7 18 ABSTRACTS OF CHEMICAL PAPERS.A n a1 y t i c a 1 Chemistry.A New Condensation Hygrometer. By A. CROVA (Compt. rend.,94, 1514--1516).-The hygrometers now in use are subject to severalcauses of errors more or less considerable. To obviate these, the authorhas contrived an instrument consisting of a, tube of highly polishednickel, closed at each end by discs of glass, one clear, the other groundANALYTICAL CHEMISTRY. 119Through this tube the air is slowly drawn, and the moment of depo-sition of moisture can be very sharply observed by viewing the inte-rior of the tube through a lens. The cooling is effected by passing acurrent of air through bisulphide of carbon contained in a vesselsurrounding the tube. R. R.Direct Estimation of Chlorine in Presence of Bromine andIodine.By G. VORTMANN (Monatsh. Chem., 3,510-530) .-Two yearsago the author published a short notice on the detection of chlorine inpresence of bromine and iodine, depending on the reactions of chlo-rides, broniides, and iodides with the peroxides of lead and manganese,in presence of acetic acid of various degrees of dilution (Abstr., 1880,509) ; and more recently he has given a sketch of the application ofthese reactions to quantitative analysis (ibid., 1882, p. 1230). I n thepresent paper, the reactions concerned in these processes are morefully discussed, and the methods of estimation are described indetail, and illustrated by numerous examples.Estimation of Chlorine in Presence of Bromine.-When the quantityof bromine present is but small, it is sufficient to heat the mixture ofchloride and bromide with lead dioxide and acetic acid of 2-3 percent.two or three times on the water-bath. With larger quantitiesof bromine, complete separation is somewhat difficult ; the method ofefiecting it will be further considered in connection with the separationof chlorine from bromine and iodine together.Estimation of Chlorine in Presence qf Iodiiie.-This is effected in thesame manner as in the last case, lead dioxide being used when thequantity of iodide present is but small, manganese dioxide being pre-ferable when it is large. I n this case also, the evaporation with diluteacetic acid must be repeated several times. The expulsion of theiodine may be accelerated by first boiling the liquid for a few minutesin a small flask ; this, however, can be done only when lead dioxide isemployed, as the use of manganese dioxide quickly gives rise toviolent percussive ebullition.In the latter case, the liquid must beheated in a beaker on the water-bahh, while a stream of air is passedthrough it. The estimations come out sharp, even when large quan-tities of iodine are present. I n using lead dioxide when small quan-tities of chlorine are to be estimated in presence of much iodine,the results are apt to come out too high.Estirnictiom of Bronaine in Presence oj' Iodine.-This estimation is veryeasily performed by eva,porating down the mixture of bromide andiodide with manganese dioxide and dilute acetic acid several times onthe water-bath, the evaporation being accelerated, if desired, by pass-ing a stream of air through the liquid.Estinaation of Chlorine in Presence of Bromine and Iodine together.-This may be effected either by boiling with lend dioxide and diluteacetic acid, whereby the iodides and bromides are decomposed simul-taneously ; or by first expelling the iodine by evaporating down withmanganese dioxide and acetic acid, and then the bromine by repeatingthis operation after addition of lead dioxide.I n operating by thefirst method, the mutual action of iodine and bromine gives rise tothe formation of iodic acid, to prevent which, as far as possible, it i120 ABSTRACTS OF CHEMICAL PAPERS.advisable to add the lead oxide to t,he boiling solution by small portionsa t a time.The liquid having been boiled for about half an honr, andthe water as it evaporates renewed from time to time, the dissolved leadis precipitated by hydrogen sulphide, without previous filtration ; andthe liquid, after being once more treated with hydrogen sulphide, iswarmed for some time on the water-bath and filtered. The filtrate isthen evaporated to complete dryness on the water-bath, the residuedrenched with dilute acetic acid, and the liquid evaporated down afteraddition of a small quantity of lead dioxide. The evaporation to dry-ness is then once more repeated, the residue finally dissolved in water,and the chlorine precipitated from the filtrate by nitrate of silver. Inworking by the second of the methods above mentioned, the mixtureof the halogen-compounds is several times evaporated down on thewater-bath with lead dioxide and acetic acid, and the chlorine in theresidue is estimated in the usual way.This method is preferable tothe former, in so far as it affects the expulsion of all the iodine andbromine without formation of oxy-acids, and may also afford themeans of estimating these two halogens a t the same time. Moreover,it gives more exact results than the first method; but, on the otherhand it has the disadvantage that in decomposing the iodides bymanganese dioxide, manganese passes into solution and is precipitatedin the subsequent treatment with lead dioxide, in the form of manga-nese dioxide, or rather of a compound of this oxide with dioxide oflead, Mn02,4Pb0,: this precipitate is difficult to wash, and it isonly after prolonged treatment with boiling water that filtrates areobtained which no longer become opalescent on addition of silvernitrate.The numerous analyses given in the paper show that the methodtherein described is applicable in all cases to the separation of chlorinefrom bromine and from iodine.Moreover it gives satisfactory resultsin the estimation of relatively large quantities of chlorine in presenceof small quantities of bromine. When on the other hand muchbromine is present, the results, even with careful working, come outtoo high by several units per cent.Fiitally the author observes that it is not necessary to briug thechlorine into combination with an alkali-metal by decomposing thelead chloride obtained in the process with potassium sulphnte, inas-much as the entire process is performed with hot dilute solutions, andthe solubility of the lead chloride is very considerably increased bythe presence of the dilute acetic acid and solution of lead acetate, sothat an incomplete solution of the lead salt is not to be apprehended.The manganese dioxide and lead dioxide added i n excess are very easyto wash, and the filtrates after a short time give not the slight,estturbidity with silver nitrate.The bromine or iodine given off in tthese processes of separatingchlorides from bromides and iodides may be collected and estimated.With regard to the estimation of iodine i n presence of chlorine orbromine, the author has already obtained satisfactory results.Forbromine, the numbers hitherto obtained are less satisfactory ; but hehopes soon to arrive at more exact results, which may form thesubject of a further communication. H. WANALYTICAL CHEMISTRY. 121Estimation of Carbonic Anhydride in the Air at Cape Horn.By A. M ~ N T Z and E. AURIN (Cow@. .rend., 94, 1651). -The scientificmission to Cape Horn has been provided with an apparatus for esti-mating the carbonic anhydride in the air. There are two sheet-ironaspirators, representing 300 litres of air ; and drawn-out tubes con-taining potassium hydroxide are fixed in metallic cases, to guardthem from accident, and are so arranged that they need not be re-moved from the cases when the air is being drawn through them.R.R.-Estimation of Phosphoric Acid. By 0. v. n. PFORDTEN (Ber.,15, 1929-1 930) .-This method depends upon the conversion of phos-phoric acid into ammonium phosphomolybdate, and subsequent esti-mation of the molybdenum (see p. 122).Estimation of Sulphur in Iron and Steel. By G. E. CRAIG (Clhem.News, 46, 199).-The method now recommended is more rapid andquite as accurate as that in which potassium chlorate and hydrochloricacid are employed ; 100 grains of the metal are placed in a 10 oz. flask,with 9 oz. water, 1+ oz. hydrochloric acid is added by means of astoppered funnel ; the gas evolved is passed by means of tubes, &c.,through an empty flask or test-tube (to condense vapours) into a nitro-gen bldb containing oz.hydrogen peroxide, and + oz. ammonia;when the action becomes sluggish heat is to be applied. After blow-ing air through, the contents of the nitrogen bulbs and the precedingcondensing flask are washed out into a beaker, and barium chlorideis added after acidifying the solution with hydrochloric acid and boil-ing. A blank experiment should be made with each new sample ofhydrogen peroxide. The presence of copper has no influence on theresults. E. W. P.A. K. 31.Estimation of Oxygen and Carbon in Iron. By A. LEDEBUR(Ding7.polyt. J;, 245,293).-The author found oxygen in many kindsof malleable iron, wrought iron Containing ferrosoferric oxide as amixture principally, whilst ingot iron contains ferrous oxide, either inthe dissolved state or as an alloy.The oxygen, especially in the latteycase, has a marked influence on the properties of the iron: hence itsdetermination in ingot iron is almost as important as that of thesulphiir and phosphorus. For analysis, clean dry iron filings free fromfatty constituents should be employed. For the removal of the lasttraces of moisture and organic matter, the filings are heated in acurrent of pura dry nitrogen gas, obtained by heating a mixture of1 pt. sodium nitrite, 1 pt. ammonium nitrare, 1 pt. potassium dichro-mate, and 10 pts. water, passing the gas through a solution of ferroussulphate, and over red hot copper turnings, and finally drying it bymeans of phosphoric anhydride. The hrdrogen gas is made fromzinc and eulphuric acid.It is passed through soda-lye and analkaline solution of lead, then through a heated tube filled with plati-nised asbestos, and eventually dried over concentrated sulphnric acidand phosphoric anhydride ; ?5 grams of iron borings are placed in aporcelain boat, and pnshed into a glass tube of which one end isconnected by means of a T-piece with the nitrogen and hydroge122 ABSTRACTS OF CHEMICAL PAPERS.tubes, whilst the other end is drawn out and communicates with hheabsorption- tube containing phosphoric anhydride. In con1 m encingthe analysis, the tube with the copper turnings is heated, and a slowstream of nitrogen passed through the apparatus. After two hours thetube wit,h the iron borings is heated, nitrogen gas being passed over con-tinuously, in order to expel all volatile constituents.The absorption-tube is then attached to the apparatus, the current of nitrogen stopped,and hydrogen passed through. After 30-45 minutes’ heating, theapparatus is cooled slowly, hydrogen still being passed over. Theabsorption-tube is then removed, and after expelling the hydrogenin the tube by means of air dried over phosphoric anhydride, it isweighed. The porcelain boat and contents are also weighed, and theweight of the oxygen of the water absorbed should agree with theloss in the weight of the porcelain boat. Analyses of a variety ofsamples of iron are given, the high percentage of oxygen in wroughtiron being explained by the admixture of slag. The determination ofoxygen is therefore said to afford a means of determining approxi-mately the quantity of slag present.For determining the carbon, t,heauthor recommends M’Greath acd Ullgren’s method. D. B.Electrolytic Estimation of Zinc. By A. MILLOT (BUZZ. SOC. Chin%.[el, 37, 339-341) .-2*5 grams of the mineral are dissolved in 50 C.C. ofhydrochloric acid, and a small quantity of potassium chlorate is addedto the boiling solution in order to precipitate the iron. If the mineralcontains much silica, it is previously evaporated to dryness with hydro-chloric acid. The liquid is cooled, diluted, and mixed with 100 C.C. ofammonia and 50 C.C. of a saturated solution of ammonium carbonatein order to precipitate the lead and calcium. The liquid is diluted to500 c.c., filtered, and 100 c.c., corresponding to 0.5 gram of the mineraland containing from 0.2 to 0.3 gram of zinc, are mixed with 1 gram ofpure potassium cyanide and placed in a beaker in which is suspendeda cylinder of platiaum gauze which acts as the positive pole, and aplatinum cone like that.in Riche’s apparatus which acts as a negativepole. Two Bunsen cells or a Clamond thermo-electric pile of 1.50 ele-ments may be used to effect precipitation, which is complete in aboutten hours. The firmly adhering deposit is washed with water, thenwith alcohol, and dried. If the mineral contains copper the latter isdeposited with the zinc. The deposit on the cone must in this case bedissolved in nitric acid and the copper precipitated from the acid solu-tion. If cadmium is present, it must be removed by treatment withhydrogen sulphide.The potassium cyanide should be used in theproportion given above ; if more is added, the metal is deposited veryslowly, whilst if less is added the deposit of zinc is not adherent. Anyaction on the electrodes may be prevented by mixing the solution withammonium aceta’te or nitrate. The latter, however, retards the pre-cipitation of the zinc.Reduction of Molybdenum Compounds. By 0. I-. D. PFORDTEN(Ber., 15, 1925--1929).-The author has examined Pisaiii’s methodfor estimating molybdenum. He finds that the end-product of thereduction of molybdic acid with zinc and hydrochloric acid is notC. H. BANALYTICAL CHEMISTRY. 133Mo,O, but Mo50,(= 2M0203 + MOO). This, however, becomesoxidised, by exposure to air, to BIo,O,.A method for the volumetricestimation of molybdic acid is founded on this reduction, and subse-quent oxidation with standard potassium permangauate.A. K. M.Otto's Method for the Estimation of Fuse1 Oil in Brandy.By C. KRAUCH (Bied. Centr., 1882,718).-Krauch does not find Otto'smethod of any use. The oxidation-products, which according to Ottocontain valeric acid, the author finds to be acetic acid.E. W. P.Estimation of Glycerol in Fatty Matters. By J. DATID(Compt. rend., 94, 1477--1479j.-lOO grams of the fat are melted;65 grams of barium hydrate, BaO,SH,O, are added with briskstirring ; when most of the water has been expelled, the heating isdiscontinued ; 80 C.C. of alcohol of 95" are poured on the mass, andthe whole is well stirred ; 1 litre of water is then added, and the wholeboiled for an hour.The barium soap remains insoluble, whilst theglycerol is dissolved by the water, which is freed from the excess ofbarium, reduced in volume by boiling, and finally evaporated in avacuum a t a low temperature ; or, preferably, the quantity of glycerolis inferred from the density of the solution. The barium soap afterbeing boiled with water is decomposed by hydrochloric acid and thefatty acid separated and weighed. Its melting point will indicateapproximately the proportions of stearic and oleic acids it contains.R. R.Estimation of Dextrose, Maltose, and Dextrin in Starch-sugar. By H. W. WILEY (Chent. News, 46,175-177).-The methodemployed was as follows:-(1) 10 grams undried sugar dissolved in1000 C.C.water ; (2) 10 grams dissolved in 100 C.C. and polarised in200 mm. tube ; (3) 10 C.C. of the solution (2) is treated with excess ofmercurous cyanide (120 grams HgCyz and 120 grams NaHO per litre)boiled, and excess of strong hydrochloric acid added and made up to50 C.C. ; this solution is polarised in 500 mm. tube and the angular rota-tion multiplied by 2. Solution (1) reduced by Fehling gives the totalpercentage of reducing matt'er, viz., dextrose with reducing value of100, maltose 62. The first polarisation gives the apparent specificrotation due t o all optically active bodies present, viz., dextrose = 52,maltose = 139 ; dextrin = 195. The second reduction (3) leaves onlydextrin unaffected, and the amount of this is determined by the secondpolarisation.If solid starch-sugar is employed, it must be boiled forsome time to destroy birotation; after reduction by cyanide, it isunnecessary to use charcoal, as the addition of the acid destroys thered colour generally present.Calculation of the results: from (1) we obtain the reducing percent. of dextrose d, + that of maltose w, which latter, compared withthe former, is only 0.62.(1.) R = d + 0.62 m.(2.) P = 5 2 d + 139m + 193Z.(3.) P' = 193d'. (Second polarisation)124 ABSTRACTS OF CHEMICAL PAPERS.To find d and m.Multiply (1) by 52 and substract from (4).(4.) P - P' = 5 2 d + 139 m.(5.) P - P' - 52 R = 106.76 m.P-€"-552R106.76 '(6.) Whence rn =(7.) d = R - 0.62 nz.. I P' (8.) d' = - 193'This process agrees well with that proposed by Allen. Severalanalyses are given. E. W. P.Diffusion of Sugar in Beet. By G. MARECK (Dingl. polyt. J.,245, 345--350).-1n determining the differences in the sp. gr. ofwhole roots and of roots cut into sections, the author found that theweight of the former was generally below that of the separate partsof the root, the differences being greater the smaller the sp. gr. of thewhole root. The valuation of beet according to the density of thejuice is said to give more accurate results than the determination ofthe sp. gr. of the roots. A table is given showing the results ofexperiments on the distribution of sugar in the beet. D. B.Estimation of Rice-starch. By F.SALOJION (J. pr. Chem. [el,26, 324-333).-The results of a series of experiments are givenwhich show tjhat while potato-&arch, when heated with hydrochloricacid, yields the full theoretical amount of glucose (111.11 per cent.),rice-starch cannot be made to yield more than about 107 per cent.The sp. gr. of the invert,ed solution is, however, identical with thatobtained from potato-starch, proving that about 4 per cent. of sub-stances are formed which have no reducing action on Fehling'ssolution. 0. H.Volumetrical Estimation of Phenol. By T. CHANDELON (BUZZ.SOC. Chiq~z. [ S ] , 38, 69-71).-The introduction of phenol as an anti-septic has necessitated a rapid and easy method for its estimation.Koppershoor has proposed to act on phenol with a standard solution ofbromine in potassium bromide so as t o convert the phenol into tribromo-phenol.The excess of bromine used may be determined by sodiumthiosulphate ; or a mixture of potassium bromide and bromate acidifiedby hydrochloric acid may be substituted for bromine-wat,er. Giacosaused a solution of bromine-water which has been standardised by aphenol solution of known strength, but as the precipitate of tri-bromophenol invariably retains a, certain quantity of bromine, theresults are far from being exact.The author proposes potassium hgpobromite which, like bromine,converts the phenol into tribromophenol. The method of operation is asfollows :-The hypobromite solution is prepared by dissolving 14-15grams pure potassium hydroxide in 1 litre of water and adding gra-dually to it 10 grams bromine.The solution is then diluted until iANALYTICAL CHEMISTRY. 125is of such a strength that 50 C.C. corresponds to 10 C.C. of a normal solu-tion of phenol of 10.5 per cent. or 0.05 gram of pure phenol. In orderto ascertain the strength of any phenol solution, 50 c.c, of the hypo-bromite is placed in a flask and the phenol solution is added until adrop of the solution gives no blue coloration with potassium iodide andstarch solntion. The method is sufficiently exact for clinical purposes,the error being about 1.2 per cent. If it is required to estimatephenol in urine, the latter is distilled with dilute sulphuric acid : in thecase of lint or cotton, the vapours of water, slightly acidulated withhydrochloric acid, are passed over them and subsequently condensed,and the phenol in the distillate estimated after rieutralisntion withpotassium hydroxide.V. H. V.Ammoniacal Alkaline Silver Solution as a Test for Form-aldehyde. By B. TOLLENS (Ber., 15, 1828-1830). - Salkowskihaving noticed the formation of fulminating silvei. in a solution ofsilver wliich had been treated witb caustic soda and ammonia, theauthor states that this has not occurred with his solution when pre-pared in the way described. If, however, it is allowed to evaporate ina shallow dish, small quantities of a detonating compound are formed,but not when it is kept in a stoppered bottle. The best way is tokeep the component parts separate, and mix them when required.With regard to the quantitative determiriation of aldehydes by theabove solution, for each molecule of formaldehyde 2 atoms of silveringeneral are thrown down ; but irregnlarities have occurred whichhave not yet been avoided. The best proportions for the solutionas yet have been found to be equal weights of a solution of 1 partsilver nitrate in 10 parts water, and 1 part caustic soda in 10 parts ofwater mixed together, and the ammonia added drop by drop untilcomplete solution of the silver oxide is effected.Examination of Fat.By H. Y. DE SCHEPPER and A. GEIRET,(Dingl. polyt. J., 245, 295-302).-1n order to ascertain the value ofa fat, it is necessary to determine the non-fatty constituents, i.e., theamoiint of water, sand, fibrous matter, &c., the total quantity of fattyacids and glycerol, and the amount of " candle material," Le., the solidfatty acids present in the latter.The water is determined by placing50-60 grams of the substance in a tared beaker and drying it a t110" for one hour, stirring the mixture occasionally with a tared glassrod. The temperature is then increased t o 125", and after two hours'heating the beaker with contents is weighed, the loss in weight givingthe amount of water. The sand and other substances are determinedby filtering 50-60 grams of the fat through a tared filter at 60--70",washing with hot benzene, drying and weighing. I f , owing to thepresence of glycerol, the dried product is deliquescent, it is washedwith alcohol before weighing.The process for determining the fattyacids and glycerol is based on the following conclusions. From theequation indicating the decomposition of fats in general :J. K. C.it follows that (calling a the molecular weight of the fatty acid126 ABSTRACTS OF CHEMICAL PAPERS.an equivalent of fat, ex.pressed by (3a + 92 - 3 x 18)g, gives3u grams fatty acids and 92 grams glycerol (9). Tf a represents themean molecular weight of the different fatty acids contained in fat,we obtain the following formulae for the quautity of fatty acids andglycerol calculated as per cent. on the fat :-To ascertain the factor a in a fat, 50 grams of the latter are saponifiedwith 40 C.C. potash-ley of sp. gr. 1.4 and 40 C.C. alcohol ; boiled in alitre of water for about an hour, decomposed with snlphuric acid, andthe fatty acids, after washing and drying, titrated with standardpotash-ley, 10 C.C.neutmlising exactly 1 gram margaric acid, or100 C.C. = 1000 : 270 = 37.037 C.C. standard acid. By calling a thetenths of C.C. necessary to neutralise I gram of fatty acid, we obtainthe following ratio between a and at :-a = 270 - ___ loo. Todetermineathe amount of neutral fat N, 1 gram of dried and filtered fat istitrated with the above solution. By taking r~, to indicate the tenthsof c.c., and assuming that the various triglycerides are all decomposedin the same manner, the amount of free fatty acids is equal t o thefollowing :-E'=- 1o09L, also N = 100 - ___ ''On per cent. of the fat.a aBy making use of these formulse we obtain the following generalequations :-3C)0a + 100 !! and 100 [l - (9%: a)]f = 100 3n + 38 a100 [l - (n : a ) ] 9200 -____300 .3a + 38' 9 =It was found that the molecnlar weight of fatty acids from tallowsranged between 280 and 274-hard tallows being nearest t o the highernumber. For palm oils, the molecular weight is 270 : without, there-fore, causing too serious an error, this number may be taken to repre-sent the molecular weight of tallows and palm oils, or mixtures ofboth, which simplifies the above method. The percentage of neutralfat in this case is N = 100 - n, whilst the above formule may bemodified as follows :-f = 95.52 ~F!!?.ZP + n, per cent., and g = 10.85 ?!!kZ? per cent. 100 100By taking into consideration the fact that all fats contain from 1 to1.5 per cent. albumin or cellulose, and that the percentage of fattyacids in tripalmitines is less than 95.5, and the glycerol more than10.85, the above formulae require a further alteration, viz. :-100 - n + 32, and g = 10.5 -100 '100 - n100f - 945 ___TECHSIC AL CHEMISTRY. 127or, f = 94.5 and y = 10.5 if n = 0, ondf = 100, and g = 0 if ?z =100, i.e., an increase in the percentage of neutral fat N from 0 to 100raises the percentage of fatty acids 5.5, whilst the glycerol is reducedby 10.5 per cent. ; so that the following formulae are obtained whichmay be used for determining the percentages of fatty acids andglycerol :-f = 100 - 0.055 N, and g = 0,105 N.The value of the fatty acids is ascertained by deterzining the crys-tallising points. It is, however, necessary to test, the acids for neutralfat by dissolving 1 gram in hot alcohol and adding strong ammoniato the solution. I n the presence of mere traces of neutral fat, thesolution is rendered turbid ou the addition of ammonia. D. B.Occurrence of Organic Bases in Commercial Amy1 Alcohol.By L. HAITINGER (Honatsh. Clzern., 3, 688-692) .-With reference tothe use of amyl alcohol in testing for alkaloi'ds, the author calls atten-tion to the fact that pyridine and other bases are frequently present,even in the commercially " pure " alcohols. A. J. G