by the action of dimethylaniline in the presence of phosphoryl chloride, dyes wool in slaty-blue shades on all mordants. 3 45 7-Tetrahydroxy-2-m-p dihydroxyphenyl: 4-benzopyran, prepared by the reduction of quercetin by sodium amalgam in alcoholic hydrochloric acid solution, dissolves in alcohol with a magenta colour, and in potassium hydroxide to a green solution, but is very readily oxidised to quercetin.

Several other derivatives of quercetin and 2-phenylI : 4-benzopyran were also prepared.

312. "An Improved Apparatus for the Determination of Molecular Weight by the Landsberger-Sakurai Method." By WILLIAM ERNEST STEPHEN TURNER and CORNELIUS

THEODORE Pollard.

The great convenience of the Landsberger-Sakurai method of determining molecular weights, especially as a time-saver (see Note 1), has led to its adoption in principle in a number of pieces of apparatus devised by subsequent investigators. Most of these forms, however, are unsuitable for accurate work, and in a paper which discussed the various sources of error in the Landsberger-Sakurai method (Turner, Trans., 1910, xcvii., 1184) an apparatus was described which enabled rapid and accurate measurements to be undertaken.

During the course of some three years' work with this apparatus, several improvements have suggested them selves, and have been collected in the new form figured below.

The most important alteration of the original is the use of the boiler as the constant temperature jacket (see Note

T2

T

tube EF, the entrance E being 12 cm. from the lower end of the molecular-weight tube. The outer jacket CD, of approximately 5 cm. diameter in the cylindrical portion, fits the molecular-weight tube at a second ground joint at b, and carries a safety-tube GH, provided with a tap TI. Although shown in one piece, the tube sealed into the boiler reached only the level of the tubulated stopper, and the tap was connected by rubber tubing, the ends of the tubes being in contact. Whilst a determination is in progress, Ti is usually closed in order to drive a steady stream of vapour through E, but when the molecularweight tube has been removed for weighing, a cork is inserted in the mouth of the boiler, and T is opened to admit dry air, drawn in through drying tubes connected at G. Hygroscopic solvents are thus protected.

Vapour escapes by the tube K, of length as short as possible, connexion with the condenser being made at a third ground joint, so that the molecular-weight tube, together with the thermometer and escape tube, can be lifted away bodily for weighing, during which process the tap T2 is closed. A small cork may be inserted at E if desirable, and complete protection from moist air thus

secured.

Of other points it may be remarked that the molecularweight tube is graduated with marks 2 mm. apart to assist in determining the correction for the use of boiling-point due to increasing head of liquid (alternatively the graduation may be in cc.); that the thermometer used is one of the Beckmann type specially made by Baumbach, of Manchester; and that it is advantageous to protect the cylindrical portion of the outer jacket by two or three thicknesses of asbestos paper, and so diminish the loss of heat by radiation.

The following results were obtained during the course of little more than an hour from the time of setting up the apparatus:—

Pyrogallol in Ethyl Alcohol. C6H3(OH)3 = 126'0.
(Weight of Solute, o 8090 grm.).

D

B

2). For one of the disadvantages of the Landsberger and Sakurai forms of apparatus is that the molecularweight tube fills so readily as to make it difficult, with easily condensable vapours, to obtain more than three or four readings in a series. By using the boiler as outer jacket and making the entrance for the vapour stream high up in the molecular-weight tube, the cooler solvent or solution in the tube is heated considerably before the entrance of vapour, and the amount of condensation thus diminished.

The molecular-weight tube AB, 175 cm. in length, 2.8 cm. diameter in the main portion, and 3'5 cm. at the mouth, carries a ground-glass stopper with two tubulures, fitting flush with the mouth of the tube at a. To the solvent or solution under measurement, vapour is admitted through two perforations at the bottom of the

Several readings did not mark the limit of the capabilities of the apparatus with this solvent, and as no more than five could, as a rule, be obtained with the earlier form, the efficiency of the new apparatus is considerably greater. Ethyl ether, carbon disulphide, and water are easier to handle than alcohol, and with them quite a long series of readings is obtainable. When, however, the boiling point of the solvent exceeds 100°, the number of readings becomes less, so that with amyl alcohol, for example, only four readings were obtained. Above 150° the apparatus may conveniently be employed only in measurements where it is sufficiently accurate to assume that the volume of the solution is the volume of the solvent; for in such a case it is unnecessary to detach the molecular-weight tube, thus avoiding the inconvenience of handling it at a high temperature.

For a discussion of the details requiring attention and the corrections to be applied in accurate work, reference should be made to the previous paper on the subject.

Note 1.-Besides saving time, the rapidity of the process permits the investigation of substances which would decompose during the prolonged boiling involved in the ordinary Beckmann method. Such substances are triethylsulphonium salts (see Turner, Trans., 1911, xcix., 880) and amylamine hydrochloride (Turner, loc. cit.; compare Hantzch and Hofmann, Ber., 1911, xliv., 1776).

(We are indebted to the Chemical Society for permission to reproduce the accompanying woodcut.)

(To be continued).

SOCIETY OF CHEMICAL INDUSTRY. (LONDON SECTION).

Ordinary Meeting, January 5th, 1914.

Dr. W. R. E. HODGKINSON in the Chair.

THE following papers were read and discussed:

"Viscosity of Oils." By J. L. STREVENS. The author, after emphasising the importance of the determination of absolute viscosity and its relation to temperature for any particular lubricant, proceeds to correct certain figures previously published by Dunstan and himself on the basis of the more correct case for the viscosity of phenol at various temperatures due to Dunstan and Thole.

I. Increase of molecular weight favours increase of Zt, e.g., rape oil (trienium) > olive oil (triolein).

2. Hydroxyl formation favours increased Zt, e.g., castor oils and the blown oils.

3. Solution of solid bodies in oils (e.g., "soaping") tends to a high viscosity temperature coefficient.

4. Combined high molecular weight and conjugated double bonds mean a high Zt, e.g., tung oil.

5. Unsaturation tends to lower the viscosity, and broadly the greater the iodine value the lower Zt, e.g., linseed and perialla oils less than nut or olive oil.

Experimental results are given and the recent work of Higgins at the National Physical Laboratory is referred to.

"Oxygen Content of the Gases from Roasting Pyrites." By LEWIS T. WRIGHT.

The author calls attention to the discussion that arose forty years ago between Scheurer-Kestner and Friedr. Bode on the deficiency noted by the former in the oxygen of the gases from roasting pyrites if the pyrites were fully oxidised according to the equation 2FeS2 + 110 = Fe2O3 + 4SO2. The matter was not satisfactorily cleared up by this discussion in which Lunge took part, because the amount of SO3 formed either free or as metallic sulphate in the cinder did not explain the full extent of the deficiency. Lunge suggested there might also be some error in the absorption of the oxygen hy the pyrogallate in the course of the gas analysis. The author, on examining a number of analyses of "burner" gas from various sources, noticed that the oxygen "deficiency" is the greater the greater dilution of the gas, and this suggests that there is, in addition to the well-known production of SO3 and metallic sulphates, some other cause, such as a constant error in the analyses which influences these. In any case the evidence of these gas analyses shows that the manner in which the oxygen is disposed of would prevent the "burner" gas from containing more than about 12 per cent of SO2 as a maximum when all the oxygen of the air applied was used up, and the author states that his various attempts to obtain more than this in practice by keeping burner gas long in contact with incandescent pyrites have failed.

"Electrical Conductivity of Milk during the Concentra tion, with Suggestions for a Practical Method of Determining the End point in the Manufacture of Sweetened Condensed Milk." By L. C. JACKSON, LESLIE MCNAB, and A. C. H. ROTHERA.

The authors have studied the variation of the electrical conductivity of milk during the process of concentration. They find that although the measurement of electrical conductivity is of no value in determining the degree of concentration of a separated unsweetened milk, it can

An ingenious device in which the resistance of sweetened milk in the vacuum pan is compared with that of an approved sample of condensed milk maintained at exactly the sample temperature is described.

Numerous experiments on which the authors base their method are described in the paper.

"Monazite from some New Localities." J. JOHNSTONE.

By SYDNEY

In this paper are given the results obtained and the methods of analysis employed in the examination of twenty-one samples of pure monazite from new localities in Ceylon, Travancore, Nyassaland, Malaya, Northern and Southern Nigeria. The results show that wide variation may occur in the quantity of thoria present in samples; notable amongst these are ranges shown by those from Ceylon whose thoria percentage varies from 9.5 to 28.2, from Malaya 3'4 to 9'4, and from Northern Nigeria 2.3 to 8.0.

Comptes Rendus Hebdomadaires des Séances de l'Académie des Sciences. Vol. clvii., No. 20, November 17, 1913. Thermochemistry of Acetylenic Compounds. Charles Moureu and Emile André.-The heat of hydrogenation of acetylenic compounds is considerable. It is of the order of 80 cal. in the fatty series and decidedly less in the aromatic series. The fixation of the first half of the hydrogen (2 atoms) to give an ethylenic compound generally liberates more than half of the heat disengaged in the complete saturation. The excess of energy of the triple bond over the simple bond is about 70 cals. in the first members of the series. The fixation of water by an acetylenic hydrocarbon, with formation of a ketone, liberates a quantity of heat of the order of 40 cals. Influence of Silicon on the Solubility of Carbon in Iron. Georges Charpy and André Cornu. Silicon gradually diminishes the solubility of carbon in iron until it becomes practically equal to zero at 900°, when the amount of silicon present exceeds 4 per cent. It also vanishes at 1000°, when the percentage of silicon exceeds 7.

Complex Salts of Uranium.-Paul Pascal. -When uranyl pyrophosphate is dissolved in a solution of the sodium salt the solidification point rises, reaches a maximum, when the two salts are present in the proportion 3PO,Na4: P207(UO2)2. and then descends to a minimum at 2P2O7Na4: P207(UO2)2. Up to this point the solution exhibits none of the properties of uranyl salts. When the solution is saturated with uranyl pyrophosphate the ratio is PO8Na4: P207(UO2)2, and the liquid then exhibits the properties of a strongly diluted uranyl salt. The intermediate pyrophosphate must be regarded as a normal uranyl pyrophosphate of formula (UO2)2(P2O7)4 Nas. The evaporation of its solution gives a gummy mass which treatment with methyl alcohol transforms into a hygroscopic powder, [(UO2)2(P207)3] Nag + 6H2O at 15° (+4H20 at 100°). The saturated solution of uranyl pyrophosphate contains the uranyl sodium salt of the acid. It is a yellow unstable powder, readily transformed into the isomeric double pyrophosphate P2O7(UO)2Na2. H2O. All these complex salts are dissociated when their solutions are heated. Alcoholic solutions of potassium cyanate and uranyl nitrate when mixed give a yellow microcrystalline precipitate of formula [UO2(CNO)4] K2. This complex salt is analogous to the anhydrous cobaltocyanate, Co(CNO), K2. In water it gradually dissociates, giving an orange double salt, 2UO2(CNO)2 + KCNO. An excess of alkaline cyanate causes the formation of the salt UO2(CNO)2 + KCNO, while

ture) is filled up to the mark and boiled as described in the case of water. It has been found convenient to take

VOL. CIX., No. 2826.

THE ESTIMATION OF ALCOHOL IN BEER BY
MEANS OF MALLIGAND'S EBULLIOSCOPE.
By JOHN CANNELL CAIN.

In view of the fact that the methods usually employed for the estimation of alcohol in beer involve the use of more or less complicated chemical apparatus, including often a delicate balance, and the employment of gas and water, it is rather remarkable that determinations with the simple apparatus invented by Vidal and improved by Malligand and Mile. E. Brossard-Vidal (Comptes Rendus, 1874, lxxviii., 1470) seem hitherto not to have been published in this country.

This is the more surprising as it has been abundantly proved that the apparatus furnishes very exact results, and it combines the great advantage that a determination can be carried out in a few minutes, and no chemical appliances are required for its use. Thus Dumas, Desain, and Thenard (Comptes Rendus, 1875, lxxx., 1114) presented a report to the French Academy in which they showed that by means of the ebullioscope the estimation of alcohol in wine could be carried out with great exactness, and Griessmayer (Dingler's Polytech. Journ., 1875, ccxviii., 262) proved the same in the case of German beers. Further, in a long report to the Royal Finance Department, Kristiania, Waage ("Ebullioskopet og dets Anwendelse ved Beskatning af Ol efter dets Alkoholgehalt"; see also Zeit. Anal. Chem., 1879, xviii., 417) made an exhaustive examination of this method, and demonstrated that it yielded accurate results, which he illustrated by comparative analyses of a large number of Scandinavian beers.

It therefore appeared to be of interest to apply this method of estimating alcohol to the case of English beers, and the results confirm those of the authors already mentioned in that the method is shown to be exact, rapid, and capable of being used by anyone who can read a thermo

meter.

The principle of the method depends of course on the fact that the boiling-point of an alcoholic liquid is lower than that of water in proportion to its alcohol content, and it has been proved by the above investigators that the solids present in beers and wines exert no practical influence on the boiling-point as their molecular weights are so great.

In making an experiment distilled water is first placed in the lower part of the vessel (up to the mark inside), the lid screwed on, the thermometer (graduated in figures showing directly the alcohol-content in volume per cent) inserted in place, the condenser filled with cold water and screwed into the corresponding opening in the lid, and the water heated to boiling by means of the spirit-lamp, which should be shaded by cardboard to prevent draughts. When the water has boiled steadily for a short time the mercury thread of the thermometer (which will have moved gradually along the tube) remains at one point, and the scale of the thermometer is moved by means of the controlling screw until the zero point is coincident with the position of the end of the mercury thread. The water may be boiled until steam begins to escape through the condenser, by which time the zero point will have been accurately ascertained. This determination is required on each day that the instrument is to be used, as this renders any other correction for barometric pressure unnecessary.

After having fixed the zero point the estimation of alcohol in beer may be made. The beer (at any tempera

the regular boiling commences. The true boiling-point is not reached until the beer has been boiling steadily (easily recognised by placing the ear near the top of the condenser, when a continuous bubbling is heard) for some time, when the mercury remains stationary. By this time the condenser will have become hot, and if the boiling be continued the mercury advances again and vapour escapes from the top of the condenser. The following readings, taken at fifteen seconds interval, are quoted in order to show the course of the experiment.

A determination by the distillation method gave 5.8 per cent. In order to prove the accuracy of the method in the case of English beers, estimations were made of the alcohol-content of a number of various samples purchased at random. The alcohol was determined first by the ebullioscope and secondly by the distillation method. It will be seen that the agreement is quite satisfactory. The alcohol-content is given in volume per cent (cc. of absolute alcohol per 100 cc. of beer; the weight percentage could of course be obtained by a determination of the specific gravity of the beer and a simple calculation).

Total Silicon.-Weigh I grm. in a porcelain basin, and add to it 20 cc. strong HNO3 (sp. gr. 1'45), and keep it covered on the hottest part of a hot plate. The metal will dissolve with evolution of nitrous fumes. The cover is then removed, and the solution is allowed to evaporate down to dryness. The basin is taken away from heat and allowed to cool. When cool, 15 cc. of HCl is added and the solution is evaporated down and baked, and again dissolved in 15 cc. HCl. When the solution becomes clear, 15 cc. of water is added. This is now boiled and taken away to cool, and then filtered through Swedish filter-paper and washed with warm hydrochloric acid and cold water, and afterwards with hot water till the last trace of acid is removed. The residue is ignited wet in a silica crucible and weighed. It is then treated with HF and H2SO4 in the usual way.

(N.B.-15 cc. HCI and 5 cc. H2SO4 were also tried; the result was not different).

Iron.-Dissolve in a beaker I grm. of drillings in 30 cc. of a 25 per cent solution of chemically pure NaOH. As soon as the solution is clear filter it and wash with hot water till the last trace of alkali is removed. Dissolve the residue in warm HCl (1 : 1), and wash the filter-paper with hot water.

Precipitate Fe with ammonia and estimate gravimetrically. (When there is much copper it requires to be dissolved in dilute HNO3 warm and added to the HCI solution, then precipitate Fe with HN,HO; Cu will show blue filtrate).

Copper.-Pass SH2 through the filtrate from SiO2, and estinate Cu as CuO.

Aluminium.-The filtrate, after the removal of the copper sulphide as above, is boiled to remove SH2, and then diluted to make up 1 litre, 100 cc. of which is taken for estimation of aluminium. Add 1 grm. Na3PO4; dilute to about 200 cc., neutralise with ammonia to precipitation, and just re-dissolve with a drop or two of HCl. Boil and add sodium thiosulphate solution in excess. (N.B.-It thiosulphate is added to the cold solution it bumps violently when boiling). Boil till the smell of SO2 is no longer perceived, and then allow to settle. Filter and wash with boiling water till the washings are free from alkaline phosphate (to be tested with AgNO, solution). The precipitate filters and washes quickly. Ignite in a silica crucible, and weigh as AIPO4. (Factor 22:22 per cent Al). The following are the reactions that take place :

I. Al2C162Na3PO4 = Al2P208 + 6NaCl.

II. A2P208+6HCI - Al2C16+ 2H3PO4.

down to dryness. Dissolve the residue in 25 cc. hot water, and add a drop of NH4HO to see if there is any Al present. Convert to chloride and weigh, &c.

The following experiments may prove of interest :— I O'I grm. aluminium (99 per cent purity) and o'i grm. Zn (99.95 per cent purity) were weighed in a beaker and dissolved in HCl, and aluminium was separated, washed, dried, and ignited; from the filtrate Zn was estimated as pyrophosphate. The following figures were obtained :Al=0.099 grm.; Zn = 0.09995 grm.

2. 0.1 grm. aluminium (99 per cent purity) was weighed in a beaker and dissolved in HCl, and Al was estimated by the method. o'099 grm. was the figure obtained. The filtrate was free from aluminium.

3. 01 grm. zinc was weighed in a beaker and dissolved in HCl. The excess acid was boiled out, diluted, added sodium phosphate, the precipitate cleared with HCI, boiled, and added hypo solution, boiled, filtered, and washed. The filter paper is ignited and weighed.

Ash.-0.00056 obtained, which was the ash of 11 cm. Swedish filter-paper, showing that Al precipitate is not contaminated with Zn. The filtrate is treated for Zn. Zn weighed o‘09995

grm.

The above experiments show the accuracy of the method. Carbon. This has been determined by the sodio-copper chloride process. But this reagent must be added gradually drop by drop over water (25 cc.) in which the aluminium drillings are placed, otherwise a too violent action will take place and some C will be lost. Filter the carbon through asbestos, and estimate it by a wet combustion as usual. Gun and Shell Factory, Ishapore, Bengal.

PASSIVITY.*

By GERHARD C. SCHMIDT.

1. Introduction.

ALTHOUGH the abnormal chemical and electrochemical behaviour of iron and certain other metals has been known

for more than 100 years, up to the present no explanation of the cause of passivity has met with general acceptance. In course of time many theories have been suggested and subsequently given up. At the present time only two theories deserve serious consideration: (a) the oxide theory of Faraday, according to which passivity is due to still finds supporters in spite of the many serious objeca coating of oxide or oxygen on the metal-a theory which tions put forward against it-and the hydrogen theory, according to which the metals in question are normally passive and become active only in the presence of dissolved

IV. Al2C16+2Na3PO4 + Na2S2O3 + 2HCl =

Al2P208+8NaCl +S+ SO2 + H2O. Zinc.-The filtrate from the precipitation of Al is made aminoniacal and boiled till the solution shows amphoteric reaction to litmus-paper. Filter and wash with boiling water. Ignite and weigh as Zn2P2O7 (Zn = 42·91 per cent). The following reactions take place :

I. Na2S2O3 + 2HCl = 2NaCl + H2O+S+SO2. II. Zn3(PO4)2+6HCl -3ZnCl2 + 2H3PO4. III. H3PO4+3NH,HO - (NH4)3PO4 + 3H2O. IV. ZnCl2 (NH4)3PO4 NH4ZnPO4 2NH4Cl. V. 2(NH4)ZnPOZ2P2O7+ H2O+2NH3. Sodium. This is estimated by the usual method, viz., solution in strong nitric acid (1:45 sp. gr.) is evaporated to dryness and baked; 5 cc. HNO3 is added again, and again dried and baked, then ignited to drive off nitrous fumes, cooled, and crushed to powder; added boiling water, boiled, the clear liquid decanted, and the residue washed with boiling water. The whole is evaporated

tions.

recently been strongly supported by my pupils, Grave (Zeit. Phys. Chem. 1911, lxxvii., 513) and Adler (Ibid., 1912, lxxx., 385), on the basis of experimental investigaZeit. Elektrochem., 1912, xvi., 335) does not regard their Flade ("Habilitationsschrift," Marburg, 1910; results as conclusive, and considers that his own investigations lend support to the oxide theory. In the present Paper the results of further experiments designed to

decide between these two theories are described.

2. Criticism of Flade's Results.

Flade found that iron rendered passive by anodic polarisation in dilute sulphuric acid immediately returns to the active state when the polarising e.m.f. is removed. When, however, it is in a circuit and forms the anode. under the influence of an external, not too great, e.m.f., the potential gradually becomes more negative, until at a certain potential, the "transition point," it suddenly

*A Contribution to the General Discussion on "The Passivity of Metals" held before the Faraday Society, November 12, 1913. (Experimental Part by W. RATHERI.)