416 THE ANALYST, INORGANIC ANALYSIS. A Delicate Colour Reaction for Copper, and a Miero-chemical Test for Zinc. Harold C. Bradley. ( A w z e T . Jozm. Sci., xxii., No. 130; through Chem. News, 1906, vol. 94, p. 189.)--It has long been known that hsmatoxylin (the unoxidized extract of logwood) gives a dark blue colour with copper salts. The author has found that the reaction is exceedingly delicate, solutions containing only 0~0000001 per cent. of copper still giving a blue colour, whilst the limit with ferro- cyanide is 0.001 and with potassium iodide 0.0001 per cent. The absolute quantities taken are not stated. Zinc can be readily identified in the presence of other bodies by its character- istically crystalline nitro-prusside, all the other heavy metals giving amorphous, slimy nitro-prussides.Thus, zinc was recognised in the blood of certain molluscs by incinerating a small quantity, preparing a roughly 10 per cent. solution of the ash, removing copper by means of sulphuretted hydrogen, evaporating the filtrate to a small bulk, and digesting 1 drop of the solution on a microscope slide with 1 drop of a fresh solution of nitro-prusside. On cooling, the rectangular plates and prisms of zinc nitro-prusside could be readily seen under the microscope, A. G. L. Estimation of Copper by Titanium Chloride. Ezra Lobb Rhead. (Journ. Chewz. SOC., 1906, vol. lxxxix., p. 1491:)-1n presence of potassium thiocyanate, copper is quantitatively reduced from the cupric to the cuprous state by a solution of titanium trichloride. The reaction takes place in either sulphuric acid or hydrochloric acid solutions.As indicator a little ferrous sulphate is added, which is oxidized by the cupric salt to the ferric salt, the latter giving the well-known red colour of ferric thiocyanate to the solution; the disappearance of this colour marks the end of the reaction. The titration must be carried out below 30" C. and in an atmosphere of hydrogen or other inert gas, as titanium trichloride oxidizes readily in air. Ferric salts must he absent or else their amount known and allowed for ; nitric acid must also be absent. The results given by the method are good. A. G. L. The Determination of Arsenic Acid. L. Rosenthaler. (Zeit. uml. Chem. , 1906, xlv., 596-599.)-The method is based on the following reaction (the converse of that in general use in the determination of arsenious acid) : 2H,AsO, + 4KI + 4HC1= As,O, + 41 + 4KCl+ 5H,O.I n the presence of a large excess of acid, the reaction begins without the application of heat, and is complete in ten to fifteen minutes, after which the liberated iodine can be titrated with FG thiosulphate solution, each C.C. of which corresponds to 9 mgms. of potassium arsenate. I n making a determination, the potassium arsenateTHE ANALYST. 417 is dissolved in water, 2 grams of potassium iodide introduced, and hydrochloric acid (25 per cent. strength) added until a precipitate results. This is dissolved by the addition of the smallest possible amount of water, and the liquid allowed to stand ten to fifteen minutes before titration of the iodine.Sulphuric acid can be used in place of hydrochloric acid with equally satisfactory resalts. For the simultaneous determination of arsenic and arsenious acids, the latter is first titrated in the usual way with iodine solution. The liquid is then treated with 1 to 2 grams of potassium iodide, and fuming hydrochloric acid and 50 per cent. sulphuric acid (proportions not given) introduced until the precipitate forms, after which the determination is completed as above described. The method can also be used in the determination of arsenic, the latter being first oxidized into arsenic acid. C. A. M. Determination of Minute Quantities of Arsenic. W. Thomson. (Chem. News, 1906, vol. xciv., pp. 156, 157, and 166, 167.)-The author points out that all nitrous compounds formed in the destruction of organic matter must be completely removed by repeated evaporation in order to obtain the proper size and depth of the arsenic mirror when the electrolytic process is employed far the determination.With the Marsh-Berzelius method the presence of some nitrous compounds does not materially affect the results. Results of experiments are also given to show that l e d cathodes are the most efficient in reducing arsenic compounds, in the presence of nitric or nitrous acid, to the state of arseniuretted hydrogen. From experiments with insensitive zinc in the Marsh-Berzelius process the author concludes that the addition of cadmium sulphate does not always have the effect ascribed to it by Chapman and Law (ANALYST, 1906, p. 3). w. P. s.Investigations on Red Phosphorus. A. Siemens. (Arb. am d. &is. Geszindheitsamte ; through Chem. Ztg., 1906, XXX., Rep., 349.)-The author shows that every sample of red phosphorus can be made to give indications of the presence of yellow phosphorus in the Mitscherlich test, if only the boiling is sufficiently rapid (under somewhat reduced pressure), and a current of hot oxygen is employed. The '' glowing " of phosphorus appears to be due to the formation of a volatile lower oxide, which is formed from red phosphorus at 90" C. at the same rate as from yellow phosphorus in the cold. The author has worked out a method for the detec- tion of yellow in red phosphorus, which depends on the reduction to metal of certain metallic salt solutions by means of the dissolved phosphorus.He also shows the view that concussion changes red into yellow phosphorus to be untenable; the red phosphorus is only changed into a, more finely divided modification, in which state it is more easily acted on by reagents and more readily soluble. A. G. L. The Detection of Ozone. F. Fischer and H. Marx. (Bcrichte, 1906, xxxix., 2555-2557.)-Paper dipped in an alcoholic solution of tetramethyl-p-diamido-diphenyl- amine becomes violet when brought in a moist condition into contact with ozone. Oxides of nitrogen colour this 6' tetramethyl base " paper a straw yellow, whilst mixtures of nitrous oxides and ozone give dirty-brown intermediate shades. It is418 THE ANALYST. essential to have the test-paper moist, since the dry paper is coloured yellow by the long-continued (thirty minutes) action of ozone.If either the nitrous oxide or the ozone largely predominates in a mixture of the gases, the paper gives the colour characteristic of the gas that is in excess. Paper that is coloured violet by ozone becomes yellow if subsequently exposed to the action of nitrous oxides, and vice cemd. A trace of nitric oxide in ozone may be detected by conducting the mixed gases into liquid air, when the ozone dissolves and the nitric oxide solidifies in flakes which can be separated by filtration, while the ozone remains in the filtrate. C. A. M. The Examination of Liquid Carbon Dioxide. Werder. (Cherm. Ztg., 1906, xxx., 1021.)-According to the author, the spread of the total abstinence movement has led to a considerable increase in the quantity of liquid carbon dioxide used in the manufacture of aerated waters and fruit syrups.The quantity of liquid carbon dioxide made in Switzerland is estimated for the year 1906 at 300,000 to 350,000 kilogms., the greater part being €or human consumption. The carbon dioxide is made in three ways : from coke, potassium bicarbonate being an intermediate product ; by heating magnesite or other carbonates ; and from breweries. The author proposes the following specification for the liquid : There should be no pungent or empyreumatic odour. The quantity of carbon dioxide present should not be less than 98 per cent., that of carbon monoxide not more than 0.5 per cent. Sulphurous and nitrous acids must be absent. Or, leading the gas for fifteerl minutes through 100 C.C.of a warm potassium permanganate solution containing sulphuric acid, no appreciable decolorization should take place ; nor should it produce a precipitate in the same time in 100 C.C. of a & silver nitrate solution acidified with nitric acid. In taking a sample of the gas, the bottle should always be in the horizontal position ; otherwise the quantity of air in the sample taken is excessive. The carbon dioxide, oxygen, and carbon monoxide present are best determined in an Orsat apparatus, the measuring vessel of which is narrowed in its upper part, so that 20 C.C. can be read to qD C.C. The measuring vessel is filled ten or twenty times with the gas, the carbon dioxide being absorbed each time, and the united residues are then analyeed for oxygen (0.01 to 0.05 per cent.) and carbon monoxide (0.05 to 0.20 per cent.).Communication between the Orsat apparatus and the reducing valve of the bottle is conveniently made by a tube carrying a three-way tap. The taste should be purely acid. A. G. L. Determination of Nitrates. Frank Sturdy Sinnatt. (Proc. Chenz. Xoc., 1906, vol. 22, p. 255.)-Knecht and Hibbert's method for the determination of picric acid (Bey., 1903, xxxvi., 1549) was applied to the determination of potassium nitrate as follows : Ten C.C. of a 0.1 per cent. solution of potassium nitrate mere evaporated to dryness, and the residue heated for thirty minutes in a steam-oven with 5 C.C. of a solution of phenol-sulphonic acid. The whole was then washed into a flask, hydro- chloric acid added, and titrated with titanium trichloride.Three titrations carried out in this way gave 0*009788, 0.009948, and 0.009817 gram of potassium nitrate respectively, instead of the 0.01 gram taken. A. G. L.THE ANALYST, 419 The Determination of Hydrosulphurous Acid in Hydrosulphites and in their Combinations with Formaldehyde. A. Seyewetz and &loch. (BztZZ. SOC. C'him., 1906, xxxv., 293-297.)-TThe method is based upon the fact that hydro- sulphurous acids and hydrosulphites reduce silver halides to metallic silver. The hydrosulphite is dissolved in boiled distilled water, with precautions to exclude atmospheric oxygen, and is immediately introduced into an ammoniacal solution of silver chloride containing four or five times the theoretical amount of silver salt for the reduction, The precipitated silver is collected, washed with ammoniacal water, ignited, and weighed.The reaction may be expressed by the equation :++ 2AgC1+ 4NH, + Na,S,O, + R,O = 2NaCl+ 2(NH,),SO, + Ag,. The results quoted show that the method gives better results thau those obtained by titration with indigo carmine. The commercial product hyraldite (a mixture in varying proportions of sodium hydrosulphite-formaldehyde and sodium bisulphite-formaldehyde) cannot be titrated with indigo carmine. In aqueous solution it only reduces amnioniacd silver chloride slowly, but the reduction is immediate and complete at 80" C. The solution of hyraldite sliould be boiled for four or five minutes with four times the required amount of the reagent, and the precipitated silver weighed as in the case of hydrosulphites.Assuming that a molecule of pure hyraldite (NaHSO, + CH,O + 2H,O) reduces 2 molecules of silver chloride, the commercial hyraldite, C, of the Lyons manufac- turers contains from 84.42 to 86 per cent. of the pure compound. The formaldehyde has no action upon the ammoniacal silver chloride, and is itself transformed in the reaction into hexamethylene- tetramine. C. A. M. The Action of Sulphides on Nitroprussides. J. F. Virgili. (Zeit. a n d Chenz., 1906, xlv., 409-439.)-A soluble sulphide acting on sodium nitroprusside, or, in general, on any soluble or insoluble nitroprusside, yields a blue substance, which the author concludes to be a molecular addition compound. The presence of free alkali or of alkaline earths prevents the formation of this blue coloration, and certain salts of weak acids which can yield alkalies on hydrolysis-e.g., silicates, phosphates, and borates-also interfere with the reaction.Ammonia, and its salts have less influence. If an excess of ,z soluble nitroprusside act upon a soluble sulphide in the presence of a sufficient quantity of free alkali, a red coloration is obtained. The author attributes the reddish-yellow, red, purple and violet colorations that are frequently obtained to the simultaneous production of the blue compound between the sulphide and nitro- prusside, and of a yellow substance formed by the action of alkali or alkaline earth upon the nitroprusside. As regards the use of nitroprusaides in analysis, the conclusion is arrived at that they are not reagents for the sulphide ion, but for the unionized sulphide molecule, and are hence less sensitive than the solutions of certain metallic salts--e.g., neutral or alkaline lead salts-which act upon the sulphide ion.The sensitiveness of sodium nitroprusside as a reagent for sulphides is increased by the use of solvents-e.g., glycerin or alcohol-which prevent or check the ionization, and also by the addition of an excess of the reagent, or by the intro- * Assuming that, in the indigo process, one molecule ( 3 ) Na,S,04 -k SH,O (forrnula of I3ernthsen) liberates I€,.420 THE ANALYST, duction of other ions, for which purpose neutral salts, and in particular alkali carbonates, are suitable. The addition of alkali greatly increases the sensitiveness of the test, fixed alkalies being more effective than ammonia.If the sulphide solution containing the nitroprusside, but not an excess of alkali, be cooled to the freezing-point, the sensitiveness of the reaction is almost doubled, but even then falls short of that shown by reagents acting directly upon the sulphide ion. The maximum of sensitive- ness is reached with an excess of alkali in the frozen solution, but the presence of the alkali prevents a distinction being made between sulphides and hydrogen sulphide. I t is not possible to differentiate sulphides from sulphydrates by means of the nitro- prusside reaction ; nor can a coloritnetric determination of sulphides be made, owing to the difficulty of obtaining coniparabie coloured solutions and the want of sensitive- ness of the reagent. C.A. M. The Determination of Halogen. James Moir. (Proc. Chem. Xoc., 1906, V O ~ . 22, p. 261.)-The substance to be analysed is weighed into a tall nickel crucible, 10 drops of water and 5 to 7 ‘( pastilles ’’ of pure potassium hydrate are added, and the mixture stirred with a platinum wire whilst it is heated on a steam-bath ; when it is uniform, 0.5 to 1 gram of finely-powdered potassium pernianganate is gradually stirred in; the heating is then continued until the mixture is dry, care being used to prevent frothing; the crucible is next heated to redness over a burner, until the reaction between carbon and manganese dioxide is over. The cool crucible is next placed in a warm dilute solution of potassium bisulphite; the solution obtained is acidified with acetic acid, and filtered into a solution of silver nitrate, the silver halide precipitated being treated as usual.Or the contents of the crucible may be dis- solved in water, and acetic acid added until the manganate has been converted into perrnanganate ; the latter is destroyed by adding barium dioxide, and the filtered solution is neutralized with sodium bicarbonate and titrated, using chromate as an indicator. A. G, L. Modification of the Hanging-Drop Fluoride Test. C. D. Howard. ( JouY?~. - h e r . Chem. SOC., 1906, xxviii., 1238, 1239.)-A test-tube of small bore, about 2 inches long, is fitted with a, rubber stopper, into the bottom of which passes a small piece of glass-tubing closed at one end, its open end extending about 3 mm.into the tube. Tho precipitate of CaCO, and CaF,, ignited until nearly free from carbonate, is well mixed with about 0-1 gram of precipitated silica, and introduced into the dry test- tube, the glass tubing nearly filled with two or three drops of water, the bottom of the stopper thoroughly dried, and the stopper inserted immediately after the addition of 1 to 2 C.C. strong H,SO,. The tube is then immersed in a beaker of boiling water for fifteen to thirty minutes, when, if the solution under examination contains any appreciable quantity of fluoride, a heavy gelatinous ring will be formed in the small tube projecting from the stopper. By comparison with tubes containing known amounts of fluoride, the process may be made roughly quantitative. W. H. S.Back Reactions in Iodine Titrations. J. H. Davies and E. P. Perman. (Chem. News, 1906, 93, p. 235.)-The ‘( back reaction ” often observed in the titrationTHE ANALYST, 421 of iodine liberated from potassium iodide, using starch as indicator, is found to depend on the concentration of the potassium iodide in solut.ion, and is most marked when the iodine is liberated by copper sulphate or potassium bichromate, but only very slight when bleaching-powder or permanganate is used. I t appears to be due to the comparative slowness and incompleteness of the liberation of iodine, and may be overcome by increasing the concentration of the iodide solution to 1 gram in 10 C.C. water for 25 C.C. & K,Cr,07 and 2 grams jn 20 C.C. water for 1 gram of CuSO,SH,O in 50 C.C.water. W. H. S. The Separation of Silica in the Determination of Citric-Acid-Soluble Phosphoric Acid. J. Hasenbaumer. (Clzem. Zeit., 1906, xxx., 665, 666).-The author finds that the magnesium pyrophosphate obtained in determinations of the citric-acid-soluble phosphoric acid in Thomas slag does not contain silica, and consequently the high results found when the silica has not been separated previously are not due to occluded silica. Further experiments make it appear probable that the higher results obtained when the silica has not been separated are due in part to the possible retention of phosphoric acid by the gelatinous silica, and also to the fact that the magnesium - ammonium phosphate has a slightly varying composition, depending on the previous separation of the silica.The ratio of magnesia to phosphoric acid is 1 : 2.02 when the silica has not been separated previously, and 1 : 1.95 in cases where the silica has been removed. w. P. s. New Method for Preparation of Standard Solutions. S. F. Acree and R. F. Brunel. (Amer. Chenz. Joz~m., 1906, vol. 36, pp. 117-123.)-For the prepara- tion of solutions of HC1 and NH,OH the authors recommend passing the purified, and well-dried gases into a tared graduated flask, nearly filled with conductivity water, until the necessary weight is dissolved, the solution being then made up to the mark with water. The titration of solutions of HC1 and H,SO, with recrystallized NaHCO,, weighing the resultant NaCl or Na,SO,, is considered preferable to the silver chloride and barium sulphate methods, while the unknown strengths of an acid and of an alkaline solution can be determined simultaneously by neutralizing one with the other and evaporating to dryness, the weight of the dry salt and the- volumes of acid and alkali solutions used furnishing the necessary data.w. H. s. On Nessler’s Reaction. A. Buisson. ( J o z m . Pharm. Chim., 1906, xxiv., 289-294.)-It is shown experimentally that the reaction of ammonia with Nessler’s reagent is incomplete, and that a state of equilibrium is established between the different substances. The animonia is not completely precipitated, and on filtering off’ the precipitate and distilling the filtrate, the distillate will give a fresh reaction with Nessler’s solution. Thus, in one experiment in which a solution containing 0.189 gram of ammonium chloride in 40 litres was treated in this way, the distillate contained 0.040 gram of amiiiouia, representing 21 per cent. upon which the reagent had not acted in the first instance.The author also finds that the precipitate has not the composition usually attributed to it. The simplest formula corresponding422 THE ANALYST. with his analytical results is Hg,N,I,. The compound is a brown amorphous body, insoluble in neutral solvents, but soluble in excess of potassium iodide solution, which liberates the whole of its nitrogen in the form of ammonia-HgN,I, + 12KI + 12H,O = 9Rg1, + 4NH, + 12KOH. I n the colorimetric method only a portion of the ammonia contributes to the formation of the colour, and, in the author's opinion, it is probable that the intensity of the resction in the standard solution and in the unknown solution is liable to vary in different fashion under the influence of different factors that affect the equilibrium, such as heat, dilution, etc.C. A. M. The Differentiation of Natural and Artificial Mineral Waters. D. Negreano. (Comptes Rendzcs, 1906, cxliii., 257, 258.)-The electrical resistance of a natural mineral water is almost a physical constant for each kind of water, and serves to distinguish it from other waters. Thus the author obtained the following results in ohms-c.c. at 18" C. : Caciulata water (Roumania), 328 ; Slanic water, 48 ; Vichy water, 140; Vittel water, 500; and Gvian water, 1,280. I n determining the resistance at different temperatures, it was found that it diminished with the tempera- ture. When the interval between the temperatures was not too great the resistance R, at a temperature t could be calculated to 18" C. by means of the formula- where a represents the coefficient of variation, and is approximately 0.02. The values of a in the case of waters mentioned above ranged from 0.019 to 0.027. The resistance of an artificial water made up to imitate the natural mineral water differs materially from that of the latter at the same temperature. For instance, water from the Cblestine spring at Vichy showed a resistance of 140 ohms at 18' C., as against 112 ohms by an artificial Vichy water. The resistance of water of natural Avian water was 1,280 ohms, whilst that of an artificial water was only 1,120 ohms at 18" c. C. A. M. R,=RIS [I -a(t - l8)],