ANALYTICAL CHEMISTRY.THIS report follows the same general plan as that of last year, and,if the plan is not immediately obvious, the brevity of the reportmakes unnecessary any long explanatory introduction. It may besaid, however, that the metals are grouped as they group themselvesin the mind of an analyst rather than by reference to MendelBeff’stable, whilst in the short section relating to organic analysis thegeneral arrangement is the same as that adopted in earlier reports,but without cross-headings, which give an appearance of exhaustive-ness to which this report does not pretend. That the division ofspace between organic and inorganic analysis would be unequal wasclear t o the writer before he finally arranged his notes; theinequality could easily have been redressed, but only, in the writer’sopinion, by omitting more interesting for less interesting matter.Apology is scarcely needed for giving rather extended referenceto one or two matters of importance to analysts with commercialinterests and to some departmental inactivities.General.A new and interesting application of precise thermometry toquantitative analysis has been made.The method depends on theobservation of the precise temperature a t which a solution attainsthe same density as a given, previously calibrated float, and itsapplication is, of course, limited to cases where not more than onesubstance, in addition to the solvent, is present. Since the tempera-ture of floating equilibrium is an almost linear function of theconcentration, when a few points on the curve connecting the twovariables have been determined, the concentration of an unknownsolution may be deduced from the temperature a t which the cali-brated float just sinks in it.Conversely, the method may beemployed to calibrate thermometers by means of solutions of knownconcentration. The exact density of none of the solutions need beknown, since the method depends only on differences of density.Lest it be thought that the method has no advantage over methodsREP.-VOL. IX. 193 1% ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.which correlate concentration with actual density, the followingnumbers may be given. A particular float just sank in alcohol of98.99 per cent. strength a t 15-391O. On adding water so that thepercentage of water in the alcohol was increased from 1.01 to 3-44per cent., the temperature of floating equilibrium was 24*093O, andintermediate experiments showed that the increase of temperatureof floating equilibrium was very nearly a linear function of theconcentration of water.1 Since the float employed was so sensibivethat the temperature a t which it sank or rose could be determinedwithin O*O0lo, a simple calculation shows that the experimentalerror of the met,hod is only about one-tenth that of a methoddepending on the determination of the density of the spirit, withan accuracy of 5 5 units in the sixth decimal place, the highestdegree of accuracy with which densities can be determined withreasonable rapidity.To obtain the greatest accuracy of which themethod is capable, a thermostat capable of being adjusted to anyneeded temperature and of being kept there within 0*005° is neces-sary, and in this connexion reference may be made to a paperon t,he control of temperature in the operations of analytical chem-istrg,2 which should not be overlooked by anyone who may haveoccasion t o design buildings intended for use as laboratories.Electric heating has been made use of in ebullioscopic determina-tions for many years, but Bigelow,3 who first made use of it, pointedout that difficulties due to electrolysis would certainly be encoun-tered if the method were extended to aqueous solutions, and Beck-mann 4 described experiments which confirmed this view.Con-sideration might have convinced these authors that electrolysiscould be avoided by choosing a heating coil of resistance so lowthat the necessary heat could be imparted t o the liquid withoutthe drop in potential across the coil exceeding some assigned quan-tity That their general warning was unsound and their conch-sions only valid for their own coils of relatively high resistance, hasnow been shown experimentally by two authors, who appear to haveworked in ignorance of the work of Bigelow and Beckmann.Thenew apparatus owes its origin to difficulties encountered in %heuse of the ordinary Beckmann apparatus with substances of highmolecular weight and sparing solubility, but it is said to possessthe advantage that readings constant to O*O0lo are obtained withoutspecial precautions, and its general use where the necessary currentis available is t o be expected.6 Of less general interest, but impor-ii, 599.T.W. Richards and J. W-. Shipley, J. Anrer. Chon. SOC., 1912, M, 599 ; A.S. L. Bigelow, Amer. Chem. J., 1899,22, 208 ; A., 1900, ii, 9.E. Beckmann, Zeikch. physikal. Chem., 1908, 63, 177; A., 1908, ii, 663.T. W. Richards, 8th Inter. Cong. Appl. Chem., 1912, 1, 403.ti E. Knecht and J. P. Ratey, T., 1912, 101, 1189ANALYTICAL CHEMISTRY. 195tant in view of the very considerable difficulties which its authorshave overcome, is an apparatus for ebullioscopic determinations a ttemperatures from -36‘3 to -830 with solvents such as hyarogensulphide, chloride, bromide, and iodide.6The work of Thorpe and Rodger is proof that so long ago aa 1894it was possible t o measure viscosities with great exactness.7 Yetuntil recently viscosities of oils have been generally measured inanalytical laboratories by one or other of the commercial instru-ments which are quite unsuitable for scientific research, althoughpossessing an ease of manipulation very desirable in such work.More rarely use has been made of Ostwald’s viscometer, which ismuch easier t o set up than Thorpe and Rodger’s “glischrometer,”an instrument which demands much preliminary work, whilst themeasurements themselves are much more complicated than with theOstwald type of instrument with vertical capillary.Several investi-gations mads within the last few years have, however, pointed tothe fact that the Ostwald type of instrument holds serious sourcesof error, by far the greatest arising from the fact that no correc-tion is made for the kinetic energy imparted to the liquid on enter-ing the capillary,* that is t o say, for the term vd/8?rZt in Pouseuille’sformula 7 = ar4pt / 8Zv - vd/ 8 d t .In the course of an investigationon the viscosity of emulsions, the authors found it possible tosimplify the apparatus9 of Thorpe and Rodger, and one of themhas since described a still further modified form of instrument,which marks a very great advance, not in precision of viscometry-in point of precision the work of Thorpe and Rodger left nothingto be desired-but in the ease and rapidity with which viscositiesmay be measured with a degree of accuracy not far short of thehighest attainable.The instrument is empirical only in the sensethat it is so constructed that the exact dimensions of the capillarycannot be determined, and it must be standardised on some liquidof approximately the same viscosity as the liquids for which it isto be used, and of which the absolute viscosity is known. Thegeneral adoption of an instrument of this type in work on oils,always selecting a, capillary of such dimensions that the kineticenergy correction was less than 0.5 per cent. of the total viscosity,would give a value to future numbers which few of those publishedup to the present possess. The extra demand on the analyst’s timeE. Beckmann and W. Weber, Zeitsch.morg. Chem., 1912, 74, 297; A., ii,Sir T. E. Thorpe and J . W. Rodger, Phil. Trans., 1894,185, A, 410.M. R. Schmidt, BaZti7nore Dissertation, 1909 ; E . C. Bingham and T. C.Durham, Amer. Chem. J., 1911, 46, 278; A., 1911, ii, 968,E. C. Binghain and G. F. White, J. Amer. Chem. SOC., 1911, 33, 1257 ; A.,1911, ii, 858.0 2431196 ANNUAL REPORTS ON TEE PROGRESS OF CHEMISTRY.as compared with the use of one of the commercial instruments isnot very great, and the numbers obtained would be available forthe purpose of correlating viscosities with other physical andchemical properties, whereas the numbers obtained by the commer-cial instruments and in many citses even by Ostwald’s viscometerare insufficiently simply related t o the true viscosities to be of anyuse for such a purpose.10The most important improvements to be recorded in quantitativemicrochemical analysis are the construction of the beam and pointerof the Nernst balance in a single piece, with a glass counterpoisefused on to the beam at the end near the pointer, the substitutionof spongy platinum for asbestos fibre in the construction of theDonau filter, and an ingenious arrangement for conducting precipi-tations and transferring the precipitate to the filter without loss.11A paper which describes the application of Carius’ method to thedetermination of quantities of sulphur or halogens of the order of1 milligram is perhaps deserving of notice even by those who haveso far found no use for microchemical methods.Since 1 milligramof sulphur or halogen can be determined with an error well under2 per cent., the use of somewhat larger quantities should giveresults of sufficient accuracy for many purpoBes, with greatlyreduced risk of serious accident as compared with Carius’ method asusually carried out, besides much saving of space, time, and glass.12Gas Analysis.Last year attention was directed to a new and accurate methodfor the estimation of nitric oxide in the presence of nitrous oxide,a problem which until that date had not found a satisfactory solu-tion.13 An entirely different method, but one which is said to beequally accurate, has since been described.It depends on Raschig’sobservation that the oxidation of nitric oxide proceeds in twostages, the first of which, corresponding with the formation of thetrioxide, proceeds with great rapidity, whilst the second, correspond-ing with the formation of the tetroxide, requires an appreciabletime €or its completion.The method now described consists inoxidat.ion of nitric oxide to nitrogen trioxide and absorption ofthe latter by potassium hydroxide in the form of sticks, the forma-tion of any tetroxide being inhibited by confining the gas togetherwith some potassium hydroxide over mercury, and then admittinga measured and sufficient volume of air. I n this way the nitrogenlo G . F. White, Biochem. Zeitsch., 1911, 37, 482 ; A., ii, 61 ; also J. Ind. Eng,l1 J. Douau, Monatsh., 1911, 32, 1115 ; A., ii, 199.l2 Ibid., 1912, 33, 169 ; A . , ii, 384.Chem., 1912, 4, 106.l3 Am.Beport, 1911, 156ANALYTICAL CHEMISTRY. 197trioxide is transformed into potassium nitrite as fast as it isproduced.14A recent paper on the estimation of hydrogen and methanedescribes no new methods, of which there is already a wide choice,but it does offer very valuable suggestions in regard to four knownmethods. The author shows that the Drehschmidt capillary may bereplaced by it quartz capillary containing a short length of platinumwire of such diameter that it nearly fills the tube, one advantageof the arrangement being that there is no risk of propagation offlame from the heated portion of the capillary to an explosivemixture contained in the gas burette to which it is connected.When using the Winkler-Dennis combustion pipette there is oftendifficulty in burning the last of the methane, so that the operationextends over a considerable period, during which such an amountof mercury may be oxidised as to give rise to serious error.Thissource of error may be avoided by enclosing the electrically-heatedspiral in a quartz tube, which becomes the outer limb of thecapillary U-tube of a simple mercury pipette. The gas is passedslowly to and fro between the burette and pipette over the heatedspiral. It is shown that fractional combustion of hydrogen bymeans of palladium asbestos is successful only if the passage of thegas is slow and the temperature of 400° not exceeded, and aningenious yet easily constructed device is described for controllingthe temperature.Finally, the comparatively new method, depend-ing on the absorption of hydrogen by means of a palladium sol,together with a soluble picrate,15 is dealt with. Although it givesexcellent results, as originally described, the method is renderedtedious by the persistent frothing of the liquid. The froth isinstantly destroyed by a few drops of alcohol, but alcohol makesthe reagent useless for further determinations. For this reason,and because the absorptive capacity of the reagent falls off some-what rapidly on storage, even in the dark, the author has alteredthe whole technique of the method, which, as modified, is economicalboth of the analyst’s time and of the expensive reagent.16I n 1910 a new interferometer was described, by means of whichit was possible to solve simple problems of technical gas analysis.The authors confined themselva, however, to relatively simpleproblems, such as the determination of methane in mine air andthe examination of technical hydrogen.17 It has since been shownthat the instrument may be wed for the analysis of flue gases. I nl4 0.Baudisch and G. Klinger, Ber., 1912, 45, 3231 ; A . , 1913, ii, 74.l5 C. Paal and W. Hartmann, Ann. Report, 1910, 160 ; also 0. Brunck, C?zem.Zeit., 1910, 34, 1313, 1331 ; A., 1911, ii, 149.l6 W. Henipel, Zciitsch. angew. Chewz., 1912, 25, 1841 ; A., ii, 987.l7 Hsberaad Luwe, ibid., 1910, 23, 1393198 ANXUAL REPORTS ON THE PROGRESS OF CHEMISTRY.the absence of carbon monoxide, the results are exact, whilst thepresence of that gas may be inferred if the sum of the percentagesof oxygen and carbon dioxide is appreciably less than it shouldbe, having regard to the composition of the fuel and to the excessof air above that necessary for combustion.18Inorganic Analysis.Qualitatzve.-Since the reference in the report for 1908 to thesystem of qualitative analysis devised by Noyes and Bray,lg twofurther papers have appeared, one by each author.The earlierone dealt with the separation of the alkali earths and the membersof the alkali group,20 whilst the last deals with acidic constituents.21The whole series of papers occupies some two hundred pages, and,since they have not all appeared in the same journal, would be ofmuch greater use if re-issued as a textbook of qualitative analysis.Necessarily these papers contain much that is familiar t o everyone,but it is obvious that they embody an immense amount of originalwork.The number of new methods is so large that the authorsare certain to incur the charge of unnecessarily multiplyingmethods, and analysts of the older school may be expected toexpress impatience of the detailed directions given, but the avowedobject of the authors was to select or devise tests of known androughly equal sensitiveness, and preference has been given tomethods which enable the experienced analyst to form an approxi-mate estimate of the quantities present. Such aims cannot beachieved without close attention to detail.The writer does not share the enthusiasm for methods of qualita-tive analysis which circumvent the necessity for the use of hydrogensulphide, because he finds a hydrogen sulphide apparatus indis-pensable for quantitative separations and for effecting reductions.The ingenuity which has been shown in avoiding the use of thegas, however, is sufficient, even if other evidence were lacking, toshow that there are circumstances where an alternative method isdesirable.Attention was directed last year to the work of Eblerin this connexion.22 The year under review has witnessed theappearance of another comprehensive scheme by an author whomakes no reference to Ebler or any other worker. No new reactionis made us0 of, so that it is possible to appraise the method roughlywithout trial, and it may be said that it might be expected to effect18 0.Mohr, Zeilsch. angezu. Chem., 1912, 25, 1313.I9 Ann. Report, 1908, 181.21 A. A. Noyes, J. Amer. Chem. Soc., 1912, 34, 609 ; A . , ii, 599.a Ann. Report, 1911, 157W. C. Bray, Tech. Quart. 1908, 21, 450; A., 1909, ii, 431ANALYTICAL CHEMISTRY. 199fairly sharp separations, whilst no unduly tedious operations areintr0duced.s It may be specially commended to the considerationof those who are impatient of lengthy descriptions. I n yet anotherpaper dealing with inorganic qualitative analysis as a whole, alarge part is occupied with a description of a scheme which dispenseswith the use of hydrogen sulphide.24It has been shown that tetramethyldiaminodiphenylmethane,previously recommended as a reagent for the *detection of lead andmanganese,25 is a very delicate reagent for gold, solutions contain-ing 1 part of gold in 5 millions developing a distinct purple colouron treatment with it.The metals of the platinum group do notinterfere.26 A microchemical test for manganese, capable of detect-ing 0.005 milligram, has also been described,27 but perhaps the mostnotable achievement in the domain of qualitative analysis has beenthe improvement in the well-known cobaltinitrite test for potassium,the sensitiveness of which has been increased a hundred-fold. Theimprovement consists in making the test in presence of silvernitrate, and depends on the much lower solubility of the doublecobaltinitrites of potassium and silver as compared with that ofpotassium cobaltinitrite or sodium dipotassium cobaltinitrite. Ifthe test is applied after removal of the heavy metals and ofammonium salts, nothing can interfere except rubidium, msium,and thallium, which will seldom be present in quantity where smallamounts of potassium call f o r detection.28&uarttitative.-Chlorate is usually determined to-day by a directiodometric method, but disagreement among different chemists hasled to periodical proposals to return t o the more troublesome distiI-lation method of Bunsen.It has now been shown that the directmethod is exact provided the chlorate be dissolved in air-free waterand the hydrochloric acid used be also freed from air.29 The exactdetermination of chlorite in admixture with chlorate, hypochlorite,and chloride is perhaps not a common problem, but should it arisea new gravimetric method depending on the insolubility of leadchlorite in 80 per cent.alcohol will prove useful, and greatlysimplify the separate estimation of the other constituents of themixture.30H. Trspp, Zeitsch. anal. Chem., 1912, 51, 475 ; A., ii, 685.24 A. Purgotti, Gatzetta, 1912, 42, ii, 58 ; A . , ii, 984.25 A. Trillat, Compt. rend., 1903, 136, 1205 ; A . , 1903, ii, 512.26 R. J. Carney, J. Amr. Chem. SOC., 1912, 34, 32 ; A . , ii, 298.27 M. Wagenaar, Pharm. Weekblad, 1911, 49, 14 ; A , , ii, 206.2* L. L. Burgess and 0. Kamm, J. Amer. Chem. Xoc., 1912, 34, 652 ; A., ii'L9 A. Kolb, Zeitsch. angew. Chem., 1912, 25, 1168 ; also Chem. Zeff., 1912, 36,30 G.Lashgue, Bull. SOC. chim., 1912, [iv], 11, 884 ; A., ii, 988.601.635200 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY,It has been shown that, in alkaline solution, the reaction :NeHa + 4K3Fe(CN), + 4KOH = 4N9 + 4K,Fe(CN)6 + 4H20is quantitative until either the whole of the hydrazine or the wholeof the ferricyanide has disappeared, and on this reaction an exactgasometric method for the estimation of ferricyanide has beenbased.31 The method is equally exact when applied to the estima-tion of hydrazine, but the technique is then somewhat complicatedby the action of t h e excess of ferricyanide on the mercury usedas confining liquid, so that, so far as the determination of hydrazineis concerned, the method scarcely challenges the iodine method ofSto116,32 which leaves nothing to be dtxired.Although there is no doubt that Feld’s method for the estima-tion of ferrocyanide i s capable of yielding accurate results inexperienced hands, there is abundant evidence that the hands mustbe experienced, not merely capable, or, in other words, that theconditions which determine success in carrying out Feld’s processhave not yet been defined with sufficient exactness.% I n these cir-cumstances, a new method, apparently less dependent on thepersonal equation, is welcome.It depends on the fact that inpresence of a small quantity of cuprous chloride and excess ofsulphuric acid, all ferrocyanides-ven insoluble ones-are decom-posed with the evolution of the whole of their cyanogen as hydrogencyanide, which may be distilled and determined in the usualmanner.The simplici€y of the method, which is shown t o be exact,should recommend it even t o those who are skilled in the use ofF eld’s met hod .34The usual methods for the determination of fluorides are farfrom exact when very small quantities have to be estimated, andspecial methods which enable a food analyst to assert with confi-dence that an article contains not less than Pome stated proportionof fluorine, although useful to such analysts, are insuffieiently exactfor the purpose of those who are engaged in determining the distri-bution of fluorine in nature, whether it be in living tissues or inminerals of which fluorine is a minor uonstituent. Three papersdealing with the estimation of minute quantities of fluorine musttherefore be accounted important, as they will undoubtedly givegreater value to future statistical work of the kind mentioned.35Having regard to the immense gums paid by chemical manufac-turers for nitrate of soda, the value of which depends solely on itsY1 F’.R. F!&y and H. K. Sen, Zeitsch. anwy. Chem., 1912, 76, 380 ; A., ii, 817.32 R. StollB, J. p. Chem., 1902, [ii], 66, 332; A., 1903, ii, 100.33 Ann. Eeport, 1910, 164.34 H. E. Wilhms, J. Xoc. Chem. Ind., 1912, 31, 468 ; A., ii, 704.y5 A. Gautier and P. Clausmann. Cownt. ~ c n d . . 1912. 154. 1469. 1670. 1763 : A , .ii, 681, 805, 806ANALYTICAL CHEMISTRY. 201nitrogen content, it is remarkable that it should still be paid foron the basis of analytical reports which, in their turn, are based onthe determination of chloride and sulphate (calculated as sodiumsalt), moisture, and insoluble matter, and the subtraction of thesepercentages from 100.The method invariably overestimates thesodium nitrate, the average error being about I per cent., but itmay reach 3 per cent. Not only do all samples contain soluble saltsof calcium, magnesium, iron, and aluminium in small amount, butsodium perchlorate is occasionally present up to nearly 1 per cent.,and the percentage of potassium nitrate may exceed 8 per cent.The suggestion made a t the recent Congress of Applied Chemistrythat the time has arrived for the general adoption of a directmethod of valuing nitrate Will commend itself to all chemists, andsince t.he suggestion came from the spokesman of an importantgroup of American consumers, it may have weight even with thosewho benefit by the present arrangement.There will be lessunanimity among chemists concerning the particular method ofanalysis recommended, namely, reduction t o ammonia in alkalinesolution by means of Devarda metal and distillation from a veryelaborate apparatus.36A method for the estimation of nitrite, depending on the readi-ne,w with which nitrous acid is converted into its methyl ester, andthe low boiling point of the latter ( - 1 2 O ) , is worthy of note. Themethod affords a ready means of determining nitrite in presenceof nitrate, and, although one may not agree with its authors thatno other satisfactory method of effecting this separation exists, thesimplicity and proved accuracy of the new method commend it.Allthat is necessary is to mix the neutral solution with excess ofmethyl alcohol and a known excessive quantity of acid, to bubbleair through the mixture, and then titrate the remaining free acidwith alkali.37The observation that even a small fraction of a milligram ofselenium per litre suffices to bring about rapid destruction of theliquors of sulphite cellulose factories, with production of sulphuricacid and gypsum,38 has led to an investigation on the estimationof minute amounts of selenium in brimstone and pyrites. Brim-stone is burned in a combustion tube in a current of air, an asbestosplug sarving to catch all but an unweighable trace of the seleniumwhich passes forward as a fog with the small quantity of sulphurtrioxide which is always formed; the selenium is dissolved out ofthe tube by means of potassium cyanide, precipitated m dioxide,37 W.M. Fischer axid N. Steinbach, Zeitxh. nnorg. Chem., 1912, 78, 134; A.,38 P. Klasoii, Yereia. Zcll.Jo$ u. P~pitrchcnzikc, 1909, 61.W. S. Allen, 8th. Inter. Cong. AppI. Chenz., 1, 19.ii, 1093202 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.freed by an ingenious device from traces of sulphur, and obtainedas pure dioxide, which is then determined iodometrically.sgThe work of Allen and Johnston on the exact determination ofsulphate made our knowledge of the barium sulphate precipitationnearly complete,40 but it did not provide an easy solution of thepractical problem of valuing pyrites, the most important rawmaterial of chemical industry. A method recently described, if itdoes not solve this problem, goes so far that i t is likely to becomeand remain for some considerable time a standard method.Allenand Johnston’s work made it appear unlikely that an exact methodwould be discovered, which would be generally applicable andrequire no correction, and yet not owe its accuracy to a compensa-tion of errors. The new method is avowedly a compensation method,but it differs from earlier methods of this kind in having beenworked out after and with full knowledge and appreciation of thework of Allen and Johnston. The “barium sulphate” as broughtto the balance always contains barium chloride, but the proportionof this is never less than 0.27 per cent., nor more than 0.30 percent., and so nearly balances the negative errors that the resultsare accurate to 0.05 per cent. How far we have travelled fromtextbook methods of determining sulphate will be appreciated fromthe statement that precipitation is made in the cold, that a Goochcrucible is used, and that the quantity of barium sulphate handledapproximates 5 grams.41 It should be added that the whole methodis as simple as, if not simpler than, any method yet described forthe valuation of pyrites.A rapid yet exact method for the estimation of the free sulphurin spent oxide has long been wanted.A method recently described,depending on the conversion of the sulphur into sodium thiocyanateby digestion with an alcoholic solution of sodium cyanide andtitration of the thiocyanate, is not only rapid but more exact thanmethods depending on the oxidation of the carbon disulphideextract and determination of the resulting sulphate because someof this sulphate may be derived from compounds of sulphur solublein carbon disulphide, but useless to the sulphuric acid maker.42Miiller’s benzidine method 43 for the volumetric estimation ofsulphate has disappointed the expectations that were raised by itsappearance and by Raschig’s44 early investigation of it.For theP. Klason, Arkiv. Kern. Min. Geol., 1911, 4, No. 18, 1 ; No. 29, 1 ; A., ii,201, 990 ; Zeitsch. angew. Chem., 1912, 25, 514.40 Ann. Report, 1911, 159.41 W.S. Allen and H. B. Bishop, 8th. Inter. Cong. Appl. Chem., 1912, 1, 33.42 C. Davis and J. L. Foucar, J. SOC. Chem. Ind., 1912, 31, 100; A., ii, 384.43 W. Miiller, Ber., 1902, 35, 1587 ; A., 1902, ii, 425.44 F. Raschig, Zcitch. angew. Chem., 1903, 16, 617; d., 1903, ii, 572ANALYTICAL CHEMISTRY. 203routine analysis of a number of very similar samples it may some-times be employed with a saving of time; but in other cases thenecessary investigation to determine whether it is applicableoccupies more time than a gravimetric determination. I n manycases the composition of a solution is known within certain limits,and in such cases reference t o two recent papers may inform theanalyst a t once whether the solution is one in which sulphate canbe determined with tolerable exactness by means of benzidine.45Although copper can now be estimated with a degree of exactnessapproaching that of a mint assay of gold bullion,46 the analyticalchemistry of copper and its ores is far from being a book the lastpage of which has been written. On the contrary, the number andvalue of papers dealing with the subject tend t o increase, andsome recent ones must be referred to.Rhead's method47 for thetitration of copper by means of titanous chloride has been greatlysimplified, and shown to be as exact as a volumetric method canbe.48 A new gravimetric method, depending on the selective reduc-ing action of hydroxylamine in presence of alkali tartrate, promisesto be useful, since it is available in presence of antimony, 'zinc,bismuth, lead, iron, arsenic, and reasonable amounts of ammoniaor nitrates.49 The estimation of copper in pyrites has been simpli-fied by the discovery that the aqua regia solution may be directlyelectrolysed if certain conditions, easily realised, are compliedwith.60 A paper and discussion on the limits of accuracy attainablein copper and brass analysis must also be adjudged important, as itshould lead to a diminution in the number of certificates disfiguredby insignificant figures.51 Such figures are not peculiar t o reportson copper and brass, but extend to every branch of analytical chem-istry, and raise a very grave question, which the profession as awhole will be compelled to consider unless amendment is moreprompt than the present writer ventures to hope for.Finally, withregard to copper, attention must be directed to a paper on thedetermination of oxygen in refined copper. It was shown sometime ago that the loss of weight on heating copper drillings inhydrogen is not due entirely to oxygen from cuprous oxide andsulphurous acid, but also includes gases derived by the metal fromthe fuel and other sources, and it is usual now to heat the drillingsto constant weight in a current of carbon dioxide before proceedingk5 K. K . Jarvinen, Ann. Acad. Sci. Fennicae, A, 2, 1910, No. 4 ; 1911, No. 16 ;46 Ann. &port, 1911, 168. Ibid., 1906, 207.48 L. Moser, Chcm. Zeit., 1912, 36, 1126 ; A., ii, 1097.49 A. Baycr, Zeitsch. anal. Chcin., 1912, 51, 729 ; A., ii, 1212.5o W.D. Treadwell, Chem. Zeit., 1012, 36, 961 ; A . , ii, 998.61 E. A. Lewis, J. SOC. Chem. Ind., 1912, 31, 96.C'hem. Zentr., 1912, i, 526 ; A., ii, 486204 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.to the determination of oxygen. It is now shown that similartreatment with carbon dioxide should follow the treatment withhydrogen, as copper may occlude so much hydrogen that an oxygendetermination, as commonly practised, may be 20 per cent. inerror.52As every analytical chemist knows, the technique of the Marsh-Berzelius test for arsenic was greatly improved, and the method madevery approximately quantitative in the years immediately succeedingthe unfortunate epidemic of arsenical poisoning in 1901, and every-one who has occasion to determine small quantities of arsenic is nowpractised in this method.The recent statement that for purposesof works control the Marsh test was obsolete, because of its toogreat tedium and expense, would undoubtedly have encounteredlively opposition had it not been the opening sentence of a paperwhich seems to establish the simpler Gutzeit test a t last on a trust-worthy basis. Since 1901 many authors have advocated the use ofthe Gutzeit test, but not even Sanger and Black,53 on whose workthe present authors have largely built, provided for the interferingeffect of small quantities of iron, which are, or may be, present inso large a proportion of the substances which have to be tested forarsenic. The success and general applicFbility of the method nowdescribed depend mainly on the observation that ferrous iron $usa trace of stannous chloride actually increases the sensitiveness ofthe test, and t o keep the conditions uniform, ferrous iron is addedas a matter of routine.54An unusual number of papers dealing with the estimation ofglucinum have appeared within the year.Two volumetric methods,depending on the fact that glucina is so weak a base that solutionsof its normal salh may be titrated with alkali hydroxide usingphenolphthalein as indicator, or made to liberate iodine from asolution of iodide and iodate, have been describedts whilst theconditions which determine the quantitative precipitation ofglucina have been the subject of careful study, with the result thatthese seem t o have been discovered and some older methods shownt o be 20 per cent.in err0r.M A further paper on the separation ofglucinum, aluminium, chromium, and iron should help to makefuture determinations of glucinum easier and more accurate.5752 G. I,. Heath, J. Id. Eng. Chem., 1912, 4, 402 ; A., ii, 1091.53 C. R. Sanger aiid 0. F. Black, J. Soc. Chcm. Id., 1907, 26, 7115 ; A . , 1908,54 W. S. Allen and R. M. Palmer, 8th Intern. Cong. App?. Chem., 1, 9.s5 B. B1eyt.r and A. Moormann, Zcitsch. anal. Chem., 1912, 50, 360 i A . , ii,56 R. Rleycr arid I<. Roshart, ibid., 1912, 51, '748 ; A., ii, 1211.67 M. Wundcr and T. Wenger, ibid., 470 ; A , , ii, 687.ii, 64.491ANALYTICAL CHEMISTRY. 205The difficulties which attend the estimation of chromium inchrom-vanadium steel are admirably summarised in the openingparagraph of a short paper, which proceeds to describe a goodmethod based on the fact previously established by the author inhis work on vanadium,58 that chromium can bO precipitated com-pletely by boiling the nearly neutralised (ferrous) solution withbarium carbonate, and on the observation of Noyes and Bray 59 thatchromate may be separated from vanadate by precipitation as leadchromate if certain precautions are observed.60Ferrous salts may be oxidised to ferric by iodine, provided thereverse reaction be inhibited by suppressing the ferric ion in thesolution as it is formed, and a method based on this fact, althoughlacking the accuracy of the permanganate method, should proveuseful because it is available in cases where organic matter andferric salts are simultaneously present.61 The separation of ironfrom titanium by volatilisation of the former as chloride may beaccomplished at a lower temperature, and with simpler apparatusby substituting a carefully regulated stream of sulphur mono-chloride vapour for the chlorine usually employed.@ Nitrosophenyl-hydroxylamine ( r r cupferron ”), as a precipitant of copper andiron,m has met with comparatively little use, probably on accountof its cost and of the excellent methods already available for thedetermination of copper and iron.I n special cases, however, it hasproved useful, notably as a means of separating iron from phos-phates, for example in plant ashes, as a preliminary to a colori-metric determination of the iron, with which phosphates interfere.64Among the usual host of papers dealing with the non-metallicconstituents of steel and steel-making alloys, three call for notice.It has been shown that Ledebur’s method for the determinationof oxygen in iron and steel may discover none when 0.1 per cent.ispresent, or a9 little as 0.07 per cent. when 0.23 per cent. is present.An exact method is described, but it depends on the use of a 14 kw.furnace.65 In the discussion of Blount and Levy’s paper on thedirect combustion method for determining carbon in steel,66 Arch-butt drew attention to the tendency of the results so obtained t o68 Ann. &port, 1911, 166.69 A. A.Noyes and W. C. Bray, Techn. Quart., 1908, 21, 14.J. R. Cain, J. Ind. Eng. Chem., 1912, 4, 17; A , , ii, 692.61 G. Romijn, Chem. Zeit., 1911, 35, 1300 ; A., ii, 94.a F. Bourion, Compt. Tend., 1912, 154, 1229 ; A., ii, 691.63 Ann. Report, 1910, 165 ; also R. Fresenius, Zeitsch. anal. Chem., 1911, 50,F. E. Nottbohm and W. Weissmange, Zeitsch. Nahr. Genocsm., 1912, 23,W. H. Walker and W. A. Patrick, 81h. lntcr. Cong, Appl. Chem., 1912,21, 139.35 ; A., 1911, ii, 386.514 ; A., ii, 690.66 B. Elount and A. G. Levy, Analyst, 1909, %, 94 j A., 1909, ii, 346206 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.exceed those yielded by wet methods, and other authors have noteda like tendency.67 The difference appears t o be due to the presencein the steel of dissolved carbonaceous gases, the carbon of which isincluded in the values given by the direct combustion method, andalmost wholly lost in the wet meth0ds.a Finally, two papers on theuse of tho calorimetric bomb in the estimation of carbon in steeland steel-making alloys make it appear probable that this is theonly method a t present available for the exact determination ofcarbon in certain alloys, notably high-grade f errochromes andf errosilicons.69The bismuthate method for the estimation of manganese shouldbe so firmly established by now as to need no defence, although itcontinues t o meet with but slighting reference in German text-books, which rank it with the lead peroxide method.However, amiscalculation by one of its critics last year has led to two newcritical investigations70 of the method, which is shown to be atleast as accurate as any known method for manganese, as it isincomparably simpler than any other method with any pretence toaccuracy.A useful note has been published on its application inpresence of much vanadium or chromium.71 It has been shownthat some of the disadvantages of Volhard’s method disappear whenit is applied t o solutions of manganous nitrate instead of to solutionsof the sulphate.72The estimation of magnesia in limestone is referred to amongnew electrochemical methods, of which it is a good type, whilst theestimation of potassium hydroxide in fertilisers, soils, and plantashes is the subject of extended notice among agricultural methods.A spectroscopic method for the estimation of small quantities ofrubidium in presence of much potassium, originally designed onlyto determine the order of magnitude of the rubidium percentage,has proved to be accurate to + l o per cent.73The exact determination of platinum and the preservation ofplatinum apparatus are intimately related problems.As the formerinterests comparatively few, a critical paper on the analysis ofplatinum ores,74 and another on the estimation of small amounts167 H. Isham, J. Ind. Eng. Chen~, 1911, 3, 577 ; A., ii, 387.6* A. G. Levy, AnaEyst, 1912, 37, 392; A., ii, 995.69 P. Mahler and E. Goutnl, Compt. rend., 1911, 153, 649 ; A , , 1911, ii, 937 ;70 W. Blum, J. Amer. Chcm. Soc., 1912, 34, 1379; A . , ii, 1214 ; also 8th Inter.ibid., 1912, 154, 1702; A., ii, 807.Cony.Appl. Chem., 1912, 1, 61; H. F. V. Little, Analyst, 1912, 37, 554.D. J. Demorest, J. Ind. Eng. Chem., 1912, 4, 19 ; A., ii, 690.T2 L. Karaoglanoff, Jahrbuch Univ. Sojia, 1910-11 ; A., ii, 1214.73 E. Wilke-Dorfurt, Zeitsch. anorg. Chem., 1912, 75, 132; A., ii, 686.74 E. V. Koukline, Rev. MLt., 1912, 9, 815ANALYTICAL CHEMISTRY. 207of platinum (and gold) in silver can receive no more thanmention here. A paper on the error of the ordinary parting assayof platinum and silver by means of concentrated sulphuric acid,with a practical suggestion for reducing the error,76 emphasises theconcexion between the two problems, a circumstance which is stillfurther brought out if this paper be read in conjunction with amore recent one on the protective action of a small quantity ofsulphur when evaporating sulphuric acid in platinum vessels.77 Therecent work of Crookes on the volatility of the platinum metals 7* (ortheir oxides) confirms the conclusion of Mylius and Hiittner 79 thatthe improvement in the mechanical properties of platinum by alloy-ing it with iridium is accompanied by the disadvantage thatapparatus made from such alloys suffers greater changes of weightthan purer platinum when subjected to the usual treatment ofchemical platinum apparatus. For the most exact.analytical work,therefore, pure platinum is still t o be preferred, but for manypurposes the mechanical properties of the alloy outweigh its dis-advantages.This can scarcely be said of the alloy of rhodium withplatinum. For fifty years Crookes has made the subject of thepreservation of platinum apparatus so specially his own that hisrecommendation of rhodium alloys 80 must carry weight; but againstit must be set the very careful investigation a t the Reichsanstalt ofa crucible which showed abnormal lose of weight each time it washeated, a property which the investigation appears to correlate withan abnormally high, although absolutely small, percentage ofrhodium .81a s long ago as 1899 Eaufmann showed that thorium could beprecipitated quantitatively, even from very dilute and strongly acidsolutions, by sodium hypophosphate. It has since been found thatthis reaction may be used for the determination of thorium inmonazite sand, an operation which it greatly simplifies.92 So far ithas received little application, and this is in part explained by thefact that the reagent is not yet procurable commercially.It is saidthat this difficulty will shortly be removed, but meanwhile it may75 F. P. Dewey, J. 2nd. Eng. Chem., 1912, 4, 257; also Chem. News, 1912, 106,8 ; A., ii, 810.v6 A. Steinmann, Schweiz. Woch. Chem. Pharm., 1911, 49, 441, 453 ; A., 1911,ii, 1035.?7 L. W. McCay, 8th. Inter. Cong. Appl. Chem., 1912, 1, 351.78 Sir W. Crookes, Proc. Roy. SOC., 1912, A, 86, 461 ; A , , ii, 563.79 F. Mylius and C. Hiittner, Tatigkeitsber. d. Physik. - Tech. 12eichsundalt ;Zedsch. Eleelrochem., 1911, 17, 38.Ann. Report, 1908, 195.81 3lylius and Hiittner, loc.eit.A. Rosenheim, Chem. Zeit., 1912, 36, 821 ; A., ii, 869 ; F. Wirth, Zcitsch.angew. Chem., 1912, 25, 1678 ; A., ii, 948208 ANNUAL REPORTS ON THE PROGRESS O F CHEMISTRY.not be amiss to call attention t o a paper which gives references tothe methods available for its preparation.83Last year an accurate method was recorded for the determina-tion of vansdium in the presence of chromium, but that methoddepended on the use of electrolytic appliances which are not avail-able in every 1aboratory.M Another exact method, available inpresence of chromium, nickel, titanium, manganese, and molyb-denum, and depending on the quantitative inclusion of vanadiumby the phosphomolybdate precipitate, has since been described.85A new method f o r vanadium in presence of ten times its weightof chromium appears to be exact if applied with close attention todetail,86 but how close that attention needs to be is brought outin a joint paper by its author and J.R. Cain.87I n 1909, Weiss and Landecker described a number of newreactions of tantalum and columbium, and on two of them foundeda new method for the separation of these two elements fromtitanium and from each other.88 Their test numbers showed themethod to be more accurate than other more complicated ones. Itwas clear from their paper, however, that success was dependent onthe use of not too large a quantity of nitre in a fusion which wasone step of the process, and one English abstract laid stress on thisfact, and pointed out that the authors did not define this quantitywith the exactness which might have been expected.With thisexception the paper was so convincing that it was with surprise thatthe writer read last year that two other authors had investigatedthe method and found it wholly untrustworthy.89 It has now beenre-investigated with the result that the precipitation by carbondioxide proposed by Weiss and Landecker is shown to be a colloid-chemical phenomenon, a fact which explains the influence of morethan traces of nitrate and the results of the American critics of themethod, and at the same time makes it possible to define therequisite conditions for obtaining accurate results.90 Satisfactoryresults are obtained when the nitrate is omitted altogether.A recent paper on the volumetric estimation of titanium is worthyof note, but it contains no reference to the very careful work ofGernmel1,gl which left little to be desired save increased speed.This83 11. KOSS, Chem. Zeit., 1912, 36, 686; A . , ii, 809.84 Ann. Report, 1911, 165.85 J. R. Cain and J. C. Hostetter, J. Ind. Eng. Chem., 1912, 4, 250 ; A . , ii,87 J. R. Cain and D. J. Demorest, ibid., 256 ; A . , ii, 1101.88 Ann. Report, 1909, 147.89 H. W. Foote and R. W. Langley, Amcr. J. Sci., 1910, [iv], 30, 401; A.,* 0. Hanser and A. Lewite, Zeitsch. angew. Chem., 1912, 25, 100 ; A . , ii, 262.91 Ann. Aport, 1910, 168.1101. 86 D. J. Demorest, ibid., 249 ; A . , ii, 1100.1911, ii, 72ANALYTICAL CHEMISTRY. 209the latest contribution t o the subject accomplishes.9~ For theestimation oE very small quantities of titanium, a colorimetricmethod depending on the use of thymol in concentrated sulphuricacid solution has been proposed.The depth of colour produced isstrictly proportional to the amount of titanium present and twenty-five times as intense as the colour produced by hydrogen peroxide.93As it has been shown that all phenols, especially polyhydric phenolswith adjacent hydroxy-groups, give a similar reaction with quadri-valent titanium, it may be that a reagent even more sensitive thanthymol might be selected.94Electrochemical Analysis.The perpetual rise in the price of platinum in part informs twoother paragraphs in this report. I n no branch is it more felt thanin electrochemical analysis, and a paper on practical electro-analysiswith the weight of the platinum apparatus reduced to 1 gram,which might some years ago have been passed by as a freak, mustbe regarded as a serious and valuable contribution to-day.95 Furtherevidence of the utility of the mercury cathode,96 and of the ground-lessness of the objections that have been raised to it, is adduced,and it is shown that for ordinary purposes it is unnecessary toprovide for the determination of the cathode potential.97 I n otherrespects, too, this paper is inspired by the spirit of reaction againstunnecessary complexity of apparatus which was discernible inseveral directions last year.When a sojution of iron and man-ganese in ammonium oxalah is electrolysed in the presence ofhydrazine sulphate, iron is deposited quantitativeIy, whiIst theh‘ydrazine prevents the deposition of manganese dioxide a t theanode.08 Copper, tin, antimony, and bismuth can be depositedquantitatively and fairly quickly from hydrochloric acid solutions,provided a reducing agent, such as hydroxylamine hydrochloride,is present.99 The electrolytic estimation of zinc in presence ofammonium salts must take place in acid solution, and has beenpossible hitherto only when rotating electrodes are employed.Itis now found that stationary electrodes may be employed providedthat lactic acid and an alkali lactate are present.l92 1’. W. and E. B. Shinier, Slh. h t e r . Poag. AppL Chem., 1912, 1, 4 4 5 .y3 V. Lenhcr and W.G . Crawford, ibid., 285.y4 0. Hauser and A. Lewite, Ber., 1912, 45, 2180 ; A , , i, 847.95 F. A. Gooch and W. L. Burdick, Zcitsch. anorg. Chent., 1912, 78, 213 ; Anaer.96 Ann. Report, 1910, 170 ; ibid., 1911, 168.g7 P. Bmmann, Zeitsch. anory. Chern., 1912, 74, 315 ; A., ii, 489.93 R. Belasio, Ann. Lab. GabeZEe, 1912, 6, 207 ; A., ii, 1097.J. Sci., 1912, [iv], 34, 107 ; A., ii, 986.E. P. Schoch and D. J. Brown, 8th I d c r . Cony. AppZ. Chem., 1912, 21, 81.R. Helasio, Ann. Lab. Gabelle, 1912, 6, 239 ; A . , i i , 1096.REP -\‘Or,. l X 210 ANh’UAL REPORTS ON THE PROGRESS OF CHEMISTRY.An interesting example of another type of electro-analyticalmethod is one which seeks t o determine magnesium in presence ofcalcium by titretion of the magnesium, making use of the factthat when alkali hydroxide is added to a neutral solution containingsalts of calcium and magnesium the concentration of hydroxyl ionsincreases very slowly until all the magnesium is precipitated ashydroxide and then increases rapidly t o the point a t which calciumhydroxide begins to eeparate.2 A somewhat similar method appliedto an easier problem is that which determines nitric acid in amixture of sulphuric and nitric acids by neutralising the solutionwith barium hydroxide, and then titrating it with sodium carbonateuntil the sharp rise in the conductivity announces the completeprecipitation of the barium.3 Applied to mixtures of sulphuricand nitric acids containing only traces of other substances, themethod gives accurate results, but the method of Dutoit andDuboux, for the determination of sulphates in wine,4 which sug-gested it, has now been shown to be untrustworthy.The resultsof Dutoit and Duboux are not impeached, but their accuracy isshown to be quite fortuitous.6 On the other hand, it has beenshown that calculation of the sulphate in water from the knownchloride and nitrate content and the conductivity, whateverobjection may be raised to it, is capable of yielding fair results.6Organic Analysis.Three papers on qualitative methods may be mentioned, namely,one in which the colour developed on treatment with potassiumdichromate and sulphuric or nitric acid is used to divide amino-compounds into four main groups,’ one which provides a ‘systemfor distinguishing a number of the simpler phenols: and one whichserves for the detection of small quantities of maltose, lactose, andmelibiose when these occur together.9Early this year Hibbert issued a warning that the methoddevised by himself and Sudborough for the estimation of hydroxy-,amino-, and imino-derivatives by means of Grignard’s reagent,1°was not EO generally applicable as had been supposed,11 but it hasJ. H.Hildebrand and H. S. Harned, 8th Inter. Cong. Appl. Chem., 1912,1, 217.H. Corvazier, dlon. Sci., 1912, [v], 2, I, 322 ; A, ii, 1092.Ann. lteport, 1908, 207.Chim. anal., 1912, 17, 243.Li A. Bruno and P. T. D’duzay, Coinpt. rend., 1912, 154, 984 ; A., ii, 600 ; Ann.‘I F. Dienert, ibid., 1701 ; A., ii, 807.ti J. A. Sanchez, ibzd., 1911, [iv], 9, 1056; A., ii, 209.H.Agulhon and P. Thomas, Bull. Xoc. chim., 1912, [iv], 11, 69 ; A . , ii, 308.C. Neuberg and S. Saneyoshi, Zeitsch. Ver. deut. Zuckerind, 1912, 559.lo A m . Report, 1911, 170. l1 H. Hibbert, T., 1912, 101, 328ANALYTlCAL CEEMISTRY 21 1since been shown that Zerewitinoff’s modification, using pyridine assolvent, gives theoretical results with each of the substances whichbehaved anomalously in Hibbert’s hands. The method has nowbeen extended to diamines of the aliphatic, benzene, diphenyl,stilbene, and naphthalene series, and it is found that such sub-stances react with two atoms of “active hydrogen” a t room tem-perature, and with three atoms a t 85O, but that the fourth atomdoes not react a t all with magnesium methyl iodide.The oneexception so far recorded is malonamide, which reacts with fouratoms. The hydrocarbons of the indene and fluorene type do notreact a t the ordinary temperature, but a t 8 5 O a quantity ofmethane is obtained corresponding with one atom of “ activehydrogen.” 12This report provides an opportunity for calling attention to amethod for the determination of benzaldehyde, which might other-wise escape attention, owing to its inclusion in a paper with a moregeneral title. It depends on the use of p-nitrophenylhydrazine,and is shown to be far more accurate than any other availablemethod. The same method yields even better results withanisaldehyde add vanillin, but these aldehydee can be determinedwith equal accuracy by other means.13 A new colorimetric methodfor the estimation of vanillin, simpler than th0 methods usuallyemployed, is shown to give accurate results when applied to genuineextracts,l4 but since it depends on the us0 of a reagent which yieldsa similar colour with all substances containing phenol groups,15 itcannot be applied to extracts of unknown origin.The estimationof furfuraldehyde is of importance in the estimation of pentoses andpentosans, and the description of a new method,lO which avoided theuse of phloroglucinol and promised to be more expeditious thanthe phloroglucinol method, was welcome. It depends on the reduc-tion by furfuraldehyde of Fehling’s solution, but its author erredin supposing that the ratio of furfuraldehyde to copper reducedwas independent of the concentration.However, the reducingpower over a range of concentrations covering those which occur inpractice has since been determined, and the method at the sametime simplified.17Konig and Hiihn, having applied each of the principal methotlsfor the determination of cellulose in woods and textile fibros toa number of typical raw materials, reach the conclusion thatl2 T. Zerewitinoff, Ber., 1912, 45, 2384 ; A . , i, 841.l P 0. Folh and W. Denis, J. Id. Ewg. Chem., 1912, 4, 670.B. G. Feinberg, 8th Inter. Cony. Appl. Chenz., 1912, 1, 187.0. Folin and W. Denis, J. Bid. Chem., 1912, 12, 239 ; A . , ii, 1011J. T Flohil, Chem Weekblnd, 1910, 7, 1057 ; A . , 1911, ii, 160.L. Eyno~l and J . H. Lane, AnnEyst, 1912, 37, 4 1 ; A ., ii, 305.r 212 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Konig’s method,’* in spite of the low yield of cellulose obtained,is the best one for the quantitative determination of that con-stituent. Next to Konig’s, the method of Tollens-Dmochowski 19 ispreferred, according to which hydrolysis by the Weende-Henneberg(“ crude fibre”) process is followed by oxidation with nitric acid.It is said that Cross and Bevan’s chlorination method fails toremove most of the hemi-hexosans and pentosans.20 Cmss andBevan remain convinced that the chlorination process, properlycontrolled, will remain the standard method for the estimation ofcellulose.21 Where experts differ so widely, i t is not for the presentwriter to express an opinion, but it is permissible to say that theissue seems to turn on the definition of “cellulose.”Innumerable methods have been described for the detection andapproximate estimation of small quantities of methyl alcohol.Forsome of these, based on the oxidation of the methyl alcohol toformaldehyde and detection of the latter by some old or newreagent, an extraordinary degree of sensitiveness has been claimed.I n practice, however, most of them have proved to be a t the mostbut new colour reactions of formaldehyde, for which other tests ofgreat delicacy have long been known, whilst they did nothingto advance the practical problem of detecting methyl alcohol inthe presence of ethyl alcohol, since, unless the oxidation is carefullyregulated, ethyl alcohol gives rise to acetaldehyde, which interfereswith some of the tests, or even to formaldehyde, which upsets themall.As a result of a careful study of the conditions necessary topreserve the ethyl alcohol from oxidation, one of the later proposalsof DenigSs 22 has been so far improved upon that it is now possibleto detect with certainty 0.05 per cent. of methyl alcohol in 90 percent. alcohol, and to determine larger quantities with an error notexceeding 4 per cent. of the amount of methyl alcohol present.23 Anovel proposal for the estimation of alcohol in pure aqueous solutiondepends on the fact that when sufficient potassium fluoride is addedto such solutions, two phases separate. If water be now addedgradually, a point is reached when the solution again becomeshomogeneous, and this point can be determined very closely if alittle solid spirit blue is present, since this makes a very thin filmof alcohol easily visible.The paper is accompanied by tables forJ. Konig, Ber., 1908, 41, 46 ; 1908, ii, 236.I9 R. Dmochowski and B. Tollens, J. Landw. 1910, 58, 1 ; A., 1910, ii, 554.~0 J. Konigand F. Huhn, Zcitsch. Farb. Itid., 1911, 10, 297, 236, 344, 366 ;21 C. F. Cross and E. J. Bevan, ibid., 1912, 11, 197 ; A,, ii, 1105 ; also 8th Inter.2? G. DenigBs, C‘ompt. rend., 1910, 150, 832 ; A., 1910, ii, 461.23 C. Siminonds, Analyst, 1912, 37, 16 ; A., ii, 208.19 2, 11, 4, 17, 37, 57, 77, 102 ; A , , ii, 1005.Cong. Appl. Chem., 1912, 13, 101ANALYTICAL CHEMISTRY. 213the use of the analyst.Although potassium fluoride cannot beused for the production of alcohol of a strength exceeding 97.5 percent., its dehydrating action is more rapid than that of most dryingagents, and it absorbs as much as 62 per cent. of its weight ofwater against 32 per cent. taken up by lime.24An interesting development of Walden's observation that, undercertain conditions, uranyl salts produce a marked change in thespecific rotation of Z-malic acid, is a practical method for theestimation of malic acid in fruit juices.25 The determication ofsmall quantities of salicylic acid in presence of much p-hydroxy-benzoic acid is less frequently required than the converse p r oposition, but a method by which it may be effected is interestingas an example of a new type of biochemical method of analysis.It depends on the fact that Penicillium glaucum can utilise p-(andalso m)hydroxybenzoic acid for its nutrition, whilst salicylic acid,in quantities of more than 1 per cent., greatly retards, if it does notwholly inhibit, the growth of the organism.26Attention may be directed to two papers on the volumetricestimation of azo-dyestuffs by means of hyposulphite. I f thelimitations of the method are obvious, so are its practical advan-tages in cases where the results of the titration are free fromambiguity.27The selection, for notice in this report, of methods relating tothe analysis of foods, drugs other than definite chemical substances,and still more of natural fats and oils, is attended with difficulty.Of the papers that one can point to as unquestionably of value,the majority concern themselves with constants and ratios, andalthough in the aggregats they advance our knowledge considerably,i t is only occasionally that any one or any group carries our know-ledge of some problem of general interest to a point where a sum-mary of progress made can be usefully attempted.Of new methodsexperience forbids one to expect that more than a very few willprove at the best more than alternatives for existing ones, whilstby nothing short of actual trial can one distinguish between testslike that of Halphen for cottonseed oil, which has proved specificand approximately quantitative, and the hundreds of colour testswhich have been described since, and since forgotten.EvenHalphen and those who have used his test for fifteen years have24 G. B. Frankforter and F. C. Frary, 8th Inter. Cong. Appl. Chem., 1912, 22, 87.25 P. B. Dnnbar and R. F. Bacon, J. Ind. Eng. Chem., 1911, 3, 826; A., ii,699.26 J. Roeseken and H. J. Waterman, Proc. K. Akad. Wetensch. ilmslo.dnm, 1911,14, 604 ; A . , ii, 306.27 E. Grandmougin and E. Havas, Chem. Zeit., 1912 36, 1167 ; A., ii, 1220 iW. Siegmund, Monalsh., 1912, 33, 1431 ; A., 1913, ii, 82214 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.been lucky if the recent statement proves correct that Halphen’sreaction fails if the amyl alcohol employed is a carefully purifiedsample.28 Fortunately, even if true, this does not touch theprinciple of Halphen’s test, as Veley’s work did absolutely destroythe whole basis of the Girard-Cuniasse method for the determinationof the higher alcohols in potable spirits, and as Chapman’s worklimited the conclusions to be drawn from the so-called tests forcreatinine and hydrogen cyanide by means of picric acid.Thehistory of these three methods, however, one good, one bad, andone indifferently good, but all a t some time in daily use by manyfood analysts, g-oes to show that the early literature of a colourreaction is not likely to provide internal evidence by which it maybe appraised.It was stated by the reporter of 1910, whose opinion on thissubject is entitled to special respect., that the methods of Jollesand Lcmeland €or the estimation of sucrose in presence of reducingsugars, even if trustworthy, which was open to doubt, had noadvantage over existing methods.29 That neither is as accurateas might be desired has since been demonstrated, but the formeris shown to afford a ready means of determining sucrose in con-densed milk with a satisfactory degree of accuracy,30 whilst thelatter has been so far improved that, in its modified form, it isrecommended as a substitute for the Clerget-Herzf eld inversionmethod in the analysis of sugars and syrups, and especially as aroutine method for the determination of the sucrose-content oflow-grade products like hobroom goods.31Reference was made last year to a cryoscopic method for deter-mining whether a sample of milk, abnormally low in non-fattysolids, owed its abnormality to addition of water or to naturalcauses.% For three years the method has been applied to alldoubtful samples in the Government Laboratory a t Brisbane,% butunfortunately it seems to form no part of the duty of any officialin this country to investigate a method a knowledge of which mightperhaps have reduced the number of cases of miscarriage of justicein the autumn of 1911.Richmond’s annual paper on the coni-position of milk shows that the average for non-fatty solids fell inAugust, 1911, almost to the legal minimum, whilst 27 per cent. ofall the samples examined were below the standard.% In only a28 E. Gastaldi, Ann. Lab. Gabellc, 1912, 6, 601 ; A., ii, 1108.29 A m . Report, 1910, 174.H. Nowak, Zeitsch. anal.Chem., 1912, 51, 610 ; A . , ii, 1004.W. E. Cross and W. G. Taggart, Zntem. ASugar J., 1912, 14, 444.32 Ann. Report, 1911, 173.33 J. B. Henderson, AILS. Assoc. Adv. Sci., 1912, 13.3A H. D. Richniond, Aitnlyst, 1912, 37, 298ANALYTICAL CHEMISTRY. 215few cases, however, was evidence of this kind available a t the time,and in not a11 of those cases did i t carry the necessary weightbecause based on comparatively few samples.Fiehe and Stegmiiller have examined most of the methods whichhave been proposed for the examination of honey, and give detaileddirections for carrying out the tests they have found most useful.35Fiehe’s test for artificial invert sugar 36 is now firmly established,but has perhaps not reached its final form, as another new reagenthas been recommended,37 whilst useful suggestions have been madefor avoiding the interference of formaldehyde and some other sub-stances which may be present,38 and a Ihird paper gives hints onpreczutions to be observed in carrying out the process, some ofwhich may save the analyst trouble.% Langer’s serological test 40has been further studied, and i t is found that the quantity ofthe precipitate is not diminished by previously heating the honeyunless a temperature approaching looo is reached, and maintainedfor ten minutes or longer.41A new method for the estimation of shell in cocoa powdersdepends on the precipitation of cocoa-red by means of ferric chloridefrom an acetic acid extract of the cocoa.It is claimed that i t willdetect with certainty the addition of 10 per cent.of shell to“normal cocoa.” In conjunction with other tests, it may proveuseful, ’but the figures given in the paper show that, if this testwere the sole criterion, i t would be possible to add nearly 40 percent. of shell to cocoa from one part of the world, and thenplead as excuse, for its low content of cocoa-red, the normal contentof the latter in cocoa grown elsewhere.42 T. Macara has developedFilsinger’s levigation method for the determination of shell to suchan extent that the authors of a recent b00k,43 as yet the only placeof publication of the modified method, express the opinion that itis of more value than any chemical method yet available for thepurpose. Another mechanical method depends on the use of amixture of chloral hydrate, glycerol and water, having a specificgravity of 1.5, and the use of a centrifuge to separate the heavierportions of the shell (about a third of the whole) from the rest ofthe mixture.44rs Arhtit.Kaiscrl. Gesiindhcitsnmte, 1912, 40, 305.36 Ann. Report, 1910, 174 ; 1911, 174.37 G. Armani and J. Bartoiri, Ann. Lab. Gabelle, 1912, 6, 85.39 L. Stoecklin, ibid., 116 ; A . , ii, 499.40 Ann. Report, 1910. 1 7 i ; 1911, 1 7 4 .41 J. Thoni, Mitt. LcbcnPmilteltLnters~~ch. 16. Hyg., 1912, 3, 74 ; Chem. Z i t . ,42 C. Ulrich, Arch. Phnmz., 1911, 249, 524.4i “ Fatty Foods,” by E. R . Rolton and C. Revis (Churchill), 303.44 L. Kalnsky, Zeitsch. Nahr. Genwsm., 1912, 23, 654.G. lidphen, Ann.Fulsif., 1912, 5, 105 ; A . , ii, 498.1912, 11. 151216 ANNUAL REPORTS ON THE PHOGRESS OF CHEMISTRY.Concerning the best method for the determination of nicotinein tobacco extracts, there is still disagreement, but the method ofBertrand and Javillier,l5 depending on the precipitation of nicotineby means of silicotungstic acid, is now admitted by T6th-theauthor of a rival method-to leave nothing to be desired in pointof a c c u r a ~ y . ~ ~ Thus at least three methods have survived thecriticism which has been so active during the last three years.47There can be no doubt, however, that that criticism has soundedthe death-knell of the method of Ulex after a useful life of twenty-five years, although there are commercial reasons for believing thatits burial may be deferred.**Attention may be directed to a new method for the separatedetermination of citronellal and geraniol in citronella oiI,49 and toanother for the direct determination of the geraniol, which dependson the conversion of the citronellal into its oxime, which passesinto the nitrile on boiling with acetic anhydride to acetylate thegeraniol, the nitrile being unaffected by the potassium hydroxidesubsequently used to saponify the acetylated oi1.60A recent method for the determination of the preservative, softresins, in hops is probably the most accurate yet described.Thatit will prove too tedious for commercial purposes is recognised byits authors, but they have made use of it to compare the accuracyof the quicker methods in general U S ~ .~ I They find that th‘e latestform52 of Lintner’s volumetric method gives results within a fewtenths of 1 per cent. of those yielded by their own, whilst that ofBriant and Meacham,53 the first practical method evolved for thepurpose, and still the prevailing one in this country, always under-estimates the soft resins, and may in some cases underestimate therdby as much as 50 per cent. The paper is written with a clearerappreciation of earlier work than is usually shown by writers onthis subject, the literature of which is difficult to piece together.Mention is made earlier in this report of the use of the calori-metric bomb in the determination of carbon in steel. A similarmethod is now employed in the United States Bureau of Mines45 G .Bertrand and M. Javillier, Bull. SOC. chim, 1909, [iv], 5, 241 ; A . , 1909,46 J. T6th, Chem. Zeit., 1912, 36, 937 ; A . , ii, 1010.47 Ann. Report, 1911, 174.49 See, for instance, F. Porchet and P. Tonduz, Chem. Zeit., 1912, 36, 843;49 V. Boulez, Bull. Soc. chim., 1912, [iv], 11, 915 ; A . , ii, 1105.50 J. Dupont and L. Labaune, Ann. Chim. anal., 1912, 17, 210 ; A . , ii, 697.O1 H. V. Tartar and C. E. Bradley, J. lnd. Eng. Chewi., 1912, 4, 209.52 C. J. Lintner, Chem. Zeit., 1908, 32, 1068 ; see also D. Neumaun, Wuchensclb.53 L. Briant and C. S. Rleacliam, J. Insl. Brcwing, 1897, 3, 233.ii, 450 ; Ann. Chim nnnl., 1911, 16, 251 ; d., 1911, ii, 827.R. Kissling, ibid., 1321.f. Brauerei, 1910, 27, 281ANALYTICAL CHEMISTRY.217for the determination of sulphur in oils and all other combustiblematerials. Of theother methods available for this determination in oils, only thatof Carius and those depending on combustion in a current ofoxygen are accurate. None of the methods depending on combus-tion in oxygen contained in a large glass bottle is trustworthy,nor are methods depending on fusion with alkalis and subsequentoxidation, fusion with alkalis and an oxidising agent, or wet treat-ment with alkalis and oxidising agents, although the three last typesof method, of which there are many varieties, are excellent forsolid fuels.54 Ten methods, all of which find use for the determina-tion of water in petroleum, have been subjected t o a critical exam-ination, and it is found that the most accurate is that dependingon the volume of hydrogen evolved when the sample is treated withfinely-cut sodium.55Agricu l tural Analysis.Subject to correction by those exclusively devoted to agriculturalchemistry, the writer would characterise as the most importantpaper affecting British agricultural analysts, one on the estimationof potassium.56 Yet there is little in the paper that is new.It isshown that the perchlorate method is 'far superior to the platinummethod, not only on the ground of cost, but as being freer fromerror, simpler, and less likely to give rise to differences betweendifferent analysts. All this information was available a t least aslong ago as 1909, when Precht stated that the perchlorate methodhad almost superseded the platinum method throughout Germany.57The present paper includes an improvement on Neubauer 's simpli-fied method, which will prove of special value in dealing with smallquantities of precipitate, but the importance of the paper as a wholedepends less on the new information it conveys than on the fact thatit is dated from Rothamsted, and may therefore expect to resultin action by the Board of Agriculture, which apparently is notadvised of analytical progress in foreign countries.The case isurgent. Analysts under the Fertilisers and Feeding Stuffs Act arerequired to determine potassium in fertilisers by the platinummethod, t o take large quantities of substance for analyfis, and tomake declaration that the analysis was made in accordance withthese and other directions of the Board,a and these regulations neces-54 I.C. Allen and I. W. Robertson, 8th Inter. Cong. App!. Chem., 1912, 10, 25.55 I. C. Allen and W. A. Jacobs, ibid., 17.56 W. A. Davis, J. Agric. Sci., 1912, 5, 52.57 Precht, 7th Inter. Cong. Appl. Chem., 1909, I, 146 ; also 6th Congress, 1906.58 Board of Agriculture and Fisheries, Statutory Bales and Orders, 1908, No.The method is rapid and as accurate as any.9642'18 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.sarily affect the practice of other analysts besides official agriculturalchemists. I f the platinum method is to be used, the use of largequantities makes for accuracy, yet quite recently two of the bestknown London analysts differed by nearly 3 per cent.in the caseof a sample of 90 per cent. sulphate of potash.59 I n view of factssuch as these, there is no necessity or excuse for deferring theofficial dethronement of the platinum method until the perchloratemethod, which admittedly falls short of exactness, has been furtherrefined A notable feature of the paper is the frankness with whichnumbers leaving much to bo desired in point of accuracy are usedto sustain the argument and yet not overstate the case. Sincefrequent alterations of official methods of analysis are attendedwith inconvenience, the Board might a t the same time invite privatecriticism of their other methods of analysis with a view to there-casting of these. So unfortunately are they worded that thewriter has heard it suggested that they must be the work of achemist with so strong an objection to the principle of officialmethods that he drafted this set t o discredit that principle.Thewriter must not be held to endorse this opinion, but that i t can beexpressed, even in jest, reflects seriously on a document which hasthe force of law.Modern legislation has made official methods necessary, but insome cases the legislative enactments or statutory rules and ordersmade under them have been in advance of the knowledge necessaryto frame satisfactory methods. A case in point is the use of astandard solution of ammonium citrate. Such a method has beenofficial in the United States for some years, and the Americanchemists did their best to define the hydrion concentration-onwhich the results obtained by the use of the solution are closelydependenbso far as this could be done with chemical indicators.Yet i t is possible for two chemists, experienced in the method, toreport, one 3 per cent., the other 5 per cent.of insoluble phosphatein one and the same sample. If electrical appliances are availablefor determining the hydrion concentration of the standard solution,satisfactory concordance is easily established, but many labora-tories lack the necessary equipment.60 Fortunately, this test isnot official in Great Britain, but in a t least one of our colonies everyfertiliser must be registered with a declaration, which must alsoappear on invoices, that i t contains not less than so much " citrate-soluble" phosphoric acid.A proclamation by the Governor of theColony 61 defines the strength of the citric acid solution to be used59 Davis, Zoc. cit.60 A. J. Patten and C. S. Robinson, J. Ind. &zg. C h c ~ , 1912, 4, 443 ; A . , ii,61 Proclamation by tha High Commissicner for South Africa, 5 Dec., 1910.1094ANALYTICAL CHEMISTRY. 219with basic slag, which is identical with that in use in nearly allcivilised countries, but is silent concerning the more or less neutralsolvent to be applied to all other fertilisers, which demanded rigiddefinition, if indeed the use of such a solution for official purposescan be recommended a t all in the present state of our knowledge.A discussion on the methods available for the rapid yet suffi-cienhly exact determination of phosphoric acid in soils 62 was chieflyremarkable for the absence of any reference to the lead molybdatemethod?! which, if not quite so rapid as some, is probably the mostexact of tho approximate methods, and possibly the most exactmethod available where the quantity to be determined is verysmall.Water Analysis.It is on the actual concentration of hydrogen ions, and not onthe potential acidity or alkalinity as given by ordinary titration,that the rate of action of water on metals depends.Electrioalmeasurements of this concentration are difficult in some circum-stances-for example, on board ship-and require appliances nota t the disposal of every analyst. Modifications of Siirensen’smethod of colour comparison with standards prepared from thesalts N+HPO, and KH,PO,, water and standards alike being tintedwith a suitable indicator, are free from this objection, and the useof his method made possible the collection of much useful informa-tion during a cruise made for hydrographical purposes from theBaltic to the Black Sea.64 A recent proposal to state the resultsof such a comparison in a new way, although certain to arousecriticism, will commend itself to the water analyst, and do much toestablish the method as it recognised part of a complete wateranalysis. Briefly, the proposal is to state the acidity or alkalinityof a solution in terms of the acidity or alkalinity of water at thesame temperature; thus a water with a hydrion concentration of1 * 3 6 ~ 1 0 - ~ a t 1 8 O would be said to have a relative acidity of 2 orrelative alkalinity of 0.5. One advantage of the method is that theresults thus stated are nearly independent of temperature. Anotherand by no means insignificant one, is that the client without scien-tific knowledge will have less difficulty in appreciating the signi-ficance of the analyst’s numbers. The proposal is accompanied bytables which do away with all calculation, and place the method a tthe disposal of anyone who will make up two standard solutions.6563 S. J. M. Auld, Anal@, 1912, 37, 130.63 H. Rrearley and F. Ibbotson, Analysis of Eteel- Works Materials (Longnians),64 Palitzsch, Compt. rend. des Travaux tlu Lab. de Carlsberg, 1911, 10, 85 ; also55.Bwchent. Zeitsch., 1911, 37, 116; A., ii, 39.J. Walker and S. A. Kay, J. Soc. Chent. lnd., 1912, 31, 1013 ; A . , ii, 1215220 ANNUAL REPORTS OX THE PROGRESS OF CHEMISTRY.A paper which draws attention to the fact that, in carbonatedwaters, ammonia may be much underestimated, unless special pre-cautions be taken, should not be overlooked, although othermethods for securing accurate results, besides that proposed by theauthors, may readily suggest themselves.66 Since the last referenceto water analysis in these reports,67 Chamot and Pratt, in collabora-tion with Redfield, have brought to a conclusion their importantresearches in connexion with the estimation of nitrates by theGrandval and Lajoux method. Those researches have laid bare themechanism of the Grandval and Lajoux reaction, hitherto whollymisconceived, and establish on sound principles the precautionsnecessary to attain the best results of which the method is capable,but these precautions and the results are such that the authors maybe said to have assisted to destroy the method to the investigationof which they have given so much labour. To known waters themethod may often be applied, and in such cases reference to thework cited will save the analyst trouble and increase the accuracyof his results; but as a general method it cannot compare withothers which are nearly if not quite as quick and far more trust+wort hy.68G. CECIL JONES.66 G. D. Elsdon mid N. Evers, Analyst, 1912, 37, 173 ; A . , ii, 601.67 Ann. Xeeport, 1910, 181.E. M. Chrtmot, D. S. Pratt, and H. Mr. Redfield, J. Amcr. Chenz. Soc., 1911,33, 366, 381 ; A , , 1911 ii, 331