ANALYTICAL CHEMISTRY. 97 An a l y t i c a1 Chemistry. Estimation of Hydrochloric Acid in the Gastric Juice. By J. BOAS (Chem. Centr., 1891, ii, 357; from Centr. med. Wiss., 1891, 509).-The author has adopted Bourget’s method (evaporation and incineration with barium carbonate, extraction with water, and pre- cipitation of the barium chloride formed by sodium carbonate), with the slight modification that, after dissolving the precipitate in decinormal acid, the liquid is boiled until all carbonic anhydride is expelled, and the excess of acid is then neutralised with decinormal alkali ; the indicator is phenolphthalei’n. Estimation of Free Oxygen by means of Nitric Oxide. By L. L. DE KONINCK (Zeit. ang. Chem., 1891, 78--80).-The author’s attention was drawn to an article by Wsnklyn and Cooper on the estimation of oxygen by means of nitric oxide.These chemists con- sider that this process, which was originated by Priestley, but which now only possesses a,n historical interest, is not only verj accurate but even one of the best methods known, and they communicated a few satisfactory test analyses. The author some time ago had occasion to try the process, but his experience agreed with the results obtained by Berthelot, Lunge, and Winkler, which were very irregular and unsatisfactory. He therefore thought it advisable to once more thoroughly investigate the matter. The modus operandi was practically identical with that employed by Wanklyn and Cooper, except, perhaps, as regards the preparation of the nitric oxide, which was obtained by the action of nitric acid (sp.gr. 1.1) on a, spiral of sheet-copper contained in an old-fashioned Hempel’s hydrogen pipette. On account of its simplicity and easy execution, the following plan was adopted :-The gas (oxygen or air) was introduced into a Winkler-Hempel burette, the nitric oxide introduced into a like apparatus, and thoroughly shaken with the water to saturate this with the gas, and also to remove traces of nitric peroxide. After a few minutes rest, the two volumes of gas were respectively recorded, and after connecting the two burettes with a capillary tube filled with water, the oxygen mixture was forced into the nitric oxide (not the reverse way), and the burettes were now disconnected. After thoroughly shaking t o facilitate the ab- sorption, and waiting for twelve or fifteen minutes, the volume of the gas was read off with the usual precautions, and the contraction noted.The results obtained once more proved the process to be absolutely useless, as the quantity of oxygeii found varied between 77.5 and 122.j per cent. of the amount operated on. From a scien- tific point of view, however, the results were not uninteresting. The ratio between the volume of the oxygen and the contraction after the action of the nitric oxide may theoretically vary between 1 : 2.33 and 1 : 5, as shown by the following equations :- VOL. LXII. h J. W. L.98 ABSTRACTS OF 0 dEMIOAL PAPERS. 4N0 + 30, + 2Hz0 = 4HNO; ratio, 1 : 2.33, 2N0 + 0 2 + R2O = HNOZ + RON,; ,, 1 : 3, 4N0 + 202 + H20 = 4HN0 ,, 1 : 5. A good deal depends on the celerity with which the gases have been mixed and shaken with the water.As a rule, a large excess of nitric oxide or admixture with an inert gas, such as nitrogen, seems to favour the formation of nitrous acid. L. DE K. Estimation of Free Oxygen dissolved in Water. By W. KISCH (Zed. ang. Chem., 1891, 105-108) .-The author compared the various processes in use for the estimation of free oxygen in water. (a,.) Bunsen-Tiemann’s process.-The water is boiled, and the gases collected over hot solution of caustic potash. The oxygen is then estimated by explosion with hydrogen or by absorption with alkaline pyrogallol. ( b . ) Mohr’s process -The sample is mixed with an acld solution of ferrous sulphate of known strength, then with aqueous soda to throw down the ferrous hydroxide.After remain- ing for a few hours (air, of course. being rigidly excluded), the -precipitate is redissolved in sulphuric acid, and the ferrous sulphate titrated with permanganate. (c.) The Schiitzenberger-Itisler process. -The water is allowed to act on a solution of sodium hydrindigotin- disulphonate, which then passes into the blue compound. Standard solution of sodium hyposulphite is then run in till the liquid is again decolorised. ( d . ) Winkler’s process.-T he water is mixed with a solution of manganous chloride, and then with potassium iodide and potassium hydroxide. The precipitated rnanganous hydr- oxide rapidly absorbs the oxygen, and passes into a higher oxidised state. On adding hydrochloric acid, iodine is liberated, which is then estimated by a solution of sodium thiosulphate.On applying these processes to the same sample of water, the results obtained by Bunsen-Tiemann’s process were decidedly lower than those ob- tained by Winkler’s process, whilst those obtained by both Mohr’s and the Schutzenberger-Rider methods corresponded very well with Winkler’s figures. The author, however, strongly recommends the latter process as being the safest, most trustworthy, and the easiest Schiltzenberger’s Process for the Estimation of Free Oxygen. By T. ‘K~NIG (Zeit. ang. Chem., 1891, 108, llO).-The author, not always getting satisfactory results with this process, investigated the cause. Success depends in a great measure on the quality of the indigo solution employed. If the sample is poor in quality, it does not absorb the oxygen rapidly enough, and a portion of this is swept away by the current of hydrogen. Indigo-carmine is obtained by precipitating a solution of purified indigo in sulphuric acid with salt, If simply collected, a pasty mass is obtained which forms the commercial indigo-carmine ; but, if well washed and dried, solid sodium indigodisulphonate is obtained.These preparations vary, however, considerably in strength, sometimes from 79.08 to 91.58 per cent. in the so-called crystallised and sublimed product, whilst ot.her in manipulation. L. DE K.ANALYTICAL CHEMISTRY. 99 samples vaaied in strength from 9.41 to 72.42 per cent. It is no use to increase the quantity of indigo solution when dealing with poor samples. The colour will generally be very dirty, and it will be almost impossible to notice the change from blue to a yellow or pale yellow.Analysts wishing t o use the process must, therefore, get the best quality of indigo-carmine. L. DE K. Analysis of Phosphates. By R. JONES (Zeit. ang. Chem., 1891, 3-41) .-The author is very satisfied with Giaser’s alcohol process, but thinks it may be much improved by lessening the amount of sulph- uric acid. Glaser recommends 45 grams of the strongest acid ; but even if the sample consisted solely of calcium oxide, 2 grams would be quite sufficient. The author also thinks that the quantity of the sample actually used for the analysis (0.4 gram) is far t,oo small, and that at least 1 gram should be taken. He finally recommends the following process :- 10 grams of the sample, which must be free from organic matter, is dissolved in nitro-hydrochloric acid, and diluted to 500 C.C. 50 C.C.of the solution (= 1 gram of the sample) is evaporated t o 25 C.C. and, while still hot, mixed with 10 C.C. of dilute sulphuric acid (1 : 5). After adding 150 C.C. of alcohol, the mixture is allowed t o remain for a t least three hours (Glaser thought half an hour sufficient). The calcium sulphate is collected and washed with alcohol until the washings, after being diluted with water, do not show the faintest acidity with methyl-orange. I f it is desired to estimate the amount of calcium (the process is really devised for the accurate estimation of theiron and nluniiiiium), the filter is put into a platinum dish, the alcohol burnt off, and the calcium sulphate finally ignited and weighed, Tho alcoholic filtrate is distilled, the residue rinsed into a beaker, and a slight excess of ammonia added, but it must again be completely boiled off. This is of particular importance when the phosphate con- tains magnesia ; the precipitate at first almajs contains magnesia, but this is redissolved on boiling. In order to get a clear filtrate, it is as well to wash the precipitate, consisting of ferric arid aluminic phos- phates, with water containing a little ammonium nitrate.If it is pre- ferred to weigh the iron and aluminium as pure oxides, the precipitate may be treated according to Stutzer’s plan, namely, removal of the phosphoric acid by ammonium molybdate and precipitation of the filtrate with ammonia.L. DE K. Physiological Detection of Carbonic Oxide in a Medium containing only 1 part in 10,000. By N. GRBHANT (Compt. rend., 113, 289--X90).-When air containing only 1 part of carbonic oxide in 10,000 is passed for half an hour through 50 C.C. of defibrinated, filtered dog’s blood, the respiratory power falls from 23.7 to 23.0. Under a pressure of 5 atmospheres, with the same gas and the same blood, the respiratory power falls to 17.2. This fact may be utilised for the detection of minute quantities of carbonic oxide. It is clear that the absorption of carbonic oxide by the blood is not determined by the percentage of this gas by volume, but by the mass of the gas present in a given Tolume. C. H. B.100 ABSTRAOTS OF OHEMICAL PAPERS. Separation of Barium from Calcium.By R. FRESENIUS (Zeit. anal. Chem., 30, 583-595).-Methods based on th)e Diferential Action of Alkaline Carbonates on the mixed Sulphates.-Boiling the freshly- precipitated sulphates with a mixture of potassium carbonate and sulphate, and subsequent treatment with hydrochloric acid, gives fairly accurate results, the comparatively small errors compensating one another. The errors are milch larger when cold ammonium carbonate solutioa is used for the decomposition of the calcium sulphate, the barium sulphate being also attacked to a considerable extent, whilst weighable amounts of calcium are left undissolved. Fleischer’s method gives results sufficiently accurate for most pur- poaes. The solution is precipitated by a mixture of 3 parts of potas- sium sulphate and 1 part of potassium carbonate.Since this at once throws down the calcium as carbonate, there is no tendency for the barium sulphate to retain calcium sulphate. After 12 hours’ diges- tion, the precipitate is washed with ammonium carbonate, then dried, weighed (burning the filter and treating the ash with ammonium carbonate), and treated with a measured excess of standard hydro- chloric acid, the excess being titrated with alkali. To complete the action of the acid on the calcium carbonate, prolonged digestion is necessary. Sidersky’s process likewiee affords serviceable results. The solution is precipitated by a mixture of ammonium sulphate and oxalate (200 grams of the former and 30 grams of the latter in a litre). The washed precipitate, consisting of bariiim sulphate and calcium oxalate, is dissolved in dilute hydrochloric acid, and the oxalic acid titrated by permaaganate.The barium sulphate can then be collected (after partially neutralising the acid by ammonia), or the joint amount of the two metals may be found in a separate portion. Rose’s proposal to separate calcium sulphate from strontium sulpliate by digestion with concentrated solution of ammonium sulphate, ap- peared likely to succeed with a mixture of barium and calcium sulph- ates, but prolonged digestion leaves much of the calcium undissolved, whilst the portion dissolved cannot be. accurately estimated by am- monium oxalate, since the ammonium sulphate solution (whether concentrated or dilute) is capable of dissolving a very notable amount of calcium oxalate.M. J. S. Combination of Wet and Dry Methods in Chemical Analysis. By W. E. ADENEY and T. A. SHEGOG (Chem, News, 64, 174-175, 185-187, and 192--193).-The problems which the authors wished to solve by direct experiments were : 1. Whether, when fused in the reduc- ing flame on charcoal with borax and sodium carbonate, the metals anti- mony, tin, lead, arsenic, silver, tismut,h, copper, nickel, and zinc could be completely reduced from their salt,s containing either volatile or non-volatile acids. ‘2. Whether, when sirnilurly treated, aluminium, chromium, manganese, cobalt, and iron are wholly non-reducible from their compounds; the cobalt and iron more especially from their arsenates and phosphates. 3. Whether, when a mixture of salts is similarly treated, the constituent metals thereof respectively behave as in simple co~ripounds, or whether their behaviour is modified in any way.In their present paper, the authors mainly deal with theANALYTICAL CHEMISTRY. 101 problems Nos. 1 and 2. The details of manipulation are as fol- lows :- The Charcoal Support.-The pieces should be about 15 inches in length, and about 1s inches in diameter, and should be tolerably free from fissures. The cavity into which the substance is to be intro- duced should be bored conical, the sides being slightly curved. The size, of course, varies with the quantities of the materials used in the experiment ; but, in the authors’ assays, the diameter was l h inches, and the greatest depth If inches. The Source of Heat.-Although an ordinary paraffin oil lamp, or one burning solid paraffin, may be ad- vantageously used, titill, when coal gas is procurable, it is by far the most convenient source of heat, as it can be burnt at the end of a flattened tube; and, if this be pivoted so that it can rotate in a vertical plane, it will be found of advantage in dealing with readily oxidisable met,allic beads.Phe blowpipe used was an ordinary mouth- blowpipe, furnished with a platinum jet, and fixed in a clip on a retort-stand. An india-rubber hand-blower was used for obtaining the blast. The boi-ax used was fused in a platinum dish, powdered, and kept in a dry aud well-stoppered bottle. The silver was at first used in the form of nitrate, but, this not answering well, it was re- placed by a, mixture of argentic chloride arid borax.In some cases it was used in the form of wire, or a butkon, after the substance and the fluxes had been first fused in the oxidising flame. The s o d i m carbojtafe, which the authors now only employ in the analysis of silicates, was the ordinary dried and powdered commercial article. Quantities.-The most convenient quantities to work on generally are 0.3 gram of substance and 1.2 grams of borax; but, when dealing with ores rich in nickel or zinc, not tilore than 0.1 gram should be taken. While the fusion is proceeding, the charge must be worked round and round the cavity, the metallic bead being made t9 run ‘round the glass and pick up the metals as reduced. ‘1’0 prevent oxidation, a stream of coal gas is directed into the cavity while the melt is cooling. As regards the first problem, the authors found that when fusing oxygenated compounds of antimony, tin, lead, arsenic, bismuth, and copper, or nickel chloride (mixed.with metallic arsenic), with a mix- kure of borax, sodium carbonate, and argentic chloride, the reduction of the metals was practjically complete, the glass bead being free from metal. Zinc, however, was foutid in both beads, showiug the incom- pleteness of the reduction. Auother series of experiments was now made. The ores and fluxes were fused together in a shallow cavity in charcoal with the aid of the oxidising flame. After cooling, the beads were transferred to a cavity of the usual shape a n d size, and, after the addittori of a button of metallic silver, hetted in the rsducing flame.Satisfactory result3 were obtained with antimony, lead, arsenic, copper, and nickel; but in the case of tin and bismuth the reduction was incomplete. As regards the ditficultly reducible metallic oxides, th3 authors operated as follows :-O*S gram of the substance was fused with 1.2 grams of borax or charcoal in the oxidising dame, 0.9 gram of silver button (1.2 grams when dealirig with cobalt) was added, and .the fusion102 ABSTRACTS OF CEIEMIOAL PAPERS. repeated i n the oxidation flame. The results were qualitatively satis- factory with cobalt, aluminium, chromium, and manganese ; of iron, only a mere trace passed into the silver button. The investigation as regards mixed metallic compounds will ba carried on by one of the authors, L.DE K.. Assay of Commercial Aluminium. By F. REGELSBERGER (Zeit. ung. Chem., 1891, 20, 52).--Klemp has proposed to measure the volume of hydrogen which is evolved by dissolving the metal in potash-ley ; but as commercial aluminium invariably contains silicon, which also evolves hydrogen, his process cannot give the exact amount of pure metal. A sample submitted to the author contained 98 per cent. of aluminium, but as it also contained 1.5 per cent. of silicon, it showed 99.9 per cent. of metal by Klemp’s process. The author thinks it by far the best plan to carefully estimate the percent- age of impurities in the sample and take the aluminium by difference ; but if a direct estimation of the metal is desired, he proceeds as follows :- Two grams of the sample is dissolved in a platinum basin, in water containing 15 grams of Fotassium hydroxide, and the whole is finally made up t o 200 C.C. 50 C.C.of the alkaline solution (= 0.5 gram of the sample) is now boiled with a slight excess of neutral ammonium nitrate which throws down the alumina with more or less silica. After washing, igniting, and weighing, any silica must be estimated in the usual way by fusion with potassium hydrogen sulphate, and if there is any reason to believe that t h e reagenta used contain alumina or silica, the usual check mnst be made. L. DE K. Direct Estimation of Aluminium in Iron and Steel. By T. M. DROWN and A. G. MCKENNA (Chem. News, 64,19+196).-The alumina obtained dur.ing a quantitative analysis often contains ferric oxide and phosphoric acid.As the direct estimation of the alumina is as yet very unsatisfactory, the analyst, as a rule, contents himself with estimating the iron and phosphoric acid, and taking the alumina by difference. This is a very good plan when the iron and phosphorus are present in relatively small quantities, but when, as in the modern alloys of aluminium and iron, the aluminium may be present only to the extent of a fraction of I per cent., nothing short of the isolation of the alumina can satisfactorily prove its presence. The authors, after many experiments, finally adopted the following electrolytic process :-About 5 t o 10 grams of the iron or steel is dissolved in sulphuric acid, and the solution heated until white fumes of aulphuric anhydride begin to come off, Boiling water is added to dissolve the iron salt, and the liquid filtered off from the silica and carbon, which are washed with acidified water.The filtrate is nearly neutralised with ammonia, and, if necessary, diluted to 300-500 C.C. The beaker in which the electrolysis is to be made contains from 500 to 1000 grams of mercury, which consti. tutes the cathode, It is connected with the battery or dpamo current in such a manner that about 2 amperes may pass through theANALP l‘ICAL CHEMISTRY. 103 solntion over night, which is best accomplished by using three lamps of 32-candle power, arranged in parallel on an Edison circuit. After, as far as possible, neutralising the acid which has been set free, the electrolysis is continued until the liquid gives no reaction for iron.It generally turns reddish from the formation of permanganic acid. When all the iron has amalgamated, the liquid is removed by means of a pipette while the cnrrent is still passing, and the mercury is repeatedly washed. The platinum anode, which will generally be slightJy coated with manganese dioxide, is now taken out, and the mercury again washed with water until the last traces of solution have been removed from it. After filtering to remoFe any suspended manganese dioxide, excess of sodium phosphate is added, and also 10 grams of sodium acetate. After boiling for at least 40 minutes, the aluminium phosphate is collected, washed, ignited, and weighed. The formula is not, as generally believed, AlzP20a, but 7AlZO3,6 P,O,. It should, of course, have a pure white colour, but it must, be remem- bered that the presence of even 4 per cent.of ferric oxide will give a, decidedly reddish product. In case of doubt, it may be fused with potassium hydrogen sulphate, and the solution once more submitted to electrolysis, but the authors hare always found this t o be super- -8uous. The test analyses are satisfactory. L. DE K, Separation of Iron from Cobalt, Nickel, and Manganese. By A. C. CAMPBELL (&it. anal. Chern., 30, 616-617; from J. arbaZ. Chem., 2, 291 ).-Ferric salts are precipitated by lead carbonate, whilst those of cobalt, nickel, manganese, and ferrous iron are not decomposed. Some lead chloride should be present to neutralise any traces of alkali. Warming promotes the reaction, and as oxida- tion of the cobaltous, manganous, and ferrous salts must be prevented, nitrates should be absent.The washed precipitate is treated with sulphuric acid, whereby the iron is redissolved and separated from the lead. To test the iron precipitate for cobalt and nickel, it is dissolved in hydrochloric acid and the concentrated solution reduced by tin foil. Traces of cobalt or nickel can then be recognised by the colour they communicate to the colourless ferrous chloride solution. M. J. S. New Methods of Quantitative Analysis. Part I. By A, BAUMANN (Zeit. ang. Chenz., 1891, 135--142).-When chromic acid is dissolved in dilute sulphuric acid and mixed with hydrogen peroxide, oxygen is evolved. According to some investigators, the reaction is not quantitative, but the author’s experiments prove that 1 mol.of chromic acid liberates exactly 2 mols. of oxygen and 1 mol. of potassium dichromate 4 mols. of that gas. The most suit- able apparatus is a Wagner’s azotometer, or a Knop’s apparatus con- nected with a Wagner’s gas flask, which is an ordinary flask into which a small glass cylinder has been sealed. Estimation of Chromic Acid.-The liquid, which must not be too concentrated and not exceed 50 C.C. in bulk, is mixed with 10 C.C. of dilute sulphuric acid (1 : 5) in the outer chamber of the flask. The104 ABSTRACTS OY CHEMCCAL PAPEHB. little gl~ss cylinder is filled with 5 to 10 C.C. of commercial hydrogen peroxide. After allowing the hydrogen peroxide to run into the chromate, the liquid will at first assume a fine blue colour, and then gradually evolve oxygen. The bulk of this gas will be given off in a few minutes, but the remainder will be only expelled after five minufeB’ brisk agitation.When shaking, the operator must open the stopcock about every half minute to let the oxygen gradually escape into the measuring tube. The liquid i n the non-graduated tube ought to stand a little lower during the evolution of the gas. When no more gas is given off, the apparatus is put, for about 15 minutes, into water of the temperature of the room ; the water in the tubes is levelled, and the volume of the gas is read off; 1 C.C. of oxygen at norms1 temperature and pressure = 0.002246 gram of chromic acid (Cr03). It is not advisable to work with hydrochloric instead of sulphuric acid, as there is always a risk of chlorine being evolved. Traces of free nitric acid do not interfere, but when the liquid contains more than 0.2 gram the oxygen found will be some- what too low.Acetic and succinic acids do not affect the results, but other organic acids or organic substances decidedly interfere. In standardising a solution of potassium dichromate, it must be remem- bered tha,t only three-eighths of the oxygen evolved is derived from the chromic acid. Estimation of Chromic Oxide.-This is very readily converted into a chromate by treating its alkaline solution with hydrogen peroxide, the excess of which may be driven off by boiling. After neutralising the solution with sulphuric acid, the chromic acid is determined as directed, and calculated to Cr203. In the assay of chrome-iron ore by this process, the use of nitre as an oxidising flux must be avoided, and 0.3 gram of the finely powdered mineral must be fused with a mixture of 3 prams of sodium carbonate and 3 grams of barium peroxide for half an hour.The mass must afterwards be decomposed with sulphuric instead of hydrochloric acid, Estimation of Combined Sulphuric Acid.-The liquid, which should contain no excess of hydrochloric acid and but traces of nitric acid or nitrates, is put into a 100 C.C. measuring flask, and precipitated in the cold with a solut’ion of pure barium chromate in hydrochloric acid. As the commercial salt often contains alkali chromate, it must before use be thoroughly washed with water until the filtrate is practically colourless. The residue i s then dissolved in an insuffici- ency of hydrochloric acid containing 3 per cent.HC1, and, after filtering, preserved for use. After diluting with water to about 90 c.c., the liquid is rendered faintly alkaline with ammonia, made up t o the mark, and filtered. An aliquot part of the filtrate, say 25 or 60 c.c., is then treated in the apparatus with sulphuric acid and hydrogen peroxide. 1 C.C. of oxygen = 0.001787636 gram of sulphuric anhydride. The various test analyses are very satisfactory. L. DE K.ANALYTICAL CHEMISTRY. 97An a l y t i c a1 Chemistry.Estimation of Hydrochloric Acid in the Gastric Juice. ByJ. BOAS (Chem. Centr., 1891, ii, 357; from Centr. med. Wiss., 1891,509).-The author has adopted Bourget’s method (evaporation andincineration with barium carbonate, extraction with water, and pre-cipitation of the barium chloride formed by sodium carbonate),with the slight modification that, after dissolving the precipitate indecinormal acid, the liquid is boiled until all carbonic anhydride isexpelled, and the excess of acid is then neutralised with decinormalalkali ; the indicator is phenolphthalei’n.Estimation of Free Oxygen by means of Nitric Oxide.By L.L. DE KONINCK (Zeit. ang. Chem., 1891, 78--80).-The author’sattention was drawn to an article by Wsnklyn and Cooper on theestimation of oxygen by means of nitric oxide. These chemists con-sider that this process, which was originated by Priestley, but whichnow only possesses a,n historical interest, is not only verj accuratebut even one of the best methods known, and they communicated a fewsatisfactory test analyses.The author some time ago had occasion totry the process, but his experience agreed with the results obtainedby Berthelot, Lunge, and Winkler, which were very irregular andunsatisfactory. He therefore thought it advisable to once morethoroughly investigate the matter.The modus operandi was practically identical with that employedby Wanklyn and Cooper, except, perhaps, as regards the preparationof the nitric oxide, which was obtained by the action of nitric acid(sp. gr. 1.1) on a, spiral of sheet-copper contained in an old-fashionedHempel’s hydrogen pipette. On account of its simplicity and easyexecution, the following plan was adopted :-The gas (oxygen or air)was introduced into a Winkler-Hempel burette, the nitric oxideintroduced into a like apparatus, and thoroughly shaken with thewater to saturate this with the gas, and also to remove traces ofnitric peroxide. After a few minutes rest, the two volumes of gaswere respectively recorded, and after connecting the two buretteswith a capillary tube filled with water, the oxygen mixture was forcedinto the nitric oxide (not the reverse way), and the burettes werenow disconnected.After thoroughly shaking t o facilitate the ab-sorption, and waiting for twelve or fifteen minutes, the volume of thegas was read off with the usual precautions, and the contractionnoted. The results obtained once more proved the process to beabsolutely useless, as the quantity of oxygeii found varied between77.5 and 122.j per cent.of the amount operated on. From a scien-tific point of view, however, the results were not uninteresting.The ratio between the volume of the oxygen and the contractionafter the action of the nitric oxide may theoretically vary between1 : 2.33 and 1 : 5, as shown by the following equations :-VOL. LXII. hJ. W. L98 ABSTRACTS OF 0 dEMIOAL PAPERS.4N0 + 30, + 2Hz0 = 4HNO; ratio, 1 : 2.33,2N0 + 0 2 + R2O = HNOZ + RON,; ,, 1 : 3,4N0 + 202 + H20 = 4HN0 ,, 1 : 5.A good deal depends on the celerity with which the gases havebeen mixed and shaken with the water. As a rule, a large excessof nitric oxide or admixture with an inert gas, such as nitrogen,seems to favour the formation of nitrous acid.L. DE K.Estimation of Free Oxygen dissolved in Water. By W.KISCH (Zed. ang. Chem., 1891, 105-108) .-The author compared thevarious processes in use for the estimation of free oxygen in water.(a,.) Bunsen-Tiemann’s process.-The water is boiled, and the gasescollected over hot solution of caustic potash. The oxygen is thenestimated by explosion with hydrogen or by absorption with alkalinepyrogallol. ( b . ) Mohr’s process -The sample is mixed with anacld solution of ferrous sulphate of known strength, then withaqueous soda to throw down the ferrous hydroxide. After remain-ing for a few hours (air, of course. being rigidly excluded), the-precipitate is redissolved in sulphuric acid, and the ferrous sulphatetitrated with permanganate. (c.) The Schiitzenberger-Itisler process.-The water is allowed to act on a solution of sodium hydrindigotin-disulphonate, which then passes into the blue compound.Standardsolution of sodium hyposulphite is then run in till the liquid isagain decolorised. ( d . ) Winkler’s process.-T he water is mixedwith a solution of manganous chloride, and then with potassiumiodide and potassium hydroxide. The precipitated rnanganous hydr-oxide rapidly absorbs the oxygen, and passes into a higher oxidisedstate. On adding hydrochloric acid, iodine is liberated, which isthen estimated by a solution of sodium thiosulphate. On applyingthese processes to the same sample of water, the results obtainedby Bunsen-Tiemann’s process were decidedly lower than those ob-tained by Winkler’s process, whilst those obtained by both Mohr’sand the Schutzenberger-Rider methods corresponded very well withWinkler’s figures.The author, however, strongly recommends thelatter process as being the safest, most trustworthy, and the easiestSchiltzenberger’s Process for the Estimation of Free Oxygen.By T. ‘K~NIG (Zeit. ang. Chem., 1891, 108, llO).-The author, notalways getting satisfactory results with this process, investigated thecause. Success depends in a great measure on the quality of theindigo solution employed. If the sample is poor in quality, it doesnot absorb the oxygen rapidly enough, and a portion of this is sweptaway by the current of hydrogen. Indigo-carmine is obtained byprecipitating a solution of purified indigo in sulphuric acid withsalt, If simply collected, a pasty mass is obtained which formsthe commercial indigo-carmine ; but, if well washed and dried, solidsodium indigodisulphonate is obtained.These preparations vary,however, considerably in strength, sometimes from 79.08 to 91.58 percent. in the so-called crystallised and sublimed product, whilst ot.herin manipulation. L. DE KANALYTICAL CHEMISTRY. 99samples vaaied in strength from 9.41 to 72.42 per cent. It is no useto increase the quantity of indigo solution when dealing with poorsamples. The colour will generally be very dirty, and it will bealmost impossible to notice the change from blue to a yellow or paleyellow. Analysts wishing t o use the process must, therefore, get thebest quality of indigo-carmine. L.DE K.Analysis of Phosphates. By R. JONES (Zeit. ang. Chem., 1891,3-41) .-The author is very satisfied with Giaser’s alcohol process, butthinks it may be much improved by lessening the amount of sulph-uric acid. Glaser recommends 45 grams of the strongest acid ; buteven if the sample consisted solely of calcium oxide, 2 grams wouldbe quite sufficient. The author also thinks that the quantity of thesample actually used for the analysis (0.4 gram) is far t,oo small, andthat at least 1 gram should be taken. He finally recommends thefollowing process :-10 grams of the sample, which must be free from organic matter,is dissolved in nitro-hydrochloric acid, and diluted to 500 C.C. 50 C.C.of the solution (= 1 gram of the sample) is evaporated t o 25 C.C.and,while still hot, mixed with 10 C.C. of dilute sulphuric acid (1 : 5).After adding 150 C.C. of alcohol, the mixture is allowed t o remain fora t least three hours (Glaser thought half an hour sufficient). Thecalcium sulphate is collected and washed with alcohol until thewashings, after being diluted with water, do not show the faintestacidity with methyl-orange. I f it is desired to estimate the amountof calcium (the process is really devised for the accurate estimation oftheiron and nluniiiiium), the filter is put into a platinum dish, the alcoholburnt off, and the calcium sulphate finally ignited and weighed, Thoalcoholic filtrate is distilled, the residue rinsed into a beaker, and aslight excess of ammonia added, but it must again be completelyboiled off.This is of particular importance when the phosphate con-tains magnesia ; the precipitate at first almajs contains magnesia, butthis is redissolved on boiling. In order to get a clear filtrate, it is aswell to wash the precipitate, consisting of ferric arid aluminic phos-phates, with water containing a little ammonium nitrate. If it is pre-ferred to weigh the iron and aluminium as pure oxides, the precipitatemay be treated according to Stutzer’s plan, namely, removal of thephosphoric acid by ammonium molybdate and precipitation of thefiltrate with ammonia. L. DE K.Physiological Detection of Carbonic Oxide in a Mediumcontaining only 1 part in 10,000. By N. GRBHANT (Compt.rend.,113, 289--X90).-When air containing only 1 part of carbonic oxidein 10,000 is passed for half an hour through 50 C.C. of defibrinated,filtered dog’s blood, the respiratory power falls from 23.7 to 23.0.Under a pressure of 5 atmospheres, with the same gas and the sameblood, the respiratory power falls to 17.2. This fact may be utilisedfor the detection of minute quantities of carbonic oxide. It is clearthat the absorption of carbonic oxide by the blood is not determinedby the percentage of this gas by volume, but by the mass of the gaspresent in a given Tolume. C. H. B100 ABSTRAOTS OF OHEMICAL PAPERS.Separation of Barium from Calcium. By R. FRESENIUS (Zeit.anal. Chem., 30, 583-595).-Methods based on th)e Diferential Actionof Alkaline Carbonates on the mixed Sulphates.-Boiling the freshly-precipitated sulphates with a mixture of potassium carbonate andsulphate, and subsequent treatment with hydrochloric acid, givesfairly accurate results, the comparatively small errors compensatingone another.The errors are milch larger when cold ammoniumcarbonate solutioa is used for the decomposition of the calciumsulphate, the barium sulphate being also attacked to a considerableextent, whilst weighable amounts of calcium are left undissolved.Fleischer’s method gives results sufficiently accurate for most pur-poaes. The solution is precipitated by a mixture of 3 parts of potas-sium sulphate and 1 part of potassium carbonate. Since this at oncethrows down the calcium as carbonate, there is no tendency for thebarium sulphate to retain calcium sulphate.After 12 hours’ diges-tion, the precipitate is washed with ammonium carbonate, then dried,weighed (burning the filter and treating the ash with ammoniumcarbonate), and treated with a measured excess of standard hydro-chloric acid, the excess being titrated with alkali. To complete theaction of the acid on the calcium carbonate, prolonged digestion isnecessary. Sidersky’s process likewiee affords serviceable results.The solution is precipitated by a mixture of ammonium sulphate andoxalate (200 grams of the former and 30 grams of the latter in alitre). The washed precipitate, consisting of bariiim sulphate andcalcium oxalate, is dissolved in dilute hydrochloric acid, and theoxalic acid titrated by permaaganate.The barium sulphate can thenbe collected (after partially neutralising the acid by ammonia), or thejoint amount of the two metals may be found in a separate portion.Rose’s proposal to separate calcium sulphate from strontium sulpliateby digestion with concentrated solution of ammonium sulphate, ap-peared likely to succeed with a mixture of barium and calcium sulph-ates, but prolonged digestion leaves much of the calcium undissolved,whilst the portion dissolved cannot be. accurately estimated by am-monium oxalate, since the ammonium sulphate solution (whetherconcentrated or dilute) is capable of dissolving a very notable amountof calcium oxalate. M. J. S.Combination of Wet and Dry Methods in Chemical Analysis.By W.E. ADENEY and T. A. SHEGOG (Chem, News, 64, 174-175,185-187, and 192--193).-The problems which the authors wished tosolve by direct experiments were : 1. Whether, when fused in the reduc-ing flame on charcoal with borax and sodium carbonate, the metals anti-mony, tin, lead, arsenic, silver, tismut,h, copper, nickel, and zinc couldbe completely reduced from their salt,s containing either volatile ornon-volatile acids. ‘2. Whether, when sirnilurly treated, aluminium,chromium, manganese, cobalt, and iron are wholly non-reducible fromtheir compounds; the cobalt and iron more especially from theirarsenates and phosphates. 3. Whether, when a mixture of salts issimilarly treated, the constituent metals thereof respectively behaveas in simple co~ripounds, or whether their behaviour is modified inany way.In their present paper, the authors mainly deal with thANALYTICAL CHEMISTRY. 101problems Nos. 1 and 2. The details of manipulation are as fol-lows :-The Charcoal Support.-The pieces should be about 15 inches inlength, and about 1s inches in diameter, and should be tolerably freefrom fissures. The cavity into which the substance is to be intro-duced should be bored conical, the sides being slightly curved. Thesize, of course, varies with the quantities of the materials used in theexperiment ; but, in the authors’ assays, the diameter was l h inches,and the greatest depth If inches. The Source of Heat.-Although anordinary paraffin oil lamp, or one burning solid paraffin, may be ad-vantageously used, titill, when coal gas is procurable, it is by far themost convenient source of heat, as it can be burnt at the end of aflattened tube; and, if this be pivoted so that it can rotate in avertical plane, it will be found of advantage in dealing with readilyoxidisable met,allic beads.Phe blowpipe used was an ordinary mouth-blowpipe, furnished with a platinum jet, and fixed in a clip on aretort-stand. An india-rubber hand-blower was used for obtainingthe blast. The boi-ax used was fused in a platinum dish, powdered,and kept in a dry aud well-stoppered bottle. The silver was at firstused in the form of nitrate, but, this not answering well, it was re-placed by a, mixture of argentic chloride arid borax. In some casesit was used in the form of wire, or a butkon, after the substance andthe fluxes had been first fused in the oxidising flame.The s o d i mcarbojtafe, which the authors now only employ in the analysis ofsilicates, was the ordinary dried and powdered commercial article.Quantities.-The most convenient quantities to work on generallyare 0.3 gram of substance and 1.2 grams of borax; but, when dealingwith ores rich in nickel or zinc, not tilore than 0.1 gram should betaken. While the fusion is proceeding, the charge must be workedround and round the cavity, the metallic bead being made t9 run‘round the glass and pick up the metals as reduced. ‘1’0 preventoxidation, a stream of coal gas is directed into the cavity while themelt is cooling.As regards the first problem, the authors found that when fusingoxygenated compounds of antimony, tin, lead, arsenic, bismuth, andcopper, or nickel chloride (mixed.with metallic arsenic), with a mix-kure of borax, sodium carbonate, and argentic chloride, the reductionof the metals was practjically complete, the glass bead being free frommetal. Zinc, however, was foutid in both beads, showiug the incom-pleteness of the reduction. Auother series of experiments was nowmade. The ores and fluxes were fused together in a shallow cavityin charcoal with the aid of the oxidising flame. After cooling, thebeads were transferred to a cavity of the usual shape a n d size, and,after the addittori of a button of metallic silver, hetted in the rsducingflame.Satisfactory result3 were obtained with antimony, lead,arsenic, copper, and nickel; but in the case of tin and bismuth thereduction was incomplete.As regards the ditficultly reducible metallic oxides, th3 authorsoperated as follows :-O*S gram of the substance was fused with 1.2grams of borax or charcoal in the oxidising dame, 0.9 gram of silverbutton (1.2 grams when dealirig with cobalt) was added, and .the fusio102 ABSTRACTS OF CEIEMIOAL PAPERS.repeated i n the oxidation flame. The results were qualitatively satis-factory with cobalt, aluminium, chromium, and manganese ; of iron,only a mere trace passed into the silver button.The investigation as regards mixed metallic compounds will bacarried on by one of the authors, L. DE K..Assay of Commercial Aluminium.By F. REGELSBERGER (Zeit.ung. Chem., 1891, 20, 52).--Klemp has proposed to measure thevolume of hydrogen which is evolved by dissolving the metal inpotash-ley ; but as commercial aluminium invariably contains silicon,which also evolves hydrogen, his process cannot give the exactamount of pure metal. A sample submitted to the author contained98 per cent. of aluminium, but as it also contained 1.5 per cent.of silicon, it showed 99.9 per cent. of metal by Klemp’s process. Theauthor thinks it by far the best plan to carefully estimate the percent-age of impurities in the sample and take the aluminium by difference ;but if a direct estimation of the metal is desired, he proceeds asfollows :-Two grams of the sample is dissolved in a platinum basin, inwater containing 15 grams of Fotassium hydroxide, and the whole isfinally made up t o 200 C.C. 50 C.C.of the alkaline solution(= 0.5 gram of the sample) is now boiled with a slight excess ofneutral ammonium nitrate which throws down the alumina withmore or less silica. After washing, igniting, and weighing, any silicamust be estimated in the usual way by fusion with potassiumhydrogen sulphate, and if there is any reason to believe that t h ereagenta used contain alumina or silica, the usual check mnst bemade. L. DE K.Direct Estimation of Aluminium in Iron and Steel. By T.M. DROWN and A. G. MCKENNA (Chem. News, 64,19+196).-Thealumina obtained dur.ing a quantitative analysis often containsferric oxide and phosphoric acid.As the direct estimation of thealumina is as yet very unsatisfactory, the analyst, as a rule,contents himself with estimating the iron and phosphoric acid, andtaking the alumina by difference. This is a very good plan whenthe iron and phosphorus are present in relatively small quantities,but when, as in the modern alloys of aluminium and iron, thealuminium may be present only to the extent of a fraction of I percent., nothing short of the isolation of the alumina can satisfactorilyprove its presence. The authors, after many experiments, finallyadopted the following electrolytic process :-About 5 t o 10 grams ofthe iron or steel is dissolved in sulphuric acid, and the solutionheated until white fumes of aulphuric anhydride begin to come off,Boiling water is added to dissolve the iron salt, and the liquid filteredoff from the silica and carbon, which are washed with acidified water.The filtrate is nearly neutralised with ammonia, and, if necessary,diluted to 300-500 C.C.The beaker in which the electrolysis is tobe made contains from 500 to 1000 grams of mercury, which consti.tutes the cathode, It is connected with the battery or dpamocurrent in such a manner that about 2 amperes may pass through thANALP l‘ICAL CHEMISTRY. 103solntion over night, which is best accomplished by using three lampsof 32-candle power, arranged in parallel on an Edison circuit. After,as far as possible, neutralising the acid which has been set free, theelectrolysis is continued until the liquid gives no reaction for iron.Itgenerally turns reddish from the formation of permanganic acid.When all the iron has amalgamated, the liquid is removed bymeans of a pipette while the cnrrent is still passing, and the mercuryis repeatedly washed. The platinum anode, which will generally beslightJy coated with manganese dioxide, is now taken out, and themercury again washed with water until the last traces of solutionhave been removed from it. After filtering to remoFe any suspendedmanganese dioxide, excess of sodium phosphate is added, and also10 grams of sodium acetate. After boiling for at least 40 minutes,the aluminium phosphate is collected, washed, ignited, and weighed.The formula is not, as generally believed, AlzP20a, but 7AlZO3,6 P,O,.It should, of course, have a pure white colour, but it must, be remem-bered that the presence of even 4 per cent.of ferric oxide will give a,decidedly reddish product. In case of doubt, it may be fused withpotassium hydrogen sulphate, and the solution once more submittedto electrolysis, but the authors hare always found this t o be super--8uous.The test analyses are satisfactory. L. DE K,Separation of Iron from Cobalt, Nickel, and Manganese.By A. C. CAMPBELL (&it. anal. Chern., 30, 616-617; from J.arbaZ. Chem., 2, 291 ).-Ferric salts are precipitated by lead carbonate,whilst those of cobalt, nickel, manganese, and ferrous iron are notdecomposed. Some lead chloride should be present to neutraliseany traces of alkali. Warming promotes the reaction, and as oxida-tion of the cobaltous, manganous, and ferrous salts must be prevented,nitrates should be absent.The washed precipitate is treated withsulphuric acid, whereby the iron is redissolved and separated from thelead. To test the iron precipitate for cobalt and nickel, it is dissolvedin hydrochloric acid and the concentrated solution reduced by tinfoil. Traces of cobalt or nickel can then be recognised by the colourthey communicate to the colourless ferrous chloride solution.M. J. S.New Methods of Quantitative Analysis. Part I. By A,BAUMANN (Zeit. ang. Chenz., 1891, 135--142).-When chromic acidis dissolved in dilute sulphuric acid and mixed with hydrogenperoxide, oxygen is evolved. According to some investigators, thereaction is not quantitative, but the author’s experiments prove that1 mol.of chromic acid liberates exactly 2 mols. of oxygen and1 mol. of potassium dichromate 4 mols. of that gas. The most suit-able apparatus is a Wagner’s azotometer, or a Knop’s apparatus con-nected with a Wagner’s gas flask, which is an ordinary flask intowhich a small glass cylinder has been sealed.Estimation of Chromic Acid.-The liquid, which must not be tooconcentrated and not exceed 50 C.C. in bulk, is mixed with 10 C.C. ofdilute sulphuric acid (1 : 5) in the outer chamber of the flask. Th104 ABSTRACTS OY CHEMCCAL PAPEHB.little gl~ss cylinder is filled with 5 to 10 C.C. of commercial hydrogenperoxide. After allowing the hydrogen peroxide to run into thechromate, the liquid will at first assume a fine blue colour, and thengradually evolve oxygen.The bulk of this gas will be given off in afew minutes, but the remainder will be only expelled after fiveminufeB’ brisk agitation. When shaking, the operator must openthe stopcock about every half minute to let the oxygen graduallyescape into the measuring tube. The liquid i n the non-graduatedtube ought to stand a little lower during the evolution of the gas.When no more gas is given off, the apparatus is put, for about15 minutes, into water of the temperature of the room ; the water inthe tubes is levelled, and the volume of the gas is read off; 1 C.C. ofoxygen at norms1 temperature and pressure = 0.002246 gram ofchromic acid (Cr03). It is not advisable to work with hydrochloricinstead of sulphuric acid, as there is always a risk of chlorine beingevolved. Traces of free nitric acid do not interfere, but when theliquid contains more than 0.2 gram the oxygen found will be some-what too low. Acetic and succinic acids do not affect the results, butother organic acids or organic substances decidedly interfere. Instandardising a solution of potassium dichromate, it must be remem-bered tha,t only three-eighths of the oxygen evolved is derived fromthe chromic acid.Estimation of Chromic Oxide.-This is very readily converted intoa chromate by treating its alkaline solution with hydrogen peroxide,the excess of which may be driven off by boiling. After neutralisingthe solution with sulphuric acid, the chromic acid is determined asdirected, and calculated to Cr203. In the assay of chrome-iron ore bythis process, the use of nitre as an oxidising flux must be avoided,and 0.3 gram of the finely powdered mineral must be fused with amixture of 3 prams of sodium carbonate and 3 grams of bariumperoxide for half an hour. The mass must afterwards be decomposedwith sulphuric instead of hydrochloric acid,Estimation of Combined Sulphuric Acid.-The liquid, which shouldcontain no excess of hydrochloric acid and but traces of nitric acid ornitrates, is put into a 100 C.C. measuring flask, and precipitated inthe cold with a solut’ion of pure barium chromate in hydrochloricacid. As the commercial salt often contains alkali chromate, itmust before use be thoroughly washed with water until the filtrate ispractically colourless. The residue i s then dissolved in an insuffici-ency of hydrochloric acid containing 3 per cent. HC1, and, afterfiltering, preserved for use.After diluting with water to about 90 c.c., the liquid is renderedfaintly alkaline with ammonia, made up t o the mark, and filtered.An aliquot part of the filtrate, say 25 or 60 c.c., is then treated inthe apparatus with sulphuric acid and hydrogen peroxide. 1 C.C. ofoxygen = 0.001787636 gram of sulphuric anhydride.The various test analyses are very satisfactory. L. DE K