ANALYTICAL CHEMISTRY.THE general plan of this Report is identical with that followed inthe corresponding Reports of the last three years. The arrange-ment of the matter falling under the head of organic analysispresents some difficulty, as may be seen by the paragraphs relatingto the estimation of sucrose and other carbohydrates beingseparated by matter cognate to neither. As far as possible,organic methods of general interest are dealt with first, and arefollowed by short paragraphs dealing with the analysis of foods,drugs, fats, fermentation products, and other materials of specialtechnical interest, but the number of papers referred t o undereach of these heads is too small to justify cross-headings. Wateranalysis, not referred t o last year, is given a good deal of space,but the separate section on physiological methods is omitted thisyear, such physiological methods as have been selected for noticefinding a natural place elsewhere in the volume.General.The solubility, and especially the curve expressing change ofsolubility with change of temperature, are valuable criteria indeciding the identity of a substance, and also in some cases indeciding whether i t is accompanied by impurities, many of whichhave a marked influence on the solubility.Whereas other physicalconstants are invariably determined for new substances, concern-ing many there are no solubility data, or at most a single valuefor some one temperature. This is, no doubt, partly due to thedifficulty attending solubility determinations by the methods gener-ally in use, usually modifications of Noyes’ or of Victor Meyer’smethod.When using Koyes’ mebhod, great difficulty is experi-enced in obtaining accurate results a t high temperatures, andalthough V. Meyer’s method gives results of sufficient accuracy, itrequires the use of a very large thermostat, which must be main-tained constant within 0*lo. A method and apparatus haverecently been described which greatly simplify the accurate deter-mination of solubilities a t high temperatures. The principle ofM 161 REP.-VOL. X162 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.the method is the determination of solubilities a t the boiling pointof the saturated solution, the boiling point being varied by chang-ing the pressure. Conducting the determination a t the boilingpoint ensures thorough agitation and eliminat'es the necessity forusing a thermostat, since a t constant pressure the boiling pointis also constant.1The work of Bingham and his collaborators in improving themeans of determining viscosities accurately, yet quickly, and indrawing attention to a fundamental error which vitiated muchrecent work in this field, was referred to two years ago.2 In 1912they described a viscometer with which absolute viscosities couldbe measured with great certainty.That form of apparatus iseasily made, but since its dimensions must be accurately knownfor absolute measurements, the time consumed in calibration isconsiderable.3 For general purposes, it is preferable to calculatethe absolute viscosities from measurements that are only relative.By this procedure the calibration is simplified, and the apparatusmay be made simpler and less delicate t o handle.Many visco-meters have been devised for relative measurements, but relativemeasurements are comparatively useless unless the results can becalculated to absolute units. That measurements obtained by theuse of instruments of the Ostwald type could be calculated toabsolute units without difficulty has been shown to be an assump-tion not generally true, this being mainly due to the fact thatthe formula almost universally used takes no account of thO loss ofkinetic energy of the liquid within the capillary, this energy dis-appearing outside the capillary without helping t o overcome viscousresistance within i t .4 Viscometers of the ordinary type aredeficient, because the pressure producing the flow through thecapillary is not variable a t will. The result is that, with very fluidsubstances, the kinetic energy correction becomes large unavoid-ably, and with rather viscous substances the time of flow becomesintolerably long, necmsitating the use of several instruments.With a very long period of flow, the difficulties due t o cloggingwith dust particles become very great.5 The necessity of theknowledge of the exact density of the liquid a t each temperaturewhere a viscosity measurement is desired lessens the convenience ofthis type of instrument. Both the untrustworthiness and the1 L.A. Tschugaev and W. Chlopin, Zeitsch. anorg. Chem., 1914, 86, 154;A., ii, 348.Ann. Report, 1912, 195.E. C. Bingham and G. F. White, Zeitsch. physikal. Chem., 1912, 80, 670 ;E. C. Binghsm, T., 1913, 103, 959.M. P. Applebey, ibid., 1910, 97, 2000.A., 1912, ii, 1144ANALYTICAL CHEMISTRY. 163inconvenience of these instruments may bs avoided by usingvariable pressure, and an instrument working on this principle hasnow been described which makes it possible to determine theviscosity of any liquid with an error not exceeding 0.1 per cent.For more accurate work, absolute measurements must be made, asdata do not exist for standardising a relative instrument for amuch higher degree of precision. The instrument is standardisedon water at several temperatures, with hexane for very fluidliquids and with sugar solutions for very viscous ones.The instru-ment is so designed that the correction for kinetic energy is alwayssmall, and can therefore be calculated with sufficient accuracy fromroughly approximate measurements. Certain minor correctionsare necessary, one of them involving the density of the liquid, asdoes the correction for kinetic energy, but, as both these correctionsare very smaIl in a properly proportioned instrument, the densityneeds only t o be known approximately, which constitutes a great,advantage of the method.’jThe most useful contributions t o microchemical analysis duringthe year have been two papers dealing with such widely differentsubjects as elementary analysis of organic substances and theelectro-deposition of metals.The latter paper needs to be supple-mented by another, which is promised, dealing with such detailsas current density, time required for complete deposition, and com-position of electrolyte ; but the apparatus described leaves little tobe desired, since i t can bet easily and cheaply constructed, andserves for the deposition of 0.2 mg. of some metals with an errornot exceeding 1 per cent., and that in the presence of 100 timesas rnucli of another metal.’ Methods such as this cannot fail tolead to the adoption of microclieinical methods even in cases wherethere is no lack of material available for analysis, since suchmethods effect economy, not only of materials, but of the aiialyst’stime, and often, as in this case, of platinum. The author of theother paper cited, which describes and illustrates Pregl’s apparatusfor micro-elementary analysis, goes so far as t o predict Chat Pregl’smethods will become the normal methods of combustion analysis,the results being as exact as those obtained by careful work withlarger quantities, whilst the saving of time is very great.Incident-ally, this paper describes certain improved absorption tubes andother apparatus, the principles of which may find useful applicationeven in laboratories where i t may not be decided to adopt micro-met hods.*E. C. Bingham, J. Ind. Eng. Chem., 1914, 6, 233 ; A., ii, 342.7 R. Heinze, Zeitsch. angew. Chem., 1914, 27, 237 ; A., ii, 482.* J-.V. Dubsky, Chem. Zeit., 1914, 38, 505, 510; A., ii, 486164 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY,Gas Analysis.The iodine pentoxide method for the estimation of carbon mon-oxide has received much attention this year, one object commonto all the authors being reduction of the volume of gas which mustbe taken for the test. This object is most successfully achievedby Graham and Winmill,” who use a 20 C.C. Haldane burette and yetget results accurate to 0.02 per cent. Where the total quantity t obe estimated is of the order of 0.1 per cent., a more suitable methodis that of Sinnatt and Ctamer,lo which will determine such quanti-ties within about 0*001 per cent. when 2500 C.C. of gas are used inthe test. These authors determine, by Pettenkofer’s method, thecarbon dioxide produced by the oxidation of the carbon monoxide.Their results do not confirm the statement of Seidell11 and earlierworkers that higher and more accurate results are obtainable bydetermining the iodine liberated from the pentoxide.Eachmethod must be applied with care where the amount of carbonmonoxide to be estimated is small, but it is probable that theestimation of the carbon dioxide, using Sinnatt and Cramer’smethod, offers least difficulty where the concentration to be esti-mated is of the order stated and the analyst inexperienced in eithermethod. I n skilled hands, however, the estimation of the iodineby means of N/lOOO-thiosulphate can be used to determineaccurately a much smaller quantity of carbon monoxide.12 Thefurther statement of Sinnatt and Cramer, that there is no risk oferror in conducting the oxidation a t 160°, is unfortunate, although,no doubt, true when the carbon dioxide produced, and not theiodine liberated, is determined.That their results by the iodine-titration method were accurate, however, was solely due to thefact that hydrogen was absent from the gas undergoing analysis.Although it is known that hydrogen alone does not react withiodine pentoxide until a much higher temperature than looo isreached, occasional erratic results have led most users of the iodinepentoxide method to conduct the oxidation a t a temperature wellbelow looo. It has now been shown that, in the presence ofcarbon monoxide, hydrogen reacts with iodine pentoxide even a t90°, and the exact conditions under which the velocity of reactioncan be kept negligibly small, without unduly reducing thetemperature and increasing the time required for the oxidation ofthe carbon monoxide, have been determined.13 Yellow mercuricJ.I. Graham and T. F. Winmill, T., 1914,105, 1996.lo F. S. Sinnattand B. J. Cramer, Analyst, 1914, 39, 163; A , , ii, 383.l1 A. Seidell, J. Ind. Eng. Chm., 1914, 6, 321 ; A . , ii, 489.l2 G. N. Huntly, Analyst, 1914, 39, 169.l3 J. I. Graham and T. F. Winmill, Zoc. citANALYTICAL CHEMISTRY. 165oxide, contained in dark glass apparatus, has also been shown tobehave like iodine pentoxide as a selective oxidising agent,oxidising carbon monoxide quantitatively, but not attackingmethane a t lOOO.14A paper on the analysis of complex gas mixtures, issuing fromthe United States Bureau of Mines, introduces no novel principle,but combines several old ones in such a manner as to increasematerially the precision of technical analysis without great loss ofspeed.I n particular, it makes it possible to estimate accuratelythe hydrogen and paraffins in coal gas, and to determine the meanmolecular weight of the paraffins, which may. differ so much fromthat of methane that methods which assume the absence of higherparaffins may underestimate hydrogen by 3 per cent., and over-estimate the paraffins (calculated as methane) by a like amount.Carbon dioxide, olefines, and oxygen having been removed in theusual manner, hydrogen and carbon monoxide are oxidised bycopper oxide a t about 250°, and separately estimated from thecontraction due to combustion and further contraction on treat-ment with potassium hydroxide.The paraffins are then estimated,and their mean molecular weight determined by slow combustionwith oxygen, and measurement of the contraction and of thecarbon dioxide produced.15 It has been shown that the simul-t,aneous presence of cerium dioxide, already recommended as acatalyst in ordinary combustion analysis,lG has the property ofgreatly increasing the speed of oxidation of hydrogen by copperoxide, whilst i t does not bring about the oxidation of methane at,270O.17A recent paper on errors in gas analysis, due to assuming thatthe molecular volumes of all gases are alike, might have passedunnoticed not many years ago on the ground that the unavoidableerrors of experiment-at least in technical gas analysis-were largein proportion to those t o which attention is now directed.Technical methods of gas analysis have now, however, been soperfected that refinements in calculation, such as are suggested bythe authors of the paper referred to, are fully justified.18Iizorgaaic d nalysis.Qualitative.-Statements in the literature as to the seiisitivenessof various reagents for lead being conflicting, experiments havel4 L.Moser anci 0. Schmid, Zcitsch,. anal. ChC??Z., 1914, 53, 217 ; A . , ii, 384.l5 G. 13. Taylor, J . l n d . Eng. Chcm., 1914, 6, 845 ; A . , ii, 814.lS Ama. Repod, 1913, 178.17 J.P. Wibnut, Chcni. Wcekblad, 1914, 11, 498; A , , ii, 585.l8 G. A. Burrell and F. M. Seibert, U.S. Bureau of Mines, Technical Paper, 54166 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.been conducted with a number of such reagents under conditionsas nearly as possible identical, and the results tabulated in a usefulform.19 The reaction of copper with 1 : Z-diaminoanthraquinone-3-sulphonic acid, described by UhlenhuthFO is not specific, but isgiven by nickel and cobalt also. At the same time, it has beenfound that a simple extension of the test, which is of ext'raordinarydelicacy, suffices to differentiate between these three metals.21A method that will detect aluminium in presence of 1000 timesits weight of iron is noteworthy. It depends on the solubility ofbarium aluminate, and i t is not unreasonable t o suppose that itmight be developed into a method for the approximate estimationof traces of aluminium.22Quantitative.-A new method of estimating bromide in presenceof chloride has been based on the fact that, under certain coil-ditions, telluric acid liberates bromine from bromides, but notchlorine from chlorides.The test results show that the method iscapable of yielding excellent results, and the authors have beencareful t o indicate thO conditions necessary to success, but, likemany other methods due to Gooch, a very slight departure fromthe conditions laid down may give rise to large errors, and already,in our own Journal, the method has been misdescribed in anessential particular.23I n view of the fact that every year slight modifications of oldmethods for the estimation of phosphoric acid are recommended,modifications which seem to serve no useful purpose, attentionmay be directed to a recent paper dealing with the Pemberton-Kilgore method, which depends on the precipitation of phosphoricacid as ammonium phosphomolybdate, solution of the washed pre-cipitate in excess of standard alkali, and titration of the excess ofalkali, using phenolphthalein as indicator.I n spite of manysources of error, the method gives results sufficiently accurate formany purposes if carried out by a chemist accustomed to use it,and saves much time when many determinatims have to be made.For occasional work it is likely to, prove slower, as well as muchless accurate, than the gravimetric (molybdate-magnesia) method.The paper referred to should show any chemist whether the methodis one suited to his circumstances.Some fifty cognate papers arecited.24l9 E. Eegriwe, Zeitsch. anal. Chem., 1914, 53, 420 ; A , , ii, 579.2o Ann. Report, 1910, 163.21 G. Malatesta and E. di Nola, Boll. Chiin. farm., 1913, 52, 819 ; A., ii, 220.22 G. H. Petit, J. Pharm. Chinz., 1914, [vii], 9, 66 ; A., ii, 221.z3 F. A. Gooch and H. I. Cole, Zeitsch. nnorg. Chem., 1914, 86, 401 ; Amer.24 P. L. Hubbard, J. Id. Eng. Chem., 1913, 5, 998 ; A., ii, 145.J. Sci., 1914, [iv], 37, 257 ; A., ji, 379ANALYTICAL CHEMISTRY. 167Several valuable papers dealing with the estimation of silica werepublished almost simultaneously early in the year.All the authorsare agreed that about 1 per cent. of silica is lost if only a singleevaporation with acid is made, that repeated evaporation withoutintermediate filtration is useless, but that all but a negligibletrace of the lost silica can be recovered by a single evaporation ofthe filtrate. This has perhaps been the prevailing view for someyears, but i t was not universally held. One paper deals with theinfluence of other acids than hydrochloric acid,25 whilst anothershows that the difficulty in igniting silica to constant weight isnot due to the obstinate retention of water, but to the fact that,at the temperature of the blowpipe flame, the solid impurities areslowly volatilised or partly decomposed.The chief impurity isusually sodium chloride, which, so far as it is not volatilised duringignition, is decomposed, with formation of sodium silicate, and i thas been propwed that when the silica has been finally treated withhydrofluoric and sulphuric acids, any sodium sulphate remainingshould be calculated to sodium oxide, and this amount deductedfrom the weight of crude “silica.”26 I n view of the volatility ofsodium chloride, a better course is to treat the silica before ignitionwith sulphuric acid to decompose chlorides. The weight of theresidue then obtained after treatment with hydrofluoric andsulphuric acids is the exact measure of the impurities weighed withthe silica.27The determination of silver in bullion, especially bullion high ingold and low in silver, by the cupellation method, is not as satis-factory as might be desired.The silver, being estimated by thedifference between the result of the total fine metal assay and thegold assay, bears the errors of both assays, which may not becompensating, and may be large in proportion to the silver present.Thus, differences of 5 fine are common bet’ween the results of twolaboratories, and even duplicates may differ by this amount, andthat on bullion only 100 to 200 fine in silver. Gay Lussac’s methodhas the advantage that, given proper equipment, a large amountof work can be completed daily; but, for accurate work, somewhat elaborate equipment is required, and the method is quiteunsuited f o r occasional work.28 A preliminary difficulty in apply-ing any wet method to bullion of the character referred to aboveis the solution of the silver.I n view of all these facts, consider-able interest attaches to a method which has been, and is still85 bl. Wunder and A. Suleimann, Ann. Chim. mznl., 1914, 19, 45 ; A . , ii, 292.?6 F. A. Goocli, F. C. Reckert, and S. 13. Kuzirian, Zeitsch. mzorg. Chem., 1914,27 S. B. Kuzirian, ibid., 430 ; Amw. J. Xci., 1914, [iv], 37, 61 ; A . , ii, 218.28 Ann. Report, 1913, 173.85, 230168 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.being, investigated in the United States Mint with promisingresults, although its author is careful to state t h a t many detailsyet require investigation. The principle of the method is alloyingthe bullion with cadmium, as suggested a generation ago byBalling,29 solution of the silver in the alloy by means of nitricacid, and its estimation by Volhard's method.30The estimation of silver in colloidal silver preparations and inorganic tissue fluids presents some difficulty, which, however,appears t o be overcome by two methods recently de~cribed.3~A method recommended nearly twenty years ago for the separa-tion and estimation of arsenic32 has been modified, and the modifi-cation shown to be of very general application, neither bismuth,cadmium, tin, iron, chromium, nickel, cobalt, manganese, nor zincinterfering.Copper and aluminium in large amounts do interfere,but by a slight modification the method is applicable even t osolut.ions containing much copper.The method depends on theprecipitation of the arsenic in the elementary condition by additionof sodium hypophosphite t o the hydrochloric acid solution, withfiltration of the precipitate and estimation of the arsenic by aniodometric method. The recent modification consisk in treatingthe precipitate with an iodate-iodide mixture. Owing t o a traceof free iodine always present, the arsenic dissolves, slowly a t first,but more and more rapidly, since iodine is one of the products ofthe series of reactions which follow the solution of the arsenic.The net result of these reactions is that each atom of arsenic setsfree one atom of iodine, and the iodine thus liberated is titratedwith thiosulphate.33 It has been suggested that arsenic may belo& by volatilisation as chloride when following the abovemethod,34 but this fear is shown to be groundless.35 For the esti-mation of small quantities of arsenic in the presence of large quan-tities of copper, a purpose which the method just described does notfulfil, the methosd of Avery and Beans36 can now be confidentlyrecommended.It is far more rapid than any other availablemethod, and depends on the fact that the copper complex formedwhen an alkali tartrate and alkali hydrogen carbonate are added t o2D Compare Crookes' " Select Methods of Cheniical Analysis," 1886, 443.XI F. P. Dewey, J. Ind. Eng. Chem., 1914, 6, 650, 728; A., ii, 778.F. Lehmann, Arch. Pharm., 1914, 252, 9; A . , ii, 578 ; P. W. Danckwortt,ibid., 69 ; A., ii, 578.32 R.Engel and J. Bernard, Compt. rend., 1896, 122, 390 ; A . , 1896, ii, 448.L. Brandt, Chem. Zcit., 1913, 37, 1445, 1471, 1496; A , , ii, 68 ; ibid., 1914,38, 461, 474 ; A., ii, 383.34 L. W. Andrews, ibid., 295; A., ii, 291.35 L. Brandt, ibid., 295 ; A . , ii, 291.36 S. Avery and H. T. Beans, J. Amer. Chem. SOC., 1901, 23, 485; A., 1901,ii, 623ANALYTICAL CHEMISTRY. 169a cupric solution is without action on either potassium iodide oriodine.37An entirely novel method of estimating glucinum depends onthe fact t h a t basic glucinum acetate is readily sublimed a t160-170°/19 mm., whilst the basic acetates of iron and aluminiumare non-volatile in these circumstances.38 A paper embodying nonew principle, but of considerable practical utility, is one dealingwith the analysis of commercial aluminium and its light alloys. Aprimary difficulty in all such work is the separation of the largequantity of aluminium which must be taken in order t o get weigh-able quantities of the minor constituents.By a single operation,the author gets a clean separation of aluminium from everythingexcept tin and nickel, and the preliminary separation of nickel, ifpresent, and the subsequent separation of tin from aluminium,present no difficulty.39 Some years ago, Brandt proposed the useof diphenylcarbazide as internal indicator in the dichromate titra-tion of iron.40 The end-point is excellent, but the indicator-ofwhich i t is necessary t o use an appreciable amount-is itself areducing agent.The authors who made this observation succeededin establishing the method on a sound basis.41 A more recent pro-posal of Brandt to dispense with a proper correction for the re-ducing power of the indicator, and to make the method empirical,cannot be endorsed except f o r routine work on nearly uniformmaterial, but his proof that cornparatively large amounts of ter-valent arsenic do not interfere is important.42The methods for the separation of iron, aluminium, andchromium from manganese and zinc, based on the fact that thesalts of the former group of elements are strongly hydrolysed,whereas those of manganese and zinc are not, or have been sup-posed not t o be, have been studied by means of the ethyl diazo-acetate method for estimating the degree of hydrolysis.Solutionsof zinc chloride are not perceptibly hydrolysed, but manganesechloride, in millinormal solution at 2 5 O , is said to be hydrolysedto the extent of 10 per cent. It is said to follow from this thatacetate separations of manganese from iron can never be sharp, butthat the iron precipitate is always contaminated with manganese.43Whilst this may be literally true, and thus comforting t o thosewho have failed t o acquire skill in acetate separations, the fact37 G. D. Lander and J. J. Geake, Analyst, 1914, 39, 116 ; A., ii, 292.:EI A. Kling slid E. Gelin, Bull. SOC. chim., 1914, [iv], 15, 205 ; A., ii, 295.39 I<. Belnsio, Ann. Chim. Applicnta, 1914, 1, 101 ; A., ii, 388.-10 L. Brandt, Zeitsch.anal. Chem., 1906, 45, 95 ; A., 1906, ii, 309.41 0. L. Barnebey and S. R. Wilson, J. Amcr. Chenz. Soc., 1913, 35, 156 ; A.,4) G. van Pelt, Bull. SOC. chint. Belg., 1914, 28, 101 ; A , , ii, 492.1913, ii, 248. -Iz L. Rrandt, Zeitsch. nnal. Chem., 1914, 53, 1 ; A., ii, 71170 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.remains that many chemists do make such separations withsufficient sharpness for all practical purposes. That others fail, how-ever, when attempting t o follow the directions given by successfulusers of the method, is sufficient evidence that those directions areincomplete in some essential detail, and the paper under reviewshould be useful t o the chemist, who may some day be expectedto describe the essential features of this method as successfully asmany practise it to-day.Although manganese is generally estimated to-day by volumetricmethods, which leave nothing to be desired in point of accuracy,it is sometimes necessary t o separate manganese from other metals,and a recent paper on the conditions which determine the separa-tion of manganese sulphide in the dense green form, which canbe easily filtered and washed, should prove usefu1.44 A very oldmethod for the estimation of zinc in coinage bronze, which wassupposed t o have been finally discredited half a century ago, hasbeen revived a t the Royal Mint and shown t o be as accurate asother available processes, whilst the saving in time is very great.The method consists in volatilising the zinc, while protecting themetal against oxidation.The use of proof assays is necessary, asis a high temperature ( 1 3 7 5 O ) , such as was not so readily obtain-able in laboratory furnaces a t the date when the method fell intodi~use.~5A new method for the estimation of cobalt in steel depends onthe fact that when iron is precipitated by zinc oxide, as inVolhard’s method for manganese, but avoiding any unnecessaryexcess of the reagent, any cobalt (and nickel) present remains insolution, whereas chromium, vanadium, molybdenum, titanium,aluminium, copper, and silica are wholly precipitated. Cobalt isfinally separated from nickel and manganese by precipitation withnitroso-&naphthol in presence of so much hydrochloric acid thatnickel is held in soIution.46 Round the dimethylglyoxime methodfor the estimation of nickel47 a considerable literature has sprungup, and attent’ion may be called to a summary, by its author, ofthe more useful suggestions, and a reply to some criticisms whichprove t o be unfounded.4*Owing t o the increasing value of platinum-iridium alloys andto the large number of industrial purposes t o which these alloys arenow being applied, the accurate estimation of iridium is of con-siderable importance, and attention may be directed to two entirely44 A. Villiers, Conzpt.rend., 1914, 159, 67 ; A . , ii, 658.45 1’. li. Rose, J. Soc. Chem. Ind., 1914, 33, 170 ; A . , ii, 385.G. Slawik, Chem. Zeit., 1914, 38, 514 ; A . , ii, 494.47 Anm. Report, 1907, 205.48 0. Brunck, Zeitsch.angew. Chem., 1914, 27, 315 ; A . , ii, 583ANALYTICAL CHEMISTRY. 171satisfactory methods which have been reported on under theanalytical investigation scheme of the Society of Public Anal~st;s.~~One of the authors of this report is also partly responsible forsome very interesting notes on the differentiation of platinum-iridium and platinum-rhodium alloys, advantage being taken ofthe fact that, in the presence of small quantities of rhodium, thetendency of silver cupellation leads to spit or vegetate is greatlyincreased.50 A useful method for the determination of the purityof platinum ware depends on the fact that the thermoelectromotiveforce of platinum against many of its alloys has been determinedwith considerable exactness. Unlike the very exact method cle-pending on measurement of temperature-coefficient of electricalresistance, which is only applicable to wires, the new method canbe applied to vessels of any form, and without defacing them.61As one of the two rapid methods for the estimation of thoriumis not generally available, owing to the fact that the reagent-sodium hypophosphate-cannot be purchased, more importancethan would otherwise be the case attaches t o a new method whichdepends on the insolubility of thorium pyrophosphate in diluteacids, the pyrophosphates of the other rare earths, includingcerous (but not ceric) cerium, being soluble.52 It has been observedthat when the rare earths are precipitated as hydroxides andignited to oxides, the results are always somewhat higher, andsometimes much higher, than when they are precipitated as oxalatesand ignited.Moreover, the former method gives less concordantresults, suggesting the possible formation of basic salts, the amountof which might be expected to vary with the experimental condi-tions. On the other hand, there was a doubt whether the oxalateprecipitation wils complete in presence of a trace of mineral acid.This doubt is finally dispelled by a recent research, which showsoxalic acid to be a better precipitant than ammonium oxalate.The latter tends to be carried down as a complex oxalate, which,on washing, is hydrolysed, with the production of rare earth oxalateso finely divided that it passes through the filter. Ammonia doesgive rise to small quantities of basic salt, but it is only whenthe hydroxides of the fixed alkalis are used that serious errors areencountered, and these are mainly due to the retention of alkaliby the precipitate.&F o r the estimation of quantities of titanium of the order of49 C.0. Bannister and E. A. Du Vergier, Analyst, 1914, 39, 340 ; A . , ii, 748.50 C. 0. Banni3ter and Patchin, Inst. Mining and Me~nllurgy, 1913, Bull. 111.51 G. 1<. Burgess and P. D. Sale, J. Ind. Eng. Chem., 1914, 6, 452 ; A . , ii, 585.52 R. J. Carney and E. I). Ccimpbell, J. Amer. Chem. sbc., 1914, 36, 1134;6s T. 0. Smith and C. James, Chem, News, 1914,109, 219 ; A , , ii, 492.A . , ii, 583172 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.10 mg. with an accuracy of 2 per cent., a method described someyears ago in an obscure journal has been subjected t o a criticalinvestigation, and shown to be serviceable.It depends on theprecipitation of titanium phosphate in acid solution, after reducingiron t o the ferrous state.54 For the inore exact separation of largerquantities of titanium and iron, Gooch's method was until recentlymuch used. It depended on precipitation of the iron as sulphidein presence of tartaric acid, which had then to' be destroyed beforethe titanium could be precipitated by the usual reagents. Destruc-tion of the tartaric acid by pernianganate, as Gooch directs, gener-ally leads to co-precipitation of manganese when the attempt ismade t o precipitate the titania by the hydrolysis of the acetate,and a second separation is thus necessitated.This tedious pro-cedure is no longer necessary, since i t has been found thatBaudisch's reagent (the ammonium salt of nitrosophenylhydroxyl-amine) will precipitate titanium from solutions containing largequantities of tartaric acid.55 A similar method serves for theseparation of zirconium from iron and al~minium.5~Tungsten of more than 99 per cent. purity is now a comrnoiiobject of commerce, and is use'd in so large a proportion in somehigh-speed steels that the exact determination of the impurities isimportant. When the total impurities range from 5 per cent.down to 0.2 per cent., large quantities of metal must be taken forthe estimation of the minor constituents, and bringing these largequantities of tungsten into solution introduces a preliminarydifficulty.Several alternative methods of bringing the metal intosolution are described in a recent paper, which also shows the greatvariety of impurities that may be present, and how best to estimateeach .57Am'ong many papers dealing with the estimation of carbon insteel and iron, attention may be directed t o a critical investigationof the comparatively little used method which depends on directcombustion, absorption of the resulting carbon dioxide by means ofbarium hydroxide, filtration and washing of the precipitatedbarium carbonate out of contact with air, solution in a measuredvolume of standard hydrochloric acid, and titration of the excessof acid with alkali. The manipulation of the precipitate requiresspecial apparatus, but this is readily assembled, and the publishedresults show t h a t the method is as exact as the gravimetric methods.54 G.S. Jamieson, J. Ind. Eng. Chem., 1914, 6, 203 ; A., ii, 298.55 W. M. Thornton, jun. and E. M. Hayden, jun., Zeitsch. anorg. Chem., 1914,56 Ibid., Chenz. News, 1914, 110, 153 ; Amer. J. Sci., 1914, [iv], 38, 137;57 H. Arnold, Zeitsch. nnorg. Chcm., 1914, 88, 74; A., ii, 679,86, 407 ; Amer. J. Sci., 1914, [iv], 37, 407; A., ii, 553.A . , ii, 779ANALYTICAL CHEMISTRY. 173Its advantages, as compared with the use of weighed potash bulbs,are obvious, whilst i t is not subject t o error from access of sulphurtrioxide, as is the gravimetric barium carbonate method, which has,nevertheless, found extended adoption.58Electrochemical A nalysis.General interest attaches to new devices which make it possibleto conduct electrochemical analysis with a minimum expenditureon platinum app-aratus.Attention may be directed to a cheapform of rotating cathode and anode suitable for the rapid estinia-tion of copper and zinc. Not only is it so designed as t o re'ducethe necessary amount of platinum to about 1 gram, but the shapeof the anode ensures thorough mixing.59 Fine-meshed brass gauzehas also been shown to be suitable material for the construction ofcathodes for t$e estimation of copper, zinc, and nickel, about1 gram of platinum wire serving as an0de.m Tantalum electrodes,recommendeld two or three years ago,61 have been stated t o becomebrittle in use62 and t o give unsatisfactory results.63 It is nowshown that they do not become brittle unless relatively highcurrent densities are used, and that the results with tantalumgauze electrodes, now obtainable, are in no way inferior to thoseobtained with platinum gauze.64It was stated some years ago that copper could be separatedfrom an equal weight of arsenic either from ammoniacal or nitricacid solution.65 It is now shown that this is only true when thearsenic is present in the arsenic condition,66 and that it is prefer-able to work with ammoniacal solutions, as much higher currentdensities may then be used, and no sensible error is introduced byprolonging the operation long after all the copper has been de-posited, in marked contrast to the behaviour of nitric acid solu-t i o n ~ .~ ~ The conditions requisite for the deposition of bismuth ina satisfactorily adherent condition, and for the separation of thatmetal from arsenic, cadmium, and lead, have now been defined withsome precision. The author of this work, in which measurements58 J. R. Cain, J. Ind. Eng. Chem., 1914, 6 , 465 ; A., ii, 577.59 E. A. Lewis, J. SOC. Chcm. Ind., 1914, 33, 445 ; A . , ii, 483.D. F. Calhane and T. C. Wheaton, kfet. and Chcm. Ewg., 1914,12, 87.0. Brunck, Chem. Zeit., 1912, 36, 1233 ; A., 1912, ii, 1128.65 G. Oesterheld, Zeitsch. EZektrochem., 1913, 19, 585 ; A., 1913, ii, 823.6* G. Wegelin, Chem. Zeit., 1913, 37, 989 ; A . , 1913, ii, 880.65 D. S . Ashbrook, J.Amcr. Chem. SOC., 1904, 26, 1285 ; A . , 1905, ii, 64.66 B. P. Richardson, Zeitsch. anorg. Chem., 1913, 84, 277 ; A., ii, 148 ;0. Blunck, ibid., 1914, 38, 565; A . , ii, 482.A. Sieverts and W. Wippelmann, ibid., 1914, 87, 169 ; A . , ii, 580.Sieverts and Wippelmann, loc. cit174 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.of cathode potential proved invaluable, suggesk that in many cases-once a separation has been shown to be possible and has beenstudied with the aid of cathode potential measurements-it shouldbe possible t o dispense with the elaborate and costly appliancesrequired for measuring cathode potential, and where possible thisis certainly desirable.G8The electrolytic reduction of ferric iron as a preIiminary t otitration with permanganate has been made the subject of study,and it is shown that quantitative reduction can be secured in afew minutes by simple means.Chlorides do not interfere, and awide range of acidity is permissible.69Attempts have been made to estimate small quantities ofalkaloids by an electrochemical method of the type made familiarby.the work of Dutoit and Duboux. A solution of the alkaloidcontaining an excess of hydrochloric acid is titrateld with sodiumhydroxide solution of known strength, the electrical conductivityof the solution being measured after each small addition. Theresults when plotted on a system of rectangular co-ordinates givea curve consisting of three well-defined portions, the first repre-senting the neutralisation of the free acid, the second the dis-placement curve characteristic of the alkaloid, and the third theincrease of conductivity due to excess of sodium hydroxide.Bycomparison with curves obtained with solutions of known composi-tion, small quantities of alkaloids may be estimated.70Organic Analysis.Experience shows that few qualitative tests sustain the claimsmade for them. This is particularly the case with tests whichare a t first claimed to be specific. Tests for which less is claimed,on the other hand, frequently prove useful. Such a one is thatwhich depends on a colour reaction with trichloroacetic acid. Sofar as the evidence goes, it appears that trichloroacetic acid con-stitutes, equally with tetranitromethane, a reagent for cyclic doublelinkings.71 A test which differentiates between morphine on theone hand and codeine and dionine on the other, and which servesto detect 0.02 mg.of morphine, is deserving of notice, more especi-ally as it has been shown that a dozen or more other commonalkaloids do not give a similar reaction.72 A reaction apparentlycharacteristic of malonic acid and its esters, and certainly notgiven by acetoacetic and dicarboxyglutaconic esters, depends on68 B. P. Richardson, Zoc. cit.6y H. C. Allen, J. Amer. Chem. Soc., 1914, 36, 937 ; A , , ii, 581.7.1 K. Goubau, Bull. Acad. roy. Belg., 1914, 63 ; A . , ii, 394.‘il K. V. Charitschkov, J. Russ. Phys. Chem. Soc., 1914, 46, 76 ; A , , ii, S90.72 T. H. Oliver, Chm. and Drug., 1914, 85, 249ANALYTICAL CHEMISTRY.175the development of an intense blue fluorescence when the substance,in alcoholic solution, is heated with hydrochloric acid, and themixture then neutralised, mixed with an alcoholic solution ofw-bromomethylfurfuraldehyde, and finally rendered alkaline withalcoholic pota;sium hydroxide.73For the combustion analysis of organic compounds by Dennstedt’smethod, a furnace electrically heated internally by means ofplatinum wires has been designed.74 A volumetric method hasbeen described for the determination of total carbon in aliphaticsubstances in the wet way. It is based on oxidation of the sub-stance by potassium dichromate and phosphoric acid to carbondioxide, or a mixture of carbon dioxide and acetic acid, the carbondioxide being measured, and the acetic acid distilled from theresidual liquid and titrated with baryta.An advantage of themethold is the small quantity of the substance required.75 Thedirect method for the estimation of oxygen in organic compounds,referred to last year,76 has been found to give erroneous resultswhen applied to substances of comparatively low oxygen content,but a new method of universal application has now been workedout by the same author. Unfortunately, it is somewhat elaborate,but the value of a direct and exact determination of oxygen issometimes very great.77A novel method for the estimation of halogens in organic com-pounds consists in dissolving the substance in alcohol, adding asmall quantity of a dilute solution of an alkali hydroxide, and1-2 grams of calcium carbonate coated with finely divided palla-dium.The air is displaced by hydrogen, and the reduction com-pleted in a shaking apparatus in an atmosphere of hydrogenunder pressure. The halogen is determined by the usual methodsin the filtrate from the catalyst, which may be used several times.78Vanadium pentoxide, recommended last year as a catalystin the absorption of gaseous olefines by sulphuric acid,79 is nowrecommended in a similar capacity in Kjeldahl’s method of decom-posing organic compounds as a preliminary t o the estimation ofnitrogen.80 Attention has been called to some limitations ofKjeldahl’s method by two physiologists whose work dealt with a73 H. J. H. Penton, Proc. Camb. Phil. Soc., 1914, 17, 477 ; A., ii., 686.n C.Milchsack and MT. A. Roth, Zeitsch. nnyew. C‘hem. , -1914, 27, 5 ; A., ii, 147.75 E. C. Grey, T., 1914, 105, 2204.77 M. C. Boswell, J. Amer. Chem. Soc., 1914, 36, 127 ; A . , ii, 142.7* M. Busch, Zeitsch. angew. Chern., 1914, 27, 432.7g Ann. Report, 1913, 168.8o L. Marino and F. Gonnelli, Atti R. Accad. Lincei, 1914, [v], 23, i, 523 ;Ann. Report, 1913, 178.A . , ii, 575176 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.number of compounds not covered by Dyer in his report on thatprocess twenty years ago.81The addition of acetic anhydride or of phenol in the estimationof methoxyl by Zeisel’s method, as recommended by Herzig andWeishut respectively,82 has been stated to introduce a source oferror, for in the absence of anything but the acetic anhydride orphenol and hydriodic acid, silver iodide may be formed.83 It hassince been shown, however, that the errors thus introduced,although real, are negligible compared with the unavoidable ex-perimental error of the method.84 The estimation of methoxyl byZeisel’s method in the presence of sulphur compounds is extremelyunsatisfactory, on account of tlie formation of silver sulphide inthe absorption flasks.A method has now been devised for theabsorption and estimation of the methyl iodide, based on the factthat this substance combines with pyridine, forming pyridinemethiodide, which latter may be titrated with silver solution,using sodium chromate as indicator.B5A quantitative separation of acetaldehyde and acetone has beenbased on the fact that t’he former is quantitatively oxidised by analkaline silver solution, whilst acetone is scarcely affected.86 Anew method for the estimation of acetone in acetone-water mix-tures depends on the fact that two phases separate when sufficientpotassium fluoride is added to such mixtures.87 The method isessentially identical with one previously described for the estima-tion of alcohol in admixture with water.88 A gasometric methodfor the estimation of formic acid depends on the fact that, inthe presence of a trace of sulphuric acid, formic acid reach quan-titatively with acetic anhydride, with the production of aceticacid and carbon monoxide.89Mannitol has been estimated hitherto either by the method ofGayon and Dubourg90 or by that of Muller.91 The former consistsessentially in concentration of the solution until mannitol crystal-lises, treatment of the magma with a saturated solution ofmannitol, followed by filtration, and final extraction of theH.I). Dakin and H. W. Dudley, J. Biol. Chsm., 1914, 17, 275 ; A , , ii, 381.R. J. Manning and M. Nierenstein, Ber., 1913, 46, 3983 ; A., ii, 150.G. Goldschmiedt, ibid., 1914, 47, 389; A., ii, 223.82 Ann. Report, 1913, 179.85 A. Kirpal and T. Biihn, ibid., 1084 ; A . , ii, 497.86 E. Hagglund, Zeitsch. anal. Chem., 1914, 53, 433 ; A., ii, 592.G. B. Frankforter and L. Cohen, J. Amer. Chem. Soc., 1914, 36, 1103; A.,ii, 548.8(1 Ann. Report, 1912, 213.89 V. Hottenroth, Chem. Zeit., 1914, 38, 598 ; A., ii, 501.91 Bzcll.SOC. chim., 1894, [iiiJ, 11, 329, 1073 ; A., 1894, ii, 1 4 1 , 334.Ann. Inst. Pasteur, 1894, 8, 108ANALYTICAL CHEMISTRY. 1’77inannit01 from the dry contents of the filter by means of hotalcohol. The limitations of such a method are obvious. Muller’smethod, depending on the fact that the rotatory power of anaqueous solution of mannitol is notably raised by saturating thesolution with borax, is troublesome, since sugars must first befermented, tartrates and malates removed, and a correction madefor gly,cerol, which, if present, combines with some of the boraxand minimises the effect of a fixed quantity of the latter, but itgives excellent results with wines. I n the presence of some othersubstances, such as yeast-water, often employed as nitrogenous foodfor yeast or bacteria in fermentation experiments, the method failsutterly, and this lends additional importance to a new methodwhich is free from this objection.It is modelled on Wagenaar’smethod for the estimation of glycerol, and depends on the capacitypossessed by polyhydric alcohols of holding cupric hydroxide insolution. The method is invalidated only by the presence of otherhexitols.92 The sugars most commonly found in glucosides, apartfrom dextrose, are &galactose, dmannose, and members of themethyl pentose group, generally rhamnose, often also rhodeose.Simple pentoses are rarely met with. Dextrose may be character-ised by fermentation, galactose is precipitated almost quanti-tatively by phenylmethylhydrazine, mannose by phenylhydrazine,arabinose by diphenylhydrazine.A recent method for the estima-tion of rhamnose depends on the formation of a cyanoliydrin withhydrocyanic acid, which is easily converted by saponification intoa-rhamnohexonic acid (or its lactone), which in turn yields mucicacid on oxidation by nitric acid. The principle of the methodmight have suggested ikelf to anyone, but no confidence can beplaced in a method depending on oxidation t o muck acid until ithas been shown to give constant results. The oxidation ofa-rhamnohexonic acid to mucic acid does not appear to be quantita-tive, but there is a fixed relation between the rhamnose originallypresent and the weight of mucic acid formed, and this relation isbut slightly disturbed by the simultaneous presence of an equalquantity of rhodeose, and in many cases the disturbing influenceof rhodeose is negligible, for example, when the purpose in viewis the determination of the molecular proportions of a mixture ofmethyl pentoses, for which an exact method is unnecessary.93 Ithas been found that when dilute hydrochloric acid acts on hexoses,starch, and cellulose, o-hydroxy-5-methyl-2-furf uraldehyde isformed to the extent of 1-2 per cent.Although it is precipitatedby phloroglucinol, i t does not interfere with the accuracy of92 J. Smit, Zeitsch. ccnal. Chem., 1914, 53, 473 ; A,, ii, 683.93 E. VotoEek and R. PotrnBi5i1, BUZZ. Soc. chim., 1914, [iv], 15, 634 ; A., ii,REP.-VOL. XI. N683178 ANNUAL REPORTS ON THE PROGRESS OF CHEhfISTRY.pentosan estimations made by the phloroglucinol method, providedaniline acetate is used as indicator.This is accounted f o r on theground of the slowness with which the o-hydroxymethylfurfur-aldehyde is produced. The recognition of this substance, however,renders previous estimations of methyl pentosans of doubtfulvalue.94ThO formation of a sparingly soluble compound with xanth-hydro1 has been used f o r the detection of carbamide in extremelydilute solutions, and its low solubility, together with its high mole-cular weight-seven times that of carbamide-aff ords an excel-lent means for the gravimeti-ic estimation of carbamide by directprecipitation.95 The gravimetric method possesses the advantageover methods depending on decomposition in the fact that carb-amide may a t the same time be definitely characterised, and thepurity of the derivative confirmed by analysis.When a solutionof semicarbazide hydrochloride is treated with potassium chlorateand hydrochloric acid, i t is decomposed, with the liberation ofexactly two-thirds of its nitrogen in the elementary condition, theremaining third being fixed as ammonium chloride. As carbamide,under similar conditions, is unacted on, the reaction affords a meansof estimating semicarbazide in the presence of ~arbamide.9~I n applying the double polarisation method t o beet molasses, thedirect reading is taken in an alkaline solution (owing to the excessof basic lead acetate remaining after clarification), and the in-version reading in an acid medium; the difference between the twoobservations is not duel solely to sucrose, since the rotation of theoptically active impurities is modified by the change of reaction toan appreciable In order t o avoid this error, it has beenproposed t o make the direct polarisation in presence of the sameamount of citric acid as in the inversion observation, clarificationbeing effected by means of bromine, the excess of which is shownto be without action on the sucrose or invert sugar.98 The resultsare about 1 per cent. higher than those obtained by the ordinaryprocedure involving the direct reading in alkaline solution, but theyare substantially identical with those obtained by the improvedmethods referred t o in last year’s Report.98aThere are still no British data bearing on the cryo,wopic methodof detecting added water in milk,gQ but the method has been muchg4 Miss M.Cunningham and C. Dorde, Biochem. J., 1914, 8, 438 ; A . , ii, 788.95 R. Fosse, Compt. rend., 1914, 158, 1076 ; A . , ii, 506.g6 R. L. Datta, J. Amer. Chem. Xoc., 1914, 36, 1014 ; A . , ii, 504.97 Ann. Report, 1913, 182.98 V. StanGk, Zeitsch. Zuckerind. Bohnt., 1914, 38, 429 ; A., ii, 586.98a Ann. Report, 1913, 182.By Ibid., 1910, 177 ; 1911, 173 ; 1912, 214; 1913, 183ANALYTICAL CHEMISTRY. 179discussed in Holland,l and a single Anglo-Indian writer speakshighly of i t . 2The methods of Lendrich and Nottbohm and of Katz for theestimation of caffeine in coffee are trustworthy, but tedious, and arecent method which gives identical results, and is somewhatquicker, is therefore deserving of notice, although embodying nonew principle.3 A new method for the estimation of nicotine in thepresence of ammonia depends on the fact that in alcoholic solutionnicotine behaves towards picric acid as a monacid base, whereas inaqueous solution it forms a dipicrate.4 Until quite recently nomethod was known for the estimation of strychnine in the presenceof quinine with any approach to exactness.A method, whichleaves little to be desired in point of accuracy, has now beendescribed, depending on separation of strychnine' as f err~cyanide.~Polenske's method of detecting beef fat in lard by the differencebetween the melting and the solidification points of the fatsC hasbeen shown to depend mainly on the fact that the a-palmitodistearinof lard has a difference-value greater than 18, whilst the P-palmito-distearin of beef fat has a difference-value of less than 12.Themethod will detect 15-20 per cent. of beef fat in lard.' A recentmethod, said to be capable of detecting 5-10 per cent., depends onthe differences in the melting points of the characteristic glyceridesof lard and of beef fat as compared with the melting points of therespective fatty acids. The difference between the melting pointsof a-palmitodistearin and its separated fatty acids is more than5O, whilst in the case of P-palmitodistearin the difference is only0.lo.* This method is only a year old, and is not universallyapproved,g but the most recent report on i t by authors independentof its inventor, states that 5 per cent.of be'ef fat can usually bedetected, and justifies one in the hope that a real step has a t lastbeen taken towards the solution of a very difficult problem.1°It has long been known that when to any of the higher fattyacids any other of the higher fatty acids is added in amouiit up t o20 per cent., the depression of the melting point is proportional toCJLCWL. IVeckblnd, 1914, 11, 126, 198, 201, 201, 206, 207, 209, 323; A., ii, 169,J . W. Leather, Annlyst., 1914, 39, 432.R. Spallino, Ouzzeltn, 1913, 43, ii, 493 ; A., 1913, ii, 1086.E. Polenslie, Aybeit. a. d. Kniserl. Cesundheitsmrnte, 1907, 26, 444.A. Bomer and R.Limprich, Zeitsch. Xnhr. Genitssm., 1913, 25, 367 ; A., 1913,A. Bomer, ibid., 26, 559.K. Fischer and J. Wewerinke, ibid., 1914, 27, 361.304, 392.a G. Pendler and W. Stiiber, Zeitsch. A ' ~ L T . Geizusbin., 1914, 28, 9 ; A., ii, 757.ti C. Simmonds, Analyst, 1914, 39, 81 ; A . , ii, 307.ii, 444.lo H. Sprinkmnyer aiid A. Diedrichs, ibicZ., 571.N 180 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.the amount of the added acid, and almost independent of its kind.Advantage is taken of this fact in a methobd recently proposed fordetermining the composition of mixtures of the higher fatty acids.For the estimation of palmitic acid, for example, 20 parts of themixture to be analysed are added t o 80 parts of pure palmiticacid, and the melting point is determined.By reference to a table,the percentage of palmitic acid and of acids other than palmiticacid in the final mixture is given, and thus the percentage ofpalmitic acid, if any, in the mixture t o be analysed. Stearic,behenic, and other acids are estimated similarly. Extensions ofthe method consist in separating the solid fatty acids from a mix-ture and examining them as described, and also in hydrogenatingthe original fatty acids by the method of Sabatier and Senderens,and examining the hardened product by the melting-point method,It is possible by methods such as these to show, for example, thatthe fatty acids of cottonseed oil consist as t o 70 per cent. of un-saturated acids with 18 carbon atoms, and as t o 25 per cent.ofpalmitic acid, whilst stearic acid and unsaturated acids with 16carbon atoms can be shown to be ahsent.llA recent method for the stirnation of rosin in varnishes, oils,and soaps embodies no new principle, but the tabulated resultsshow that i t gives results a t least as accurate as any previousmethod, whilst it is simpler and much more rapid.12The proposal to estimate water in moist alcohol by observing theclouding point of mixtures with other liquids is not new, but noneof the methods hitherto based on this principle has proved satis-factory in practice. A recent method of this kind, which makesuse of a-bromonaphthalene, a substance readily obtained in asufficient state of purity for this purpose, makes it easy to det’er-mine 1-10 per cent.of water in alcohol within 0.02 per cent.l3The method employed for the determination of the originalgravity of beer is a purely empirical one, based on the observedrelation, in a great number of instances, of the “spirit indication ”to the “degrees of gravity lost.’’ It has been known for manyyears that the table contained in the First Schedule t o the InlandRevenue Act, 1880, was seriously inaccurate over an important, partof its range, and, in 1909-1911, Sir Edward Thorpe, on behalf ofthe Treasury, and Dr. H. T. Brown, representing the industry con-cerned, carried out an investigation with a view to collecting dataon which to construct a new table of Original Gravities. Thistable has just been given Parliamentary sanction,’* and full detailsl2 H.Wolff and E. Scholze, Chem. Zeit., 1914, 38, 369 ; A., ii, 393.1s Miss M. Jones and A. Lapworth, T., 1914, 105, 1804.l4 Finance Act, 1914 (Session 2).E. Twitchell, J. Ind. Eng. Chem., 1914, 6, 564; A., ii, 655ANALYTICAL CHEMISTRY. 181of Thorpe and Brown’s work have now been published,l5 whilst a tthe same time Brown has published a history of previous tables oforiginal gravity, a comparison of these with the new table, and areport on the scientific principles underlying the empirical methodof determining original gravity.16Although impossible to summarise in a paragraph, reference mustbe made to a series of papers on the analysis of mixtures of naturalwith artificial asphaltum, papers which materially add to ourknowledge of the chemistry of asphaltum, and take account of allrecent German work on the subject.17 Their value is t o someextent diminished by the fact that the author ignores the work ofClifford Richardson, and indeed of all American writers on thissubject.Agricultimxl Chemistry.Since Veitch’s method for the estimation of the lime requirenleiitof soils18 requires a t least six independent experiments, and givesmuch trouble with certain types of soil, an older and simplermethod,l9 long since abandoned as misleading, has been modified,and in its new form recommended as giving results substantiallyidentical with those obtained by Veitch’s method.20 Unfortunately,there is already evidence that this is not the case,21 and until theauthors meet the criticism which has been directed against theirmethod, that of Veitch, with all its shortcomings, is likely to bepreferred. Admittedly unsatisfactory, Veitch’s method has prob-ably been of more service than any other of the many methodsdescribed hitherto.The writer will hazard the opinion, however,that it and other methods are destined t o give way to a methodquite recently described. This method depends on treatment of thesoil with N/50-calcium hydrogen carbonate, and titration of analiquot portion of the filtrate. The principle seems unassailable,the technique is of the simplest, and the results are said to be con-firmed by pot and field experiments.21aMeasuring the capillary lift of soils is not analytical chemistry,but it seems right to call attention here to a recent paper whichshows how such measurements can be satisfactorily carried out withapparatus available in most analytical laboratories.22l5 Sir T.E. Thorpe iind H. T. Brown, J. Znst. Brewzkng, 1914, 20, 569.l6 11. T’. Browu, ibid., 645.l7 J. Marcusson, Chem Zcit., 1908, 32, 965 ; Client. Rev. Fett. Harz-Ind., 1911,18, 47 ; Zeilsch. angew. Chem., 1913, 26, 91 ; Chern. Zeit., 1914, 38, 813, 822.F. P. Veitch, J. Anaey. Chem. SOC., 1902, 24, 1120 ; A., 1903, ii, 400.11) Albert, Zeitsch. nngew. Chtm., 1888, 1, 533.1u J. A. Bizzell and T. L. Lyon, J. Ind. Eng. Chem., 1913, 5, 1011 ; A., ii, 150.21 C. R. Moulton and P. F. Trowbridge, ibid., 1914, 6, 835 ; A., ii, 828.‘Lia H. B. Hutchinson and K. MacLeniian, Chem. News, 1914, 110, 61 ; A., ii, 784.C.J. Lynde and H. A. DuprB, J. Amer. Soc. Ayronom., 1913, 5, 107182 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Almost the only papers dealing with the analysis of plantmaterial are dated from Rothamsted. Continuing their re-~earches,~3 the workers of this school have determined the cupricreducing power of xylose and arabinose,04 and have contributed avaluable paper on the estimation of starch. For the estimation ofstarch in plant materials, the modified Sachsse method, which isofficial in the United States,25 and is based on the hydrolysis ofstarch by means of boiling dilute hydrochloric acid, is quite value-less, not only because such tissues invariably contain pentosansand other substances that yield reducing sugars on hydrolysis, butbecause of the actual destruction of dextrose which occurs duringthe prolonged treatment with acid.Although ordinary diastasegives with purified starch results by O’Sullivan’s methold 26 whichare approximately correct, values 15 to 20 per cent. lower than theactual starch content may be obtained when it is applied to leafmaterial or plant tissues in general, owing to the loss of dextrinwhich occurs during the purification of the solution by means ofbasic lead acetate. The observation that basic lead acetate, whichdoes not of itself precipitate dextrin, does carry it down if certainother substances are present, is important to others besides agri-cultural chemists. For the estimation of starch in plant material,an entirely satisfactory method has been based on Hill’s observa-tion 2.7 that taka-diastase converts starch wholly into maltose anddextrose. The paper includes valuable evidence of the rela-tively enormous errors that may be introduced in work of thiskind unless the greatest care be taken to ensure that the sampletaken for analysis is a representative one.I n the estimation ofstarch in dried, ground leaf material, tipping a portion of thesample out of the bottle for analysis, instead of turning the wholesample out, may result in the starch being underestimated by 20per cent., the heavy starch granules tending to sink to the bottomof the bottle.28The preparation of neutral solutions of ammonium citrate andtheir use in the analysis of phosphatic manures, which were referredt o somewhat fully two years ag0,29 have received much attentionduring the year under review.As regards the preparation of thesolution, i t has been shown that a deviation of 0.5 per cent. fromthe ammonia: citric acid ratio given by Patten and Marti30 isnegligible,31 whilst a simple method of controlling the hydrion-con-29 Ann. Report, 1913, 185.as C. O’Sullivan, Y‘., 1884, 45, 1.29 Ann. Beport, 19’12, 218.A. J. Daish, J. Agric. Sci., 1914, 6, 255.27 A. C. Hill, P., 1901, 17, 184.U.S. Bureau of Chemistry, Bull. No. 107.W. A. Davis and A. J. Daish, J. Agric. Sci., 1914, 6, 152 ; A . , ii, 588.P. Rudnick and W. L. Latshaw, J. Ind. Eng. Chem., 1913, 5, 998 ; A,, ii, 145.Ibid., 1913, 185ANALYTICAL CHEMISTRY. 183centration has been based on colour comparison with a standardsolution of hydrochloric acid and disodium hydrogen phosphate,both solutions being tinted with a suitable indicator.The authorsof the latter method also make the suggestion that the results ofthe test would be far less liable to vary if it were agreed to adoptas the standard solution one that was distinctly acid or alkaline,as small differences in the exact composition would then have lessinfluence.32 Another proposal is the substitution of sodium citratsfor ammonium citrate.33 The preparation of the solution thenpresents no difficulty, but the results do not always agree with thoseobtained by the use of ammonium ~ i t r a t e . 3 ~ Most of this work isAmerican, but continental workers have also contributed to theattempt t o place this test on a more satisfactory basis.35Three years ago some space was given to a consideration ofmethods of analysis of the new nitrogenous fertilisers, and anattempt made to reconcile the conflicting statements made concern-ing the estimation of cyanamide.36 A recent paper goes some wayin this direction, pointing out that Monnier’s results might be ex-plained by polymerisation of a portion of the cyanamide t o dicyano-diamide under his experimental conditions.Simple means aredescribed for combating this tendency to polymerisation, whethermaking use of Caro’s or of Kappen’s 36 method of analysis, and itis shown that, with this precaution, the two methods give identicalresults.37 The accurate estimation of total nitrogen in mixtures ofcalcium cyanamide and Norwegian nitre was found to present somedifficulty, methods which might’ have been expected t o be satisfac-tory all giving results below the truth.A simple and satisfactorymethod has now been described.38Water Analysis.The last two years have witnessed the appearance of an extra-ordinary number of papers dealing with the estimation of hard-ness in water, several of them inspired by the comparatively recentmethod of Blacher, others concerning themselves with the methodsdevised by Wartha and popularised by Pfeifer about twelve yearsago, whilst others either ignore the important work of these authorsor attempt t o discredit their methods by publishing records of32 D.Eastman aiid J. H. Hildebrand, J. Ind. Eng. Chem., 1914,6, 577; A . , ii, 675.33 A. W. Bosworth, ibid., 227 ; A . , ii, 289.1’. Rudnick, W. B. Derby, and W. L. Latshaw, ibid., 486 ; A . , ii, 576.35 T. Warynski and J. Langel, Ann. Chim. nnnl., 1914, 19, 1 ; A , , ii, 216.36 Ann. h’eport, 1911, 177.:%7 G. Grube and J. Kruger, Zeitsch. nnyew. Cham., 1911, 27, 326 ; A . , ii, 593.Y8 A. Stutzer, Chem. Zeit., 1914, 38, 597; A . , ii, 485184 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.experiments which will not stand examination. Wartha's methodswere originally described in a Hungarian journal, of which thepresent writer has never been able to discover the name; but theywere published in German by Pfeifer,39 and shortly after in Englishby Proctor,40 and have given excellent service in many hands duringthe last ten years.The method for permanent hardiiess dependedon the insolubility of calcium carbonate in the presence of excessof sodium carbonate, and that of magnesium hydroxide in thepresence of excess of sodium hydroxide. A volumetric method forestimating magnesia-necessary for calculating the chemicals re-quired for softening purposes-was also baseld on the insolubilityof magnesium hydroxide in presence of excess of calcium hydroxide.A dozen authors might be cited in support of these methods, but asingle reference to a recent paper will suffice, more especially asthat paper abounds in references and shows that the one authorwho seriously challenged the accuracy of the methold for estimatingpermanent hardness did not use the Wartha-Pfeifer method a t all,but a travesty of it, due to Lunge.41 The method of Blacher andhis co-workers for the estimation of temporary and permanent hard-ness dates back to 1907,42 but was not established on a firm basisuntil last year.I n its latest form i t depends on titration of thebicarbonate hardness with standard acid, using dimethylaminoazo-benzene as indicator, followed by titration of the total hardnesswith potassium palmitate, using phenolphthalein as indicator. A tfirst potassium stearate was used, but the preparation of the solu-tion was difficult; when made, its titre varied with temperature,owing to separation of solid soap; the end-point left much to bedesired, and, in the presence of complex saline mixtures, anomalousresults were obtained.These were found to be partly due to thepresence of palmitic acid as an impurity in the stearic acid. Purepotassium stearate gave more constant results, but the history ofthe method as a practical one dates from the adoption of purepotassium palmitate as reagent.43 Since then the method has beenreported on f avourably by several independent authors,44 includingJ. Pfeifer. Zeitsch. ccngcu*. Chem., 1902, 15, 193.4o H. R. Proctor, J. SOC. Chem. lnd., 1904, 23, 8.J. Zink and F. Hollandt, Zeitsch. angew. Chem., 1914, 27, 235 ; A . , ii, 490.-12 C. Blaclier, Rignsche Indwtrie Zeit., 1907, 305 ; C. Blacher and J. Jacoby,Chm. Zeit., 1908, 32, 744; A . , 1908, ii, 897 ; C.Blacher, U. Korber, and J.Jacoby, Zeitsch. angezu. Chmn., 1909, 22, 967.C. Blacher, P. Griinberg, and M. Kissa, Chem. Zeit., 1913, 37, 56 ; A . , 1913,ii, 153.41 J. Zink and F. Hollandt, Zeitsch. angew. Chem., 1914, 27, 437 ; A . , ii, 670 ;E. Nockmann, Pharm. Zenlr-h., 1914, 55, 435 ; A., ii, 490 ; W. Pflanz, N d t . Kgl.Lundesanst. f. Wmserhyg. zu. Berlin-Dahlem, 1913, 17, 141 ; A., 1913, ii,1073ANALYTICAL CHEMISTRY. 185L. W. Winkler, who prefers it t o his own method, and makes theuseful suggestion that the use of dimethylaminoazobenzene, whichis sensitive to carbon dioxide, is best dispensed with, and that thetemporary hardness should be estimated by Hehner’s originalmethod, using methyl-orange, which is destroyed by means ofbromine before proceeding to the titration with potassium palmitateand phenolphthalein.The same author describes an improvementof his method for the estimation of hardness due to lime, depend-ing on titration with potassium oleate, as in Clark’s process, usingan alkaline tartrate solution to prevent co-precipitation of mag-nesia.45 The latter method may prove useful to those who, likeone recent author,46 find difficulty with Wartha’s method for mag-nesia. The writer has instructed many persons in the lattermethod, all of them acquiring skill in it with ease, and he prefersto estimate magnesia--usually present in least amount-directly,and lime by difference, but the paper referred t o cannot be ignored.The use of soap solutions, used in Clark’s manner, has tended tofall into abeyance ever since Hehner published his methods for theestimation of hardness, and still more rapidly since Pfeifer pub-lished Wartha’s method for estimating permanent hardness, whichwas a great advance on that of Hehner.However, the recent dis-covery that solutions of potassium myristate, unlike solutions ofsodium oleate o r of Castile soap, behave similarly towards equiva-lent solutions of calcium and magnesium, may leald to a revival ofClark‘sFor the estimation of dissolved oxygen in water, Romijn’s modifi-cation of Winkler’s method is in every respect more convenientthan the original method, but there are cases where its use maygive rise t o erroneous results, for example, with sea-water.48 I nthese cases, the use of Rochelle salt t o inhibit the formation of aprecipitate must be abandoned.This involves the abandonment ofRomijn’s pipette with its manifest advantages. Attention may,therefore, be directed to a new form of pipette for use accordingto Winkler’s original method, but which, like Romijn’s pipette,makes it possible to take samples from wells a t definite depths orfrom springs difficult of access, excludes any possibility of errordue to atmospheric oxygen, and is convenient to handle.49Whilst Winkler’s method leaves little t o be desired on the scoreof accuracy, and has been made simpler in execution by the designof apparatus such as that just referred to, circumstances frequently45 L. W. Winkler, Zeitsch. anal. Chem, 1914, 53, 409 ; A ., ii, 578.46 C. Bahlmann, J. I7d. E?ty. Ch.m., 1914, 6, 209 ; A . , ii, 294,47 Miss H. Masters and H. L. Smith, T., 1913, 103, 992.‘8 W. P. Jorissen, Zeitsch,. anal. Chem., 1910, 49, 424 ; d., 1910, ii, 749.J. J. van Eck, ibid., 1913, 52, 753186 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.arise where the convenience of a simpler test, that could be com-pleted in the field in a few minutes, would outweigh some smallloss of accuracy. The number of simple methods that have beendescribed, some of them too rough t o merit consideration, areevidence of a widespread desire for a more satisfactory field method,and attention may therefore be directed to a recent adaptation tofield conditions of a method originally due t o Linossier, who usedphenosafranine as an indicator for ferrous iron in the presence OFalkaline tartrate.The field modification of the method is empirical,but in competent hands yields very satisfactory results.50 Therecommendation of the Royal Commission on Sewage Disposal,1898, that a general statutory standard f o r effluents can properlybe fixed, itself a highly controversial proposition, could find nonotice in a report on analytical chemistry if it stood alone, butwhen the Commissioners in their Eighth Report proceed to formu-late a test--the amount of dissolved oxygen taken up from tap-water during five days' incubation-which is unnecessarily trouble-some, likely to yield discordant results, and renders useless theanalytical data hitherto accumulated, attention must be directedto the test and t o the controversy it has aroused.51E. B. Phelps suggested the use of an acid solution of o-tolidinef o r the detection of such small quantities of free chlorine as maybe present in drinking water which has been treated with hypo-chlorite. It was recently stated that the method was not adaptedt o quantitative work,52 but' a slight modification of the originalmethod is now sho'wn to be incomparably the best available for theestimation of quantities of the order of 1 part per million or less.53Less than onehundredth of this amount may be detected by itsmeans, and for some practical purposes its sensitiveness is toogreat, the comparative bluntness of the starch-iodide reactionhaving proved to be an actual advantage in devising field methodsfor use by thQse engaged in the testing and sterilisation of drink-ing water for our troops on the Continent.A recent paper on the estimation of small quantities of lithiumin the presence of large quantities of the salts of other metalsproperly falls into this section. The method substitutes isobutylalcohol for amyl alcohol as a means of separating lithium chloridefrom the chlorides of sodium and potassium. The latter salts aremuch less soluble in isobutyl alcohol than in amyl alcohol, whichis, moreover, objectionable t o use. Attention is directed t o the50 J. Miller, J. 8oc. Chew. Ind., 1914, 33, 185; A., ii, 380.51 H. T. Calvert, ibid., 1913, 32, 265.52 Dittoe and Van Buskirk, Ohio State Board of Health, 1913, Bn11, 3.53 J. W. Ellmsand S. J. Hauser, J. h i d . Eng. Chcm., 1913, 5, 914; 1914, 6,553: A., ii, 66, 669ANALYTICAL CHEMISTRY. 187fact that, in the course of a water analysis, lithium will usuallybe underestimated if calcium salts are eliminated in the usualmanner, and special methods are described for avoiding such lossof lithium.64A t intervals during the last twenty years efforts have been madeto agree on a uniform manner of reporting results of wateranalysis, efforts in which American chemists have been alwaysprominent. Hitherto little has been accomplished, but it is to behoped that the latest effort of a joint committee of the AmericanChemical Society, the American Public Health Association, andthe Association of Official Agricultural Chemists will be moresuccessful, and lead to similar agreement being reached here. Arecent paper discusses the confusion that exists, and advocatesreporting the constituents in ionic form.55 Some at least of thearguments raised against this proposal when i t was first made haveceased to have weight.G. CECIL JONES.54 L. W. Winkler, Zeitsch. anal. Chem., 1913, 52, 628; A . , 1913, ii, 877.55 R. B. Dole, J. Ind. Eng. Chem., 1914, 6, 710; A . , ii, 77