year to the importantquestion of the detection and estimation of cocoanut oil in butter,the subject has not h e n very much advanced, and there can belittle doubt that the Polenske method (either as originally publishedo r modified) and the phytosterol acetate test combined are stillcapable of furnishing the most trustworthy indications.Referenceshould, however, be made t o a communication by Hinks,Q7 who hasproposed a direct qualitative test for the presence of the cocoanutoil. This is a micro-crystallographic method, and in the hands ofits author has given distinct indications when the cocoanut oil waspresent only t o the extent of 5 per cent.This method appears t obe useful, and t?he advantage of possessing a characteristic test f o r93 Gzcmni. Zeit., 1907, 21, 1205.Qj Zcitsch. Nahr. Gentcssm., 1907, 13, 338.94 ]bid., 1229.96 Absk.., 1607, ii, 586.97 Analyst, 190’1, 32, 160216 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.this troublesome adulterant will be apparent. Most analysts knowhow greatly the distillation conditions may affect the Reichert-Meissl,and especially the Polenske, values of fats, and Goske 9* has devotedsome attention to the influence of the mode of heating on thePolenske number. This paper may be read with advantage, if onlyto emphasise the importance of adhering strictly t o a standard setof conditions.In view of the importance to analysts of the phyto-sterol acetate test, a paper on this subject by Jaeger 99 is deservingof attention. Windaus and Hauth 1 showed that Calabar beanscontained two isomeric phytosterols having very different meltingpoints, and it is probable that other isomerides exist. Jaeger showsthat the melting point curves for mixtures of cholesteryl acetate,with the acetates of the two above-mentioned isomerides, affordevidence that the melting points of such mixtures are not a t alltrustworthy criteria of composition, and it is clear that the indica-tions of the test must be interpreted with considerable caution.While dealing with this subject, attention may be drawn to a paperby Lewkowitsch3 on the determination of paraffin in the unsaponi-fiable matter from animal fats.The author confirms the accuracyof Polenske’s method: and shows that good results can be obtainedby determining the saponification value of the mixture of the ace-tates and the paraffin. The esterification of fats and oils by themethod of Haller, to which the objectionable name “ alcoholysis ”has been given, has been applied with slight modification by Hanui 5to the detection of cocoanut oil in butter, and the recorded resultsindicate that the method may a t least serve as a useful auxiliaryone. Further work on the ‘‘ silver ” process of Wijsman and Reijsthas confirmed the view that the results are untrustworthy, and thatthe values do not afford any certain criterion of the purity of butterfat.It should be noted that Siegfeld 6 and others have shown thatthe feeding of cattle with cocoanut cake and beetroot leaves mayappreciably affect the Polenske and iodine values of the butter.Several chemists, including Kreis and Canzoneri, have endeavouredt o isolate the constituent of sesame oil which produces the redcoloration with furfuraldehyde and hydrochloric acid, but withoutcomplete success. Malagnini and Armanni7 appear t o have beenmore fortunate, inasmuch as they have separated a crystalline sub-stance, melting a t 67O, having the formula C7H603, which gives,with the abovementioned reagents, a very intense coloration. Theauthors regard this substance as a methylene ester of hydroxy-Og Zeitsch. Nahr. Gertussm., 1907, 13, 491.3 Chdm.Rev. Fett, Earz, Ind., 1907, 14, 51.Abstr., 1907, ii, 315.Abstr., 1905, ii, 8701fiid., i, 129. B i d .Zeitacit, Nahr. Crenussm., 1907, 13, 18. Ibids, 613s ’ G’hemi &it., 1907, 81, 884ANALYTICAL CHEMISTRY. 217quinol, and point out that it does not exist as such in the oil, butis formed from a more complex compound (which they have alsoisolated) by the action of dilute mineral acid. This reaction forsesame oil is a very important one, especially in view of the factthat in several countries sesame oil is compulsorily added to mar-garine for the purpose of earmarking that substance and of facili-tating its recognition in butter. The literature of this test isvoluminous, and much has been written as to the conditions whichaffect the sensitiveness of the reaction.Although samples of sesameoil which do not yield the red coloration are undoubtedly veryrare,* it cannot be overlooked that several observers, including Weig-mann and Soltsien, have recorded cases in which the reaction couldnot be obtained, and it is also well known that samples of sesameoil differ very widely in the intensity of the coloration which theyproduce. Lauffs and Huismanng have been working on this sub-ject, and have recorded another instance of a rancid sample whichfailed t o respond to the test. T'hese authors appear t o be of opinionthat this is due t o the formation of a condensation product of thesesamol with the aldehydic substance present in the rancid oil, andthey recommend the addition of a proportion of cottonseed oil t othe fat t o be tested, which is said to prevent the interfering effectof rancidity and t o render the test more sensitive.However thismay be, it is clear that this question requires further study, for inview of these recorded observations chemists are not justified inconsidering a negative result when applying the Baudouin test asa conclusive proof of the absence of sesame oil.The estimation of tartaric acid in the presence of malic and suc-cinic acids is a frequently recurring problem in the analysis of food-stuffs and other products, and chemists will no doubt criticallyexamine a method recommended by von Ferentzy.10 This is basedon the insolubility of basic magnesium tartrate in aqueous alcoholof 50 per cent.strength, and the ready solubility in that solvent ofthe corresponding salts of malic and succinic acids. Two otherpapers dealing with the same subject which are worthy of attentionare by Jorgensenll and by Kunz and Adam.12 Some interestingwork has been recorded during the year by T. B. Wood13 in con-nexion with the important question of the " strength '' of wheat flour.Perhaps the most noteworthy observations are those from whichi t is sought €0 show that the strength of flour is largely conditioned8 Lewkowitsch, i n a private communication to the writer of this Report, sbtesthat he has never yet met with a sample which failed in this respect, although hehas examined many which were old and undoubtedly rancid.Chem, Zeit., 1007, 81, 1023,Ibidr, 312,lo Abstr., 1907, ii, 991,la Zeitscb.anal. Chem., 1907, 40, 261,la d b 8 h ) 1907, ii, 810, and J, Agric A'ci~, 1907, 2, 139218 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.by the ratio of protein t o mineral matter, small quantities of certainsalts having, as is well known, an appreciable effect on the physicalproperties of the proteins. The extension of this work will bewatched with interest. The estimation of creatine and creatininein meat and other extracts has attracted the attention of severalchemists during the past year. Baur and Barschall l4 appear t ohave been the first t o apply the colorimetric method of Jaff6 (basedon the red coloration produced when picric acid is added to analkaline solution containing creatinine) to the examination of meatextracts, and their results were subsequently extended by Grindleyand Woods.15 Hehner16 has shown that in applying this methodi t is necessary t o ensure the presence of an excess of picric acid, andto match the coloration in as strong solutions as possible in orderto avoid the errors produced by the dissociation of the colouredcompound. The method affords a valuable means of distinguishingbetween extracts of meat and extracts made from yeast, since thelatter are practically devoid of either creatine or creatinine.It isperhaps worthy of note that extract of crab is also, according t oAckermann and Kutscher,17 free from those bases. Winton andBailey 18 point out that meat which has undergone even slight decom-position yields volatile sulphur compounds when distilled with phos-phoric acid, as in the ordinary estimation of sulphites.These areoxidised by bromine water, and so lead to an over-estimation of anysulphurous acid present. A method for obviating this difficultyis proposed, and the matter is deserving of the attention of foodchemists, as it may apply to other subst'ances than meat, and isespecially of importance in countries like the United States ofAmerica, where the presence of sulphites in food-stuffs is regardedso seriously. Woodman and Talbot l9 call attention to the presenceof fluorine as a very general constituent of malt, and consequently ofmalt liquors, and suggests one part in one hundred thousand as amaximum limit for fluorine normally present.For distinguishingbetween natural and artificial colouring matters in wines, Jean andFrabot 2o recommend warming with formaldehyde and hydrochloricacid, and the subsequent addition of ammonia. A colourless filtrateis said to be obtained with all natural wines. Of a number ofpapers which have been published dealing with the analysis ofpotable spirits, reference need only be made to one by Bedford andJenks21 on the estimation of the higher alcohols. These authorsare probably correct in stating that, whilst t'he results obtained byl4 Arb. Kaisl Gesun,dhcitsnmt, 24, 562.l6 Pharm. J., 1907, 78, 683.' 8 J. Amw. Chena. Soc., 1007, 29, 1499.l.5 Abstr., 1907, ii, 187.l7 Abstr., 1907, ii, 283.Ibid., 1362.ao Abstr., 1907, ii, 320.31 Jbid., 405ANALYTICAL CHEMTSTRY. 219the Allen-Marquardt process are good in the case of amyl alcohol,the intermediate alcohols, such as butyl and propyl, largely escapedetermination, and it is obvious that isopropyl alcohol cannot beestimated a t all. The method they propose resembles that ofBeckmann, in that the alcohols are extracted by carbon tetra-chloride, and then converted into nitrous esters. Instead, however,of estimating these by oxidation with permanganate, the authorstreat them with potassium iodide, and determine the liberatediodine. This communication is worthy of attention, but theauthors' claim to distinguish between amyl alcohol, on the one hand,and propyl and butyl, on the other, will scarcely be admitted byanalysts generally.Many papers have been published during the year describing newalkaloidal reactions (usually colour reactions), but these are scarcelyof sufficient importance to merit special reference.An exceptionmay, perhaps, be made in favour of two papers by ReichardF2 whodescribes somes new reactions of scopolamine and yohimbine.Knorr23 appears to have been the first t o suggest the use of picro-lonic acid (4-nitro-l-p-nitrophenyl-3-methyl-5-pyrazolone) as a preci-pitant for alkaloids ; Warren and Weiss 24 have experimentedwith this substance, and find that it is a more delicate test thanpicric acid for coniine, strychnine, and morphine, and Mattbes and.Rammstedt 25 have applied this reagent t o the assay of nux vomica,hydrastis, and jaborandi, and their galenical preparations.Theresults appear, as a rule, to be good, and the method is likely t o beuseful. The same authors26 have also obtained good results whenusing this reagent in the estimation of narcotine, codeine, and mor-phine. Tschirch and Edner 27 have investigated the constituents ofEnglish and' French rhubarb, and have shown that a diazotisedp-nitroaniline solution constitutes a useful reagent for the evaluationof that drug. The results when working with pure emodin wereaccurate, and the method appears t o mark a distinct advance on thecolorimetric process. It may be pointed out that the found per-centages of hydroxymethylanthraquinones (expressed for conveni-ence in terms of chrysophanol) range from 4.24 per cent.in the caseof the Shanghai t o 1.58 per cent. in the case of the Frenchproduct.Bougault28 has shown that the reagent (a solution of hypophos-phorous acid in hydrochloric acid) which he previously recom-mended for the detection of cacodylic and methylarsonic acids isapplicable t o the detection and estimation of arrhenal (disodium23 Abstr., 1907, ii, 915.26 Ibid., ii, 592.23 B i d . , 1897, i, 314. 24 Ibid., 1907, i, 869.26 Ibid., 1908, ii, 75.27 h3id,, 1907, ii, 501, 525. T&bz'Cl., 828220 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.methylarsonate) and atosyl (sodium anilinoarsonate). The authorfinds that the black substance which is produced when the above-mentioned reagent is allowed t o act on arrhenal has the formula(MeAs) and that it is quantitatively oxidised by nitric acid or byiodine to methylarsonic acid in accordance with the equation:MeAs + 41 + 3H20 = MeAsO(OTi),+4HI.The atoxyl compound issimilarly oxidised, and may be estimated in the same way. The sameauthor 29 finds that methylarsine di-iodide and oxide may be esti-mated by titrating their aqueous solutions with a solution of iodinein potassium iodide.Toxicological Analysis.Calvi and MalacarneM show that alcohol materially retards thedecomposition of cyanides, so that hydrocyanic acid may be detectedin organs which have been preserved in alcohol, even a t the end ofsome considerable time. Schaefer 31 has investigated the naturaloccurrence of arsenic in the human body, and has detected veryappreciable traces in the hair, skin, liver, kidneys, brain, and thyroidglands, and has thus confirmed the conclusion arrived at by otherworkers that arsenic in traces must be considered a ' I normal "constituent of the human organism.He expresses the opinion thatthe arsenic is localised largely in the nucleins, a suggestion whichseems t o find some support in the marked readiness with which theyeast cell takes up small amounts of arsenic. These results, support-ing as they do those already arrived a t by W. Thomson and others,appear t o indicate that no special significance can be attached t othe discovery of traces of arsenic in many of the organs of the body.Bolland and Franzos,32 in an investigation of two cases of phos-phorus poisoning, where the bodies had been buried for four months,found that it was impossible to obtain any definite indications ofthe presence of phosphorus by either the Mitscherlich or the Dusart-Blondlot method. During recent years, the guaiacum test f o r oxy-haemoglobin has been studied by a number of investigators, especi-ally with reference to the production of the coloration by othersubstances than blood.Bolland33 finds that the presence of tracesof iron in the solvents which are ordinarily employed in the treat-ment of blood stains gives rise to the well-known blue colour, butthat this is inhibited by citric acid. I n the presence, however, ofoxyhEmoglobin, the blue coloration is obtained, citric acid notwith-standing. Buckmaster,s4 dealing with the same subject, points outthat the reaction is due to the iron of the haemoglobin, iron-freederivativeg of that substance, such as hamatoporphyrin, failing toAbstr., 1907, ii, 916, Eo Ibid,, 409.Ibid., 371,34 Ibid.9 60Q1 sz Chcnz, Bitaj 190'1, 31, 8, 83 Absfr,, lQO9, ii, §GOANALYTICAL CHEMISTRY. 221give it. I n confirmation of this, Lesser35 finds that the blood ofinvertebrates, which contains no hzmoglobin, does not give thereaction. The benzidine test for blood proposed by Schlesinger andHolst has been investigated by Utz,3G who finds that it is moredelicate than the guaiacum reaction, provided that the reagents arefreshly prepared. Van Rijn,37 in the course of a study of the distri-bution of morphine in the animal organism, administered 200 milli-grams of the hydrochloride t o an animal, which was killed six hourslater.By the ordinary process of extraction, about 84 per cent.of the total amount administered was recovered, and the authorshows that the alkaloid tended t o accumulate very largely in theurine.A ppuratus.The following list contains reference only to those new pieces ofapparatus which have been described in recognised journals, andwhich appear to be of real utility. The titles are not alwaysexactly those given by the authors, but have been, in some cases,more or less altered in order to indicate more clearly the nature oftho apparatus in question.- In connexion with apparatus, it may be mentioned that Heraeus 38has shown by direct experiment that hydrogen from the flame of aBunsen burner diffuses through platinum, and may bring about thereduction of ferric oxide and other reducible substances. Theseexperiments throw a good deal of light on the corrosion and ultimatedestruction of platinum crucibles.‘( The Eennicott-Sargent Colorimeter ” (Cheem. Engineer, 1907,5, 213).“ A portable milk colorimet,er.” A.Bernstein (Chem. Zeit., 1907,31, 727).“ New melting point thermometer.” G. Muller (Chem. Zeit.,1907, 31, 571).(‘ New apparatus f o r determining melting points.” J. Thiele(Abstr., 1907, ii, 330).“ A calorimeter for volatile liquid fuels.” W. H. Rawles ( J . SOC.Cl~ein. Ind., 1907, 26, 665).“ A new portable gas generator.” A. W. Browne and M. J.Brown (Abstr., 1907, ii, 678).‘‘ A constant pressure gas generator.” A.W. Browne and M. J.Brown (Abstr., 1907, ii, 679).(‘ New gas-generating apparatus.” E. Steiger ( A bstr., 1907, ii,339).3B Absts.., 1907, ii, 827. 3G Ibid., 916. 37 Ibid., 995.38 Ibid., 969222 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.“ New gas-generating apparatus.” A. Burger and M. W. Neufeld‘( New apparatus for gas analysis.” 0. Pfeiffer (Abstr., 1907, ii,(‘ A gas-developing apparatus.” A. Kleine (Abstr., 1907, ii, 446).(( An improved gas generator.” F. Southerden (Chem. News,1907, 95, 207).(( Apparatus for crystallising and filtering in indifferent gases.”W. Steinkopf (Abstr., 1907, ii, 161).‘I A new absorption and washing apparatus f o r gases.” B. Phyl(Zeitsch. a n d . Chem., 1907, 46, 150).( I New absorption vessels for elementary analysis.” 0.Carrasco(Ckem. Zeit., 1907, 31, 342).(‘New form of potash bulb.” P. Malherbe (,4nn. Chim. anal.,1907, 12, 318).( ( Drying apparatus for elementary analysis.” 0. Mittelbach(Chem. Zeit., 1907, 31, 551).(( Simple apparatus, with stirrer, for treating a liquid a t itsboiling point with two or more gases.” N. L. Gebhard (Proc., 1907,23, 34).(‘ Boiling and distillation of foaming liquids.” R. Fanto (Zeitsclt.angew. Chem., 1907, 20, 1233).Two fractionating columns for readily volatile liquids.” H.Schlemmer (Chern. Zeit., 1907, 31, 692).(‘ Apparatus for distillation and desiccation a t low temperatures.”D’Arsonval and Bordas (Ann. Chim.. anal., 1907, 12, 4).(‘ Modifications of Dennstedt’s method of combustion analysis.”R.Baumert (,4bstr., 1907, ii, 909).(( Platinum resistance furnace for melting points and cornbus-tions.”(( A simple fat extraction apparatus.” G. S. Fraps (Abstr., 1907,ii, 314).‘(New extractor for the determination of fat by means of carbontetrachloride.” F. Vollrath ( A bstr., 1907, ii, 514).‘ I An improved fat extraction apparatus.” J. van Leeuwen(Chem. Zeit., 1907, 31, 350).“Extractor for use with small quantities of material.” C. L.Jackson and J. E. Zanetti (Abstr., 1907, ii, 859).((A new eudiometer for use in the analysis of mixtures of airand inflammable gases.”(( Some new ureometers.” M. E. Pozzi-Escot (,4bstr., 1907, ii,414; compare Albert Garcia, Abstr., 1907, ii, 994).( ( Modification of Regnard’s ureometer.” M.E. Pozzi-Escot(Abstr., 1907, ii, 724).(Ahstr., 1907, ii, 339).194).S. A. Tucker (Abstr., 1907, ii, 842).J. Meunier (Abstr., 1907, ii, 989)ANALYTICAL CHEMISTRY. 223“ Standardisation of Engler’s Viscosimeter ” (Zeitsch. angew.‘ I A modified Engler viscosimeter.” L. TTbbelohde (Ch em. Zpit.,“ A new filter paper of cotton cellulose ” (Chem. Zeit., 1907, 31,“ A simple arrangement for the filtration of alkaline liquids (in“ Pressure filter for laboratory use ’’ (Chem. Tracle J., 1907, 41,‘’ A new sublimation apparatus.” V. Schworzoff (‘4 b s f r . , 1907,” A self-filling burette.”‘‘ Burette reading.”If Arrangement for reading thermometers, burettes, &c.” L. H.Zeller (Chem. Zeit., 1907, 31, 115).“ Water-jet blower of very simple ~onstruction.’~ S.M. Revingtonand I. G. Rankin (Abstr., 1908, ii, 30).“ A convenient air-bath and hot-plate.” E. D. Campbell ( A bstr.,1907, ii, 446).“Asbestos wire gauze.” €1. Carliczek (Ckem. Zed., 1907, 31,500).“ Simple lamp for monochromatic light.” E. Beckmann (Abstr.,1907, ii, 209).‘‘ A new centrifugal apparatus for laboratory use.” T. Eorner(Abstr., 1907, ii, 161).Improved Liebig’s condenser.” H. R. Ellis ( A bstr., 1907, ii,“Some new forms of apparatus.” W. M. Dehn ( A b s t ~ . , 1907,“ Portable photometer.” J. A. Evans ( J . Amer. Chem. Soc., 1907,‘‘ A mercury joint in place of cork or rubber in organic analysis.”“Device for filling bottles from carboys.” R. M. Hughes and“The use of nickel crucibles in quantitative analysis, and theR.Kriiian (Abstr.,‘‘ Apparatus for absorbing acid vapours given off during the assay‘‘ Apparatus for obtaining standard temperatures in dryingChem,, 1907, 20, 832).1907, 31, 38).465).air free from carbon dioxide).305).ii, 160).R. Rinne (Abstr., 1907, ii, 447).N. J. Lane (Abstr., 1907, ii, 390).P. Kusnetzoff (nbs.fr., 1907, ii, 809).160).ii, 755).29, 1009).J. Marek (Abstr., 1907, ii, 909).C. Barrow ( J . Amer. Chem. SOC., 1907, 29, 241).composition of the so-called ‘‘ Nickel Soot.”1907, ii, 390).of gold, silver, &c.” Dard (Abstr., 1908, ii, 72).ovens.” J. Habermann (-4 bstr., 1908, ii, 17)224 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.‘ I Collodion membranes for dialysis.” S. L. Bigelow and A. Gem-bcrling (Abstr., 1907, ii, 933).“ Simple apparatus for analytical purposes.” J. McC. Sanders(Proc., 1907, 23, 232).New mercury-drop tensimeters.” W. M. Dehn ( J . Amer. Chem.Soc., 1907, 29, 1052).“Method of obtaining the flame spectrum of metals.” G. A.Hemsalech and C. de Watteville (Chem. Zed., 1907, 31, 693).‘I An automatic vacuum regulator.” A. E. Andrews (Chem. News,1907, 96, 76).“ A simple means for the recognition of the colour of small quan-tities of faintly-coloured liquids, and its use in microchemicalanalysis.” F. Emich and J. Donau (Abstr., 1907, ii, 809).“ Apparatus for the estimation of carbon dioxide in carbonates.”P. Malherbe (Abstr., 1907, ii, 719).‘( Apparatus for the supply of carbon dioxide in the determina-tion of nitrogen in organic compounds by the absolute method.”G. Young and B. Caudwell (Abstr., 1907, ii, 394).Apparatus for use in testing substances for phosphorus.” J.Habermann (Abstr., 1908, ii, 17).‘( Apparatus for the continuous preparation of pure oxygen foruse in organic analysis.” A. Seyewetz and L. Poizat (Abstr., 1907,ii, 162).“Improved form of apparatus for the rapid estimation of sul-phates and salts of barium.” W. R. Lang and T. B. Allen (Proc.,‘‘ Apparatus for the estimation of sulphur (in iron).” VonNostitz and Jankendorf ( A bstr., 1907, ii, 393).“ New laboratory method for the preparation of hydrogen sul-phide.”“ Apparatus for the gasonietric determination of hydrogen per-oxide.” W. M. Dehn (Abstr., 1907, ii, 906).“ Apparatus for determining the coefficients of expansion of oilswhich solidify a t low temperatures (crude petroleum).” J. Gruszkie-wicz and W. Bartoszewicz (Petroleum, 1907, 2, 525).“ Electrolytic oxidation in quantitative analysis.’’ 0. Gasparini(Abstr., 1907, ii, 650).“ Apparatus for the comparative observation of fluorescence.”H. Ley and H. Gorke (Abstr., 1907, ii, 920).“ A Filter-tube.” P. W. Shimer (Chem. Engineer, 1907, 6, 197).“ Laboratory apparatus f o r the evaporation of liquids by radiation‘‘ A gas-generating apparatus.” Eugen Miiller (.4 bstr., 1908, ii,m 7 , 23,187).F. R. L. Wilson (Proc., 1907, 22, 312).from above.”129).H. J. S. Sand ( J . SOC. Chem. Znd., 1907, 26, 1225)ANALYTICAL CHEMISTRY. 225As year by year the number of published papers dealing withanalytical subjects increases, the responsibility and difficulty attach-ing to the task of presenting this report become greater. It is clearthat exigencies of space render it impossible to refer to a very largenumber of communications, many of which record quite usefulobservations, and the author is compelled to select for notice thecomparatively few which, in his judgment, are of special importancein relation either to theory or laboratory practice.ALFRED CHASTON CHAPMAN.REP.-VOL. IV