from hot saturated solution, and the fine crystals thus produced were filtered and washed upon a Buchner funnel with the alcohol. The sucrose obtained was redissolved in water again precipitated with alcohol, and washed as above. The second crop of crystals was dried between blotting paper, and preserved in a glass vessel for

use.

Purification of the grape sugar-Grape sugar (Merck, extra pure) was dissolved in distilled water and the solution was shaken with animal charcoal and filtered. The filtrate was concentrated under reduced pressure to a thin syrup, shaken with an equal volume of hot 96 per cent. alcohol and filtered from precipitated impurities. The alcoholic filtrate was evaporated under reduced pressure to a syrup and crystallised in a desiccator.

The sugar solution was shaken with animal charcoal for forty minutes and then separated from charcoal by a centrifugal machine. The amount of sugar adsorbed was calculated from the difference of sao charimeter readings before and after treatment. Ventzke readings were measured at constant temperature (20° C.) by using a glass polarization tube with metal jacket, the Zeiss spiral heater and water-pressure regulator being used for controlling temperature.

The results of the experiments may be summarised as follows:

adsorbed by

The amount of sucrose animal charcoal from its aqueous solution, increases with increase of its concentration, but reaches a maximum when the concentration of sucrose solution becomes about 0.2 mol per litre, and then decreases gradually.

When the concentration of sucrose solution is smaller than 0.11 mol per litre, Freundlich's equation holds good for the adsorption of sucrose by animal charcoal. The author has found the following empirical formula:

m

= 0.2829 C0.1364

The amount of sucrose adsorbed by animal charcoal is directly proportional to that of charcoal added. Sucrose is adsorbed considerably by vegetable decolorising carbons (Norit, Darco and King). Grape sugar is also adsorbed by animal charcoal in such amount that the error through adsorption is not negligible in sugar analysis.

GENERAL NOTES.

EXPERIMENTS WITH DRUGS IN MALAYA.

A number of drugs of minor importance, totalling about ten sorts, are being grown at the Serdang Experimental Plantation, Federated Malay States, and enquiries for several of these have recently been received from England. A sample of Ipecacuanha roots was reported on favourably, and an offer made to purchase crops produced. Three species of Chaulmoogra oil trees are under cultivation for the production of seeds in the leper asylums in Malaya. It is hoped to have 30 acres planted with each species.

INSTITUTE OF METALS.

During the past year the membership of the Institute of Metals increased from 1801 to 1903. It is anticipated by the Council that the present year the twentieth of the Institute's existence-will witness the enrolment of the 2,000th member. Particulars of membership are contained in a booklet descriptive of the work of the Institute, which can be obtained from Mr. G. Shaw Scott, M.Sc., Secretary, Secretary, 36, Victoria Street, London, S.W.1. The annual general meeting of the Institute will be held in London on March 7-8, and the autumn meeting in Liverpool on September 4-7.

FOREIGN EXCHANGES IN 1927. The principal events of the year in the foreign exchange market have been the continued advance in the value of the £ in New York from August onwards, and the recent announcement that the Italian Government had decided to return to the gold standard. The firmness of the leading South American exchanges in 1927 was noticeable, and the return of the Argentine Government to the gold standard occurred in August.

The closing quotation for the £ in New York was 4.858 on 1 January, 1927, and from this rate it weakened to 4.849 (the lowest closing quotation recorded in 1927) on the 4, 5 and 7 February. There was a firmer tendency in March, the four weekly averages of 4.852, 4.853, 4.855 and 4.857 indicating the steadiness of the exchange during that month. At the end of August the closing rate was 4.862, parity level (4.866) was reached on 20 September, and on the 24 September the closing rate (4.867)

was slightly in excess of parity. This improvement was more than maintained in October and later months; the closing rates in December varied between 4.880 and 4.884, the quotation for the last day of the year being 4.88 7-32. On balance sterling gained about 2 cents in 1927.

Italian lire, which closed on 31 December, 1926, at a rate of 108.063 to the £, depreciated during January and the first week of February, the highest closing quotation of 1927 (114.063) being recorded on 7 February. From that date there was a steady improvement, the closing rates on the last days of February, March and April being 110.850, 105.700 and 91.625 respectively. The sustained demand for lire caused a rate of 83 to be touched at one time on the 26 April, but on profit-taking sales the rate re-acted to 89.625 at the close on that date. On 2nd May the lira declined in value to a rate of 95.125 to the £, but it recovered to 88.188 on the last day of that month. June saw the rate forced down, probably by speculative influences, to below 87 on several days, the closing quotation on the 25 (88.70) being the most favourable recorded to the currency during

.1927.

SPAIN.

still

One of the few currencies which lends itself to speculative operations is the Spanish peseta, and the sharp fluctuations which occurred in this currency during 1927 were mainly due to these influences. The closing rate on 31 December, 1926, was 31.70 and on 30 April, 1927, it was 27.590. The monthly averages for the four months January to April (30.322, 28.899, 27.899, and 27.535) show that the peseta, on balance, was steadily appreciating, the impression prevailing that it was the intention of the Spanish Government to allow the rate to approach near parity and then In to return to the free gold standard. November the closing rate again rose, the quotations on the 1 and 30 being 28.50 and 29.730 respectively. December showed a further depreciation, the closing rate on the 31 being 29.89. On balance the peseta appreciated during 1927 by about 6 per cent.

FRANCE.

French francs were quoted at 122.800 to the £ at the close on 1 January, 1927, at 123.175 on 1 February, and at 123.995 on 1 March. From that date until the end of the year the rate was virtually stationary at around 124.

SCANDINAVIA.

Of the Scandinavian exchanges, the Norwegian appreciated on balance in 1927, as compared with sterling, by about 4 per cent., the most favourable quotation of the year (18.32) being recorded in November; the Swedish crown appreciated by about one-third of one per cent., while the Danish crown showed little change in value. Except for a few days early in March and at the end of April the rate for Swedish currency on London was above parity from March to December, and Danish currency was in a similarly favoured position from the middle of July to nearly the end of August. Norwegian currency was still about 1 per cent, below London parity in December.

On balance in 1927 there was no considerable movement in the rates of the Continental gold currencies on London. Reichsmarks steadily depreciated in the first five months, the quotation reaching 20.54 on 11th May, the highest figure recorded since the currency was stabilised; they recovered from June to November, being actually at a premium, as quoted in sterling, on most dates in October and November, and closed the year at an exchange value slightly below parity. Belgas depreciated until the middle of the year, subsequently improving until September.

SIR JOHN CASS TECHNICAL
INSTITUTE.

A pleasing indication that the executive control and administration of chemical works is passing into the hands of chemists is furnished by the announcement that the Sir John Cass Technical Institute has added a short course of lectures on "Law as related to Industrial "Chemistry " to its curriculum. The governing body of the Institute have wisely arranged for the lectures to be given by a professional chemist and practising barrister (Mr. G. S. W. Marlow, B.Sc., F.I.C.), and by holding them in the evenings (Tuesdays at 7 p.m.) they have made it possible for those already engaged in industrial work to attend. The necessity for some legal knowledge by anyone holding a responsible executive appointment is too obvious to require comment, and the course should prove very popular both with chemists already engaged in industrial work and with students preparing for an industrial career. At the first lecture, which is to be held on January 24, the chair will be taken by Mr. James Whitehead, K.C.

The theory of complete ionisation had its origin in a physico-chemical study of the properties of solutions; but it was very soon adopted by physicists as an explanation of the high symmetry of the crystal-structures revealed by the X-ray analysis of metallic salts, since this could not be accounted for on the traditional theory that the ions of the salt were united into molecules. The new theory has had a large measure of success in both fields of study, since, on the one hand, the conductivity of solutions of ordinary metallic salts has been calculated for the first time by means of a rationally deduced formula, which is valid up to a concentration of perhaps N/10; on the other hand, many of the physical properties of crystals, such as their linear dimensions and angles, heat of formation, elasticity, refractive index and infra-red absorption, have been calculated successfully on the assumption that the crystals are aggregates of ions of varying degrees of "hardness," which are drawn together by the electrostatic attraction between their opposite charges, but are also repelled from

66

one

another with a force varying inversely as (say) the 6th to the 35th power of the distance. On the other hand, the theory of complete ionisation (which cannot be expected to " have it both ways ") fails to cover the behaviour of those weak electrolytes which obey Ostwald's dilution Law, just as completely as the latter law failed to cover the behaviour of highly-dissociated metallic salts. Again, although many of the properties of solid ionic aggregates, such as rock salt or calcite,have been predicted successfully from fundamental data, all attempts to predict the dimensions and behaviour of the carbonate ion itself have failed, since this does not behave as a mere aggregate of carbon and oxygen ions. The object of the present paper is therefore to determine the boundaries of the region within which the theory of complete ionisation is valid, to find out the factors which make it inoperative in so many cases, and in particular to study the chemical significance of the classification of electrolytes as "strong" and "weak." The applications

of similar considerations to the study of crystal-structure will form the subject of a subsequent communication.

In both cases the principal argument put forward will be that, although the ions of a salt are generally prevented from neutralising their opposite electric charges by factors which find their simplest expression in the "octet " rule, there are many cases in which this inhibition does not operate, with the result that the ions can be converted into real molecules, as distinguished from mere aggregates of ions, by the neutralisation of their opposite charges with formation of a chemical 66 bond.' These two groups of compounds correspond broadly with the 66 66 and strong weak electrolytes; but the two methods of classification cannot be expected to lead to identical results. Thus, on the one hand, a salt which consists entirely of ionic aggregates may be placed under conditions which are so unfavourable to the independent migration of the ions as to bring it within the group of "weak electrolytes "; and, on the other hand, the bond which holds the positive and negative radicles together in a covalent molecule may be so weak as to interfere but little with the disruption of the molecule into ions, with the result that a compound of the second type may exhibit the behaviour of a strong electrolyte.

II. ELECTROLYTIC DISSOCIATION OF
STRONG ELECTROLYTES.

A "strong electrolyte " means, in the first instance, nothing more than a substance which has a high electrolytic conductivity, e.g., in aqueous solutions. In the language of Arrhenius' theory,such a substance was said to have a large "coefficient of ionisation "; but this term has no longer any real meaning when applied to a salt which is 100 per cent. ionised even in the solid state. We can, however, still interpret the decrease of equivalent conductivity in strong solutions as being due to the formation of electrically-neutral doublets, which play the part of molecules, just as in the vapour of the salt. We need therefore only modify Arrhenius' equation to the extent of writing

culated directly, as was formerly supposed, from the simple equation a = A/A. Moreover, since the readiness with which neutral doublets of oppositely charged ions are formed by the mutual attraction of their electrostatic charges must vary with the linear dimensions of the ion, with the dielectric constant and other properties of the solvent, and with the extent to which the naked ion becomes hydrated or solvated in solution, as well as with the temperature at which the association and dissociation of the ions takes place, there is ample scope, even in the theory of complete ionisation, for variations in the " degree of dissociation " of different salts, either in solution or in the fused state.

(a) Ionisation of Strong Electrolytes in Solution. Is is possible, then, for a completelp ionised salt to behave as a "" weak electrolyte " and exhibit a low "coefficient of ionisation," or, more accurately, a low degree of dissociation" of the pre-existing ions? The answer is undoubtedly "Yes," provided always that the conditions under which this effect is theoretically possible can be realised in practice.

66

а

Of these conditions the most obvious has reference to the nature of the solvent. Thus Hartley has shown that, in the case of N/500 KI,the ratio A/A falls from 0.98 in water to 0.90 in methyl alcohol, 0.81 in ethyl alcohol and 0.77 in acetone, although his solutions were all so dilute that 1 was still proportional to the square root of the concentration as required by the theory of complete ionisation. If then we accept the general validity of Walden's rule, that a dissolved salt gives a definite value for A/A at a dilution which is proportional to the cube of the dielectric constant of the solvent, it follows that we should be able to repress the dissociation of the ions to any desired extent by dissolving the salt in a medium of low dielectric-constant.

At this point, however, a fresh difficulty present itself, since the majority of metallic salts are insoluble except in media which possess marked ionising properties; in particular, most of these compounds are completely insoluble in media, such as benzene or cyclohexane, where the dielectric constant is low enough to compel the salt to behave as a weak electrolyte. Walden has, indeed, put forward a definite rule, according to which the coefficient of ionisation of a salt is approximately constant for saturated solutions in any anhydrous solvent. Thus in the case of tetraethylammonium iodide, we should not expect to find any solution in

is

=

which the "coefficient of ionisation much below 50 per cent., since saturated solutions in a dozen different solvents (but excluding water and all other solvents in which the solubility is more than 1 grammolecule in 2,500 c.c.) gave values for a A/A which were almost constant at 0.48. ceptional cases will it be possible to prepare a solution of a strong electrolyte in which the coefficient of ionisation is so small as to be negligible, and that this effect when it occurs, will probably be due to some special chemical relationship between the solvent and the solute, which enables the solute to dissolve, but without any marked disruption of the ions.

Such an effect mught be looked for in Reychler's trimethylcetylammonium cetylsulphonate

in which the single polar valency in the middle of so long a chain of atoms might very well remain almost undetected. Thus, whilst the 30 carbon atoms in the two hydrocarbon chains should suffice to make the compound soluble in hydrocarbon solvents, these solvents need not necessarily be so drastic in their action as to force the two ions apart. Fortunately, however, this phenomenon can be observed in much simpler compounds. Thus, sodium oleate at a dilution of 100 litres gives the following equivalent conductivities in water and in methyl, ethyl, and amyl alcohols

0.0140 in carbon tetrachloride to 1.57 in chloroform and 9.51 in methylene chloride; but the equivalent conductivities were abnormal in that they generally fell to a minimum value at some intermediate dilution, from which there was not only the usual increase on dilution, but an equally striking increase on passing to more concentrated solutions. The coefficients of ionisation, however, were obviously small and in some cases almost negligible, so that the "salt" was behaving quite clearly as a weak electrolyte " in all these solvents. Neutralisation of the ionic charges, however, would appear to be even more difficult than usual, in view of the fact that the positivelycharged nitrogen of the tetra-amylammonium ion is separated from the negativelycharged halogen ion by a permanent hydrocarbon "atmosphere " containing

6:

Molecular weight Density

Specific conductivity Viscosity

20

The normality of the fused salt is 4715 ÷ 143.34 = 32.9N; and the equivalent conductivity is therefore A = 4.8 ÷ 0.0329 136. The limiting value for the conductivity at infinite dilution in aqueous solution is given by the sum of the ionic mobilities as 54 + 65 119; if, therefore, a proportional allowance is made for an increase of viscosity from 0.01506 to 0.01606, the corresponding limiting value for the fused salt would be A = 119 × 0.01056 0.01606 = 78. The equivalent conductivity of a completely ionised salt can, however, also be calculated, without making use of the value for aqueous solutions, by means of Walden's relation, A M 1 1.15, which has been verified for six salts in 29 non-aqueous solvents, as well as for two "anhydrous salts in aqueous solution. For silver chloride this relation gives A

atoms of carbon. The small conductivity of these solutions then suggests that the salt must be present predominantly in the form of neutral ionic doublets, rather than as freely dissociated ions. The formation of these doublets also appears to afford a sufficient explanation of the different absorptioncoefficients which a sufficient explanation of the different absorption-coefficients which Hantzsch observed when quaternary ammonium salts of this type were dissolved in organic solvents instead of in water, since we need not suppose that the absorptioncoefficients of the ionic doublet must necessarily be identical with those of the aqueous ions.

(b) Ionisation of Fused Salts.--Since considerations of limited solubility do not apply in the case of fused salts, it is of special interest to enquire what is known in reference to the magnitude of the "coefficients of ionisation" under these conditions. It has been generally assumed that the calculation of these coefficients is impossible, since it cannot be done either by Arrhenius' method, which depends on measuring the conductivty of the salt at different stages of dilution with a solvent, or by van't Hoff's method, which implies a knowledge of the osmotic pressure of the salt in solution. A formal solution of the problem is possible, however, in the case of fused silver chloride at 600° C., for which the relevant data are all known, namely: