ANALYTfCAL CHEMISTRY. 55 An alytical Chemistry. Modified Gas-burettes. JULIUS ZHZAWY (Chevz. Zeit. 1904 28 1172).-Hempel burettes so constructed that it is possible to read the volunie of gas to 0.01 C.C. L. DE K. A New Burette for Testing Normal Solutions. A. HESSE (Clmn. Zeit. 1904 28 1172).-The burette is fitted with a beaker- shaped cup. The first 0.5 C.C. and the last 43-51 C.C. are divided to l j l 0 0 c.c. the tube being drawn out a t these places; the remainder is dividled as usual to 1/10 C.C. The parts drawn out are not calibrated It is mereIy intended for checking normal solutions. L. DE K. Determination of the Neutralisation Point by Conductivity Measurement. 11. FRIEDRICH W. KUSTER MAS GRUTERS and W. GEIBEL (Zeit. u n o ~ y Cheiiz. 1904 42 225-231. Compare Abstr.1903 ii 611).-Ten C.C. of X / l O sulphuric acid were diluted with 500 C.C. of water and then N/10 sodium hydroxide gradually added the conductivity of the solution being measured after the addition of each C.C. of alkali. The conductivity of the sulphuric acid falls regu- larly as the neutralisation proceeds until a minimum value is attained exactly at the neutralisation point ; an increase in the value occurs when more sodium hydroxide is added. By this method the end-point may be determined much more sharply than by the use of any indicator. Phosphoric acid behaves as a inonobasic acid; the break of the con- ductivity curve coincides with the results obtained by titrating phos- phoric acid with methyl-orange as indicator.56 ABSTRACTS OF CHEMICA4L PAPERS. I n order t o show how the method may be applied when indicators fail 20 C.C. of N/lO potassium dichromate were mixed with 10 C.C.N/lO sulphuric acid and then titrated with N/10 sodium hydroxide by conductivity measurement as before ; the sulphuric acid present can be titrated as sharply as if the dichromate were absent. Similar measurements were made with potassium permanganate in the s o h tion instead of potassium dichromate. The presence of precipitates also does not impair the accuracy of the method thus citric acid may be titrated by barium hydroxide. When weak organic acids were titrated by the gradual addition of sodium hydroxide the minimum of conductivity was attained too soon ; the method yields correct results however when the acid is added t o the alkali. Acetic lactic tartaric and citric acids were titrated in this manner.Determinations of the amount of acetic acid in a strongly coloured vinegar examined are quoted. The total acid in a specimen of red wine was also determined. Magnesia and various alkaloids were similarly titrated. A. McK. The Theory of Indicators and its Bearing on the Analysis of Physiological Solutions by means of Volumetric Methods. GEORGE H. A. CLOWES (Amer. J. Pharm. 1904 ’76 453-467 611-525).-This paper has for its object the outlining of the nature of the various classes of indicators used in acidimetry and alkalimetry studying their behaviour towards various weak bases and acids which may occur in the course of physiological work; further to make a series of tests on artificial and normal stomach contents pancreatic fluids blood serum &c.With regard to the behaviour of amino-acids asparagine glycine leucine tyrosine and aspartic acid towards indica- tors experiments were made in which their action was compared with that of the ammonium salts of formic acetic and lactic acids acet- amide formamide and urea Poirrier’s blue was found to be sensitive to all the free acid groups and entirely indifferent to ammonia and amino-groups. Phenolphthalein is less sensitive to acid groups and far more sensitive to ammonia than Poirrier’s blue. Alizarin is sensi- tive to both ammonia and strong acid groups b u t is indifferent to the weak acid groups in asparagine tyrosine &c. and also to the amino- group of these compounds. Phloroglucinvanillin is entirely indifferent to all weak organic acids This indicator is of great value in differentiating the weak bases being sensitive not only to ammonia but also to the amino-group in aspara- gine glycine leucine and tyrosine and indifferent to the amino-groups present in acetamide formamide and urea.From the experiments with proteids &c it seems possible that some sort of relationship obtains between the various nitrogen-containing groups in the pro- teid and the end-points of different indicators. The comparison of volumetric with gravimetric analyses of normal and pathological stomach contents showed that provided a purely proteid test meal was employed the phenolphthalein end-point indicates fairly accu- rately the total available hydrochloric acid present both free andANALYTICAL CHEMISTRY.57 combined more or less feebly with proteids. The actual free acid is given by phloroglucinvanillin. After allowing for the influence of phosphates and traces of organic acids the range of titration between the end-point of phloroglucinvanillin and that of phenolphthalein may be sa.id to be a function of the basic affinities of the proteids and their decomposition products present. Under conditions where peptic diges- tion fails to take its normal course the above rules are no longer applicable. Jn tryptic digestion the range of titration after allowing for the influence of carbon dioxide is a t first more or less proportionate to the amount of proteid brought into solution but subsequently increases in undue proportion t o the total nitrogen as the digestion proceeds.The wide variation between the end-points of phenol- phthalein alizarin and phloroglucinvanillin in estimating the alkali- nity of blood serum emphasises the necessity of certain uniform stand- ards for this purpose. With regard to urine the titration results are fairly dependent on the amount of phosphates present. I n the presence of any considerable quantity of ammonia which exerts a more powerful effect on alizarin and phenolphthalein than urea the urine will be less acid or even alkaline to phenolphthalein and more strongly alkaline to alizarin than is normally the case. w. P. s. Standardisation of Normal Solutions. MAX SCHOLTZ (Arch. Plmrm. 1904 242 575-578).-0f an approximately 0.01N sodium hydroxide solution 10 C.C.required the following volumes? I of an approximately 0.01 N hydrochloric acid ; and of the same acid 10 C.C. required the following amounts 11 of the same alkali ; according to the indicator employed Phenol- Rosolic p-Nitro- Methyl- ph thalein. Litmus. asid. phenol. Iodoeosin. orange. I. 9.68 9.83 9.87 10.00 9.80 10.70 C.C. 11. 10.08 9 92 9.92 9-99 9.68 9.28 When standard solutions as weak as these are being used it is evidently important that the indicator and the direction of the change of colour in the standardisation should be the same as are employed in the act,ual estimation. The results are discussed from the point of view of the theory of electrolytic dissociation (compare Abstr. 1904 ii 771). C. F. B. Potamium Bromide as Indicator when using Fehling's Solution.PIO BERTI (CYhern. Centr. 1904 ii 1433 ; from Bull. Assoc. Cl~ivz. Xucr. et Dist. 21 1234-1236).-A drop of the liquid is put on to a double piece of filter paper resting on a porcelain plate and after removing the paper the moist spot is touched with a particle of potass- ium bromide and a drop of sulphuric acid. I f there should still be a trace of dissolved copper a violet colour is noticed owing to the forma- tion of cuprous bromide L. DE K. Apparatus for Extraating Liquids with Chloroform. ERICR BAUM (Chem Zed. 1904 28 1172).--A modification of the van Rijn58 ABSTRACTS OF CHEMICAL PAPERS. apparatus (Abstr. 1896 ii 17) which consists in fixing the overflow tube (siphon) in the lower part of the extractor thus allowing the use of solvents heavier than water.L. DE K. New Extractor. ANDREA SANNA (Gazxelta 1904,34 ii 224-228). -This extractor consists of a flask fixed obliquely and filled about half-full with the substance to be extracted and the solvent. Through the double-bored stopper of the flask pass the bent inner-tube of an in- verted condenser and a short glass tube furnished with a cock and an india-rubber tube The upper end of the condenser tube is also bent and passes into an Xrlenmeyer flask through the stopper of which passes also a short glass tube with a cock. The latter is left open during the heating of the substance with the solvent. When the extraction is complete the condenser and Erlenmeyer flask are turned round so that the end of the bent condenser tube is now a t the lowest part of the round flask whilst the Erlenmeyer flask is a t a still lower level The cock attached to the round flask is then opened and the solution forced through the condenser tube into the Erlenmeyer flask. The apparatus is then inclined so that the round flask is lower than the Erlenmeyer and the solvent distilled off from the latter into the round flask tho cock attached to the Erlenmeyer flask being meanwhile kept closed.The extracted material then remains in the Erlenmeyer flask. These operations may be repeated until the extraction is complete The apparatus gives good results for the extraction of liquids or solids as for example in the determimtion of fat in cheese butter &c. and it acts more quickly than a Soxhlet or Garralowschy extractor.I n determining f a t in a material like cheese it is advisable to pack the substance in filter paper tied up with cotton as this prevents solid particles from being carried away. T. H P. Palladium-Hydrogen as a Reducing Agent in Quantitative Analysis. ALFRED c. CHAPMAN (Ancilyst 1'304,29 346-357).-1n an investigation undertaken for the purpose of ascertaining t o what extent palladium-hydrogen could be employed as a quantitative reducing agent it was found that ferric salts as well as potassium ferricyanicle were completely and readily reduced. Chromates in acid solutions were reduced to chromic salts. The palladium was employed in the form of foil of such thickness that a piece 2 inches square weighed about 8 grams. A piece of thick palladiuin wire was welded to this and the foil was charged by being made the negative electrode in an ordinary cell containing dilute snlphuric acid the current being obtained from the main or from accumu- lators.The charged metal was after washing suspended in the boiling solution under examination. In the case of cupric stannic arsenic manganic vanadic and moly bdic compounds the reduction was either incomplete or imperfect Potassium chlorate was only partly reduced whilst bromates and iodates mere not attacked. C'eric salts were reduced to cerous salts which latter could be titrated with ferrous ammonium sulphate solution as the lower oxide was itself capable of oxidising the ferrous solution. w. P. s,ANALYTICAL CHEMISTRY 59 Action of Hydrochloric Acid on Potassium Chlorate.A. KOLB and E. DAVIDSON (Zeit. nypu. Chena. 1904 17 1883-1887).- During the action of hydrochloric acid on potassium chlorate in the presence of potassium iodide atmospheric oxygen and especially the oxygen in solution causes oxidation of the hydriodic acid formed and the iodine thus liberated accelerates the action. The error thus caused in the estimation of chlorate may be obviated by conducting the operation in the absence of oxygen The reaction between hydro- chloric acid and potassium chlorate proceeds at the ordinary tempera- ture only in the presence of a considerable excess of hydrochloric acid and is quickly completed; chlorates may be estimated in this way. The substitution of cadmium iodide for potassium iodide has no special advantage whilst aluminium iodide and mercuric iodide are unsuitable.An acceleration of the reaction was not observed when antimony chloride or cerium sulphate was present,. The action is retarded by the addition of water. A. McK. Estimation of Iodine in Soluble Iodides also in the presence of Bromides and Chlorides. HUGO DITZ and BENJAMIN 1%. l)~ARaoSCHES (Clem. Zed. 1902 28 1191-1 194).-The solut3ion con- taining the iodide is mixed with a sufficiency of potassium iodate and then with a slight excess of dilute sulphuric acid. The liberated iodine is extracted with toluene and finally titrated with sodium thio- sulphate. Jf a known amount of iodate has been added the excess of this may be estimated in the aqueous liquid by adding potassium iodide and dilute sulphuric acid and titrating with sodium thiosulphate. The shaking with toluene may even be omitted and the iodine expelled by distillation when the distillate which is collected in a solution of potassium iodide may be a t once titrated; as however this renders the process somewhat tedious it is more convenient t'o boil off the iodine and titrate the residual iodate.If bromides or chlorides are present the same process applies if the iodine is estimated in its toluene solution and not indirectly in the aqueous liquid. If the excess of iodate is t o be estimated in the aqueous liqnid after shaking with toluene acetic acid should be substituted for sulphuric acid. 1,. DE K. Estimation of Sulphur by Aid of Sodium Peroxide. ALBERT NEUMANN and JOSEPH MEINERTZ (Zeit. ylrysiol. Chenz. 1904 43 37-40 Compare During Abstz.1 S97 ii 63).-The substance is fused in a nickel crucible with a mixture of sodium and potassium carbonates and sodium peroxide. The peroxide should be added gradually in 3 or 4 distinct portions. It is found that a gas flame may be employed without causing error. J. J. S. Estimation of Sulphates in Vegetable Products. GEORGE S. FRAPS (Chem. Centr. 1904 ii 1433 ; from IZep. Chenaist North Carolina Agric. Experim. Stat. 1902 1903).-Five grams of the substance are treated for half an hour with 50 C.C. of 1 per cent. hydrochloric acid and the residue is washed with the same acid until the filtrate measures60 ABSTRACTS OF CHEMICAL PAPERS. about 250 C.C. barium chloride as usual. This is then heated to boiling and precipitated with A number of result,s are communicated. L.DE K. Apparatus for the Estimation of Nitrogen. EMILE NICOLAS and DELAUD (Bull. SOC. chim. 1904 [iii] 31 1193--1194).-This apparatus is a slightly modified form of that described by Porcher and Brisac (Abstr. 1903 ii 179). T. A. H. Sodium Peroxide i.n Organic Analysis. FRITZ VON KONEK and ARTHUR ZOHLS (Zeit. cmgew. Chem. 1904 17 lSS7. Compare Abstr. 1904 ii 775).-As much as 90 per cent. of the nitrogen in organic substances which are difficult t o burn and which are rich in nitrogen may be converted into nitrate by means of sodium peroxide. A. McK. Rapid Estimation of Nitrogen in Steel or Iron. HJALMAR BRAUNE (C'hem. Centr. 1904 ii 1167 ; from Oestew. Zed. Befag. Hutt. 52 491).-One gram of the sample is dissolved in 10 C.C.of ammonia- free hydrochloric acid of sp. gr. 1,124 and the filtered solution intro- duced into a distilling flask containing 250 C.C. of boiling water and 20 C.C. of an alkali hydroxide solution sufficiently strong t o neutralise tho acid. The distillate is then '' Nesslerised " in the usual manner. L. DE K. Estimation of Ammonia and Amides. JEAN EFFRONT (Bey.. 1904 37 4290-4295).-The reaction between alkali hypochlorites and nitrogen compounds may be more accurately followed by determin- ing the loss of active chlorine than by measuring the nitrogen evolved. Amines imines nitriles amides and amino-acids react with sodium hypochlorite the active chlorine disappearing being proportional to the weight of organic compound added ; tetra-alkylamnionium com- pounds and betaine do not react.A detailed method of carrying out the determination is desct ibed. Ammonia is completely oxidised by sodium hppochlorite at the ordinary temperature (compare Thiele Abstr. 1893 ii 317) the nitro- gen evolved containing a trace of nitrogen chloride. The ammonia in samples of water may be determined by this method to 0.2 milligram per litre. Proteids may also be determined in water after expelling free ammonia by boiling with sodium carbonate. Proteids show a con- stant reducing power towards hypochlorites peptones also show a con- stant but lower reducing power. C. H. D. Estimation of Ammonia in Vegetable Products Beets &c. EUGENE SELLIER (Clzena. Centr. 1904 ii 1433-1 434 ; from Bull Assoc. Chim. SUCT. Dist. 21 1223-1 232).-Ammonia may be completely expelled from beetroot juice by the author's magnesia process (Abstr.1903 ii 329). I n the presence of oxamide the distillation with magnesia should take place below 40'. Liquids containing urea should be distilled below 80'. Acid ammonium uratsANALYTICAL CHEMISTRY. 61 should be treated for a short time with dilute hydrochloric acid and then distilled with magnesia a t 36-42'. As hydrochloric acid how- ever may hydrolyse urea the author recommends distilling urine or sediments containing acid urates with calcium oxide a t a moderate temperature. Methylamine if present partially passes over with the ammonia. If the products should happen to contain magnesium ammonium phosphate they should be digested for some time with dilute acids ; the ammonia will then be completely expelled by boiling with magnesia but not at a lower temperature.L. DE K. Electrolytic Estimation of Nitric Acid with a Rotating Anode. LESLIE HOWARD J N G m w (J. Ame.r. Chem. sot. 1904 26 1251-1255).-The author has tried a process suggested in 1890 by Vortmann (Abstr. 1890 1467) for the electroljtic estimation of nitric acid and obtained satisfactory results. The solution of the nitrate is mixed with a known volume of standard sulphuric acid a definite amount of copper sulphate is added and the mixture sub- mitted to electrolysis. The ammonia formed by the reduction of the nitrate neutralises a portion of the acid and from the amount of free acid left (which includes that of the copper snlyhate) the amount of nitric acid is readily calculated.The best conditions for a successful experiment were found to be 25 C.C. of copper sulphate solution ( = 0.2533 gram of copper) 0.5 gram of potassium nitrate 25 C.C. of normal sulphuric acid; 9 volts 4 amperes time 30 minutes speed of rotating anode I I? = about 230 revoluticns per minute]. L. DE K. The Evolution of Hydrogen for Marsh's Arsenic Test. CARL RIA1 and H. HURT (Zeit. anal. Chenz. 1904 43 557-559).-To accelerate the evolution of hydrogen from zinc and sulphuric acid it is usual to add a small quantity of a salt of some electro-negative metal. Many of these howevei. diminish the sensitiveness of the test copper and platinum seeming to be the only unexceptionable agents. With the former the evolution of gas is more regular than with the latter.. Iron in any form must be moat carefully avoided since in its presence comiderable quantities of arsenic may be over- looked M.J. 8. Detection of Arsenic in the Ashes of Cremated Bodies. CARL MAI (Zeit. ctiaccl. Chew. 1904 43 617-619).-0n the occasion of the examination of the ashes of a cremated body the larger bone fragments after careful cleaning were found to be absolutely free from arsenic whilst the mixed pulverulent portion of the ash in which iron was also present was highly arsenical. Some iron nails and screws found in the urn also contained much arsenic and i t is assumed that the presence of that element in the illiscellaneous powder was entirely due to the metals employed in the construction of the coffin. From a forensic point of view it is advisable that neither metal nor pigments should be used in cofTins intended for cremations. M.J. S.62 ABSTRACTS OF CHEMICAL PAPERS. Improvement of Drown and Shimer’s Method of Estim- ating Silicon in Irons. J. THrLL (Zeit. anal. Chem. 1904 43 552-553).-Tf it is attempted to accelerate the tedious evaporation of the iron solution by substituting the direct heat of a lamp for that of the water-bath there is danger of loss from bumping and spirtiog. The addition of ammonium chloride obviates this difficulty. The author’s practice is as follows a litre of concentrated sulphuric acid is mixed with an equal volume of water and (after cooliog) a litre of nitric acid of sp. gr. 1.4 and a solution of 240 grams of ammonium chloride in a litre of water are added. Of this reagent 50-70 C.C.are used for dissolving 1-2 grams of iron. Solution and evaporation can then be carried on in a beaker over a Bnnsen burner with interposed wire gauze and an estimation completed within an hour. M. J. a. Qualitative Detection of Silica. JULIUS PETERSEX (Zeit. GL~CGZ. &em. 1904 43 619-623).-A modification of the method suggested by Daniel (Abstr. 1904 ii 289) permits the detection of 2-5 mg. of silica with certainty. I n a small platinum crucible is placed a mixture of the substance (0.5 gram) with cryolite (0.25 gram) R few milligrams of magnesite and some concentrated sulphuric acid. The crucible is lowered t o the bottom of a test-tube t o which is then fitted a cork carrying a glass rod nearly reaching the crucible. The rod is flattened a t its lower end which is covered with asphalt varnish and from it a drop of water hangs.The whole of the glass rod and the interior of the test-tube have first been coated with collodion contain- ing 2 vols. per cent of castor oil. This film protects the glass com- pletely from the action of hydrofluoric acid. The bottom of t.he tube is then plunged into water of 60’. A t this temperature the formation of silicon fluoride is sufficiently rapid to give a white film on the sur- face of the drop of water in five minutes. M. J. S. Estimation and Separation of Calcium in presence of Phos- phoric Acid. K. K. JXRVINEN (Zeit. cciaccl. Chern. 1904 43 559-562). -The estimation of calcium in presence of phosphoric acid is liable to two sources of error incomplete precipitation of the lime and con- tamination of the precipitate with phosphoric acid.The following inode of operating gives fairly satisfactory results the solution which must be as free as possible from ammonium salts is mixed with ammonia until calcium phosphate j u s t begins to precipitate and the precipitate is redissolved by a drop of hydrochloric acid. The liquid is heated to boiling and poured slowly into a mixture of equivalent quantities of ammonium oxalate and oxalic acid. Ammonia not stronger than 1 per cent. is then added drop by drop until the mixture is alkaline and the estimation is finished in the usual manner. M. J. S. Estimation of Calcium Oxide in Burnt Lime. Solubility of Calcium Carbonate in Solutions of Ammonium Nitrate. GEORG BERJU and WLADTSTAUR KOSINEXKC) (Lcmclzo.Trers?tchs-Strtt 1904 60 41 9-425).-Bodenbender and Ihleo’s method for determin- ing calcium oxide in burnt lime (Zeit. Biibenxuckerind 1879) by boilingANALYTICAL CHEMISTRY. 6.3 with ammonium nitrate can give correct results only under certain conditions owing t o the varying amounts of calcium carbonate dis- solved. It is found that calcium carbonate is practically insoluble in N/5 ammonium nitrate when the mixture of varbonate and oxide contains 8 per cent. or more of the latter. According to the amount of carbon- ate present (determined with a Pcheibler apparatus) 3-5 grams of substance are treated with 1 litre of N / 5 ammonium nitrate in a rotatory apparatus turning about 40 times a minute. The calcium oxide is afterwards determined in the usual manner in a n aliquot portion of the filtered or subsided solution.N. H. J. 11. Use of a Rotating Anode in tbe Electrolytic E s t i m a t i o n of Zinc. LESLIE 13. IKGHAX (J. Amer. C'hem. h'oc. 1904 26 1269-1283). -A lengthy investigation as t o the best and quickest method of electrolytic zinc assay. The ore is oxidisecl with nitric acid and repeatedly evaporated with hydrochloric acid and the iron is removed by a doiible precipitation with ammonia. The ammoniacal filtrate after adding some more ammonium chloride is then submitted to electrolysis using a rotating anode ; for working details the original paper should be consulted. Good results are also obtained by t,he electro- lysis of solutions containing sodium acetate or formate. L. DE I<.Volumetric E s t i m a t i o n of Lead. GUIDO C'ERVI (C'r'tem. Caut~. 1904 ii 1343-1344 ; from L71ndustric~ C'hemicn 6 889-290).-A slight modification of the process recoinmended by Guess ( l'~ni2s. AULW. I y ~ t . Zin. Ethg. i 904). The nentralised solution of the lead is acidified with acetic acid sodium acetate is added the solution is heated to boiling and the lead precipitated with standard solution of potassium dichromate. When cold the liquid is diluted to a definite volume and in a n aliquot part of the filtrate the excess of dichromste is estimated as usual by adding dilute sulphuric acid and potassium iodide and titrat- ing the liberated iodine with standard sodium thiosulphate. I n the presence of antimony or iron the lend should be first separated as sulphate which is then dissolved and titmted.L. DE E(. [Analysis of] Commercial Lead. AUGUSTE HOLLARD and L. BERTIAUX (BUZZ. Soc. c l ~ i ) ~ ~ 1904 [ iii] 31 11 24-1 128. Compare Abstr. 1899 ii 523).-Five grams of commercial lead are dissolved in 86 C.C. of nitric acid of 36" B. previously diluted with water. The liquid is made up t o 300 C.C. and the lead removed as the peroxide using a current of 0.8 ampere (compare Abstr. 1903 ii 294). Five C.C. of sulphuric acid are then added to the lead-free liquid and this is evaporated down until sulphuric acid funies are copiously evolved ; it is then diluted to 300 C.C. and the copper determinetl electrolytically the precautions already indicated (Abstr. 1904 ii 6S2) being taken to obtain a copper deposit free from arsenic antimony or bismuth.Nickel and zinc are separately estimated in the ix-sidual liquid by the process previously described (Abstr. 1903 ii 335 ; 1904 ii 92 and 682) and iron by titration with iodine (Abstr 1904 ii 368 and 592).e4 ABSTRACTS OF CHEMICAL PAPERS. For the estimation of arsenic and antimony a fresh portion of lead is treated with sulphuric acid and the arsenic separated by distillation after the addition of ferrous sulphate and hydrochloric acid (Abstr. 1900 ii 438). The residual liquid is placed in a flask having the bottom coated outside with fireclay and to it is added zinc chloride solution of sp. gr. 2 ; the antimony chloride is distilled off in a current of hydrogen chloride and finally estimated electrolytically. From a hot nitric acid solution of ths lead silver is precipitated as the chloride the latter is dissolved in potassium cyanide solution and the metal determined electrolytically.Sulphur is precipitated as barium sulphate from a nitro-hydrochloric acid solution of the original material Bismuth is separated as the basic nitrate which is then converted into sulphide and freed from arsenic antimony and tin by means of ammonium hydrosulphide and is finally estimated by elect yo- lysis of a solution of the sulphate (Abstr. 1904 ii 684). The tin obtained as the sulphide in the separation of the bismuth is freed from antimony by electrolysis in presence of potassium cyanide and finally detmmined electrolytically in hydrochloric acid solution in presence of ammonium oxslate. A tabular statement showing.the comDosition of v thirteen samples of commercial lead is given in the original. T. A. H. Electrolytic Separations possible with a Rotating Anode. DONALD S. ASHBROOK (J; Amer. Chem. Soc. 1904 26 1283-1290).- A paper unsuitable for adequate abstraction. Working details are given as to the electrolytic separation of copper from aluminium anti- mony arsenic cadmium chromium cobalt iron lead magnesium manganese nickel uranium and zinc ; cadmium from aluminium chromium cobalt iron magnesium manganese nickel and zinc ; silver from aluminium. L. DE K. Electrolytic Estimation of Copper. AUGUST KUFFERATH (Zeit. angew. Chern. 1904 17 1785-1786).-About 1.5 grams of the sub- stance containing copper are dissolved in dilute sulphuric acid with the aid of nitric acid ; if an excess of the latter is carefully avoided there is no need for evaporating to dryness.After diluting to about 130 c.c. 2 grams of formaldehyde are added and the copper is deposited as usual by heating at 60-65" and using a current of 2.5-3 amperes and 2 volts. Without interrupting the current the deposit is washed first with water then with alcohol and finally with ether and is then dried in a vacuum over sulphuric acid and weighed. Iron nickel and cobalt are not deposited in the presence of formaldehyde. L. DE K. Titratiorr of Copper by Potassium Iodide and Applicability of the Method in presence of Iron and Arsenic. L. MOSER (Zed. arzal. Chem. 1904 43 597-616).-Both the original method of De Haen and Rumpler and Low's modifications (Abstr.1896 ii 450) are affected by sources of error and the variation suggested by Litterscheid (Abstr. 1902 ii 531) is useless in presence of iron orANALYTICAL CHEMISTRY. 65 arsenic. The author has reinvestigated De Haen's method employing thiosulphate for the titration of the iodine. When free acids are absent the results are only satisfactory if the solutions are sufficiently con- centrated but in presence of free sulphuric acid correct results are obtained even when the dilution and amount of free acid are largely varied. Hydrochloric and nitric acids as acidifying agents are inadmissible but acet,ic acid may be used. Although the action of potassium iodide on a cupric salt indubitably consists at first in the formation of cupric iodide in the presence of free acid the decomposi- tion into cuprous iodide and free iodine is so rapid and complete that titration may be commenced two minutes after making the mixture ; n delay of ten minutes does not materially affect the results but after an hour higher numbers are obtained.The following is the form in which the process is recommended the copper solution (50 C.C. con- taining about 0.15 gram of copper as sulphate) is placed in a 300 C.C. stoppered bottle ; 5 C.C. of 10,V sulphuric acid and 2 grams of solid potassium iodide are added and the bottle is shaken for two minutes. The free iodine is then titrated with N/10 thiosulphate adding starch when near the end. To render the method applicable in presence of iron ancl arsenic the solution of the metals (as sulphates) is treated with an excess of sodium pyrophosphate sufficient to redissolve the precipitate at first formed. If the original solution contained free sulphuric acid some sodium acetate is added then 4-5 grams of potassium iodide and 5-10 C.C.of 80 per cent. acetic acid. I n these circumstances the copper alone liberates iodine but to obtain good results it is advisable to avoid dilution as far as possible wherefore solid sodium pyro- phosphate should be used for the precipitation and if the amount of iron exceeds that of copper by more than one-third the quantity of ptassium iodide should be increased; a somewhat longer time (15 mmutes) is allowed before titrating and the thiosulphate run in very slowly. M. J. S. Detection and Elstirnation of Minute Quantities of Mercury.CONSTANTIN ZENGELIS (Zed. unal. Chena. 1904 43 544-547).-For the detection of mercury in urine the urine is feebly acidified with hydrochloric acid and left in contact for 12 hours with a spiral of copper and platinum wires. The spiral is then washed with dilute sodium hydroxide water alcohol and ether and dried completely over sulphuric acid. It is then thrust to the bottom of a narrow test-tube the upper part of which is moistened with a solution of iodine in anhydrous ether Holding the tube horizontally the coil is heated and the evolved mercury vapour is arrested by the ring of iodine solution and converted into mercuric iodide. Complete absence of water is indispensable 0.2 milligram of mercury can be detected in 200-300 C.C. of. urine. For quantitative estimation the urine is boiled with an excess of sodium hydroxide and a little of a reducing sugar.The phosphate precipitate is dissolved in nitric acid and the solution electrolyaed at 40-50" with a weighed piece of platinum foil as cathode. M. J. S. VOL LXXXVIII. ii. 5Gbi ABSTRACTS OF CHEMICAL PAPERS. Detection and Estimation of Traces of Mercury in Urine with the aid of the Nernst Balance. ERNST JANECKE (Zeit. anal. Chev~. 1904 43 547-552).-1t is often desirable to be able to estimate with accuracy quantities of mercury not exceeding 0.5 mg. yr litre of urine. The use of the Nernst balance (Abstr. 1903 ii 571) enables this t o be done with 250 C.C. of the urine. The organic matter is destroyed by heating the urine with potassium chlorate and hydrochloric acid and the met’cury is then precipitated on a spiral of clean copper mire.The spiral is dried and strongly heated in a test-tube which has been drawn out to a capillary at the open end. The mercury is thus collected in the capillary tube. This is then crushed under a mixture of nitric and sulphuric acids and after heating to dissolve the mercury a little potassium sulphate is added and the solution measuring 10 c.c. is electrolysed with a coil of gold wire as cathode and a spiral of platinum wire as anode. These coils are conveniently kept from contact during the electrolysis by placing one inside and the other outside a short plass tube open at both ends. The gold wire is weighed before and after the electrolysis by substi- tuting it for the scale-pan of the Nernst balance.Since the gold absorbs hydrogen during the electrolysis a correction is made by inserting into the circuit a second similar coil in a cell free from mercury. The actual presence of mercury on the gold should be confirmed by expelling it into a capillary tube containing a trace of iodine. The author has succeeded by this method in detecting 0.01 mg. of mercury which had been added to two litres of urine. M. J. S. A Modifled Form of the Persulphate Method of Estimating Manganese in Iron and Steel. H. PROCTER SMITH (Chem. Newe 1904 90 237).-The following modification of this process is described 0.2 gram of the sample and the same weight of a steel containing a known quantity of manganese are bot,h placed in test-tubes and each gently heated with 10 C.C.of nitric acid of sp. gr. 1.2 until dissolved. The solutions are boiled to expel red fumes and 10 C.C. of silver nitrate solution (1.7 grams per litre) and about 1 grain of ammonium per- sulphate are added. After heating until the latter is nearly dissolved the sides of the tubes are washed down with water and the contents cooled. The perfectly cold pink solutions are then transferred to china basins and titrated with sodium amenite solution. The latter is pre- pared by dissolving 5 grams of arsenious oxide and 15 grams of sodium hydrogen carbonate in 250 C.C. of boiling water and diluting to 1 litre. Forty C.C. of this solution diluted to 500 C.C. form the titration solution which for each set of analyses must be standardised on a steel con- faining a known quantity of manganese.w. P. s. Volumetric Estimation of Iron. NAZARENO TARUGI aid S. SILVATICI (Chern. Centr. 1904 ii 1341-1342 ; from Boll. Chint. Fwrn. 43 637 -641).-The solution which must contain the iron in the ferric state is carefully neutralised with sodium hydroxide and after adding some potassium thiocyanate the liquid is titrated with i V / l O potassium oxalate until the colonr changes to yellowish-green. One C.C. of theANALY‘I‘ICAL CHEMISTRY. 67 oxalate solution = 0.00186673 gram of metallic iron. to make a blank experiment using 5 C.C. N/10 fevric chloride. It is advisable L. DE K. Colorimetric Estimation of Iron in Blood. ADOLF JOLLES (Zeit. c c m l . Clbenz. 1904 43 537-539).-The author’s most recent modification of his original method (Abstr.1897 ii 126) consists in cciiiparing the colour of the iron solution obtained from 0.05 C.C. of blood with that of the glass wedge of Fleischl’s hsmometer and he regards the apparatus foy this form of estimation to which he gives the name ‘‘ clinical ferrometer,” as the simplest and most convenient for cliiiical pwposees (compare Oerum Abstr. 1904 ii 449). Whereas in general the percentage of iron in the blood is proportional to that of the hEmoglobin in many pathological conditions such as ansmia icterus &c. the relation undergoes a marked alteration. M. J. S. Analysis of Commercial Tin and its Alloys. AUGUSTE HOLLARD and 1,. BEHTIAUS (Bull. SOC. chim. 1904 [iii] 31 1128-1131).-A complete scheme for the analysis of commercial tin is given. Arsenic is cletermiried by distillation with hydrochloric acid and ferrous sulphste the distillate being titrated with iodine (Abstr.1900 ii 438). The copper bismuth and lead are separated from the tin by precipita- tion as sulpliides in presence of ammonium sulphide and the bismuth and lead from copper by precipitation as sulphicles in presence of ammonia and potassium cyanide. Finally the lead and bismuth are separated and estimated electrolytically as already described (Abstr. 1904 ii 684). Antimonyis estimated by dissolving 1 gram of the tin in nitro-hydrochloric acid evaporating down repeatedly with hydro- chloric acid adding a little sodium hydroxide to the residue and electro- lysing under the conditions previously described (Abstr. 1903 ii 455). For the estimation of copper and sulphur 5 grams of tin are treated with nitric acid and the whole evaporated to dryness; the residue is washed thoroughly with water containing a little nitric acid and in these washings the copper is estimated electrolytically and the sulphur by precipitation as barium sulphate.For lead-tin alloys 1 gram of the alloy is treated with 50 C.C. of nitric acid (water should be added if the alloy is rich in lead) in presence of 10 grams of copper. The liquid is diluted to 300 C.C. and warmed to cause the tin oxide to ngglomerate. When cold the lead is estimated electrolytically as the peroxide (Abstr. 1899 ii 523; 1904 ii 294); to obtain the last t’races of lead the anode should be withdrawn when the greater part of the lead has been deposited and the mixture shaken and left in a warm place for a time.The anode should then be replaced and the electro- lysis continued. T. A. H. Estimation and Separation of Gold in the Electrolytic Way. SARAH P. MILLER (J. Amer. Chena. h’oc. 1904 26 1255-1369).-A lengthy article unsuitable for adequate abstraction showing the conditions under which gold can be satisfactorily separated from iron cadmium silver zinc cobalt nickel and copper. The gold is deposited68 ABSTRACTS OF CHEMICAL PAPERS. from solutions containing either potassium cyanide phosphoric acid or sodium sulphide. L.. DE K. Inadmissibility of Soap for Estimating Hardness of Ferru- ginous Water. A. GAWALQWSKI (Zeit. anal. Chem. 1904 43 533-536).-A sample of well-water containing 0.036 gram of ferrous oxide per litre in the form of sulphate gave on titration with soap solution results widely a t variance with the amounts of calcium and magnesium estimated gravimetrically neither could concordance be es- tablished by including in the calculation the amount of soap consumed by the iron present.An attempt to titrate with soap after the removal of the iron by potassium ferrocyanide also failed to give a satisfactory result. M. J. S . Estimation of Ammonia and Proteid-Nitrogen in Waters. JEAN EFFRONT (Cham. Centr. 1904 ii 1253-1254; from Mon. Sci. [ iv] 18 669-674).-An accurately titrated solution of bleaching powder is mixed with a definite volume of the sample of water to be tested and the loss in active chlorine represents the nitrogen of the ammonia and proteids. The experiment is then repeated with water from which the ammonia has been expelled by evaporation with sodium hydroxide ; the result is the proteid-nitrogen only.The active chlorine is estimated by adding excess of sodium arsenite and titrating the excess with standard iodine. L. DE K. Addenda to Simple Method f o r Decarbonising Substances. Estimations in the Decarbonised Product. ALBERT NEUMANN (Zeit. physiol. Chem. 1904 43 32-36. Compare Abstr. 1903 ii 243).-Further manipulative details are given for the processes and estimations already described. J. J. S Alkaloid Reactions. IV. Morphine. C. REICHARD (Chem. Zeit. 1904 28 1102-1105. Compare Abstr. 1904 ii 791 992).-The author communicates a new reaction for morphine which is quite characteristic. If morphine is gently warmed with strong sulphuric acid containing some arsenious or arsenic acid an intense and per- manent purple coloration is developed. In practice it is best to make a solution of arsenious acid in strong aqueous sodium hydroxide; to this may be added first the morphine and then an excess of very strong sulphuric acid. Morphine is also coloured red by antimonious chloride without using sulphuric acid. Stannous chloride in conjunction with sulphuric acid also causes the reaction. L. DE K. Mechanical Andysis of Soils and Sub-soils by Centrifugal Action ; with Notes on Treatment of Samples. J. R. KILROE (&con,. Proc. Roy. Dublin Soc. 1904 1 223-230).-A centrifugal apparatus is described by means of which mechanical separations can be made rapidly and with comiderable accuracy. The amount of soil employed for a determination is 20-30 grams and very little distilled water is required N. H. J. M.