ii. 118 ABSTRACTS OF CHEMICAL PAPERS. Analytical Chemistry. Application of the Interferometer to Gas Analysis. J. D. EDWARDS (Chtem and M e t . Eng. 1919 21 56&565).-The gas interferometer measures the difference in refractivity of two gases of binary mixtures. Usually one of the components is taken as the standard gas and the difference in refractivity of the mixture and one of ita components is determined. For a change of a% of one of the componenta the change in refraotivity AB = 2731 T p/760 . a/ 100(R - R2).ANALYTICAL CHEMISTRY. ii. 119 Each scale division corresponds with a definit’e value of Ar which is the difference in refraction of the two gases. The calibration of the instrument is described (see also A. 1918 ii 47). I n measuring a small percentage of a gas which deviates appreciably from Boyle’s law a correction is necessitated as the refraotivity of the gas a t a low partial pressure differ8 from the value calculated from the pressure ratios.I n taking readings calibration curves should be drawn for different temperature and pressure intervals from which readings may be taken by interpolation. A variation of 3O in the tempera- tures of estimation and of calibration involves an error of 1%. The error due to shifting of the central white or achromatic fringe was investigated by Adams (A. 1915 ii 478). Typical illustrations are given of t-he use of the instrument for air-carbon dioxide mixtures. It may also be employed for the analysis of flue gases and the determination of helium in mixtures. A table is given showing the relative sensitiveness of the inter- ferometer for different gases. Improved Orsat Apparatus €or Gas Analysis.G. W. JONES and F. R. NEUMEISTER (Chem. and Min. Eng. 1919 21 734-736).-The modification of Orsat’s apparatus used by the Bureau of Mines U.S.A. has a Jager’s copper oxide tube for the removal of hydrogen and carbon monoxide. The gas burette can b4 put in communication either with the Orsat pipettes containing the usual reagents for carbon dioxide unsaturated hydro- carbons and oxygen or with a compensator of the Peters- sen type the mercury i n which forms contact with a platinum wire and lights a lamp when the levelling bulb is raised or lowered and thus facilitates the reading of the level. The copper oxide tube above the Orsat bulbs is heated by means of an electrical heater which is lowered over it and then rests on supports. For the deter- mination of hydrocarbons in a gas a combustion is made in the combustion pipette with a measured quantity of oxygen the pipette being cooled by compressed air to prevent breakage and the amounts of the two predominating hydrocarbons are calculated by means of the usual formula from the total contraction and amount of carbon dioxide found after the combustion.[See further J . SOC. Chem. Ind. 1920 1388.1 Theory of Electrometric Titration. W. D. TREADWELL (Helu. Chim. -4cta 1919 2 672-680).-A theoretical paper in which the determination of the end-point in the titration of silver solutions is considered. The method consists in measuring the ter- minal voltage between a standard silver chloride cell and a silver eleotrode immersed in the solution the end-point being that at whioh the voltage becomes zero.The theory is considered for the titration of solutions of pure silver salts and of the case where there is a quantity of other substances present. Electrometric Titration. W. I,. TREADWELL and L. WEISS (Helv. Ch,im. Actu 1919 2 680-697).-1t is shown that the end- W. J. W. C . A. M. J. F. S.ii. 120 ABSTRACTS OF CHEMICAL PAPERS. point of many titrations may be determined electrometrically by measuring the terminal voltage between a metal electrode whiqh is immersed in the solution being titrated and is capable of sending the same ions into the solution and a comparison electrode made of the same metal and the insoluble precipitate or product produced in the titration.The two terminals arG direct,ly connected with a voltmeter and the point of zero potential is the end-point of the titration that is the point at which the concentration of the metal ions is the same in both the solution being titrated and the comparison electrode. It is essential that during the titration the liquid should be rapidly stirred and the authors describe an appara- tus in which the comparison electrode is built up in the stirrer. I n t.he case of silver solutions the comparison electrode is Ag I AgC1. sat.[\ a silver wire furnishes the electrode in the solution and the titration is carried out with standard sodium chloride. The results are extraordinarily good. I n the presence of foreign sub- stances t.he titration values are too low but they may be corrected from the course of the titration curve. Hydrogen sulphide may be titrated with silver ions using as comparison eleotrode the combina- tion Ag I Ag,S.dil. NaN0,II. Mercurous ions may be titrated by chloride ions using as comparison electrode Pt . Ag I Hg2C1 dil. H2S04(( and a similar silver-plated platinum wire as titration electrode. Better results are obtained if the titration is carried out with bromide ions and the calomel in the comparison electrode substi- tuted by mercurous bromide. Copper ions may be titrated by ammonium thiocyanate using as comparison electrode Cu 1 CuCNS . dil. NaHSO 11. In the foregoing cases soluble electrodes have been employed; further experiments were made with insoluble electrodes and it is found that zinc ions may be titrated with ferrocyanide ion using as titration electrode a platinum gauze and as comparison electrode Pt I Zn,Fe(CN)6.HCl(5 C.C. in 300H20)II. The titration in this case is carried out a t 70°. Ferrous chloride may be titrated with potass- ium dichromate using an already completed titration as the com- parison electrode and platinum wires in both solutions. This titra- tion is also carried out in warm solutions. Uranous ions may be titrated with potassium permanganate in warm solution using platinum wires and a uranyl solution as the comparison electrode. Vanadyl solutions may be similarly titrated with permanganate using platinum wires and a vanadate solution as comparison eleo- trode. I n the case of the hydrogen-ion titration the electrode is too slow in its action for the method as at present constituted to give sharp results. Electrometric Titration.J. PINKHOF (Chem. Weekblad 1919 16 1163-11 67) .-The investigation concerned the selection of a constant electrode which should have the same potential as the electrode of the liquid at. the moment when the end-point was reached. It would then be merely necessary to determine whether the difference of potent,ial of the two electrodes was positive or nega- J. F. S.ANALYTICAL CHEMISTRY. ii. 121 tive. To ascertain which compensation electrode is suitable a poten- tial measurement was made during titration and when the desired potential was found a combination of electrode and liquid was sought which would give this potential.For the titration of halogens in presence of each other N-potassium bromide is suitable for the com- pensation electrode in the estimation of the iodine. For the bromine Ag-AgCl-O*2NCl. is used. I n titrating heavy metals with sodium sulphide a silver-halogen compensation electrode is employed. Titrations of bases with mercuric oxide electrode and the use of the hydrogen electrode are discussed. W. J. W. A Rapid Hydrogen Electrode Method for Determination of Hydrogen-ion Concentrations in Bacterial Cultures or in other Turbid or Coloured Fluids. HARRY M. JONES (J. Znfect. Dzs. 1919 25 262-268).-As compared with the colori- metric method the apparatus described has wider application is more accurate less cumbersome and only slightly less rapid.The hydrogen electrode vessel described was designed to provide a vessel accurate a t least to 0.01 p and to give a rapid saturation with hydrogen gas and yet one which is easily construct-ed. A rapid and labour-saving technique combining the indicator and the gas-chain methods is described which obviates the difficult task of preparing standard solution for the former method and of making needless repetitions by the latter. CHEMICAL ABSTRACTS. Iodometric Estimation of Acids. I. M. KOLTHOFF (Pharm. Weekblad 1920 57 53-68).-The author has investigated the iodometric method especially with reference to weak acids. The influence of dilution in the case of strong acids of catalysts and of a variation in the amount of potassium iodide added was deter- mined. Tests were carried out with hydrogen chloride tartaric acid citric acid lactic acid malic acid succinic acid acetic acid formic acid benzoic acid salicylic acid sodium phosphate and currant juice.The iodometric method givm exce,llent results with strong acids especially if diluted in which case the reagents after mixing must remain for ten minutes before titration. A deficiency of potassium iodide causes inaocuracy. Organic oxy-acids can be accu- rately estimated if a salt of calcium magnesium or zinc is added. For phosphoric acid and weak organic acids not containing a hydroxy-group the method is unsuitable but i t may be used in the case of formic acid. The applicability of the method to the estima- tion of fruit juices requires further investigation. W. J. W. Indirect Electrolytic Estimation of Anions without Platinum Electrodes.11. E. LASALA (Anal. Fis. Quim. 1919 J7 235-247).-A continuation of work already published (A. 1919 ii 239). The present communication deals also with anions which form insoluble silver salts. These are carbonate oxalate ferro- and ferri-cryanide orthophosphate arsenate chromate and thiocyanate. The anion is precipitated as insoluble silver salt.ii. 122 ABSTRACTS OF CHEMICAL PAPERS. This is dissolved in a suitable solvent and the silver estimated elec- trolytically with a nickel-plated copper cathode and an iron anode. Satisfactory results were obtained only in the case of carbonate oxalate ferro- and ferri-cyanide and thiocyanate. The method was applied to the separation of chloride and iodide. The precipitated mixed silver haloid is washed with ammonia solution (D 0.92). The silver chloride is electrolysed after the addition of ammonium sul- phate. The residual iodide is dissolved in 20% solution of potassiu-m cyanide and electrolysed.A small correction must be applied for the solubility of silver iodide in ammonia solution. W. S. M. Electrolytic Estimation of the Halogens. An Indirect Method. J. H. REEDY (J. Arner. Chem. Soc. 1919 41 1898-1902). -The decomposition voltage of .A7-sulphuric acid saturated with silver chloride between silver coated with silver chloride as anode and bright platinum it9 cathode is found to be 0.59 volt which corresponds exactly with the sum of the solution potential of silver ((3.52 volt) and the overvoltage of hydrogen on bright platinum (0.07 volt).As long as this limiting voltage of 0.59 volt is not exceeded the electrolysis of haloid solutions with silver anodes proceeds with the formation of an adherent layer of silver haloid o n the anode without precipitation in the solution. With the com- pletion of the deposition of the halogen silver ions appear in the solution with no break in the continuity of the current and with scarcely any rise in potential. An indirect method of estimation of chlorides bromides and iodides is based on the above facts. The principle of the method consists in depositing the halogen on a silver anode under a constant voltage of 0.59 volt; some silver will thereby be lost from the anode. The anode is then dried and weighed and the silver haloid is reduced to metallic silver and the loss in weight determined which gives the weight of the halogen. A lead accumu- lator of large capacity is connected across the ends of a sliding rheo- stat and a current of E.M.P.0.59 volt tapped off. The electrolysis vessel consists of a beaker of 400 C.C. capacity fitted with an anode of silver gauze coated with silver electrolytically and a bright platinum cathode and a stirrer. The electrolysis is carried out with 200 C.C. of solution for about fifteen minutes; the current flowing at first is rather large but after about fifteen minutes it becomes negligible. The potential is then raised to 0.60 volt for five minutes. The silver electrode with its halogen deposit is washed dried and weighed. After weighing it is made the cathode in a dilute solu- tion of sodium hydroxide and electrolysis carried out until hydrogen is freely liberated from the surface.The silver haloid is thus reduced t o black silver which is not very adherent and must be carefully handled. After a thorough washing the electrode is heated at 500° in an electric furnace until the deposit is white and thoroughly sintered together. It is then weighed and the loss in weight gives the weight of the halogen. The results given in the paper show that hhe method .i,s qlaibe as good as the usual gravimetric method. J. F. S.ANALYTICAL CHEMISTRY. ii. 123 Estimation of Bromide in Mineral Waters and Brines. W. W. SKINNER and W. F. BAUGHMAN (Science 1919 50 50).- Chromic acid in concentrated solution liberates bromine quantita- tively from bromides a t the ordinary temperature and the bromine is removed by aspiration.Only a trace of chlorine is liberated from chlorides. When chromic acid acts on a solution containing chloride and bromide therefore the mixture of halogens removed by aspira- tion is dissolved in a solution of sodium sulphite and hydroxide which is then evaporated to dryness and the residue again treated with chromic acid solution and aspirated a second time. The double treatment gives very trustworthy results. CHEMICAL ABSTRACTS. Estimation of Alkali Sulphide in Dilute Solution. H. G . BENNETT and W. G. BENNETT ( J . SOC. Leather Trades Chem. 1919 3 ( l l ) 190-193).-A measured quantity of the solution contain- ing not more than 0'04% of hydrogen sulphide is distilled in an atmosphere of carbon dioxide with an excess of boiling magnesiuni chloride solution.A current of carbon dioxide carries the liberated hydrogen sulphide into a receiver containing standard iodine solu- tion. A second receiver containing standard sodium arsenite solu- tion is arranged to trap any iodine vapour carried over by the gas. The method is accurate under these conditions and provided there are no rubber connections exposed to the iodine vapour. D. W. Neutralisation Curve of Sulphurous Acid. I. M. KOLTHOFF (Chem. Weekblad 1919 16 1154-1163).-The accuracy of the results when sulphurous acid is titrated with methyl-orange or dimethyl-yellow a3 indicator depends on the concentration of the hydrogen ions of the sodium hydrogen sulphite and therefore also on the second dissociation constant of the sulphurous acid.This has been determined from the degree of hydrolysis of solutions of sodium sulphite and found to be 1 x 10-7 at 15O. The hydrogen-ions con- centration of a 0.1 mol. sodium hydrogen sulphite solution calculated from the two dissociation constants gave the value [H*]=3*3 x A curve showing the neutralisation of sulphurous acid was con- structed. I n $he titration of sulphurous acid errors of +_2% and of * 6% were obtained when the indicators were dimethyl-yellow and phenolphthalein respectively. The addition of sodium chloride reduced the error with phenolpht.halein to *2%. Accuracy can only bo ensured if excess of barium nitrate is added after the pink coloration of the phenolphthalein has appeared and the titration then continued to the end-point. Estimation of Sulphates in the presence of Organic Sulphonic Acids.T. CALLAN JAS. A. RUSSELL HENDERRON and R. BARTON ( J . SOC. Chenz. lnd. 1919,38 410-411~).-The solution of the substance (containing about 0.2 gram of sulphate) is rendered slightly ammoniacal (a small quantity of calcium chloride is also added to precipitate any carbonate which is then separated by filtration) and the mixture is boiled until all excess of aniinonia W. J. W.ii. 124 ABSTRACTS OF CHEMICAL PAPERS. has been expelled. Twenty C.O. of Nlkbarium chloride solution are then added the heating continued for a few minutes and the excess of barium titrated with W / 4-potassium chromate solution using starch-iodide paper as indicator.A drop of the mixture is placed on the paper and the edge of the spot is touched with a drop of very dilute hydrochloric acid. A blue colour develops in about ten seconds when an excess of soluble chromate is present. An immediate blue colour indicates that a considerable excess of chromate has been added. w. P. s. Detection of Nitrogen in Organic Compounds by Castellana’s Reaction. J. FLIERINGA (Pharm. WeekbEad 1920 57 3-4).-Castellana’s reaction (A. 1905 ii 207) may give a positive reaction even in nitrogen-free compounds owing t o com- bination of atmospheric nitrogen if the magnesium is in excess. It is preferable to use a mixture of 2 parts of sodium carbonate and 1 part of magnesium. Sodium carbonate containing 1 molecule of H,O should be employed. Addition of sugar to the reaction mixture eliminates the error caused by absorption of atmospheric nitrogen.W. J. W. The Relative Accuracy of Colorimetric and Titrimetric Procedures for estimating Nitrogen as Ammonia. E. R. ALLEN and B. S. DAVISSON ( J . Bid. Chem. 1919 40 183-197).- It appears that the colorimetric methods are slightly the more accurate. On the other hand the conditions for accurate titrations may be attained more easily than those for satisfactory Nessler esti- mations and at the same time the titrimetric methods are applic- able t o a wide range of nitrogen values. The chief disadvantage of the Nessler process ia the fact that the reaction is imperfectly understood and the appearance of clouds and precipitates difficult to avoid. Again it is believed that the personal error is greater in colorimetric methods than in titrimetric methods.The general conclusion is therefore in favour of the titrimetric methods. J. C. D. Estimation of Nitrogen in Nitrates. F. PILZ (Zeitsch Eandw. Vers.-Wesen Deutschosterr. 1919 22 180-188 ; from Chem. Zentr. 1919 iv 767).-Comparative estimations of nitrogen in nitrates by various methods showed that reduction with Devarda’s alloy yielded the most trustworthy results and is to be recommended more than the method described by Arnd (A. 1917 ii 504). Busch’s method (A 1905 ii 418j is less trustworthy. Other methods investigated were the one proposed by Ulsch (reduc- tion in acid solution with iron powder) and the indirect method Method for the Estimation of Minute Quantities of Nitrogen in Organic Substances which furnishes a New Quantitative Method of Diagnosis in some cases of Mental Disease.R. V. STANFORD (Rep. Chem. Lab. Card@ C i t y Mental Hospt. 1919 No. 2. Compare A. 1913 i 93i)).-A small quantity (conversion of the nitrate into chloride). w. P. s.ANALYTICAL CHEMISTRY. ii. 125 of the organic substance as for example 1 C.C. of cerebrospinal fluid is oxidised by the Kjeldahl process in the usual manner. The subsequent distillation is carried out in the ordinary apparatus but the distillate is collected in three fractions in Nessler cylinders. The third fraction should show little or no coloration with Nessler’s reagent. The colour produced by the first two fractions is com- pared in the usual manner with that produced by a standard solu- tion of ammonium chloride.Considerable a r e has t o be taken to avoid errors arising from ammonia in the reagents and apparatus. The author records a number of analyses of cerebrospinal fluid from mental cases. The ‘ I nitrogen number,” that is the number of hundredt.hs of a milligram of total nitrogen contained in 1 C.C. of cerebrospinal fluid may be of value in diagnosis. On the whole density and nitrogen number of the fluids run parallel to one an0 ther . No traces of choline and only the merest traces of ammonium salt were detected in the cerebrospinal fluid in mental disease. J. C. D. Procesa for the Estimation of Arsenic. Chemistry of the Marsh-Berzelius Process. B. S. EVANS (Analyst 1920 45 8-17).-The arsenic is evolved as hydrogen arsenide in the same way as in the ordinary Marsh-Berzelius test but is conducted over a roll of red-hot copper which retains the arsenic the amount of which is obtained from the increased weight of the copper.The compound formed with the copper usually melts and resembles solder in appearance but occasionally brilliant silvery-white blade- shaped crystals are produced. Analyses of the deposits gave 67.5 and 70.5% of copper and 32.28 and 29.27% of arsenic and thus appeared to be mixtures of the arsenides Cu,As and Cu,As,. The resulk obtained with elementary arsenic and various arsenic com- pounds were invariably 1.4% too low and the cause of this loss could not be definitely traced although the rwults indicated that if the arsenic remained in the flask it was in the form of a very stable complex. I n order to prevent separation of elementary arsenic it is advisable that the amount of substance introduced should contain considerably less than 0.5 gram of arsenic.[Sw also J . SOC. Chem. Ind. 1920 February.] C. A. M. Estimation of Arsenic in Tin and Tinnings. LUCIEN VALLERY (Compt. rend. 1919 169 1400-1402) .-Marsh’s method is liable to give results which may ia some cases be wrong to the extent of 75% of the amount found when applied to the estimation of arsenic in tin. The author recommends that the arsenic should be distilled off its arsenic chloride by Hollard and Bertiaux’s method and the arsenic estimated colorirnetrically in the distillate after reduction to colloidal arsenic by means of hypophosphorous acid. In this way the results obtained are accurate to within 3% of the total a t a concentration of 1 in 100,000.W. G.ii. 126 ABSTRACTS OF CHEMICAL PAPERS. New Methodof Estimation of Carbon and Hydrogen in Organic Substances. KUNIHIKO SUMIKURA (J. Tokyo Chem. SOC. 1919 40 593-606).-The essential point of the method is the combined use of platinum and cerium dioxide. (1) Substances containing Carbom Hydrogen (and Oxygen) only.-An ordinary combustion tube is filled in order with cerium dioxide (5 cm.) porcelain boat cerium dioxide (8 cm.) (only one burner is necessary under each of these) and a coil of platinum or platinum-iridium wire (2 m.) which is electrically heated with adjustable resisb ances. The wire is first heated in a separate tube before being placed in the combustion tube.The cerium dioxide is prepared by heating copper gauze packed with asbestos soaked in a solution of ceric nitrate. (2) Substances containing Nitrogen and Sulphur.-Lead dioxide is placed between the second cerium dioxide gauze and the platinum coil and is heated a t 300-350O. (3) Substances containing Halogens.-Silver foil is placed between the second cerium dioxide gauze and the platinum wire. I f nitrogen is also present lead dioxide is placed between the silver foil and the platinum wire and is heated a t 300-320O. (4) Substances containing Phosphorus.- Lead monoxide is mixed with the substance. Trustworthy results are recorded. CHEMICAL ABSTRACTS. Rapid Estimat.ion of Carbon Monoxide in Air. ARTHUR B. LAMB and ALFRED T. LARSON (J. Amer. Chem. SOC. 1919 411 1908-1920).-Two methods for the rapid estimation of small quantit,ies of carbon monoxide in air are described.I n the first method the gas is passed over a platinum wire heated to redness and the carbon monoxide partly burned in contact with it. The gas then passes over heated copper oxide where the carbon mon- oxide is completely burnt; i t then passes through a spiral con- tained in a thermostat to bring it to its original temperature and finally over a second exactly similar platinum wire heated by the same current as the first. The two platinum wires are made arms of a Wheatstone bridge and since the first wire will necessarily be hotter than the second they will have different resistances. The change in the resistance of the first wire gives a measure of the carbon monoxide concentration.I n the second method the mix- ture is passed through a platinised platinum catalyst when the carbon monoxide is completely burnt. The temperature change is measured by a thermo-element and indicates the concentration of carbon monoxide. Both methods are rapid and the latter is par- ticularly accurate easy to instal and to operate. The catalyst with reasonable precautions remains active for many weeks. J. F. S . Apparatus for the Volumetric Estimation of Small Quantities of Carbon Dioxide by Displacement from Liquids by means of a Rapid Current of Air at the Ordinary Tem- perature and Pressure. A. COSTANTINO (Atti R. Accad. Lincei 1919 [v] 28 ii 118-121).-1n the apparatus described whichANALYTICAL CHEMISTRY. ii. 127 consists of a closed system a rotary mineral oil pump causes the repeated circulation of purified air first through the liquid con- taining the carbon dioxide and then through a definite volume of standard barium hydroxide solution this being subsequently titrated with acid in presence of phenolphthalein.The determin- ation occupies in all about forty-five minutes. Test experiments with sodium carbonate solutions cont9aining 6-4-21 mg. of carbon dioxide show that the method yields accurate results. T. H. P. Eetimation of Metals by Electrolytic Deposition without using an External Supply of Electricity. MAURICE FRANCOIS (Ann. Chtim. 1919 [ix] 12 178-192).-A more detailed account of work already published (see A. 1919 ii 34). W. G. Use of Nickel Crucibles for the J.Lawrence Smith Method in Estimating Soil Potassium. SETH 5. WALKER (J. Ind. Eng. Chem,. 1919 11 1139-1140).-The same results are obtained whether platinum or nickel crucibles are used for the heating of the soil with ammonium chloride and calcium carbonate but the nickel crucible is attacked by the mixture and the life of the crucible is short,. w. P. s. The McCrudden Gravimetric Calcium Method Modified. J. P. HALVERSON and J. A. SCHULZ ( J . I n d . Enq. Chem. 1920 12 77-78).-The calcium oxalate precipit.ate is collected on a filter of specially prepared asbestos contained in a Gooch crucible; after washing both filter and precipitate are mixed with water acidified with sulphuric acid and titrated a t 65O with perman- ganate solution. The filter is prepared by digesting ignited asbestos with dilute sulphuric acid and a slight excess of per- manpanate for one hour at 60° to 70°.and then washinp it until freeofrom acid. [See further .7. Soc. (7htem. Ind. 1920 Fkbruary.] w. P. s. Estimation of the Calcium Oxide Content of Milk of Lime. GEORG LENART (Zeitsch. Ver.. deut. Zuckem'nd. 1919 1-15 360-361) .-MategEek's table (0sterr.-Ungar. Zeitsch. Zuckerind. 1871 12 718) correlating the sp. gr. and calcium oxide content of milk of lime is regarded as far from correct whilst that published by Lunge and Blattner (Dingl. Polytechn. J. 1883 466) is also considered to be untrustworthy. Experiments are described demon- strating the difficulty of obtaining exactt results by means of the hydrometer (the method of density determination adopted in the construction of the second table) since the readings are influenced by the shape and length of the spindle by the width and depth of the column of liquid in the cylinder and also by the degree of movement imparted to the liquid a t the moment of observation.Using a pyknometer consisting simply of a cylindrical vessel having a capacity of about 500 c.c. provided with a ground-glass plate as cover a new table has been constructed by the author in which the density values range from 1.0085 to 1.2195 ( 1 - 5 O to 26.2' Be'.)ii. 128 ABSTRACTS OF CHEMICAL PAPERS. at 20° temperature corrections being stated for readings made between 20° and 40°. Milk of lime which has been vigorously shaken for about two hours assumes a viscous condition in which the particles remain almost completely in suspension for some days this being believed t.0 be due to the formation of a hydrated Compound probably Ca( OH),,H,O to which attention has already been directed by Herzfeld (Zeitsch.Ver. deut. Zuckerind. 1897 817) and by Karcz (Chem. Zeit. 1898 22 38). The Permanganate Method for Copper. LOUIS F. CLARK (2. Ind. Eng. Chem. 1919 11 1138-1139).-A modification of the procedure for the titration of cuprous thiocyanate with per- manganate solution is proposed with the object of making the estimation less arbitrary. The cuprous thiocyanate precipitate is treated in a beaker with 25 C.C. of hot 4% sodium hydroxide solu- tion until the precipitate is decomposed 25 C.C. of cold 15% sulphuric acid and a quantity of 1% ferric sulphate (anhydrous salt-) solution are added and the mixture is titrated with permanganate solution.Towards the end of the titration and when the mixture becomes whiter in d o u r 5 C.C. of concentrated hydrochloric acid are added and the titration then complet.ed. Under these con- ditions 6 atoms of copper should be equivalent t o 7 atoms of iron or a ratio factor of 0.1626; the ratio factor actually found as the result of experiment is 0.1646. [See further J . SOC. Chem. Znd. Microchemical Identification of Soluble and Insoluble Mercury Compounds. G. DENIGBS (Bull. SOC. Pharm. Bordeaux 1919; from ,4nn. Chim. anal. 1919 rii1 1 383-385).-The tests depend on the characteristic crystals obtained when a small quantity of the compound is treated with a drop of potassium bromide iodide solution.Mercurous compounds and mercuric cyanide are oxidised previously with bromine. Mercuric iodide itself is identified by the crystals obtained from its alcohol or acetone solution. w. P. s. OTTO SASSE (Pharm. Zeit. 1919 64 633).-Referring to the method described by Utz (A. 1919 ii 428) the author mentions that he himself had published a method (Pharm. Zeit. 1887) for the volumetric estimation of mercuric chloride; this method depends on a tiha- tion with potassium iodide solution and is trustworthy. J. P. 0. 1920 138~.] w. P. s. Volumetric Estimation of Mercuric Chloride. w. P. s. Application of Rotating Reductors in the Estimation of Iron. WALTER SCOTT ( J . Ind. Eng. Chem. 1919 11 1135-1137). -Gooch and the author have shown previously (A.1918 ii 373) that vanadic acid may be reduced by rapidly rotating anodes of silver or copper and the author now describes a similar method of reduction for ferric salts. The latter in sulphuric acid solution are reduced completely by rotat,ing reductors of zinc or aluminium with or without the use of an external current. w. P. s,ANALYTICAL CHEMISTRY. ii. 129 A Sensitive Reaction for Manganese. P. H. HERMANS (Pharm. WeekbEad 1919 56 1344-1346).-Caron and Raquet's reaction for manganese (RLp. Pharm. 1919 97) was independently discovered by the author and was in addition utilised as a test for nitrogen tetroxide and for nitrites. All oxidising agents are not effective in promoting the red coloration for example hydrogen peroxide potassium persulphat e chlorine bromine iodine chloric acid iodic acid potassium ferricyanide but addition to them of 0.5% solution of osmic acid produces a satisfactory result.Active oxidising agents for the test are bleaching powder potassium per- manganate and dichromate lead peroxide cerium ammonium nitrate nitrous acid and nitrogen tetroxide. The reaction is sensitive to 0.07 mg. manganese or 0.03 mg. sodium nitrite per C.C. of solut,ion. The red coloration is considered to be a complex alkali manganese oxalate. W. J. W. Lead peroxide is most suitable for the manganese test. Indirect Detection of Tin. FR. FEIGL (Chem. Zeit. 1919 43 861).-Ferrous salts give a red precipitate with an ammoniacal solution of dimethylglyoxime but the presence of ferric salts which give a precipitate of ferric hydroxide with the reagent interferes with the test.The precipitation of ferric iron as hydroxide and of ferrous iron as ferrous dimethylglyoxime may be prevented by adding an organic hydroxylated compound. such as citric or tart.aric acids or potassium sodium tartrate and the ferrous ion then gives a carminered solution. I n each case complex iron compounds are formed. I n usinq this reaction as a sensitive indirect test for tin the mixed sulphides mesumably of antimony and tin are dissolved in hydrochloric acid and one portion of the solution is tested for antimony with iron wire. I n a second portion the tin and anti- mony are precipitated by means of pure zinc and the tin extraded from the deposit by means of concentrated hydrochloric acid. The solution is poured into a hot dilute solution of ferric chloride.a little citric acid or potassium sodium tartrate added and the liquid tested with ammoniacal alcoholic dimethylglyoxime solution. The ferrous ion formed by reduction with the stannous ion will give a red coloration. the intensity of which is proportional t'o the amount of stannous chloride. C. A. M. Estimation of Thiophen in Industrial Benzenes. ANDR& MEYER (Compt. rend. 1919 169 1402-1404).-A modification of DenigBs's method (compare A. 1895 ii 332 372). Denig'es's reagent is prepared by dissolving 50 grams of mercuric oxide in 200 C.C. of sulphuric acid diluted with 1000 C.C. of water. Twenty C.C. of this reagent are mixed with 75 C.C. of methyl alcohol and 5 C.C. of the benzene under examination in a stoppered bottle.The mixture is shaken from time t o time and after t.hirty minutes is filtered. Ten C.C. of the filtrate are diluted with 20 C.C. of water a few drops of nitric acid are added and the whole is titrated withii. 130 ABSTRACTS OF CHEMICAL PAPERS. iV/10-ammonium thiocyanate solution using 1 C.C. of a 20% iron ammonium alum solution as indicator. I n view of the possible presence of acetone in the methyl alcohol a blank experiment is conducted using pure benzene free from thiophen. The difference between the two titrat,ions is a measure of the mercury precipit.ated by the thiophen as the compound (HgO),SO,,HgSC,H and the factor for the conversion of mercury to thiophen is 0.14. W. G. Titration and Partial Reduction of Nitro-compounds with Titanous Chloride.C. F. VAN DUIN (Chem. TVeekbZad. 1919 16 1111-llZZ).-An examination of the titanous chloride method for the titration of nitro-compounds. The constancy of titre of a titanous chloride solution at the ordinary temperatures and after helating was especilally investigated ; also the general applicability of the method to various classes of nitro-compounds. A t the ordinary temperatures only slight variations in titre were found after keeping for several days. Titration experiments before and after heating the titanous chloride solution were carried out in comparable conditions in presence of alcohol hydrwhlorio acid and acetic acid. It was found that tqhe assumption of an invariable titre during the heating of the titanous chloride solution led to quite erroneous results.The error is small in presence of acids especially in hydrochloric acid tderabb results being only obtained in strongly acid solutions. From experiments wit.h picric acid mdinitrobenzene and s-trinitrotoluene the author concludes that the method yields results of the same order of accuracy as t(he Dumas method. The reduction and titration of the following classes of nitrecompounds were carried out nitroamines (s-trinitro- phenylmet#hylnitroamine aminotrinitrophenylmethylnitroamine trinitromethylnitroaminophenol trinitromethylnitroaminoanisole trinitrodimetqhyldinitroaminobenzene) tetranitro-compounds (tetra- nitrophenol tetranitrophenylmethylnitroamine tetranitroaniline) and hexanitro-compounds with two nuclei (hexanitrodiphenyl sulphide hexanitrodiphenylsulphone) .I n the reduction of nitro- amines side reactions may take place in which the nitroamino-group is entirely removed from the nucleus. I n all cases blank experi- ments must be carried out in order to estimate the alteration in t-he titanous chloride solution on heating. The amount of excess of the solution necessary for complete reduction must also be deter- mined in each case. Only in those conditions can this method of nitrogen analysis be considered superior to that of Dumas. W. S . M. Detection of Methyl Alcohol in Ethyl Alcohol. M. POLINSKI (Chem,. AnaZ?yst 1919 28 ll).-Five C.C. of the sample are dis- tilled with 50 c . ~ . of water and 3 grams of sodium persulphate the distillate being collected in fractions of 2 C.C. each. The fifth fraction is treated with eight drops of a freshly prepared filtered 1% solut.ion of phenylhydrazine hydrochloride four drops of 5%ANALYTICAL CHEMISTRY. ii.131 potassium f errooyanide and 1 c. c. of concentrated hydrochloric acid. A pink coloration is produced if met,hyl alcohol is present. CHEMICAL ABSTRACTS. Complex Organic Mercury Compounds. EINAR BIILMANN (Medd. K . Vetenskupsukad. Nobel-Inst. 1919 5 No. 12 1-11). -The complex compound formed from allyl alcohol and mercuric acetate is very stable the concentration of the mercuric ion in a 1/160-molar solution being only 10-8-40 (compare A. 1917 i 123). This fact can be made use of in the electrometric estimation of allyl alcohol. The aqueous solution of allyl alcohol is put in a beaker containing mercury which forms one electrode the other being a calomel electrode.A standard solution of mercuric acetate is then run in and the E.M.F. curve plotted. The point of inflexion on the curve is very sharp and accurate results are obtained. Ally1 alcohol can also be estimated by adding mercuric acetate and potassium bromide to the solution and titrating the acetic acid set free with phendphthalein as indicator. The liberation of acetic acid is quantitative according to the equations C3H,-OH + Hg(OAc),= C,H,.O-Hg*OAc+ AcOH ; C,H,*O*Hg*OAc + KBr = C,H,*O*HgBr + OAcK. Mercury can also be estimated very accurately by making use of t-he quantitative reaction C,H,*O*Hg*OH + KBr = C,H,*O*HgBr + KOH. A weighed quantity of mercuric oxide is dissolved in 5N-sulphuric or nitric acid; the solution is diluted with water free from carbon dioxide a few C.C.of allyl alcohol added and then N/5-sodium hydroxide until the liquid is alkaline t~ phenolphthalein. The whole is then exactly neutralised with N/5-sulphuric acid 5 grams of potassium bromide added and the liquid well shaken to give a solution the entrance of carbon dioxide being carefully excluded. The liberated alkali is then titrated. Halogen acids must not be present in estimations according to this method. Maleic and crotonic acids do not give sufficiently stable complexes for estimation by the above methods but allocinnamic acid gives good remlts with the mercuric acetate-potassium bromide method. Cinnamic acid also gives mercury complexes and is to be reported on later. Mercury can be readily and accurately estimated by a modifi- cation of fncze’s method (A.1917 ii 327). To the solution of the mercuric salt or of mercuric oxide in nitric or sulphuric acid is added a large excess of alkali; the whole is then exactly neutralised with standard acid using phenolphthalein it9 indicator. Potassium iodide (5 grams) is then added and the liberated alkali carefully titrated carbon dioxide being excluded. D. D. VAN SLYKE and W. W. PALMER (Proc. Soc. Exp. Biol. Med. 1919 16 140-141). -Carbonates and phosphates are removeld by adding 2 grams of calcium hydroxide to 100 C.C. of urine and filterin after ten minutes. The acidity of 213 C.C. of the filtrate is broug fl t t o pH=8 T. S. P. Titration of Organic Acids in Urine.fi. 132 ABSTRACTS OF CHEMICAL PAPERS.(circa) by adding O.2N-hydrochloric acid until the pink colour of phenolphthalein disappears 5 C.C. of 0.02% Tropceolin 00 are added and the solution is titrated to pH=2-7 with 0-2N-hydre chloric acid the volume being brought. to approximately 50 C.C. by the addition of water towards the end of the titration. The colour is compared with that of 50 C.C. of a control solution with the same pE value and indicator. The amount of 0-2N-hydro- chloric acid required to give the end-point with a control in which water replaces urine is subtracted. The titration est-imates from 93% to 100% of each of the organic acids known to be present in urine in quantitatively significant amounts. It also includes very weak bases but of this class of substance8 only creatine and .creatinine which are estimated to nearly loo% are significant.The titration figure corrected for the amounts of these two bases represenb the organic acids in urine. Detection of P-Hydroxybutyric Acid Acetoacetic Acid and Acetone in Urine. R. LILLIG (Pharm. Zeit. 1919 64 696-697 707-709).-A collect.ion of some sixty methods which have been proposed for the purpose; a brief description is given of each method. w. P. s. Bellier’s Reaction for Vegetable Oils A. VAN RAALTE (Chem. Weekblad 1919 16 969-960).-The benzene used for this reagent must be free from thiophen otherwise a dirty green instead of a violet coloration is obtained. Relative Sensitiveness of the Fehling Phenylhydraziae and Nylander Tests for the Detection ot Dextrose in Urine. GEORGE E.GWE (Amer. J . Pharm. 1919 91 717-719).-Fehling’s test is the most sensitive of the three tests investigated; it yields a reaction with a solution containing 0.00125% of dextrose. The phenylhydrazine test is also sensitive giving a reaction after two hours with a 0.0025% solution of dextrose but the Nylander reagent (potassium sodium tartrate 4 grams; 10% sodium hydroxide solu- tion 100 c.c.; bismuth subnitrate 2 grams) fails to give a reaction unless the solution contains at. least 0.5% of dextrose. E. SCHOWALTER (Zeitsch.LATarhr. Gmusm. 1919 38 221-227).-A reply to Rupp and Lehmann (A.) 1919 ii 434); the author maintains the correctness of his former remarks The Isaacson Method for Estimating Dextroee. ELISABETTA C. PENNELL (J. Lab. Clin. Med. 1919 4 736-737).-Isaacson’s method (A.1918 ii 246) was found highly untrustworthy when compared with the Fehling volumetric method. CHEMICAL ABSTRACTS. W. J. W. W. P. S. Titration of Sugars. (A. 1919 ii 172). W. P. s. CHEMICAL ABSTRACTS. Separation of Dextrose and Lavulose. FRANZ LUCIUS ( Z e i t s c h . iVahr. Genwsm. 1919 38 177-185).-A method for the estimation of dextrose in the presence of lzevulose depends on the fact that the rotation of the latter is destroyed by heating for sevenANALYTICAL CHEMISTRY. ii. 133 hours with dilute hydrochloric acid whilst that of dextrose is not affected. Fifty C.C. of the sugar solution (containing from 2 to 4% of the sugars are mixed with 10 C.C. of 5N-hydrochloric acid and heated in a b oiling water-bath for seven hours. The mixture is then nearly neutralised with sodium hydroxide solution diluted to 100 c.c.treated with animal charcoal and filtered if necessary and polarised at 20°. About 99% of the dextrose actually present is found. The dextrose remaining after the acid treatment cannot be estimated by determining the copper-reducing power of the solution as the lzvulose is converted into reducing substances which how- Separation of ‘ Saccharin ” from Benzoic Acid. E. SCHOWALTER (Zeitsch. Nahr. Genussm. 1919 38 185-194).-The presence of “ saccharin ’’ interferes with the estimation of benzoic acid in foods etc. by the methods described by Mohler von der Heide and Grossfeld (A. 1908 ii 906; 1910 ii 359; 1915 ii 848) but the two substances may be separated from one another by taking advantage of the insolubility of ‘ I saccharin ” in carbon tetra- chloride.The residue obtained on evaporating the et-hereal solu- tion of the two substances a t a low temperature is dried under reduced pressure to remove the last traces of ether and acetic acid and is then treated with carbon tetrachloride; the benzoic acid dis- solved by this solvent may be estimated by titration and the insoluble “ saccharin ” identified by Klostermann and Scholta’s reac- tion (A. 1916 ii 586). Separation of the two substances between sublimation is untrustworthy ; “ saccharin ” begins to sublime a t 120° and 130O. Animal charcoal absorbs considerable quantities of ‘‘ saccharin,” its anhydride and sodium salt as well as benzoic acid Estimation of p-Phenylenediamine. T. CALLAN and JAS.A RUSSELL HENDERSON ( J . Soc. Chem. Id. 1919 38 408-410~).- A modification of a method described by Effront (A. 1905 ii 60) is recommended. Fifty C.C. of sodium hypochlorite solution con- taining 12-15% of available chlorine are diluted to 1 litre and 100 C.C. of this solution are diluted with an equal volume of water 1 gram of sodium carbonate is added followed by 10 C.C. of the p-phenylenediamine solution (in dilute hydrochloric acid) ; the latter solution should contain from 2 to 6% of the substance and the addi- tion should be made gradually with constant stirring. Benzo- quinonedichloroimide is precipitated as a white substance and with- out filtration the excess of chlorine in the mixture is titrated with N / 10-sodium arsenite solutioa. The difference between this titra- tion and the quantity of arsenite solution required for the titration of the hypochlorite solution itself is equivalent to the amount of chlorine removed from solution as dichloroimide; 1 C.C.of LV/ 10-arsenite solution t.herefore corresponds with 0*0018 gram of p-phenylenediamine. The method may also be used for the estima- tion of p-nitroaniline p-aminodiphenylamine benzidine and naphthyIene-l:4-diamine but with the latter the yield found is ever are optically inactive. w. P. s. and benzoates from an aqueous solution. w. P. s.ii. 134 ABSTRACTS OF CHEMICAL PAPERS. only 98-99% of the quantity of substance present owing to the slight solubility of the dichloroimide formed. The method is not applicable to substances containing a sulphonic group since these give soluble dichloroimides.w. P. s. Estimation of Carbamide in Blood and Urine. H. CITRON (Beut. med. Woch. 1919 45 975; from Chem. Zentr. 1919 iv 719).-A reply to Brahm (this vol. ii 66). H. W. Electrometric Method for the Estimation of Ferrocyanides depending on a Change in Oxidation Potential. G. L. KELLEY and R. T. BOHN (J. Amer. Chem. Soc. 1919 41 1776-1 783).-Solutions of ferrocyanides may be estimated by titraticn with potassium permanganate in the presence of sulphuric acid using the sudden change in E.M.F. observed with a bright platinum electrode to mark the end-point. The method is as follows About 1 gram of potassium ferrocyanide in 250 C.C. of water and 2.5-5-0 C.C. of sulphuric acid (D 1-58> is titrated with 0*@5N-potassium permanganate using an electrometric apparatue which will indicate changes of potential of the order of one millivolt (A.1917 ii 512). A small precipitate may form during the titra- tion but this will redissolve. The titration should be performed slowly with mechanical stirring. As the end-point is approached the permanganate should be added drop by drop and time allowed for the completion of the reaction. The end-point is taken as the point of greatest change in potential for equal additions of permanganate. Should too much permanganate be added the end-point may be approached from the other side by adding a standard solution of ferrocyanide. The presence ot' ferricyanides does not interfere with the titration neither do amounts of chloride up to 1 gram of sodium chloride.Any salts which under the conditions of experiment produce a precipitate with either ferrocyanide or ferricyanide seriously interfere with the method. J. F. S. Use of Silicotungstic Acid in the Estimation of Alkaloids. ERWIN TAIGNER (Zeitsch. anal. Chem. 1919 58 346-352).- Silicot-ungstic acid solution may be used for the quantitative pre- cipitation of atropine strychnine and cocaine from their hydro- chloric acid solutions ; the precipitates formed are collected after the addition of sodium chloride washed with 1% hydrochloric acid dried a t 40° and weighed. The precipitates have the general formula 12W0,,Si0,,2H20,4-a1ka10id. w. P. s. Colour Reactions of Emetine. ABEL LAHILLE (Arch. mkd. exp. 27 296 ; Zerttr. Biochem. Biophys. 1919 20 398) .-Emetine hydrochloride develops a yellow coloration with calcium chloride and with concentrated nitria acid. When a small crystal of ammon- ium molybdate is added to a solution of the hydroohloride in con- centrated sulphuric acid a green zone which subsequentry turns blue is formed round the crystal; the colanr disappear8 on dilutionANALYTICAL CHEMISTRY.ii. 135 with water and reappears on heating. A green coloration is pro- duced when a dichromate is used instead of ammonium molybdate. CHEMICAL ABSTRACTS. The Arsenotungstic Reagent in the Method of Bertrand and Javillier for the Estimation of Nicotine. LUIS GUGLIAL- MELLI and ULAUS HORDH (Anal. SOC Quim. Argentina 1919 7 121-123) .-The preparation of silicotungstic acid for the estima- tion of nicotine by the method of Bertrand and Javellier (A.1909 ii 450) presents difficulties. The authors recommend the use of arsenotungstic acid (Guglialmelli A. 1916 ii 584 587) as a precipitant. The procedure is otherwise unaltered. W. S. M. Ehrlich’s Reaction with p-Dimethylaminobenzaldehyde and Urorosein. E. SALKOWSKI (Biochem. Zeitsch. 1919 97 123-1 29) .-Tb e reaction of p-dimethylaminobenzaldehyde with indole indoleacetic ind olepropionic and indoleaminopropionic acids is described and discussed ~ Herter’s theory concerning the origin of urorosein in urine is also discussed. s. s. z. Estimation of Tyrosine by means of Millon’s Reaction. MORIZ WEISS (Biochem. Zeitsch. 1919 97 170-176).-A 1 /50,000 solution of tyrosine to which Millon’s reagent is added is used as a standard.The solution to be examined is then diluted until it gives with Millon’s reagent a coloration equal to that of the standard. The method can be employed for the estimation of s. s. z. New Method of Estimating Albumin in Urine. DOMENICO GANASSINI and PIETRO FABBRI (Boll. Chim. Farm. 1919 58 313-319).-A method of estimating albumin in urine is based on the fact that the sulphur of the cystinic group in the protein mole- cule is readily detached by the action of hot alkali hydroxide and in presence of lead plumbite yields lead sulphide. On then titrating the liquid with iodine solution the metaproteins are first oxidised and iodised and the lead sulphide is then converted into lead sulphate. I n applying tnhe method from 20 to 100 C.C. of urine (according to the amount of albumin indicated by a pre- liminary test) are mixed with 5% of sodium dihvdrogen phosphate (to prevent precipitation of phosphates of alkaline earth metals) and the liquid heated and stirred for a few seconds until t.he albumin has coagulated.The precipitate is separated washed and dissolved in 10 to 30 C.C. of 10% potassium hydroxide solution from 3.3 to 10 C . C . of 5% normal lead acetate solution added and the beaker immersed for five minutes in boiling water. The liquid is then cooled to 21° and titrated with N/lO-iodine solution which is added a t the rate of two drops per second with continual stirring until the liquid becomes colourless. The number of C.C. of iodine multiplied by 0.0330 gives the quantity of albumin in the amount of urine taken Standard potassium ferricyanide solution may be used instead of iodine solution for the titration but the end-point of the reaction is not so sharp.tyrosine in the products of hydrolysis of proteins. C. A. M.ii. 136 ABSTRACTS OF CHEMICAL PAPERS. Medical Analysis of Urine by the Combined Use of Tungstic Acid and Other Reagents. HUGH CAMPBELL Ross (Brit. Pat. 134710).-When to 2 C.C. of urine 0-2 C.C. of a reagent composed of a 12% aqueous solution of colloidal tungstic acid to which 10% of acetic or tartaric acid to neutralise any possible alkalinity of the urine and 2.5% of sodium potassium tartrate have been added the presence of albumin is indicated by the form- ation of a precipitate which does not dissolve on warming. If the acid is impure an excess of urates also causes a precipitate but it redissolves on boiling the solution.I n the latter case a diabetic condition is indicated and the presence of reducing sugars may be tested for by adding copper sulphate and an alkali hydroxide or preferably for the sake of convenience a tablet containing 0.02 gram of copper sulphate and 0.01 gram each of sodium carbonate and calcium hydroxide. Volumetric Analysis of Ion-protein Compounds. JACQUES LOEB (Proc. SOC. Exp. BioZ. Med. 1918 16 39).-Gelatin a t > 4.7 combines only with cations and at p < 4.7 only with anions whilst a t the isoelectric point (p,=’4.7) it combines with neither anion nor cation. Curves representing the influence of univalent anions or cations on the swelling. osmotic pressure and viscosity of gelatin are always approximately parallel wit.h the curves representing the amount of anion or cation found in chemical combination with the gelatin. Estimation of Pepsin. JOHN H. NORTHROP (J. Gen. Physiol. 1919 2 113-120).-See this vol. i 193. Influence of Acetic Acid on the Benzidine Reaction. EMMO SCHLESINGER and JULIUS GATTNER ( B e d . RZ&. Woch. 1919 56 706-707; from Chem. Zentr. 1919. iv 526).-The discordant views of different authors on the value of the benzidine reaction in the detection of blood are due to the previously unnoticed fact that the sensitiveness of the reaction is a function of the concentration of the benzidine in the acetic acid. The authors find that a blue substance is the actual product of the benzidine reaction; this is precipitated if the proportion of acetic acid to benzidine is low and is thus removed from further action and remains stable for a long period. I n the presence of an excess of acetic acid however it dissolves with the formation of a very unstable. green solution. Diagnostic conclusions cannot therefore be immediately based on the occurrence of one or other coloration since the quantity of acetic acid used is a decisive factor. The blue colour invariably becomes green on addition of acetic acid. Reduced Eosin as a Reagent. LUCIANO P. J. PALET and AMANCIO FERNANDEZ (Anal SOC. Quim. Argentina 1918 5 119-120).-Traces of blood in dilute solution may be detected by means of the fluorescence produced with an alkaline solution of the leu-derivative of essin. The substitution of fluorescein for eosin is recommended. W. S. M ’ G . F. M. CHEMICAL ABSTRACTS. H. W.