ANALYTICAL CHEMISTRY.THE problems of analytical chemistry cover such an exceedingly widefield, and are, if I may be permitted to use the expression, of such ttheterogeneous cha,racter, that it is ’ practically impossible to presentthis report in ths form of a continuous and connected narrative.Advances of distinct significance often consist merely in the detectionof some sour*ce of error in a more or less well known process and itselimination or in a slight modification of some existing piece ofapparatus. Trivial as these matters may appear when compared withsome of the discoveries made in other branches of our science, they arenevertheless often of prime importance to the analyst, and as such claimattention in this review. I n place of the unbroken and more readablestory which the author would have liked to present, he ie, from the verynature of the subject, compelled to deal largely with disconnected andisolated facts, and he can only hope that in consequence of the arrange-ment of the subject inatter which he has adopted he has been able topreserve in parts some semblance of sequence and continuity.Theadoption of a definite plan is as necessary for the Reporter as f o r theReader, and it is believed that the arrangement of the subject followedlast year is that which lends itself most readily to clear andmethodicaltreatment. The subjects referred to in this report will therefore badealt with 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 arbitrary character of this or indeed of any other subdivisionwill be evident, and in some cases there is perhaps no good reason whya particular process should have been discussed under one headingrather than under some other. Still the method has certain practicaladvantages and appears to the author to answer its purpose well:Inorganic Anulysis.It comparatively rarely happens that new reactions likely to be ofreal service in qualitative inorganic analysis are recorded, althoug200 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.there is unquestionably a large field still open for investigation in thisbranch of analytical chemistry, more especially in connexion with thedetection of some of the rarer elements and the easier identification oftraces of one metal in the presence of large quantities of another fromwhich it cannot be easily separated by existing methods.The import-ance of studying the nietallic derivatives of organic compounds fromthis point of view is now thoroughly recognised, and observations arefrom time to time recorded which are usuallyinteresting, if not alwaysof practical utility. Bradley finds that the hmnatoxylin reaction ofcopper, which has long been known, although but little used, is farmore sensitive than that given by potassium ferrocyanide or even bystarch and potassium iodide, and in the same paper points out thattraces of zinc in tissue ashes containing relatively large quantities ofcopper, iron, calcium, and phosphoric acid may readily be detected bytaking advantage of the fact that zinc nitroprussitle differs from all theother insoluble nitroprussicles in readily forming well defined crystalswhich can be recognised by means of the micrascope.According toGrossmann and Schuck dicyanodianiide constitutes a fairly delicatetest for nickel even in the presence of a layge excess of cobalt, althoughits sensitiveness is evidently very much inferior to that of Tschugaeff’sreagent referred t o below.For the detection of gold and platinum, more especially whenpresent in small quantities in the presence of other metals, Petersen 3outlines a scheme which appears to be capable of useful application,although it does not embody any new principle.The solvent action ofboiling sulphuric acid on platinum in the presence OF ammoniuirisulphate has been made the subject of several papers by Delkpineduring the past few years, and in a recent commnnication4 he statesthat platinum-iridium alloys are appreciably soluble in strong sulphuricacid at a temperature of 365O, and that on boiling the resulting soln-tion with ammonium sulphate the platinum is deposited, the iridiuiriremaining in solution. This solution, which appears t o containammonium iridium sulphate, exhibits the green colour becoming deepviolet on the addition of nitric acid, which is R characteristic reactionof iridium. The colour is so intense that i t is said to be possible byits means t o detect iridium in many samples of coiniiiercial ‘pureplatinum.’The detection of small quantities of yellow phosphorus in phos-phorus trisulphide (and in match heads) has been made a subjectof investigation by several chemists, and an interesting process hasbeen devised for this purpose by Schenck and Scharff,‘) depending on1 Amcr. J.Sci., 1906, [iv], 22, 326.3 Zeil. anal. Chem., 1906, 45, 342.a Bcr., 1906, 39, 3356.Conzpt. nnd., 1906, 142, 631.Bey., 1906, 39, 1522ANA4LYTICAL CHEMISTRY. 201the increase in the rate of discharge of an electroscope when subjectedt o the action of air in which the sample containing phosphorus hasbeen slowly oxidised. The method appears to be capable of being use-fully employed in toxicological analysis: and might be useful for thepurpose of controlling the purity of the air in rooms or factories inwhich phosphorus is used.The detection of sins11 quantities of nitric oxide and ozone in thepresence of one another is not an easy problem, and I?.Fischer andMarx * show that this may be done by leading the mixture into liquidair, which dissolves the ozone and solidifies the nitric oxide. Afterseparation by filtration, the two substances may be identified by theemployment of wet '' tetramethyl-base paper " (4 : 4'-tetramethgldi-aminodiphenylmethaae). The use of liquid air in this connexion isinteresting.The nitroprusside test for sulphides is not perhaps a very importantone, but it is occasionally useful, and it is well that its limits ofsensitiveness and the extent to which other acid radicles may interfereshould be understood.Following Reichard, Virgili 2 has studied thereaction, and has found that the blue coloration, which is preventedby the presence of free alkali, is also very considerably interfered withby the presence of silicates, phosphates, borates, and other salts whichare capable of yielding alkali on hydrolysis. It is, of course, wellknown that even under the most favourable conditions sodium nitro-prusside constitutes a less sensitive reagent for sulphides than neutralor alkaline solutions of lead salts. This, according to the author,appears to be. due to the fact that nitroprussides are not reagents forthe sulphide ion, but for the non-ionised metallic salt, any conditionwhich tends to prevent ionisation at the same time increasing thedelicacy of the reaction.In the quantitative section of this branch of analysis much goodwork has been done during the year.The electrical conductivitymethod has in several cases during recent years 'been successfullyapplied to the determination of the solubilities of very sparinglysoluble substances, and, working in this way, BGttgeu.3 finds that thesolubilities of the chloride, thiocymate, and bromide of silver in waterat 100" are respectively 153 x 10-6, 39 x 10-6, and 20 x gram-equivalents per litre. Every analyst knows from experience how greatan error may be introduced into estimations involving the weighingof silver chloride if that substance is washed with unnecessarilylarge quantities of hot water, but it is well to realise than an aqueoussolution of silver chloride saturated at 100" contains no less than2.18 mg.per 100 C.C. I n this connexion it is interesting t o note thatBcT., 1906, 39, 2555. Zcit. a d . Clicm., 1006, 45, 409.Zeit. pJqsiktsl. C h n . , 1006, 56, 83202 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Donaul recommends the application of the conductivity method tothe estimation of small quantities of gold and palladium in solutionsreduced by means of :carbon monoxide, and shows that even withexceedingly dilute solutions (0.0005 per cent.) the mean error is verysmall. Processes involving the employment of delicate electricalappliances are not as a rule well suited to the needs of the analyticalchemist, but they appear to the author to be worthy of special notice,as indicating a praiseworthy tendency to press new methods into theservice of analytical chemistry.For this reason the observation ofDrapier,2 that a galvanometer may be used as an " indicator " in thevolumetric estimation of silver is noteworthy. The estimation ofmoisture in substances containing other constituents which volatiliseon heating is a very common problem in the analytical laboratory andoften one of serious difficulty. I n a very interesting paper on thissubject P. V. Duprk3 describes a method based on the interaction ofthe water and calcium carbide. The resulting acetylene is measured,and the process appears to be not only accurate, but capable of wideapplication.Thus it can be employed for the estimation of water in crys-tallised salts and of moisture in cordite and i n mixtures containingvola-tile constituents such as camphor or naphthalene. There also appears tobe some foundation for the hope that it may enable the analyst todistinguish between moisture and combined water. The use of thenephelometer for the estimation of opalescent silver chloride pre-cipitates is dealt with by Wells4 and by Richards,5 both of whomindicat'e the chief precautions which must be observed in ~ephelometricwork. Horn,G and the same author and Sue. A. Blake,7 have made adetailed study of the question of sensitiveness in colorimetry, and havecalled attention t o a fact which many analysts overlook, namely, thatin every colorimetric process there is one dilution which gives themaximum degree of sensitiveness and consequently the most accurateanalytical results.I n their last communication the authors enunciateseveral generalisations in regard to colorimetric measurements whichdeserve the attention of analysts. The appearance of a further paperby Sorensen and AndersenS on the selection of a substance for thestandardisation of solutions employed in acidimetry goes to justify theuse in my previous report of .the word " perennial " as applied to thissubject. The authors still advocate the use of sodium oxalate inopposition to Lunge, who prefers sodium carbonate, and deal with theimportant question of the choice of indicator. Acree and Brunelghave also devoted attention t o this subject, but their paper, althoughMonatsh., 1906, 27, 59.Analyst, 1906, 31, 213.Ibid., 510.Ibid., 253. lbid., 1906, 36, 195 and 516.Zeit. anal. Chem., 1906, 45, 217.BuU. Xoc. chint. Belg., 1906, 20, 148.Amer. Chem. J., 1906, 35, 99 and 508.Anzs~. C ~ G I I L J . , 1006, 36, 117ANALYTICAL CHEMISTRY. 203recording some useful observations, does not contain much that isreally novel. A new indicator consisting of a condensation derivativeof methylfurfuraldehyde which appears t o possess useful properties isdescribed by Fent0n.l Several authors have dealt with the standardisa-tion of solutions employed in iodimetric analysis and two papers-oneby Bruhns 2 and the other by Riegler 3 may be read with advantage.I n the former the conditions necessary f o r obtaining accurate resultswhen using potassium dichromate and iodide are cliscussed, and it ispointed out that potassium permanganate may be advantageouslysubstituted for the dichromate.In the second paper Kiegler advancesclaims on behalf of ammonium tri-iodate as a fundamental substancefor use in iodimetry and in volumetric analysis generally. This sub-stance which has the formula (NH4)Hz(I0,), is crystalline, anhydrous,and non-hygroscopic, and is very stable both as a solid and in solution.Directions are given for using this salt in alkalimetric titrations andf o r ascertaining the strength of its solutions by means of hydrazinesulphate, with which it reacts with evolution of nitrogen, which canbe measured. Mathewson and Calvin 4 describe a method for estirnat-ing hydrogen peroxide by titration with a solution of ferrous ammoniumsulphate or vice versd, the chief point of interest being the employmentof a solution of titanium potassium sulphate as indicator, the peroxidegiving with the titanium solution, as is well-known, a deep yellowcoloration due to the formation of a higher oxide. I n a preliminarycommunication Jannasch and Zimmermann 5 describe a method f o r thequantitative separation of iodine from chlorine and bromine in whichthe mixture containing the three halogens is dissolved in a solution ofhydrogen peroxide strongly acidified with acetic acid, and the liberatediodine distilled over with steam and collected in an ammoniacal solutionof hydrazine sulphate.I n a subsequent paper 6 Jannasch deals withthe separation of bromine and chlorine, and shows that from solutionsacidified with sulphuric acid and containing hydrogen peroxide theformer halogen may be distilled over in a current of carbon dioxide,leaving the chloride undecomposed. These separations, more parti-cularly that of iodine, appear to be very complete, and the methods willno doubt be rigorously tested by chemists under the varying condi-tions of analytical practice. A further communication is promised bythe same author on the application of the hydrogen peroxide methodto the separation of the three halogens when present together and willbe awaited with interest. According t o B u s c ~ , ~ nitrous acid can bequantitatively oxidised to nitric acid by hydrogen peroxide undercertain conditions, and on this observation he has based a process for2 Zeit.nnorg. Chenz., 1906, 49, 277.Amer. Chenz. J., 1906, 36, 113.Proc. Camb. Phil. S'oc., 1906, 13, 298.Zeit. n?pw. Chenz., 1906, 19, 845.Bcr., 1906, 39, 196. Jbid., 3655. 7 B i d . , 1401204 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.the estimation of nitrite and nitrate when present together, theformer being determined volumetrically by the permanganate method,and the total nitrate gravimetrically (after oxidation by hydrogenperoxide) by means of the “ nitron ” method, t o which refercnce wasmade in my previous rep0rt.lSeyewetz and Bloch find that when an aqueous solution OF a hypo-sulphite is added to a solution of silver chloride in ammonia, reductionof the silver t o the metallic state occurs quantitatively, and assulphites, which are formed in the reaction, and thiosulphates, do notinterfere, the method can be used with advantage for the estimationof hyposulphurous acid.Manning and Lang3 have studied the question of the estimation ofboric acid, alone and in the presence of phosphoric acid, and althoughthe paper in which their results are recorded does not contain mucht h a t is new, it is deserving of the attention of analysts.The authorsdeal with the separation of the boric acid as the trimethyl ester, andits subsequent estimation either as the barium salt or by the ordinarymethod. Their results indicate that both processes are capable ofbeing employed with a higher degree of accuracy than is usuallysupposed.Methods for the detection and accurate estimation of traces of anyof the elements are always welcome, especially in view of theirimportance in connexion with many of the problems of metallurgicaland physiological chemistry, and in the last report on the progress ofAnalytical Chemistry* I referred to the fact that Tschugaeff hadsuggested the employment of a-dimethylglyoxime as a delicate test fornickel. Armit and Harden5 have now applied this method t o thequantitative estimation of very small quantities of nickel in animaltissues and other organic substances.The nickel having beenseparated by methods which are only slight modifications of those inconlmon use, is estiiiiatecl colorimetrically hy means of Tschugaefl’sreagent.The advantages of this method over that in which ainmoniumsulphide is employed are that t’he scarlet coloration is much morecliaracteristic, and that traces of iron do not interfere. It is alsomore sensitive, since a distinct reaction is given with the 1/1000 mg.of nickel in 30 C.C. of solution. Cobalt gives with the sawe reagenta yellow coloration which only interferes seriously with the nickelestimation when the amount of cobalt present is more than twice thatof the nickel. The estimation of very small quantities of manganeseis often of importance, and Tarugi 6 has described a new method basetlon the fact that manganoua hydroxide dissolves in glycerol and thatEtcZI.SOC. ehim., 1906, [iii], 35, 293.Ann. &port, 1905, 186.6 Gnxctln, 1906, 36, i, 332.Ann. hkport, 1905, 2, 190.J. Xoc. Chcm bd, 1906, 25, 397.Proc. Rog. Xoc., 1906, 77, E, 420ANALYTICAL CHEMISTRY. 205the resulting solution readily oxidises with the formation of a rubyred colour, which can be made the basis of a colorimetric process. Thereaction appears to be capable of indicating as little as 0*0000005gram of manganese.Hydrazine has, during recent years, been recommended in a numberof cases for the volumetric estimation of certain of the metals, andRiminil has described, for the estimation of mercury ancl also ofpersulphates, volumetric processes involving the use of that reagentwhich appear to be accurate, and, in the case of persnlphates moreespecially, useful.The estimation of cadmium as snlphide leaves much to be desired inpoint of accuracy, and Baubiguy deals very fully with the conditionsnecessary for the exact conversion of the sulphicle into sulphate, aform in which, as many analysts know, this metal may, like zinc,with advantage be weighed.C. Goldschmidt 3 records the interestingobservation that cadmium is quantitatively prccipitntsd whensolutions of its salts are boiled in aluminium vessels in the presenceof a trace of chromium nitrate or cobalt nitrate, the aluminiumacting catalytically just as nickel and cobalt do in the case of goldand silver respectively. Whilst the volumetric estimation of zincwould scarcely he resorted to when p e n t accuracy is tlesired, it issometimes capable of being employed with advantage in technicalwoi-k.Two papers on this subject may be usefully referred to, oneby decker^,^ on the sodium sulphide method, ant1 the other byillurmann,5 who has studied the conditions under which the bestresults can be obtained by titrating with ferrocyanide, using uranylchloride as indicator.Czerwek ti describes a method for the separation of tin and antimonybased upon the formation of a double compound of stannic acid andphosphoric acid. The method appears to be not only convenient butaccurate, and will doubtless be submitted to a careful examination.The various gravimetric methods for the estimation of calciumhave been criticised in a paper by Brunck,7 who recommends theconversion of the ignited oxalate into fluoride by treatment withhgdrofluoric acid solution.Several papers have been published during the year dealing with theuse of zinc as reducing agent for ferric salts prior to titration withbichromate or permanganate.I n this connexion the author may,perhaps, be permitted t o point out the great advantage of usingpalladium-hydrogen (charged palladium) for this purpose, and to-I Atti R. Accnd. Lincei, 1906, [v], 15, ii, 320.Compt. rend., 1906, 142, 577, 792, and 959.:: Zeit. anal. Chenz., 1906, 45, 344.ti &it. anal. Chenz.,~1906, 45, 174.B d l . Xoc. chim. Belg., 1906, 20, 164.Ibid., 505. 7 Ibid., 77206 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRP.express his surprise that so few analysts should avail themselves ofthis clean and useful method.All who regard chemical analysis as an important branch of thescience of Chemistry and not merely as a useful a r t will read withmuch interest a paper by Bruni and Padoa 1 on the conditions affectingthe precipitation and solution of metallic sulphides. These authorshave experimentally demonstrated the correctness of Ostwald’s pre-diction, based on a consideration of the law of mass action and thehypothesis of electrolytic dissociation, that by causing hydrogensulphide to react under pressure, i t would be possible to precipitatefrom acid solutions the sulphide of those metals, for example, iron,cadmium, cobalt, and nickel, which, in ordinary circumstances, aremore or less readily soluble in acid. On the other hand, by diminish-ing the pressure cadmium sulphide is found not to be precipitatedfrom solutions which yielded precipitates under ordinary conditions.Following Stahler and Scharfenberg,2 Moser a has studied thephosphate method for the estimation of bismuth and its separationfrom certain other metals, and his paper may be read with advantage.Funk4 has studied the separation of iron from manganese, nickel,cobalt, and zinc, and in his second communication shows that theformic acid process in which the iron is precipitated as a basic ferricformate is preferable to the acetate method, since the metals remainingin solution after the separation of the iron are amenable to moredirecttreatment .Jannasch and Gottschalk 5 have studied the use of ozone in quanti-tative analysis, and have shown that it affords an accuratle method forthe estimation of manganese, and that this metal may be quantitativelyseparated by its aid from a number of others, including zinc andcadmium. The reducing activity of hydrogen has been made thesubject of further study by the author of this report and H.D. Law,Gand i t has been shown that the efficiency of the hydrogen obtained bythe interaction of metals and acids is dependent on a number offactors, both chemical and physical in their nature, but that amongthese the question of “ potential )’ or ‘‘ supertension )’ plays animportant part. A special study has been made of the Marsh-Berzelius process as applied to the estimation of traces of arsenic,and i t has been shown that whilst certain metals, such as platinum,iron, and copper, lower the “potential’) a t which the hydrogen isevolved and so diminish the sensitiveness and accuracy of themethod, other metals, such as cadmium and tin, do not produce thiseffect.I n this way the so-called ‘ insensitiveness ’ of many samples1 Atti R. Accnd. Liqzcei, 1905, [v], 14, ii, 525.3 Zeit. anal. Chem., 1906, 45, 19.Bcr., 1905, 38, 3862.Ibid., 181 and 489.t i Analyst, 1906, 31, 3. J. pr. Cheni., 1906, [ii], 73, 497ANALYTICAL CHEMISTRY. 207of commercially pure zinc which frequently contain traces of iron isexplained. Several papers dealing with the estimation of arsenic bythe modified Marsh method have been published during the year, andwhilst these do not as a rule contain anything that is really new, someof the authors, notably Vdmossy,l Gautier,2 and Bishop,3 eitherrecommend the addition of platinum or copper salts as “ accelerators,”or the use of copper-coated zinc in the hydrogen generation flask.Of course, when comparatively large quantities (several milligrams)of arsenic are in question, the error introduced is not, perhaps,serious, but when fractions of a milligram have to be estimated i t isfull time that the inadmissibility of such additions was recognised.Two papers which are worthy of attention and which deal with thetitration of solutions of liydrofluosilicic acid have been published,the one by Sahlbom and Hinri~hsen,~ and the other by Schucht andMoller.5 The influence of dissociation on the results obtained bytitration with alkali under varying conditions is very interesting andinstructive.Glucinuni salts, like those of ferric iron, chromium, and aluminium,liberate iodine from solutions containing iodide and iodate, thehydroxide being a t the same time precipitated.Glasmann G baseson this fact a method for the estimation of glucinum which appears togive very accurate results. The method described by the same author7for ttie separation of glucinum and aluminium is sometimes a usefulone, but, as has been pointed out by Friedheim,& it is old and perfectlywell known to analysts.Titanium trichloride, which appears to have been first employed inanalytical procedure by Knecht, is, as is well known, a powerfulreducing agent, and Rhend has applied it to the volumetric estimationof copper.The end point is sharper than in the iodide method,lO andthe test results are good.In the electrochemical branch of analytical chemistry there is com-paratively little of importance that is new to record, Owing tothe many obvious advantages which electrochemical estimations andseparations present to the analyst, a large and increasing amount ofattention is now being paid t o the improvement of apparatus and tothe sharper definition of the working conditions necessary for success,but the greater part of such work, important though it is, scarcelycalls for special notice in this report. Foersterll shows that byworking at a potential (2.05 volts) below that at which hydrogen is1 Bzdl.Xoc. chiin., 1906, [iii], 35, 24.a J. Amer. Chem. Xoc., 1906, 28, 178.5 Ibid., 3693. ]bid., 3368. Ibid., 3366. Jbid., 3868.9 Trans., 1906, 89, 1491. lo Gerlinger, Zeit. amyew. Chew, 1906, 19, 520.11 Ber., 1906, 39, 3029.Ibid., 207.Ber., 1906, 39, 2609208 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.evolved, it is possible to obtain very good results in the electrolyticestimation of copper in solutions slightly acidified with sulphuric acid,and t o separate that metal from others (cadmium, cobalt, nickel, iron,and zinc) which reqnire higher voltages for their deposition. Priceand Judge 1 have investigated the deposition of zinc from zinc sulphatesolutions, using a rotating cathode, and fiacl that under certain definedconditions very good results can be obtained.The electrolytic pre-cipitation of gold from solutions of the chloride in presence of potassiumcyanide and of sodium siilphide, using a rotating anode, has been thesnbject of stitdy by Withrow,? who finds that the former is, if any-thing, the better electrolyte €or this purpose, althoiigh both arecapable of giving very good results. The same author has alsodevoted some attention to the electrolytic estimation of alkali halide,employing a method very similar t o that originally suggested by E. P.Smith,:: in which a silver anode is employed and the halogen weighedas silver halide. I n Withrow’s experiments, a rapidly rotating spiralcathode w : ~ employed and excellent resiilts were ohtninetl, the anodeconsisting of a silver-plated dish.In no branch of analytical chemistry is more ingenuity devotedto the devising of new appliances and to the improvement of old thanin that which concerns itself with the analysis of gases, and in nobranch is it more true that as a general rule the piece of apparatnswhich gives the best results in the hands of any operator is that t owhich the operator in question is thoroughly accustomed.It is there-fore very diacult in many cases to estimate correctly the value of anynew piece of apparatus without being practically acquainted with itsworking. Gautier and Clausmann* call attention to the difficulty inestimating carbon monoxide i n mixtures either by cuprous chlorideabsorption or by the combustion method, and advocate the employ-ment of iodic anhydride,5 a substance which is also recommended byNowicki? and by LBvy and P ~ c o u ~ , ~ for the estimation of smallquantities of carbon monoxide in air. FranzenS advocates the use ofan alkaline solution of sodium hyposulphite as an oxygen absorbentin gas analysis, and points out that in addition t o its cheapness itpossesses the advantage over the usual reagents of acting efficiently a tlow temperatures, and of not absorbing carbon monoxide.Collins 9describes an improved form of Scheibler’s apparatus for the gasometricestimation of carbon dioxide in carbonates, which eliminates severalimportant sources of error inherent in the original form, and permitsof its application t o problems in which a fairly high degree of accuracyJ.Amer. Chcna. Soc., 1906, 28, 1350.Conapt. rend., 1906, 142, 485.1 Trans. Fnraday SOC., 1906, 2, 85.3 Ibid., 1903, 25, 883.5 Compare Compt. rend., 1898, 126, 931.6 Oesterr. Zeit. Berg. Butt., 1906, 54, 6.8 Ber., 1906, 39, 2069,Compt. rend., 1906, 142, 162.J. SOC. Chcni. Ind., 1906, 25, 518ANALYTICAL CHEMISTRY. 209is demanded. Attention is drawn by Stock and Nielsen to the errorswhich may arise in the analysis of gaseous mixtures rich in one ormore constituents owing to the air, dissolved in the absorbing liquids,which is frequently present in quantities sufficiently great seriously tovitiate the results. Moureu 2 deals with the estimation of rare gases ingaseous mixtures obtained from natural sources, and describes theappamtns he employs.Organic Analysis.A good many papers have been published during the year dealingwith new reactions of organic compounds, but these, although oftenuseful, have, as a rule, such a very limited application that theyscarcely call for notice in this report.An exception may perhaps bemade in favour of the following. Sperling 3 describes a modification ofthe isonitroso-reaction of antipyrine and its more important deriv-atives which is said to be an improvement on the test as recommendedin the German Pharmacopceia. Direct tests for the identification of-the more important sugars in carbohydrate mixtures are greatlyneeded, and it is therefore to be regretted that Schoorl and vanK a l m t h ~ u t , ~ who have submitted to a critical study the colour reactionsdescribed by Pinoff,5 have arrived at the conclusion that these are notvery characteristic. Fenton 6 describes an extension of the bromo-methylfurfuraldehyde test for carbohydrates, which he published a fewyears ago, as he has found that this substance reacts readily withmalonic ester, yielding a strongly fluorescent product, which serves forthe detection of hexoses, or of compounds such as glucosides, whichyield those carbohydrates on hydrolysis.The qualitative detection ofsmall quantities of lsevulose in the presence of comparatively largequantities of dextrose (by other than optical means) is a particularlytlifficult problem, and one in which Pinoff's colour reactions appear tobe of little use, The results obtained by Mulliken in his study of theosazone test, more especially as regards the time required for theformation of a precipitate with the different sugars, have been con-firmed by Sherman and Williams,7 who have shown the influencewhich dilution and the presence of other sugars exert on the test asapplied to the detection of dextrose and Isvulose. Reference mayperhaps be made to a paper by Harang a on the detection and estima-tion of trehalose in fungi and other plants, since it involves the employ-ment of a specific enzyme (trehalase) and is a new instance of anindirect general method which is often of the greatest service inBcr., 1906, 39, 3389.Ibid., 1905, 38, 3308.C'ompt.rend., 1906, 142, 44.Bcr., 1906, 39, 280.Proc. Camb.Phil. Soe., 1906, 14, 24.J. PAnrm. Chim., 1906 [vi], 23, 16.3 Zeit. Oesterr. Apoth., 1906, [v],&, 51.7 J. Anter, Cheni. h'or., 1906, 28, 629.VOL. 111. 1210 AKNUAT, REPORTS ON THE PROGRESS OF CHEMISTRY.arriving a t a knowledge of the constituents of commercial carbohydratemixtures. The work of Hansen and others on the preparation of purecultures of the various species of Xccccharomycetes has, in fact, been ofthe greatest importance to the analyst as affording him a ready meansof obtaining the special enzymes necessary. I n the case of trehalaseit may be noted that advantage is taken of its secretion by the mouldAspergillus niger. Raikow and Urkemitsch point out that sodiumhydroxide gives a yellowish-brown coloration with nitrotoluene, butnot with nitrobenzene, and i t seems possible that the test may be ofservice in detecting small quantities of toluene in benzene.A test for‘( saccharin ” has been described by Kastle which is said to be verydelicate, and which appears to be worthy of notice inasmuch assalicylic and benzoic acids do not interfere and need not apparently beremoved. A delicate test €or indole is described by K ~ n t o , ~ and somecolour reactions for dietinguishing between proteins, indole, and scatoleare given by Steensma? who confirms the correctness of thosepublished previously by Rohde. The importance of such tests to thephysiological chemist is sufficient justification for referring to them inthis report.I n this connexion attention may perhaps be directed to amethod proposed by Herter and M. Louise Foster 5 for the quantitativeseparation of indole and scatole, the former being removed as auaphthaquinone compound. New alkaloidal colour tests are ever forth-coming, but the great majority of these are of little practical use, inthat they are either not sufficiently characteristic or are only definitewhen applied to the pure alkaloid, which is so rarely obtained inlaboratory practice. I n this branch of analytical chemistry Reichardis indefatigable, and it will perhaps suffice if notmice is taken in passingof papers by that author dealing with some new colour reactions ofberberine,G LZ new reaction for morphine,7 some new tests for cocaine,*for q~inoidine,~ and for thebaine,lO I n papers by Bredemann l1 on thealkaloids of the rhizome of Veratrum album, a number of reactions ofthe various alkaloids present are described.Lemaire describes somedistinctive reactions of alypine (benzoyltetrnmethyldiaminopentanolhydrochloride) which enable one readily to distinguish between thatsubstance and cocaine, stovaine, and other active therapeutic agents.The identification of artificial organic dyestuffs, more especially whentwo or more are present, is a problem which is daily becoming moredifficult. Various schemes have been proposed for this purpose,Chem Zeit., 1906, 30, 295.Zeit. physiol. Chent., 1906, 48, 186.5 J. Biol. Chem., 1906, 2, 267.7 Ibid., 247.10 Pharm. Centr.-H., 1906, 47, 623.1.2 Bep.Phnrm. 1906, 18, 385.Chem. Centr., 1906, i, 1575.IZlid., 1906, 47, 25.Pharm. Centr.-H., 1906, 47, 473,Ibid., 532.l1 Apoth. Zeit., 1906, 21, 41 and 53.8 Ibid., 347, and Pharm, Zeit., 1906, 51, 591ANALI’TICRT~ CHEMISTRY, 211notably by Witt, WeingWner and Green, and during the past year afurther contribution to the subject is made by Gulinoff,l who attachesspecial importance to the behaviour of the dyestuffs with variousreducing agents, which is generally recognised as affording one of thebest methods of classification.A number of investigators have, during recent years, attackedthe problem of the identification of small quantities of methylalcohol in the presence of large quantities of ethyl alcohol and withmore or less success.The majority of the proposed methods arebased on the production of formaldehyde, aud its recognitionby various colour tests. The difficulty is, however, increased bythe fact that these colour reactions are not always characteristic, andin addition a number of organic compounds other than methyl alcoholyield formaldehyde on oxidation, and Scndder and Riggs 2 point outthat this constitutes an objection to the recently published method ofLeach and Lythgoe, in which a hot copper spiral is used. Voisenet3has described an oxidation method in which the formaldehyde isrecognised by a test which is said to be quite characteristic if carriedout as recommended, but to what extent the second of the abovedifficulties may apply cannot be determined until the method has beenmore extensively tried.I n this connexion it may be noted that Leys 4has described a mercury reagent which is said t o admit of a readydistinction between acetaldehyde and formaldehyde in very dilutesolutions.That the method usually adopted for the combustion of organiccompounds in elementary organic analysis leaves much to be desiredin respect of convenience of working, neatness, and cleanliness isfully recognised, and a number of workers have introduced improve-ments more especially in the direction of employing “ contact ”substances and heating ^by electricity. I n niy previous report 5 refer-ence was made to papers by Dennstedt in which he advocated the useof an electric furnace and specially-devised tubes.These suggestionshave been criticised by Holde,G who has failed to obtain good resultsin certain cases with the form of furnace recommended by Dennstedt,but the latter author in a reply 7 draws attention to some practicaldetails, the observance of which is essential for success. The methodproposed by Carrasco,s and slightly modified by that author andP l a n ~ h e r , ~ is simpler, inasmuch as the combustion is brought about bya heated platinum spiral in an atmosphere of oxygen, and theJ. Amer. C?LC?IZ. Soc., 1906, 28, 1202.Ann. Chinz. anal., 1906, 11, 84.Ber., 1906, 39, 1615.Zeit. Parb. Text. Ind., 1906, 5, 337.3 BuZl. SOC. chin%., 1906, [iii], 35, 748.Ann,. Zeport, 1905, 195.7 Ibid., 1623.]Bid., 613.8 Atti A!. Accad. Lincci, 1905, [v], 14, ii, 608.P 212 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.apparatus is very compact and neat.The results are good, the work-ing expeditious, and the method is applicable to substances containingnitrogen, halogens, and sulphur. This method is similar in principle t othat devised by Morse and Taylor referred t o in my last report, andwhich has now been modified by Morse and Gray1 to adinit of thesimultaneous estimation of carbon, hydrogen, and sulphur. Thewpapers may be read with advantage by all chemists who are frequentlycalled on to make organic combustions, and who are working inlaboratories where the electric current is available.Vaubel and Scheuer 2 recommend for the estimation of halogens inorganic compounds a simple method in which the compound is heatedwith concentrated sulphuric acid in the presence of excess of sul-phurous acid.The halogen is obtained either as the element or as thehydracid, and is converted directly into the silver halide. This method,although not applicable to very volatile substances, appears t o be auseful one in many cases, and gives good results. For the samepurpose Stepanoff 3 proposes a method in which the substance isheated with metallic sodium and ethyl alcohol. The reaction proceedsaccording to the following general equation, in which " X " standsfor one of the halogens :RX + C2Hj*OH + Na, = RH + NaX + C2H,*ONa,and the method which is said to be almost nniversally applicable givesapparently accurate results.Some years ago Dnnstan and Carr found that the nitrogen inaconitine could not be accurately estimated by the ordinary Dumascombustion method, inasmuch as the nitrogen in the measuringvessel was always accompanied by methane.Haas4 finds that anumber of organic bases behave in a similar manner, but thatcorrect results can be obtained if lead chromate is substituted for thecopper oxide, and if the substance is mixed prior to combustion with asufficient quantity of cuprous chloride. I n view of the fact thatDumas' method is always regarded as the standard one, this com-munication is obviously of considerable importance. In order to avoidthe somewhat large percentage error inseparable from the weighing ofvery small quantities of carbon dioxide in potash bulbs, McFarlaneand Gregory advocate barium hydroxide absorption, as in the well-known Petteukofer's method, and the conversion of the bariumcarbonate into sulphate.When rather larger quantities of carbondioxide are in question, the authors make use of an ammoniacal bariumchloride solution for absorption. The principles involved are of coursewell known, but the method is one that might be more often employedAmey. Chna. J., 1906, 35, 451.Rer., 1906, 39, 4056.Chew. Nezcs, 1906, 94, 133.C'hem. Zcit., 1906, 30, 167.Trms., 1906, 89, 570ANALYTlCAL CHEMISTRY. 213in the estimation of small quantities of carbon, such as have to bedealt with, for example, in steel analysis.Watson Smith, jun,,l has introduced certain modifications intoStrache’s well-known phenylhydrazine method for the determination ofthe carbonyl groups in organic conipaunds which are said considerablyto improve the pyocess.The estimation of the percentage of methyl alcohol in commercialfornlalclehyde solutions is sometimes of importance, and for this pur-pose Blank and Finkeiibeiner 2 recomiiiend a process based on oxida-tion by means of chromic acid and the indirect determination of theamount of oxygen used.Vaniiio and Seitter’s method for the estima-tion of forrnsldehyde by oxiclation with permanganate iii acid solutionin the cold has been studied by Grossmtmn and Aufreclit,3 who findthat it gives good results if sufficient time is allowed for the com-pletion of the reaction. Formic acid is also completely oxidised inacid solutions by permanganate, but Rupp4 shows that the reactionproceeds more rapidly in alkaline solutions.For the estimation of acetone in wood spirit, crude acetone, aridsimilar liquids, Auld 6 has devised a new method depending on theformation of bromoform, its subsequent hydrolysis with alcoholicpotash, and the volumetric estimation of the bromine by iiieans ofsilver solution.This method appears to be capable of giving goodresults, and it possesses several obvious advantages over the variousniodificatiofis of the iodoform process.Jollesg has applied the bisulphite reaction to the estiniation ofacetone, and given a sufficiently lengthy period of contact the resultsare good.No year passes without cz number of new processes being put forwardfor the estimation of the reducing sugars.Occasionally some of theseare useful in special cakes, and frequently they are characterised byconsiderable ingenuity ; but generally speaking it may b e said that noprocess approaches in point of accuracy the gravimetric Fehlingmethod when properly carried out, and certainly none is so generallyaqplicable. C. A. Browne, j ~ i n , , ~ points out that the cupric-reducingpower of a sugar is constant for all concentrations if the excess ofcopper remaining in solution is kept constant, and that the ratio ofthe weights of two sugars which reduce the same amount of copperis a constant one at all concentrations. The author has determinedthe value of this ratio for a number of sugars (compared with dextroseas a standard) and finds, notwithstanding statements t o the contrary,that the “dextrose equivalent ” of a mixture of reducing sugars is*; Ibicl., 2455.CJ~ent. Arms, 1906, 93, 83.Zeit. anal. Chena., 1906, 45, 687.Ber., 1906, 39, 1326.J. SOC. Chem I?kd., 19Oti, 25, lG0.7 J. Amer. C h e m Xoc., 1906, 28, 439. ti Ber., 1906, 39, 1806214 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.equal to the sum of the ‘( dextrose equivalents” of the constituents,The disturbing effect of sucrose is dealt with in this paper, and also ina communication by Munson and Walker,l who have suggested a setof standard conditions for the carrying out of the reduction process,and have drawn up tables for the use of operators observing the con-ditions they lay clown.shows that rafinose can be com-pletely hydrolysed to melibiose and laevulose by citric acid undersuitable conditions, and has based on this fact a method for thepolarimetric estiniation of sucrose and raffinose when present together.Mlle. Talon 3 calls attention to a possible source of error in the estima-tion of sugar in liquids containing much alcohol, since during the acidhydrolysis process esters of dextrose are produced which appreciablyinfluence the results obtained by the reduction process or by thepolarimeter.The detection and estimation of adulterants in oil of turpentineconstitutes an analytical problem of some importance and frequentlyof great difficulty. The “bromine absorption” has often beenadvocated as a trustworthy test, but, as Utz4 has pointed out, thevariations in the genuine oil are great, and some resin oils so nearlyapproach turpentine in this respect that considerable proportionsmight be present, and yet not reduce the bromine value below that ofcertain pure oils.A paper has been published by Bohme 5 on thesame subject, in which an improved method is recommended for theestimation of paraffin and benzene hydrocarbons in turpentine by thewell-known sulphuric acid process. Valenta finds that methylsulphate readily dissolves aromatic hydrocarbons such as occur in taroils, but not paraffin hydrocarbons or resin oils, and has based onthis observation a method for the estimation of the former whenadmixed with the latter which promises to be of considerable service.It will be interesting to see whether the use of this reagent cannot beextended to t l e analytical examination of other hydrocarbon mixtures.I n connexion with hydrocarbon oils, attention may be directed to apaper by Ross and Leather 7 in which the problem of the valuation ofoils used for gas-making purposes is very fully dealt with.Theresults of the examination of a considerable number of oils by alaboratory method which the authors recommend are given and arecompared with those yielded in practical working,Much useful work has been devoted during the year to the study ofessential oils, not; the least important being the determination of theanalytical ‘constants’ of genuine samples of those oils, which on accountPieraerts1 J.Amer. Chem. &‘oc., 1906, 28, 663.2 Bull. Assoc. Chi7n. Sucr. Dist., 1906, 23, 1261.3 Ann. Chim. anal., 1906, 11, 244.4 Chem. Rev. Frtt. €€am. lnd., 1906, 13, 161.6 Ibid., 266. Analyst 1906, 31, 284.Chem. Zeit., 1906, 30, 633ANALYTICAL CHEMISTRY. 215of their high price are most liable to sophistication. Valuable as thisis to the analyst who is specially concerned with this branch ofanalysis, it is scarcely of suEcient general interest to justify referenceto it in this report.The accurate estimation of tannin in raw materials is another im-portant problem to the solution of which a considera6le amount ofenergy has been devoted. Wislicenusl has now arranged for themanufacture on the large scale of ‘‘ porous alumina” of uniformquality, and describes a form of apparatus by which the process ofanalysis can be best carried out.The absorptive power of thisalumina for colouring matters appears to constitute one objection t oits employment, and there seems to be an impression among those bestable to judge that it possesses no r e d advantage over the chromedhide-powder method, which in the form adopted by the AmericanLeather Chemists’ Association will probably be recognised as thestandard process. Those who are specially concerned with thisquestion will read with interest papers by Small,2 Parker and Bennett,3Procter and Bennett,4 and K ~ p e c k y . ~. Several papers dealing with the analysis of explosives have ap-peared during the year, but of these attention need only be drawn toone by Silberrad, Phillips, and Merriman on the direct estimation of‘ nitroglycerine ’ in cordite and allied explosives, and which is based onthe reduction of the saponification products of the ‘ nitroglycerine ’ andthe titration of the resulting ammonia.The estimation of indigotin in commercial indigo is obviously amatter of very great importance to the dyeing trade and judging fromseveral papers recently published the problem does not yet appear tohave been finally solved.Numerous methods have been from time t otime proposed, but only those depending on the oxidation or reductionof sulphonated indigos appear to be capable of yielding good results,and are in aaything like general use. I n a critical paper byBergtheil and Briggs it is claimed that the permnnganate methodas worked out by Rawson with slight modifications is capable ofyielding concordant and trustworthy figures.The use of bariumchloride as a precipitant is, however, shown to be inadmissible, andthe authors recommend the employment of freshiy precipitated bariumsulphate for the purpose of removing impurities. On the other hand,W. P. Bloxam * maintains that correct results cannot be so obtained,and recommends a process based on the separation of the indigotin aspotassium indigotintetrasulphonate. It is clear that, in spite of thelarge amount of valuable work done by the above chemists, a generally1 Collcgiu?n, 1906, 77.3 l b i d . , 1193. Ibid., 1203. Collegazm, 1906, 97 c t scp.J.Sot. Chenz. I d , 1906, 25, 296.6 J. SOC. Chem. Ind., 1906, 25, 628. Ibid., 729. Ibid., 735216 ,4KNTJA4L REPORTS ON THE PROGRESS OF CHERZISTRP.acceptable process has not yet been worked out, and the results offurther study of this subject will be awaited with interest.Am.!@ of Foods a i d Dmys.Dealing firkt with that most important food product, milk, severalpapers of importance to analysts have appeared during the gear. Themethod devised by T. E. Thorpe for the analysis of samples of milk(sour and otherwise decomposed) referred to the GovernmentLaboratory under the provisions of the Sale of Food and Drugs Acts,and which was noticed in my previous report has been submitted toa critical study by Richmond and Miller,2 who have shown that whilstin certain cases a somewhat closer approximation to the fresh milkvalues could be arrived a t by a n extension of the GovernmentLaboratory process, that method is capable of yielding results whichare substantially accurate.I n those comparatively rare instanceswhere unusually high proportions of butyric or propionic acids areformed, the method appears to be susceptible of improvement. It isgenerally admitted, however, that the process is one of greatpractical value, and it has to be borne in mind that i t could not becomplicated beyond a certain point without seriously reducing itsusefulness. I n this connexion it is interesting to note that accordingto Tice and Sherman3 the character of the decomposition changesoccurring in milk on keeping is largely determined by the nature ofthe added antiseptic when such substances have been used.I n thepxesence of 0-07 per cent. to 0.1 per cent. of formaldehyde, althoughbacterial action appeared to be entirely suppressed, extensiveproteolysis Loccurred with very little loss of milk-sugar, whilst inthe presence of sodium fluoride or salicylate a large proportion of thesugar underwent decomposition before any marked digestion of thecasein took place. Two other papers by Richmond are of interest inview of the preparation of so-called homogenised milk, and the veryextensive use which is now being made of dried milk or milk powder.In the first place, he gives a series of analyses which seem to showthat the Adam’s coil method when applied to the analysis of homo-genised milk gives results appreciably below those yielded by theGottlieb, Werner-Schmid, or Gerber methods, and in the second hecalls attention, itater &a, to the impossibility of extracting the f a tdirectly from milk powders and recommends the Werner-Schmidmethod. The danger of applying direct extraction methods to theanalysis of certain milk-containing mixtures, such as infants’ foods, isone which ought not t o be overlooked.1 Ann.Kq.mrt, 1905, 203.8 J. dmcr. Chcm. SOC., 1906, 28, 189.Analyst, 1906, 31, 317. ‘ A?zalyst, 1906, 31, 2i8, 219AN A LTTIC d L C H EXISTRY. 217I n his annual communication on the composition of milk Richmond 1states that the average percentage of fat in nearly 15,000 samplesanalysed during 1905 was 3.73 per cent.with 8.97 per cent. of non-fatty solids. These numbers are almost identical with those of theprevious year. For the determination of proteins in milk and ofcasein in cheese, Trillat and Sauton recommend a method based on thefact that formaldehyde renders the proteins insoluble without alteringtheir weight, and so permits of their gravimetric estimation. Schrott-Fiechtl3 has compared Gottlieb's, Gerbei-'s, and the Wollny refractometermethods for the estimation of fat in milk, and, as the result of a largenumber of analyses, has come to the conclusion that all three yieldquite trustworthy results. Siegfeld points out, however, that irkthe case of machine-separated milks, the cholesterol and lecithin whicliare extracted by the solvents used in the Gottlieb method introduce anappreciable error into the fat determination.If a simple chemicallriethod for the recognition of milk derived from diseased cows couldbe devised i t would obviously be of the highest importance from thepublic health point of view. The difficulties appertaining to thesolution of this problem will be apparent to all who have anyknowledge of the bacteriology of milk, and it will be interesting to seeto what extent a determination of the " catalase number )' as proposedby Lam in a paper communicated to the sixth International Congressof Applied Chemistry supplies this want. The detection andestimation of cocoanut oil in butter is a problem which is even moredifficult than was a t first supposed, alzd which cannot yet be said tohave been satisfactorily solved. I n my last report I called attentionto the work of K.Jensen, Kirschner, and 0. Jensen on this subject,pointing out that these authors had endeavoured to base, on thedifferences in the proportions of octoic and butyric acids occurring inbutter-fat and cocoanut oil respectively, and in the comparativelysparing solubility of silver octoate, a method for the detection andestimation of cocoanut oil. The hopes which were raised by thepublication of these papers have not altogether been realised, andmany analysts who have had much experience in this branch ofanalysis have been led to regard the so-called silver numbers as of verylittle value. Wijsman and Keijst have suggested a modification, orrather an extension, of the Jensen method in which the Reichert-Meissl distillate is divided into two unequal portions with the objectof obtaining inore widely separated solubility differences for the silversalts. The numbers given by the ,authors did not, however, appearC ' o q k m i d ., 1906, 142, 794, and 143, 61,Ibicl., 1.1 Annlyst, 1906, 31, 176.:$ Milchw. Zenh*., 1906, 2, 13.5 Chemist and Druggist, 1906, 68, 914.7 Zeit. ATnhr. GCnusSrn,, 1906, 11, 267.' AWL. RqmA, 1905, 204218 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.very convincing, and the process has since been adversely criticisedboth by Jean and by Luhrig.2 Lewkowitsch, in a private commiinica-tion to the author of this report, also states that he has obtainedconflicting and untrustworthy results.After all it would appear thatthe Polenske method and the phytosterol acetate test still furnishthe analyst with the most trustworthy indications, and if in additionthe iodine value and refraction numbers are determined, a very nearapproach to certainty is obtained. It may perhaps be pointed out inopposition to a rather widespread belief that the degree of fluidity ofthe insoluble volatile acids obtained in any of the distillation processesis of very little diagnostic value.Bellier4 proposes for the detection of cocoanut oil in butter amethod based on the precipitation of the fatty acids by means ofcopper sulphate solution, and claims that by its employment 5 percent. of cocoanut oil can be detected with certainty.The analyticaldifferences given by the author, and on which the detection ofadulteration depends, are very small, but the process will doubtless becarefully examined.A. W. Thorp5 has proposed to substitute for the methods ofPolenske and of Muntz and Coudon a simplified procedure which isreally only an extension of the ordinary Reichert-Wollny process.This is based on the same principle as that underlying the methodsabove referred to, but does not involve the use of special apparatus,and, judging from the test analyses recorded, is capable of giving veryuseful results. Jean calls attention to the somewhat extensiveemployment of Karitk or Fulwar butter (the fat obtained from theseeds of Bassia butyracea) for the purpose of adulterating butter, afact which still further complicates the analysis of that food product.He gives the analytical “constants ” of this fat, and indicates themanner in which the various butter values are affected by its use asan adulterant.I n connexion with the detection and estimation of preservativesin food-stuffs there is not much that; calls for special notice.It is well known that sulphur dioxide enters into combinationwith some of the constituents of certain foods, and that only aproportion of the amount actually present can in some instances beestimated by the ordinary distillation process.It is interesting tonote that in certain meats, for example, Holley7 was able t o recoveronly about one-fourth of the amount added as sodium sulphite.Forthe detection of fluorine in foods, Ville and Derriens propose a newZeit. Nahr. Genussrn., 1906, 12, 588. Ann. Chirn. apnal., 1906, 11, 121.Ann, Chirn. unul., 1906, 11, 412.Ann. Chiin. anal., 1906, 11, 201.Bull. s’oc. chiin., 1906, [iii], 35, 239.3 Compare Liihrig, Zeit. Nuhr. Genussm., 1906, 11, 11.Anulyst, 1906, 31, 173.’J; Anzer.. Chew SOC., 1906, 28, 993ANALYTICAL CHEMISTRY. 219method based on the change in the absorption spectrum of methEmo-globin brought about by the addition of fluorine compounds. Thetest is interesting on account of its novelty, but it remains to be seenwhether i t is capable of replacing t o any extent the methods now inuse. It is also of interest to note in passing that the well knowncolour reaction which formaldehyde gives with proteins in the presenceof sulphuric acid containing traces of certain oxidising agents, andwhich is so largely used in the testing of milk, depends on the presenceof the trytophan (indole) group, and that when that group is absent,as, for instance, in gelatin, no reaction is obtained.1 In view of thefact that in some countries the use of formaldehyde as a foodpreservative is either partially or entirely prohibited, the observationof Perrier that cider and various smoked foods may naturally containmore than 2 milligrams of formaldehyde per 100 grams is of import-ance, It is to be remarked that this observer has relied uponVoisenet’s colour reaction, and it is not impossible that the colorationshe obtained may have been due to some other substance than form-aldehyde.I n any case the observation is an interesting one to publicand other analysts who are concerned in the examination of foodstuffs under the provisions of adulteration Acts.The question of the detection of beef-fat in lard has occupied theattention of Dunlop,3 who has shown that the indications of theBelfield crystal test must be accepted with considerable caution, andthat the preseuce of ‘‘ plumose ” groups of crystals, as observed witha low magnifying power, is not to be taken as evidence of beef stearin,unless the characteristic form of individual crystals can be recognisedwith a higher magnification. I n view of the fact that lard obtained frompigs fed on cotton seed meal gives the Halphen reaction, this test is oflittle value as a proof of adultbration.Fnrnsteiner, Lendrich, and But-tenberg4 have made recent experiments on this point and have foundthat, although the lard from pigs fed on cotton seed meal often gave astrong Halphen reaction, the melting point of the acetate obtained byBomer’s method showed that cotton seecl oil itself could not have beenpresent. The glycogen in horse flesh apparently remains unchangedfor a long time, whilst that contained in beef, veal, and pork almostentirely disappears within a few days of the death of the animal, Onthis fact Martin5 bases a method for the detection of horse flesh insausages and potted meats, the glycogen being estimated by Pfluger’smethod. In this connexion attention may perhaps be directed to apaper by Pfliiger on the estimation of glycogen,G and to a series ofcommunications on the same subject by Desmouliiire.7Rosenheim, Biochon.J., 1906, 1, 233.J. SOC. Chaz. had., 1906, 25, 458.Ibid., 249.Qo?,zpt. rend., 1906, 143, 600.Zezt. Na7w. Genzcssm., 1906, 11, 1.PJugcr’s Arckiv, 1906, 114, 231.7 J. €’harm. Ckim., 1906, [vi], 23, 244, 281, and 332220 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.The detection and estimation of free mineral acid in vinegar is at thepresent time of academic rather than of practical interest. Attentionmay, however, be directed to a paper by Richardson and Bowen,l inas-much as these authors deal in some detail with the influence which thealkaline and earthy phosphates in the ash exert on the results obtainedby Hehner's well-known method, and indicate the nature of theapproximate correction to be applied.Much work has been devoted from time to time t o the questionOF the estimation of the higher alcohols in spirits, and muchstill remains to be done.Veley2, has recently made a specialstudy of the Rose-Herzfeld and sulphuric acid colorimetric pro-cesses, and finds that both, but inore especially the latter, areuntrustworthy. Schidrowitz and Kaye,3 in continuation of theircritical examination of some of the better known methods, havefurther studied the Allen-Marquardt process, and are of opinion thatcarried out under proper conditions it is capable of giving quite trust-worthy results. It is shown that the usual " mineral acid " titrationwith methyl-orange as indicator may be dispensed with when itamounts to less than one-tenth of the whole, as it is not due, as isusually supposed, t o hydrochloric acid, but apparently to organicacids, some of which are known to have an appreciable effect on themethyl-orange, The same conclusion has been arrived at by Alannand Stacy,4 who have also studied the influence of temperature onthe proportion OF ethyl alcohol extracted by the carbon tetrachloride.It is in the very nature of things impossible that an exact methodshould be devised for the estimation of ft group of substances such asthe so-called higher alcohols, but there can be no doubt that, of allthe methods which have been proposed, that of Marquardt, as modifiedby Allen, is for several reasons the best.It is at least based onscientific principles, and has the further great merit of yielding actualoxidation products which can be examined and, if desired, identified.Numerous papers dealing with the analytical examination of drugshave, as usual, appeared during the year, but the majority of these,whilst undoubtedly useful, are scarcely of sufficient importance orgeneral interest t o warrant a special reference.St,an&k5 finds that betaine and choline, which frequently occurtogether in plants, can be separated almost quantitatively as per-iodides, the betaine periodide only crystallising from acid solutionswhilst the corresponding choline compound also crystallises fromliquids which are neutral or slightly alkaline.The test analyses arefairly good, and the method appears to be a useful one.I J. Soc. C h m . 1/2d., 1906, 25, 836. Ibid., 398.3 Analyst, 1906, 31, 131. J. Sos'oc. C/icnt. T?icl., 1906, 25, 1125..%it. physiot. Chenz., 1906, 47, 83ANALYTICAL CHEMISTRY. 29 1The testing of glycerol for traces of arsenic is a problem of some-what frequent occurrence, and Galimard and Verdier 1 call attentionto the fact that arsenic is sometimes present in a form (1glyceryl-arsenite), in which it cannot be detected by the direct application ofthe Marsh method, preliminary treatment being necessary.Keller’s nitric acid oxidation method for the separation of brucineand strychnine has been studied by Reynolds and Sutcliffe,2 who findthat accurate results can be obtained by the process as modified byStoeder or by Gordin if certain details of procedure to which theydirect attention are observed.The testing of disinfectants, of which there are now so manyon the market, is clearly a matter of considerable importance, andthe adoption of bacteriological methods such as that of Rideal andWalker3 marks a great advance on the unsatisfactory chemicalmethods formerly in use.I n the above-mentioned process thegermicidal value of the disinfectant in relation to that of phenol(“ carbolic acid coefficient ”) is determined in water, but Lloyd * andM. W. Blyth5 have shown that in the presence of organicmatter, suchas milk or urine, quite different results are obtained, and that two dis-infectants may occupy very different positions i n a table of relative“eficiencies,” according as the test is made with water, or with, say,diluted milk.ToxicoZogicul Analysis.Only a few communications need be referred to in this branch ofanalytical chemistry.Ipsen-Innsbruck 6 has studied atropine from the toxicological pointof view, and finds that i t is rapidly absorbed by all parts of the bodyand distributed in the blood.He also finds that the alkaloid is pos-sessed of very great sta5ility in the presence of decomposing organicmatters, and that 0.03 gram of atropine, after remaining in contactwith decomposing blood for twelve years, could still be detected. Sardnand Caffart 7 describe a method of treating blood stains for purposesof identification, in which chlorohzematin is obtained, the crjstals ofwhich can be easily recognised by the aid of the microscope.Themethod is said to work well with very old stains, and to be preferableto those in ordinary use. I n this connexion, attention may be directedto a paper by CarlsonS on the guaiacum test for blood, and to one onthe catalase test for blood stains by van Itallie.g The serum test forJ. SOC. CJiena. Ind., 1906, 25, 512.J. doc. C’iicm. hi?., 1906, 25, 405.(i Zcit. aqyem. C‘henz., 1906, 19, 141.Zeit. physiol. C‘hem., 1906, 48, 69.J. Pharm. Chim., 1906 [vi], 23, 153.J. h‘aitit. Inst., 1903, 424.Analyst, 1906, 31, 150.Conzpt. m i d . , 1906, 143, 251.Proc. K. Akad. l17ctezsch. Anaatcrdnnz, 1906, 8, 628222 ANNUAL REPORTS ON THE PROGRESS OF CHESIISTRli.differentiating the blood of different animals has also been studied byPiorkowski,l who has devised a simplified method of applying it.Bettink and van den Driessen hlareeuw have dealt with the identi-fication of chloral hydrate obtained from parts of dead bodies, andhave proposed a method for its estimation in such material which issaid t o give better results than th2t of Kippenberger.T. E. Thorpe3has devoted attention to the estimation of arsenic in wall-papers,fabrics, and similar materials, and finds that the electrolytic methodas employed in the Government Laboratory, preceded by appropriatetreatment of the substance, is capable of giving good results,This report would hardly be complete without some reference to thesixth International Congress of Applied Chemistry held in Rome a tthe end of April.More than fifty papers were presented to the sec-tion concerned with analytical chemistry, snd many more were com-municated to other sections, notably that dealing with bromatology.To one of these reference has already been made, but many have notyet been officially published, and cannot therefore be discussed. Anumber of the sections have adopted resolutions, many of whichembody most useful suggestions, although others cannot be regardedas anything but aspirations towards ideals which are scarcely likelyt o be reaiised. Of particular importance to analysts is the verymarked leaning of some of the sections towards uniformity andstandardisation of analytical procedure. It cannot be denied thatthere is much to be said in favour of the standardisation of manyarbitrary methods employed in the analysis of commercial products,more especially when, as is frequently the case, the results have t oserve as the basis of a contract between buyer and seller, or perhapsas the foundation of a criminal prosecution.On the other hand,there has been a marked tendency, which many competent analystsdeplore, to prescribe rigidIy standardised conditions for the carryingout of many analytical methods which are thoroughly well known, andwhich in skilled hands are capable OF yielding highly accurate results.I n the case of all the well known and commonly used processes, theconditions essential for accuracy have been determined and laid downby numerous workers, and no capable analyst will depart very farfrom them, although he will remember that the exact set of conditionsobtaining in one analysis are not often reproduced in another, and hewill consequently recognise that a certain amount of latitude is necessaryif the best results are to be obtained.U p to a certain point themovement in favour of " unification " and ' standardisation ' is goodand worthyof support, but beyond that point it must receive the con-Ber. Dezct. Phmrm. Ges., 1906, 16, 226.Pharm. Weekblnd, 1906, 43, 487,TTans., 1906, 89, 408ANALYTICAL CHEMISTRY. 223demnation of all who regard chemical analysis as a highly importantbranch of applied chemistry and not merely as a useful art.Apparatus.I n the body of this report reference has been made to certain newappliances, and it does not appear to be necessary to refer to theseagain.The following list contains references only to those new piecesof apparatus which have been described in recognised journals, andwhich appear to be of general utility, It may be remarked thatthe titles are not always exactly those given by the authors, but havebeen in some cases slightly altered so as to indicate more clearly thenature of the apparatus in question.“A new gas calorimeter.” C. Y. Boys (PTOC, Roy. Xoc., 1906,‘( Apparatus for removing gases from aTrated liquids before deter-K. Ulrich (Chem. Zeit.,‘‘ On vessels for collecting and transporting samples of gas.” R.“Apparatus €or the continuous registration of the results of gasB.Stollberg (Clwm.‘‘ New form of absorption tube for very soluble gases.” E. P. Per-“ Gas analysis apparatus.” John S. Haldsne (J. Hygiene, 1906,“ Modification of the Orsat gas analysis apparatus.” Louis de“ Simplified measurement and reduction of gases.’’ H. Rebenstorff‘‘ New apparatus for the examination of gases, poor in certain con-C. J. Gulich (Chern. Zeit., 1906,“ A new form of gas generating apparatus.” A. W. Gregory (Chem.“ A ‘ continuous-flow ’ wash bottle ” (Analyst, 1906, 31, 34).‘‘ A combined wash-bottle and pipette.” J. W. Hogarth (CAem.‘‘ Weighing-bottle for liquids.” K. Buschmann (Chern. Zeit., 1906,“ Modification of Maquenne’s wash-bottle.’’ Antoine Villiers (Ann.77, A, 122).mining the rjpecific gravity of the latter.”1906, 30, 90).Nowicki (Oesterv.Zeit. Berg. Hutt., 1906, 54, 62).analysis (increase of weight due to absorption).”Zeit., 1906, 30, 347).man (Chern. News, 1906, 93, 213).6, 74).Saint-Martin (Ann. Chim. anal., 1906, 11, 96).Chem. Zeit., 1906, 30, 486).stituents, by absorption methods.”30, 1302).News, 1906, 93, 27)News, 1906, 93, 71).30, 1060).Chim. anaZ., 1906, 11, 211)224 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.‘‘ A burette filling device.” Edwartl French (Chem. News, 1906,‘‘ A burette top for preventing the absorption of carbon dioxitle and“ Struthers syphon pipette ” (Analyst, 1906, 31, 247).“ New antomatic pipette.”“Dougal assay tube ” (Analyst, 1906, 31, 246).“ Filter tubes for collection of precipitates.” Samuel L.Penfieldand W. M. Bradley (Anzer. J. Sci., 1906, [iv], 21, 453).“ A delivery fnnnel for introducing liquids into vessels underincreased or diminished pressure.” T. J. Bryan (.L Amer. Chens.“ A new vacuum filter for laboratory use, and a novel method forcleaning the filtering material ” (Zeit. angew. C/t,ern., 1906, 19, 95).‘( Vacuum-filter drying apparatus.” Barth (Chem. Zeit., 1906, 30,907).“ The properties of apparatus (rings, tubes, crucibles, cbc.) made ofmagnesia.” Kurt Arndt (ClLenz. Zeit., 1906, 30, 21 1 ; compareE. Wedekind, Clien2. Zeit., 1906, 30, 329).‘‘ Porcelain-lined bomb for general laboratory use.” S. F. Acree(Anter. Chem. J., 1906, 35, 309).“ A new burner for spectroscopic work.” E.H. Riesenfeld andE. H. Wohlers (Chem. Zeit., 1906, 30, 704).“ A new sodium burner ” (Chem. Zeit., 1906, 30, 835).“ Alcohol calorimeter for coal testing.” W. H. Wallace (Engineer-ing, 1906, 81, 527).‘‘ An improved form of the William Thomson calorimeter.” ThomasGray (J. SOC. Chem. Ind., 1906, 25, 409).‘‘ A. new laboratory sink,” Heinrich Giickel (Chenz. Zeit., 1906,30, 755).“ Viscosimeter for varnishes.” E. Valenta (Chenz. Zeit., 1906, 30,583).“A new form of calcium chloride tube.” A. E. Hill (Proc., 1906,22, 87).“A new form of pofash bulbs f o r estimation of carbon dioxide inorganic combustions.” S. F. Acree (8mer. Chenz. -J., 1906, 35,309).“ Two new forms of apparatus for use in organic analysis (azoto-meter and potash apparatus).” Erwin Rupp (Zed.cmal. Chem., 1906,45, 558).“ Modification of Liebig’s potash bulbs.” R. Villiers (Ann. Clkna.anal., 1906, 11, 250).“ Improved apparatus for the continuous extraction of liquids withether.”93, 71).other gases ” (Chem. Zeit., 1906, 30, 459).Stein (Chem. Zeit., 1906, 30, 967).SOC., 1906, 28, 80).R. S. Bowman (Proc., 1906, 22, 24)ANBTJYTICAL CHEMISTRP. 225’* Apparatus for the complete extraction of liquids containitigsaccharin.” Maurice Duyk (Ann. Chim. unal., 1906, 11, 82).A modification of Foerster’s fat extraction apparatus.’’ ErnestPescheck (Zeit. nngew. Chem., 1906, 19, 1513).6‘ New form of platinum parting apparatus.” A. Jarman (Trans.Inst. Mining and Metall., 1906, 15, 625).“ Thermometer for low temperatures.” A. Stock and C. Nielsen(Bev., 1906, 39, 2066).“New method for the calibration of thermometers below Oo C.”T. W. Richards and F. G. Jackson (Zeit. physikal. Chem., 1906, 56,362).“ A temperature regulator for use with the immersion refractometer.”F. Lowe (Chem. Zeit., 1906, 30, 686).‘‘A simple form of rotating electrode for use in electrochemicalanalysis.” F. M. Perkin (Tmns. Furaday Xoc., 1906, 2, 91).(( A new electrolytic apparatus (rotating propeller-like anode).”8. F. Scree (Ameft. Chem. J., 1906, 35, 313).“ A new rheostat for electrolytic analysis.” G. Pascalis (Mon. ,S&,,1906, [iv], 20, 168).“ A modified Westphal balance for solids (particularly cements andminerals) and liquids.’’ F. 39. Williams (J. Amer. Chem. Xoc., 1906,28, 185).6 ‘ New zero adjustment for chemical balances.” J. McDowall (Chem.News, 1906, 94, 104).‘ 6 Improved Beckmanii apparatns for molecular weight determina-tions.”6‘ Portable universal stand for elementary analysis ” (Chem. Zeit.,1906, 30, 1045).6‘ The production of a high vaciiurn in the Scheihler (lesiccator.”H. C. Gore ( J . Amer. Chen?. Soc., 1906, 28, 834).6‘ Apparatus for distilling solids in a vacuum.’’ Hngo Haehn (,&it.nngew. Chem., 1906, 19, 1669).“ Shortened manometer with receivable vacuum (for distillation in nvacuum, &:.).”g C Apparatus for the estimation of snlphur and carbon with single ordouble receiver.” Arthur Wilhelmi (Zeit. Cliem. Appuratenkunde,1906, 1, 155).6‘ Jmproved apparatus for estimating total sulphur in coal gas ;modification of Dreschmidt’s method.” Everhard P. Harding (J. Amer.Chern. Soc., 1906, 28, 537).“New apparatus for the estimation of sulphur and carbon (iniron).”6‘ Percolator for me in assaying diwgs.” Frank R. Eldrecl (J. A n w .Chem. SOC. 1906, 28, 157).VOI,. 111. QJ. If. Sanders (Proc., 1906, 22, 165).Leo Ubbelohde (Chem. Zeit., 1006, 30, 966).A. Kleine (Zeit. angew. Chern., 1906, 19, 171 1)226 ANNUAL REPORTS ON THE PROGRESS OE‘ CHEMISTRY.‘‘ Improved Mayer’s apparatus for the evolution of chlorine. ” (Zeit.‘‘ A new urometer ; modification of the hypobromite method.”Nahr. Genussm., 1906, 12, 221).William M. Dehn (Zeit. anal. Chern., 1906, 45, 604).In conclusion, the author mould once again remind his readersthat the task of selecting from the enormous mass of publishedmatter those communications which appear to contain observationsof special importance to analysts is one of increasing difficulty,and? within the limits of a review such as this, many papers,embodying the results of really useful work must of necessity remainunnoticed.ALFRED C. CHAPMAN.Addendum.1Since the above mas written, I have been informed that in July,1901, H. A. Danne published, in The Chemist and Druggist ofAustrukasia, a method for the estimation of moisture identical inprinciple with Dupre’s acetylene method, referred to on page 202 ofthis report.A. C. c1