54 THE ANALYST. INORGANIC ANALYSIS. The Determination of Bismuth, and its Separation from Copper, Cadmium, Mercury, and Silver. A. Staehler and W. Scharfenberg. (Berichte, 1905, xxxviii., 3862-3869.)-The method described is based upon the quantitative precipitation of bismuth by sodium phosphate, even in the presenceTHE ANALYST. 55 of hydrochloric acid. From 10 to 20 C.C. of the solution of the bismuth salt (0.1 to 0.2 gram of bismuth) are diluted to 300 or 400 c.c., and heated to boiling, any precipitate of basic salt being redissolved by the addition of the smallest possible quantity of nitric acid. The boiling liquid is treated with a boiling 10 per cent. solution of trisodium phosphate, the quantity required depending upon the amount of free acid present. I n the presence of much hydrochloric acid a larger addition of sodium phosphate is necessary.If the solution has become alkaline, it must again be rendered acid by means of a little nitric acid. After boiling for some time the precipitate is allowed to subside, the supernatant liquid tested with sodium phos- phate, and if the precipitation be complete, the precipitate is collected while hot in a Gooch's crucible, washed with hot 1 per cent. nitric acid containing a trace of ammonium nitrate, dried at 120' C., ignited for five or ten minutes over a Bunsen- burner, and weighed as BiPO,. The tabular results quoted agree within fractions of a mgm. with the theoretical amounts. Separation of Bismuth and Copper.-The bismuth is precipitated as above described, whilst the copper is precipitated from the filtrate by mcans of hydrogen sulphide, and weighed as cuprous sulphide (Cu 3S).Electrolytic determination is unsatisfactory in the presence of hydrochloric acid. Separation of Bismuth and Cudmiurn.-After separation of the bismuth as phosphate, the cadmium can be determined by electrolysis of the ammoniacal filtrate after the addition of potassium cyanide, provided hydrochloric acid be absent. In the presence of hydrochloric acid it is best to precipitate the cadmium as sulphide, which is then dissolved in'dilute nitric acid, and the solution electrolyzed after the addition of potassium hydroxide and potassium cyanide. Separation of Bismuth and Mercwy.-The filtrate from the bismuth phosphate is treated with a large amount of hydrochloric acid, and then with ammonia in excess, any precipitate formed being redissolved in hydrochloric acid, and the solution again rendered alkaline with ammonia.The liquid is next heated to the boiling-point, and immediately treated with hydrogen sulphide. After five to ten minutes the mercury sulphide separates out uncontaminated with free sulphur. I t is collected on a Gooch's filter, washed with hot water, alcohol, and ether, dried at 105" C., and weighed. Separation of Bismuth and Silver.-The bismuth can obviously only be separated in the absence of hydrochloric acid. The silver in the filtrate from the phosphate can be determined by the usual methods. Separation of Lead a d Bismuth.-It is suggested that a method of separation may be based on the precipitation of the two metals as phosphates, and the removal of the lead by boiling the precipitate with nitric acid (1 per cent.).Lead is quantitatively precipitated from a neutral solution by means of sodium phosphate ; but the precipitate consists of at least three different phosphates, which are con- verted, however, by long-continued boiling into the tertiary phosphate. C. A. M. On the Precipitation of Metallic Gold. P. E. Jameson. (Jozmz. Amer. Chem. SOL, 1905, xxvii., 1444.)-By the following method metallic gold may be56 THE ANALYST. precipitated in one or two minutes in a form resembling silver chloride, A stick of potassium nitrite weighing about 5 grams is placed in a solution of 1 gram of gold chloride in 30 C.C. of water, and 5 C.C. of concentrated sulphuric acid are at once added. As soon as the brisk reaction which takes place is over, another piece of potassium nitrite of the same size a5 the first is added, and the solution stirred until the reaction is over.The gold will then be found at the bottom of the clear liquid in the form of dark brown nodules, which, after drying, are more yellow than ignited cadmium oxide. A. G. L. Estimation of Platinum. Nordenskjold. (Chem. Zed. Rep., 1905, xxix., 293.) --Platinum is usually estimated as the metal by precipitating with sodium formate or with zinc. The former method can be carried out by treating the platinum solution with 2 grams sodium formate and a little ammonium acetate, diluting to 200 c.c., warming until the evolution of carbon dioxide slackens, and then heating €or twelve hours.The platinum is apt to pass through the filter-paper, but this can be prevented by adding a little nitric acid. The second method gives a granular precipitate of platinum. Magnesium is preferable to zinc. About 17 cm. of magnesium ribbon should be used per gram of platinum, and dilute solution used. Excess of magnesium is removed by hydrochloric acid ; any magnesium oxychloride is at the same time dissolved out. The two methods give concordant results. For iridium, ruthenium, and rhodium a slight modification is necessary, as these metals are sensibly attacked by acids in the precipitated form; only sufficient excess of magnesium to form oxychloride is used, the residue is heated, whereby the metals mentioned are rendered insoluble in acids, and the oxy chloride of magnesium dissolved out with acid.To estimate platinum in a mixture of this group, the heated metal powder 3s digested with aqua regia, when platinum (and palladium) go into solution. E. K. H, Note on an Antimonial Pigment for Enamels. P. Rasenack. (Arbeit. Kaiserl. Gesundheitsamte, 1905, vol. 22, pp. 653, 654.)-The author finds that a substance sold under the name of '' Leukonin " consists almost entirely of sodium metantimoniate, and that it is soluble to a considerable extent in water and in solutions of sodium chloride, sodium carbonate, tartaric acid, citric acid, etc., antimony in every case going into solution. The author suggests that enamels made with this substance may yield antimony to weak acid liquids. w. P. s. The Use of Potassium Periodate in the Detection of Manganese, Cobalt, and Zinc.S. R. Benedict. (Amer. Chem. Journ., 1905, vol. 34, pp. 581-585.) -On adding a solution of potassium periodate to a solution of a manganous salt a deep-red precipitate (or coloration with very dilute solutions) is immediately pro- duced. This precipitate is soluble in hydrochloric acid, and is converted by sodium peroxide into a black hydrated manganese dioxide. Tv solution of the potassium salt is capable of detecting 1 part of manganous chloride in 10,000, the test being more sensitive in %t warm ammoniacal solution than in a neutral warm As a reagent,THE ANALYST. 57 solution. Solutions of pure zinc salts yield a white periodate, but the presence of ammonium chloride and ammonium hydroxide prevent the formation of this, unless there is a very large amount of zinc.On boiling the filtrate from the precipitated manganese the zinc is precipitated. Conversely, a solution of manganese chloride may be used for the detection of periodate in the presence of iodate, iodide, etc. Iodates give no coloration with manganese in the presence of ammonium chloride and hydroxide. Solutions of pure nickel salts react with potassium periodate to give a light- green precipitate, which becomes slightly grey on boiling, whilst cobalt salts yield a dark-brown precipitate, which becomes greenish on adding an excess of the reagent, and dissolves, forming a greenish-black solution on heating. If cobalt and nickel are present together, the precipitate will become more or less olive green on boiling, and the test is capable of detecting as little as 0.1 per cent.of the former. It is advisable, however, to make a comparative test with a solution of a pure nickel salt of the same strength. If the precipitate turns black, either in the cold or on boiling, traces of manganese are present. Zinc does not interfere with the reaction, C. A. M. Separation of Iron and Zinc by Means of Ammonia. W. Funk. (Zeits. nngeiu. Chem., 1905, xviii., 1687.)-In the separation of iron and zinc by means of ammonia, the author shows that good results can be obtained if a large excess of ammonia is used in the presence of much ammonium chloride, the precipitate being washed with a 5 per cent. solution of ammonium chloride, and the precipitation being carried out twice.Practically the whole of the zinc.is then found in the two filtrates. A. G. L. Estimation of Metallic Iron in Presence of Ferrous and Ferric Oxides. B. Neumann. (Chem. Zeit. Rep., 1905, xsix., 310.)-Metallic iron can only be directly estimated in presence of ferrous oxide when no ferric oxide is present, otherwise the latter is reduced to ferrous oxide, using up hydrogen. The author therefore proposes the following indirect method : the substance is very finely powdered, and if only the content of metal is required, this is simply extracted with a magnet. In other cases the finely-divided substance is warmed with a measured quantity of a standard copper sulphate solution, and the precipitated copper directly estimated or, better, calculated from a determination of the copper left in the solution.This gives the metallic iron. Next the hydrogen given off on treatment with dilute sulphuric acid is measured. This represents only part of the hydrogen, some being used in reducing ferric oxide to ferrous oxide. The solution is then titrated with standard permanganate, which gives the original ferrous oxide dis- solved and the reduced oxide. From these figures the original content of ferrous oxide can be calculated. The total iron is then estimated in the usual manner, and the ferric oxide obtained by difference. E. K. H. Determination of Zinc in Zinc Aluminium Alloys. R. Seligman and F. J. Willott. (Journ. Soc. Chem. Ind., 1905, xxiv., 1278.)-A rapid method for the approximate determination of zinc (within 0.2 per cent.of the gravimetric result)58 THE ANALYST. consists in dissolving the alloy in caustic soda, precipitating the zinc as sulphide, redissolving in hydrochloric acid, and titrating the zinc with potassium ferrocyanide. Half a gram of the drilled alloy is dissolved in a 400 C.C. beaker in 25 C.C. of a 25 per cent. caustic soda solution, the liquid warmed till rapid evolution of gas takes place, and allowed to stand until solutioniis complete-usually five to ten minutes. The solution is then diluted to 300 C.C. with boiling water, any undissolved iron, copper, lead, tin, or nickel allowed to settle, and the clear liquid poured off, the residue being washed twice by decantation. This residue generally contains a small amount of zinc, which is removed by dissolving in a few drops of concentrated hydrochloric acid, diluting to 20 c.c., neutralizing with caustic sods, and adding 2 C.C.of the latter solution in excess. After warming the precipitated hydroxides are filtered off and washed, and the filtrate added to the main bulk of the zinc and aluminium solution. The zinc is DOW precipitated by sulp huretted hydrogen, which is passed till a skin forms at the point where the bubbles of gas burst, the precipi- tate allowed to settle, the clear liquid decanted, and the precipitate thrown upon a filter, whence it is dissolved in 8 C.C. hydrochloric acid, washed into the beaker used for precipitation, and diluted to 250 C.C. with boiling water. The presence of a small amount of alumina in the zinc is harmless, so that it is unnecessary to wash the precipitated sulphide.Five grams of ammonium chloride are then added, and the zinc titrated with a solution containing 22 grams of potassium ferrocyanide per litre, using uranium nitrate or acetate or ammonium molybdate as indicator. For the spot-tests a s1a.b of paraffin wax or a porcelain plate covered with a thin coat of wax is recommended. W. H. S. The Determination of Barium as Chromate, and its Separation from Strontium and Calcium. A. Skrabal and L. Neustadtl. (Zeit. anal. Chem., 1905, vol. 44, pp. 742-?55.)-The authors from a study of the different methods find that good results are obtained by precipitating the barium by means of ammonium bichromate from a neutral or slightly acid solution in the presence of ammonium acetate.The precipitation may be made from a hot or cold solution, but the filtration and washing of the precipitate must be done in the cold. Barium can be separated from calcium with approximately correct results by means of a single precipitation with ammonium bichromate from a cold, very dilute acetic acid solution in the presence of ammonium acetate. Better results are obtained, however, by a double precipitation. I t is impossible to effect a complete separation of barium from strontium by means of a single precipitation with ammonium bichromate. The following combined method is recommended for the separation of barium from calcium and strontium: The neutral or slightly acid solution of the salts (about 0.135 gram of the respective oxides) is mixed with ammonium acetate in excess (10 C.C.of a solution containing 300 grams per litre), brought to the boiling-point, and treated with ammonium bichromate solution (5 C.C. of a solution containing 100 grams per litre), added drop by drop, with continual shaking. The precipitate is allowed to subside, and when cold washed by decantation with a dilute solution of ammonium acetate (20 C.C. of the above solution diluted to a litre) until the washings passing through the filter are colourless. The small amount of the precipitate thatTHE ANALYST. 59 has been retained by the filter is dissolved in warm dilute nitric acid, the solution and washings returned to the original beaker, and sufficient nitric acid added to dissolve the whole of the precipitate. The clear solution is now treated drop by drop with ammonium hydroxide until the first signs of a permanent turbidity appear, after which ammonium acetate is added (10 C.C.of the strong solution), and the liquid boiled and allowed to cool gradually. The precipitate is washed by decanta- tion and on the filter with cold dilute ammonium acetate solution, then dried, ignited in a platinum crucible, and weighed. Strontium and calcium are then deter- mined in the united filtrates in the usual manner, C. A. M. The Determination of Titanium in Soil and the Ash of Plants. H. Pellet and C. Fribourg. (Ann. de Chim. anal., 1905, vol.10, p. 413-416.)-I. Direct Colori- metric Method-Half a gram of the dry earth, or 2.5 grams of ash, are introduced into a platinum crucible containing 15 grams of pure hydrofluoric acid and 1 C.C.of sulphuric acid, and evaporated to dryness. The residue is powdered, mixed with 5 grams of potassium bisulphate and fused, and the resulting product taken up with water containing 15 C.C. of sulphuric acid per 100 c.c., at a temperature not exceeding 60' C. The liquid is made up to 100 C.C. (there should be no insoluble residue), 1 to 10 C.C. taken and made up to 100 C.C. with distilled water, treated with 5 C.C. of hydrogen peroxide (12 volume solution), and compared colorimetrically with solutions of pure titanic acid containing 0.1 to 1.0 gram per litre. 11. Gruvimetric Methods.-In the case of soils containing about 2 per cent. of titanic acid, 3 grams of the dried and finely-powdered sample are introduced little by little into a, platinum crucible containing 30 grams of hydrofluoric acid, the mixture evaporated to dryness on the water-bath, after the addition of 3 C.C.of sulphuric acid, and the residue powdered and fused with 15 grams of potassium bisulphate. After cooling, the mass is powdered and dissolved in 200 to 250 C.C. of water at a temperature of about 60" C., and the solution, when cold, made up to 300 C.C. and filtered; 250 C.C. of the liquid is next transferred to a 400 C.C. beaker, and 40 C.C. of the remainder titrated with a solution of potassium hydroxide, 10 C.C. of which neutralize 5 grams of potassium bisulphate. From the result can be calculated the amount of potassium hydroxide solution to be added to the 250 c.c., so as to leave 5 grams of potassium bisulphate still unneutralized.After this addition, the liquid is treated with 50 C.C. of a freshly- prepared solution of sulphurous acid (1.020 to 1.025 specific gravity) and boiled for two hours, in the course of which two fresh additions of 50 C.C. each of the sulphurous acid solution are made. I t is then filtered, and the precipitate washed with boiling water and ignited, the residue being nearly pure titanic acid, with a little phosphoric acid. It is fused with pure potassium carbonate, the mass taken up with boiling water, and the insoluble portion washed with a 2 per cent. solution o potassium carbonate. The titanic acid remains behind in the form of an insoluble titanate, and the traces that pass into solution can be determined by a blank experiment with pure titanic acid.This titanate is ignited, fused with 1 gram of potassium bisulphate, and treated as above described, the residue of titanic acid being ignited and weighed,60 THE ANALYST. I n the case of soils containing less than 1 per cent. of titanic acid, 2 portions of 5 grams each of the finely-powdered and dry sample are ignited in separate platinum crucibles to destroy organic matter, and then fused with a mixture of 10 grams of sodium carbonate and 10 grams of potassium carbonate. The products of the fusion are taken up with very dilute hydrochloric acid, so as to form one solution, the silica separated by evaporating the liquid to dryness and igniting the residue, this being again taken up with dilute hydrochloric acid and the liquid filtered. The residual silica is treated with a mixture of hydrofluoric acid and sulphuric acid, which leaves a sulphated residue, whilst the filtrate is treated with ammonia, and the resulting precipitate separated from the filter and dried (the filter-paper being ignited alone).The united residue from the silica and ammonium precipitate is then fused with 15 to 20 grams of potassium bisulphate, and the titanic acid determined as described. I n the case of the ash of plants containing about 0.2 per cent. of titanic acid, 50 grams are treated with dilute hydrochloric acid, the silica separated, and the titanium determined in the filtrate. C. A. M. Rapid Method for the Valuation of Fluor-spar. A. W. Gregory. (Chem. News, 1905, vol. 92., pp. 184-185).-The following method may be found useful for the determination of calcium fluoride in fluor-spar when carbonates and silica are present in the mineral, all the determinations being carried out on the sample previously dried at 120" C.The carbon dioxide is determined by heating 2 grams of the sample to a red heat, until no further loss in weight takes place. Two grams of the sample are treated in a platinum basin with pure hydrofluoric acid, evaporated, ignited, and weighed. The calcium carbonate is thus converted into fluoride, and the loss in weight obtained is due to the silica present after allowance has been made for the quantity of carbonate previously found. (When the silica is present partly as silicate, this determination is not strictly accurate.) Another quantity of 2 grams of the sample are now heated in a platinum basin with concentrated sulphuric acid, hydrofluoric acid being first added if the amount of silica present be large.The excess of sulphuric acid is then driven off and the residue ignited and weighed. In this operation there will be an increase in weight, due to the conversion of calcium carbonate and fluoride into sulphate, and a loss of any silica present. The amounts of carbonate and silica being known, the fluoride may be calculated. w. P. s. Electrolytic Calcium. Joseph H. Goodwin. (Journ. Amer. Chem. SOC., 1905, xxvii., 1403.)-Metallic calcium can be easily and quickly prepared from calcium chloride by using the apparatus described. It consists essentially of a hollow cylinder of Acheson graphite, which is used as anode, the bottom being formed of a cooling copper coil, insulated from the graphite by asbestos.The cathode is a $-inch iron rod mounted on a, screw arrangement, by which it can be raised or lowered. The calcium separated at first acts as cathode during the remainder of the operation, and, by gradually raising the iron cathode, sticks of metallic calcium are easily obtained. The reaction is started by filling the anode with cold calcium chloride, andTHE ANALYST. 61 producing an arc between the cathode and anode ; a8 the calcium chloride melts, the cathode is lowered into it until electrolysis takes place, the heat generated being sufficient to keep the salt fused. The temperature must be kept within rather narrow limits, otherwise the calcium separates in a spongy instead of a coherent form.As, with the heavy currents used, the asbestos does not completely insulate the copper coil from the anode, a cell is inserted between the copper and the graphite so as to cause a, small current (0.04 amp8re) to flow in the other direction; this device effectually prevents contamination of the calcium by copper. In six experi- ments, with voltage varying from 14 to 22, and amperage from 105 to 185, a current efficiency of 21.5 to 41.9 per cent. was obtained, the average figures being: Volts, 17.7 ; ampAres, 163.0 ; efficiency, 26.6 per cent. I n six hours about 150 grams of calcium can be obtained. In one experiment the separated metal had the following composition: Ca, 98.00; Si, 0.03; Fe, 0.02; AI, 0.03; Mg, 0.11 ; C1, 0.90; and 0, by difference, 0.91 per cent. This sample had a conductivity of 3.43 microhms per C.C.at 0" C., the temperature coefficient being 0.00457, and a tensile strength of 8,710 lbs. per square inch, the elongation being 23.0 per cent. for 1 cm., 15 per cent. for 2 crn., 11.0 per cent. for 3 cm., and 6-6 per cent. for 5 cm. Calcium was found to be harder than sodium, lead, or tin, almost as hard as aluminium, but softer than zinc, cadmium, or magnesium. I t s specific gravity was found to be 1.5446 at 29.2" C. To clean the metal, after breaking off most of the calcium chloride, it is placed in 95 per cent, alcohol for some time, the loss owing to reaction with water and alcohol not being very great. The solid metal can be heated to a red-heat in air without igniting. I t is not hardened by quenching in water. At 300" to 400" C.it is as soft as lead. When cold, the bright surface of the metal rapidly dulls in air, but when hot its surface may be polished, and will keep bright as long as the metal remains hot. By placing the hot metal in a bottle heated to 150" C., and closing the bottle while at this temperature, the surface may be preserved quite bright. A number of the cylinders obtained were fused together in a closed iron tube and quickly cooled, when a mass of large reddish-violet cubical cystals was obtained, having the following composition : Ca, 91.28 ; gangue, 0.03 ; SiO,, 0.77 ; FelO,, 046; A1,0,, 0.77; &$g, 0.11; C1, 1.28; C, trace; N, trace; 0, by difference, 5.30. The specific gravity of the crystals was 1.5425 at 28.1" C ; they were quite soft, and could be hammered into sheets as thin as paper, often exploding with a slight flame under the impact of the hammer.A. G. L. Estimation of Percentage of Nitric Aeid in the Concentrated Acid by the Specific Gravity. Veley and Manley. (Chem. Zed. Rep., 1905, xxix., 1207.)-The authors publish some figures and details of their methods of purification, etc., in consequence of the divergence shown between the results published by Lunge and by Winteler. The authors come to the conclusion that Winteler's acid was not sufficiently pure. They purified their acid as follows : The purest commercial acid was distilled, and the distillate was redistilled over silver and barium nitrates to remove sul- phuric and hydrochloric acid. A current of ozonixed oxygen was then passed62 THE ANALYST.through, and finally the acid was fractionated in vacuo at the lowest possible tempera- ture. In this way an acid was obtained containing 99.8 per cent. ENO, with only 1 part per million of nitrous acid, 4 per million of sulphuric acid, and 3 per million of halogen acids. This acid was further freed from water by a special apparatus (Phil. Trans., 1890, A. 365)) and an acid finally obtained which contained 99.97 per cent. HNO,. The authors used for titrating the various dilutions of this acid a sodium hydroxide solution prepared from sodium and steam, and a N-sulphuric acid solution standardized against sodium carbonate. The following table shows the figures obtained by various observers : Specific gravity: 1,485 1.490 1.495 1.500 1.505 1510 1.615 1.520 Percentage according to- 1.Lunge and Rey ... ... 87.7 89.6 91'6 94.1 96'4 98.1 99'1 99.7 2. Vcley and Manley ... ... 87% 89.6 91.1 94'0 96.5 97'8 98 7 99'8 4. Mean of fisst three series ... 87.6 89.6 91.4 93.9 96'4 97.9 98.9 99.8 5. Winteler ... ... ... 86'3 88.2 90.2 91'8 93.6 95.6 97'3 99.7 8. Ferguson .. ... ... 87.6 89.7 91.4 93'7 - -. - I Difl'erence between 4 and 5 ...- 1'3 -1.4 -1.2 -2.1 -2.8 -2.3 -1'6 -- For further comments on these figures, reference should be made to the original ; but the authors express the view that the clearing up of the question lies with Winteler rather than with the other observers. E. K. H. Estimation of Percentage of Nitric Acid in the Concentrated Acid by the Specific Gravity.H. Putzer. (Chem. h i t . Rep., 1905, xxix.)-Compare paper by Veley and Manley, (see previous abstract). He adds his contribution to the figures already published. Great care was bestowed by the author on the purifica- tion of the acid. One hundred and fifty grams of this on evaporation in a platinum dish left a scarcely visible residue which was not weighable; the acid was quite free from nitrous acid (when diluted with water (5 : 1) and 1 drop of TG potassium per- manganate added it remained pink for an hour), sulphuric acid, hydrochloric acid, and iodine. The author having found the ordinary method of obtaining sodium carbonate from so-called chemically pure sodium bicarbonate unsatisfactory, he therefore pre- pared Na,CO,.H,O by boiling a solution of the bicarbonate and re-precipitating.This was then heated in a platinum dish by a burner whose flame-tip was about 5 cm. below the bottom of the dish. Na2C0, was used, with methyl orange, very dilute, as indicator. The specific gravity was determined by a hydrometer showing the first three decimal places exactly, and the fourth by estimation. The author's figures are : Specific Gravity $ (Vac.) Yc. HNO,. I Specific Gravity % (Vac.) Pc. HNO,. 1.4878 1.4908 1.4923 1 *4943 1.4958 1.4973 1.4988 1.4998 89-19 90.40 90.98 91.79 92.52 93.22 93.93 94.53 1.5018 1 -5033 1.5043 1.5058 1.5073 15103 1.5128 1.5163 95.51 96-13 96.59 97.21 97-76 98.62 99.20 99.75THE ANALYST. 63 By graphic interpolation these figures give results differing uniformly by 0.5 from those of Lunge and Rey.The author considers that these differences are due to the fact that his acid was even purer than that used by Lunge and Rey, whilst Winteler’s results may be rejected, as the acid used was, on his own showing, not pure. E. K. H. On the Determination of Sulphuric and Nitric Acids in 64Nitrous Vitriol.” G. Lunge and E. Berl. (Zeits. angew. Chem., 1905, xviii., 1681.)-As the result of determinations made, partly on a synthetic, partly on a commercial acid, the authors recommend that nitro-sulphuric acid should be examined for (a) total acidity, ( b ) permanganate consumed, (c) total nitrogen by means of nitro- meter, The nitric acid may then be calculated by subtracting (b) from (c), and the sulphuric acid by subtracting (c) from (a). Gravimetric determinations of sulphuric acid as barium sulphate tend to give high results; determinations of the nitric acid by the ‘ 4 nitron ” method easily lead to low values. ’ Evaporation of the nitric acid and titration of the residual sulphuric acid generally gives results slightly too low for the sulphuric acid, and consequently a little too high for the nitric acid. A. G. L. Modification of W. Winkler’s Method for the Determination of Oxygen in Waters. (Zeits. angew. Chem., 1905, xviii., 1767.)-The author believes that with ordinary waters Winkler’s method gives more exact results if the correction made by Winkler for the iodine consumed by the organic matter present is not made. When a, correction must be made, as in the case of sewage effluents, he proceeds as follows : A mixture of 2 C.C. of 40 per cent. sodium hydroxide solution, 2 C.C. of 50 per cent. manganous chloride solution, and 20 C.C. of distilled water is shaken in a large flask until it becomes brown in colour; 50 c.c of concentrated hydrochloric acid are then added, and the whole made up to 300 C.C. with distilled water. Of this solution, 20 C.C. are added to 100 C.C. of the water under examina- tion, and also to 100 C.C. of distilled water. After five minutes 10 C.C. of a 5 per cent. potassium iodide solution are added to each, and the iodine liberated is deter- mined with a solution of sodium thiosulphate, 1 C.C. of which corresponds to 0.1 C.C. of oxygen. The difference between the two values found gives the quantity of iodine taken up by the organic matter in the water, and may be applied as a correc- tion to the results found by Winkler’s method. Results obtained in this way on several waters agreed better with gasometric determinations of the oxygen in the waters than the values found using Winkler’s method of correction. Hermann Noll. A. G. L.