silk is to the production of threads which are hexagonal in section. The sections become angular in consequence of the pressure exerted by one thread upon the other. Similar sections are to be noted in many of the natural fibres from a similar cause. The resemblance to the hexagonal section is determined by the extent to which the fibres press upon one another in all directions. There is a greater tendency towards this hexagonal section in the case of multiple threads produced with rotating nozzles in the coagulating bath. This angularity of the cross section does not detract from, in fact it may rather add to, the sheen and lustre of the product.

Dreaper and Davis, in the above cited communication, show that when fine threads (in the form of multiple threads), varying between o'492 and 3:56 denier, are tested for breaking strain, the breaking strain per denier is greater as the denier becomes less. Which, being put in another way, comes to this-that, within the limits above cited, the strength per unit of cross-sectional area increases as the diameter of the thread diminishes. This is attributed to a solid skin produced upon the surface of the thread, which skin is more or less of a uniform thickness irrespective of the diameter of the thread. Consequently, the proportion of the thick skin bears a much greater relationship to the total cross-sectional area in the smaller than in the larger threads, and it can only be in the case of small denier that the influence of the thick skin, in regard to the increase of strength and so forth, can make itself apparent. From this one is forced to the conclusion that, with ordinary commercial artificial silk, the influence of the "solid skin," as it is called, upon the strength of the product is almost infinitesimal. At any rate, its influence is so small as to be altogether masked by the other various factors which go to determine the physical qualities of the product.

In order to determine whether there could be discovered any relationship between the denier, percentage stretch at break, and the breaking strain, a number of monofils varying between 400 and 330 were produced in a factory under our own supervision, but entirely according to the commercial practice, and the finished skeins were tested by us with the results as given in Table II., which also gives in the last column the breaking strain in grms. per denier.

0'0402

300. O'0200

O'1597

0.1623

1933

1.48

494

0.1650

1'34

481

1'33

40. 00026

350

22.0

488

I'40

0.1674

50. 0'0033 0'0652

340. 00226

330

20'0

506

1.38

O'1724

70. 0.0046

Manufactured September 10, 1912.

260

313

1.36

I'40

0'0921

390. 0'0260

0.1820

1'34

400. 00266 0'1843

The diameter in the above table is given on the assumption that the threads are perfectly cylindrical, whereas this is seldom the case, more particularly in the ase of artificial silk, consisting of a number of filaments twisted together. The tendency in the case of artificial

Mean 237

These monofils were made in February, 1911. As we wanted to extend the series we arranged with the factory to send us lower denier down to 190. These we tested in a similar way. It will be observed that the average grms. per denier of the threads made in September, 1912, was 1.36 (varying in deniers between 280 and 190), and

those produced by us in February, 1911, of the higher denier gave an average of 138. The average specific breaking strain is practically identical in the two sets.

It will here be noticed that the breaking strain in grms. per denier is remarkably constant, being in the neighbourhood of 1.40, and that therefore within the limits above recorded there cannot be said to be any influence as the result of variation in denier. The stretch at break shows a greater variation than the specific strength, but these variations do not appear in any way to be fixed or influenced by the denier. As will hereafter be shown, there are many factors which have a marked effect upon the specific strength and elasticity of artificial silk and monofils, and these are more than likely to preponderate against any change which might be brought about by variation of diameter. These remarks, however, were addressed either to single or multiple threads which are in the neighbourhood of 10 denier and upwards.

It is frequently asserted that, if part of an iron nail or wire is rendered passive, the remainder of the nail or wire also assumes this state. That this is not the case is shown by the following experiment.

An iron nail, three inches long, was immersed to a depth of one inch in strong nitric acid. The nail was allowed to drain, and the other end was now immersed to a depth of one inch in dilute nitric acid. It was found that this end of the nail was in the active, or ordinary, state, while the end which had been immersed in the strong nitric acid was found, on testing, to be still passive. If, however, we render passive part of a nail, by immersing it partly in strong nitric acid, and then lower the nail, passive end first, into dilute nitric acid until it is wholly immersed, it will be found that, after a brisk evolution of gas, which ceases after a second or two, from the part of the nail not rendered passive in the strong nitric acid, that the whole of the nail is in the passive state.

This experiment was repeated with a piece of iron wire shaped like a V. Part of one limb of the V was made passive by immersing it for about half its length in strong nitric acid. The wire was allowed to drain. The part of the limb rendered passive was was now immersed in dilute nitric acid. The other limb of the V was now brought into the dilute nitric acid until about half its length was immersed. After a brisk evolution of gas, which stopped in from two to four seconds, the part of the limb immersed in the dilute acid assumed the passive condition. The apex of the V was now dipped into dilute nitric acid and was found to be active, the parts which had been rendered passive still remaining so.

The same things were observed when a passive iron nail was connected by means of a platinum, or copper, wire to a nail of ordinary iron.

Two iron nails were pushed through a cork which fitted into a test-tube containing dilute nitric acid. One of the nails was rendered passive by strong nitric acid, washed, and allowed to drain. Each of the nails was connected to a terminal of a mirror galvanometer. The nails were now placed into the test tube containing the dilute nitric acid. A deflection of the galvanometer occurred, which decreased until it was zero, when the evolution of gas from the ordinary iron ceased. The direction of the deflection showed that the passive iron was the positive pole and that the ordinary iron was the negative pole.

It was found that, in addition to passive iron, carbon, platinum, and gold, which acted as positive poles, had the power of producing the passive state as described in the previous paragraph. When, however, poles of passive iron obtained by dilute nitric acid, and aluminium or zinc were used, it was found that passivity was destroyed. The aluminium and zinc in this case were the negative poles.

From these experiments it is evident that the passivity produced in the dilute nitric acid is simply due to anodic polarisation. The condition produced when iron is immersed in strong nitric acid may be due to some definite arrangement of the particles at the surface of the iron, as in magnetism. This would explain why passivity vanishes in a magnetic field, inasmuch as the arrangement of the particles would be disturbed. This view is quite probable, because a piece of magnetised iron, where a definite arrangement of the particles at the surface occurs, dissolves more slowly, especially at the poles, than does a piece of ordinary iron.

THE TENDENCY OF ATOMIC WEIGHTS TO
APPROXIMATE TO INTEGRAL AND
SEMI-INTEGRAL VALUES.*

By ERNEST FEILMANN.

THE fact that the atomic weights of the elements so often approximate to integral values, when measured in terms of the usual units, must strike any observer, and has, indeed, given rise to many speculations on the nature of the elements themselves.

The present communication deals with the relations between the deviations of the atomic weights from integral values on the one hand, and the number of elements exhibiting such deviations on the other.

The atomic weights of the International Table for 1913 were rounded off to the nearest tenth of a unit, and then sorted into ten groups according to the magnitude of the decimal portion of the figures so obtained. Where the figure in the table was exactly midway between two values, that is, where the second decimal place had the value 5, the corresponding element was considered to be shared by the two adjacent groups, each of which was credited with

half a unit.

The groups were :

9

3

ΤΟ

Value of decimal 0'0 0'1 02 03 04 05 06 0.7 0.8 0.9 Number of elements... 23 4 7'5 45 7'59'5 5 From this table the upper curve (continuous line) in the figure is plotted. The symmetry of this curve it remarkable, considering how many gaps still remain in the periodic table of the elements and other irregularities, such as errors in the atomic weight determinations of the less known elements. The other obvious characters of the curve are, first, the very marked rise as integral values are approached, and, secondly, the smaller maximum in the neighbourhood of 0.5.

Although all the figures given in the atomic weight table might be supposed to be significant, it was thought advisable to plot a further curve, using those atomic weights

• Read before the Chemical Society, November 21, 191,

NEWS

The iodide solutions were prepared by solution of the hydroxides in strong hydriodic acid, evaporation, extraction of the free iodine with carbon disulphide, and solution of the resulting iodides in acetone. The bromide solutions were prepared in the same way, omitting, however, the carbon disulphide treatment. The nitrates were dissolved directly in acetone.

In certain cases reactions do not take place with the nitrates or bromides, but do with the iodides; or do not with the nitrates, but do with the bromides; hence some of the general reactions have frequently been tried with more than one salt dissolved in acetone.

2. GENERAL REACTIONS.

A. Reactions with Common Acids and Study of the Halides. Hydrochloric acid yields insoluble earth chlorides. The acid may be added either by passing the dry gas into the earth solution or by employing a solution prepared by diluting concentrated aqueous hydrochloric acid with a considerable volume of acetone. Upon addition of hydrochloric acid white precipitates of the earth chlorides appear immediately in the form of an emulsion from which crystallisation proceeds slowly and incompletely. precipitates are soluble in large excess of the reagent. This solvent action is best shown by continuing the passage of gas for some time when complete solution is effected. Dilution with acetone re-precipitates the chlorides.

The

Yttrium group chlorides are not precipitated by such salts as cupric, stannous, and ferric chlorides in acetone. However, acetone solutions of a number of common chlorides dissolve the solid yttr um group chlorides, which in the absence of chlorides of this character are practically insoluble. Zinc, bismuth, ferric, cupric, antimonous, stannous, cobaltous, and mercuric chloride solutions in acetone dissolve the earth chlorides quite readily. Cadmium, arsenious, and uranyl chlorides act slowly but appreciably. Upon evaporation the copper, cadmium, and cobalt clorides give crystalline products.

Mercuric chloride in acetone added to an acetone solution of yttrium group iodides gives no precipitate imme

diately, but in a short time a white precipitate of the

chloride forms which becomes heavier as more mercuric chloride is added. With excess of mercuric chloride the precipitate re-dissolves to a clear solution. Addition of yttrium iodide solution again causes a precipitate to form, but, on the other hand, an excess of the iodide also gives a clear solution. If the solution is concentrated mercuric iodide precipitates, but this can be avoided by dilution with acetone, in which it is soluble. To ascertain the molecular proportions existing between HgCl2 and YI, standard solutions of the two were prepared in acetone. By numerous titrations the following ratios were indicated to exist, although the results are only approximate inasmuch as the end-points were rather obscure: 2YI3.HgCl2, 2YI3.3HgCl2, YI3.2HgCl2.

The formation of double chlorides indicated above was tried on mixed rare earth chlorides of both the cerium and yttrium groups. Cupric, bismuth, stannous, and ferric chloride solutions in acetone were found to dissolve large quantities of the solid chlorides, although only a very slight amount was soluble in acetone alone. This strengthens the view that double chlorides are formed in solution. On account of the ready solubility of the chlorides in the corresponding solvent it has not been possible to effect a fractionation of the earth by this means. Crystallisation of double bromides was attempted by evaporating solutions of calcium, cadmium, sodium, and bismuth bromides with the rare earth bromides in acetone solution. The products were of syrupy consistency, and not crystalline. Attempts to crystallise double iodides gave similar results.

Silver nitrate in acetone gives a complete precipitation of the halogens from acetone halide solutions, although the first few drops of reagent cause no precipitate to form. Volhard's method for the estimation of silver or determi

nation of chloride can be carried out in acetone by adding in the above reaction an excess of standard silver nitrate solution in acetone and titrating the excess of silver present with a standard acetone solution of ammonium sulphocyanate, using a ferric salt as indicator. Ferric nitrate crystallised from fuming nitric acid, then dissolved in acetone, can be employed. Thus prepared the ferric solution is only moderately permanent, the iron gradually precipitating. One can prepare a dilute solution of ferric chloride or sulphocyanate and use measured portions for the indicator. Naturally under these conditions a “blank" must be subtracted for the volume of silver nitrate necessary to react with the measured volume of indicator.

Hydrobromic acid precipitates the bromides from concentrated iodide, but not from the nitrate earth solutions. When sodium iodide in acetone is added to a rare earth bromide acetone solution sodium bromide precipitates, the earth iodide remaining in solution.

Hydrofluoric, sulphuric, oxalic, citric, and mucic acids precipitate the earths completely as the corresponding salts when acetone solutions of the acids are added to yttrium, neodymium, cerium, or lanthanum nitrates dissolved in

acetone.

Tartarie and phosphoric acids precipitate the earths almost completely. The phosphates are soluble in large excess of acid. Malic acid gives almost complete precipitation upon standing.

Formic, lactic, maleic, and succinic acids give incomplete precipitation of the four earths studied. Lactic acid precipitates the earths completely from the iodide solutions.

cinnamic, or stearic acids do not precipitate the corre Hydrosulphuric, propionic, benzoic, salicylic, hippuric, sponding salts of yttrium, neodymium, lanthanum, and and stearates are but very slightly soluble in acetone. cerium. However, the cinnamates, benzoates, hippurates, Stearic acid gives partial precipitation of the earths from

the iodide solutions.

The

Lactic acid yields white gelatinous lactates with the nitrates of the yttrium earths. The precipitation is almost complete. The lactates are very soluble in water, and very soluble in dilute ammonia. Stronger ammonia (sp. gr. o'9) dissolves the lactates, forming two immiscible layers both containing yttrium. Warming re precipitates the earths from the ammoniacal solutions. When lactic acid in acetone is added to a moderately strong solution of cerium nitrate no precipitate appears. Dilution and prolonged standing cause the lactates to settle out. solubility of the cerium precipitate in water and its deportment with ammonia resemble yttrium. Neodymium precipitates like cerium. Lanthanum is more difficult to throw down with lactic acid, the precipitate forming much more slowly than with cerium or neodymium. The yttrium group precipitates so readily and the others so slowly that without doubt lactic acid can be applied as a rapid fractionation agent for the concentration of this group away from the others.

B. Reactions with Bases. The reaction of a number of organic bases were tried in acetone with the earth nitrates, but they gave no precipitates or general appearance of reaction. Among these were aniline, ethylaniline, acetamide, naphthylamine, diphenylamine, phthalamide, pyridine, quinoline, and urea. Benzylamine gives partial precipitation. Phenylhydrazine added to a concentrated nitrate solution of the earths gives two immiscible layers, the lower layer containing practically all of the earths. The lower layer is slightly pink and is miscible with acetone, hence is not obtained in dilute

solutions.

Ammonia Reactions.

When anhydrous NH3 is passed into an acetone solution of yttrium nitrate, lanthanum nitrate, cerium nitrate, or neodymium nitrate, a heavy white precipitate forms which contains varying amounts of earths, nitric acid, ammonia, and acetone, the first and last named being the highest in percentage composition. These precipitates are difficult to handle, inasmuch as in many cases when desiccation is utilised to free the precipitates from the acetone held in loose combination, decomposition occurs, yielding a dark coloured mass which contains considerably more nitrogen than corresponds to the nitric acid and ammonia content. The nitrates are not well adapted for the study of this reaction, due to the difficulty of keeping the conditions sufficiently anhydrous. Naturally if water is present in appreciable amounts the hydroxides are formed in part at least. Decompositton due to oxidation is another factor with the nitrates. Some of the compounds obtained were semi-explosive when heated, hence the oxide value could not be obtained by direct ignition. This reaction is being further investigated.

Reactions with the Alkaloids.

Acetone is a good solvent for the alkaloids. When a number of the alkaloids are dissolved in acetone and the acetone solution added to an acetone earth solution, compounds are formed which contain the earth nitrate and the alkaloids, hence they appear to be a new type of alkaloidal compounds.

The quinine compounds of cerium, lanthanum, neodymium, and yttrium are precipitated by the addition of an excess of acetone solution of quinine to the acetone earth solution as white amorphous bodies (the nitrates were employed). The precipitates in the case of yttrium, lanthanum, and neodymium are soluble in the earth nitrates. The cerium compound is precipitated with the first drop of alkaloidal solution, hence is not soluble in excess of earth nitrate. All are soluble in water.

Solutions of lanthanum, cerium, neodymium, and yttrium earth nitrates in acetone were treated separately with acetone solutions of quinine, and the precipitates handled like those with brucine. The acetone is very difficult to remove from the quinine. The compounds 4LaONO3.C20H24N2O2, 4NdONO3.C20H24 N2O2, 4Y1ONO3.C20H24N2O2, and 4CeONO3.C20H24N2O2 were indicated to exist.

Cinchonidine is sparingly soluble in acetone, hence ethyl alcohol was used as the solvent. An excess of reagent precipitates yttrium, lanthanum, and neodymium slowly. Warming hastens reaction and precipitates cerium immediately. The precipitates obtained with the first three are soluble in excess of earth nitrates, but not so with cerium. All are white compounds, soluble in water. Cocaine dissolved in acetone yields with the earth solutions white precipitates soluble in water.

Sanguinarine is readily soluble in acetone and yields yellow compounds, forming readily with lanthanum, more slowly with cerium and neodymium, and still more slowly with yttrium. The compounds are soluble in water. The precipitates have a tendency to turn red on standing.

Chelerythrine is readily soluble in acetone and gives yellow precipitates with the earths, with lanthanum quickly, with cerium not so rapidly, with neodymium more slowly

and with yttrium still more slowly. The precipitates are soluble in water.

Piperidine is readily soluble in acetone. It gives white precipitates with the earths, which are almost completely insoluble in acetone but are soluble in water. Hyoscyamine is soluble in acetone. With the earths in excess no precipitate is formed, but with alkaloid in excess white precipitates are obtained. All are soluble in water. Brucine is readily soluble in acetone. With lanthanum, cerium, and neodymium precipitation ensues immediately with the first drop of alkaloid added and gradually becomes heavier, although complete precipitation does not occur for several days with an Yttrium excess of alkaloid. requires an excess of alkaloid to start precipitation. The products are all white, except that of neodymium, which has a pink tinge. All are soluble in water. Weak acetone solutions of the nitrates of lanthanum, cerium, neodymium, and the yttrium group were treated with an excess of brucine dissolved in acetone. The pre cipitates were filtered and washed with acetone by suction, then dried in vacuo over potassium hydroxide for several days. The precipitates retained a large amount of acetone which could be satisfactorily removed only by prolonged desiccation under diminished pressure.

Upon analysis the following ratios were found to obtain: La(NO3)3. C23H26N2O4. 2Nd (NO3)3.C23H26N2O4, Yt(NO3)3. C23H26N2O4, and Ce(NO3)3.C23H26N2O4. alkaloid may be functioned as indicated or as basic earth alkaloidal nitrates.

The

Morphine in acetone gives a white precipitate with the earths, requiring an excess of reagent in case of yttrium, lanthanum, and neodymium, while with cerium the pre cipitate forms immediately with the first drop of alkaloid solution. The first three compounds are soluble in excess of earth nitrate. All are soluble in water.

Coniine is miscible with acetone in all proportions. Coniine gives a white precipitate with yttrium nitrate, requiring a larger excess of alkaloid than with the other three earths studied. When added to the earth solution in quantity just sufficient to give a permanent precipitate, the precipitate is soluble in water. When added in larger amounts so that the alkaloid is in excess the precipitate is not soluble in water. Neither precipitate is soluble in alcohol. This deportment indicates the formation of at least two compounds.

With lanthanum the reaction is similar except that the precipitate forms with less excess of alkaloid. With cerium the reaction is also similar to that of lanthanum, but does not show the water-soluble compound to such a marked degree.

With neodymium the reaction is analogous to that of lanthanum.

Strychnine is almost insoluble in acetone, hence a solution in ethyl alcohol was employed. With all four of the earths white precipitates are slowly formed. They are all soluble in water.

Leucine dissolved in warm ethyl alcohol (in which it is only slightly soluble) yields white precipitates soluble in

water.

Cinchonine, narcotine, and piperine give no precipitates with the four earths studied.

(To be continued)

Rotatory Power of Camphor Dissolved in Carbon Tetrachloride.-A. Faucon. - The rotatory power of camphor dissolved in carbon tetrachloride is independent of the time which has elapsed between solution and the observation. It increases with the concentration and also with the temperature. The variation of the rotatory power under the influence of the temperature depends upon the concentration; thus an increase of 1° increases the angular deviation by a larger amount in the case of concentrated solutions than in that of dilute solutions, and the augmentation is greater at about 12° than at 40°.-Bull. Soc. Chim. de France, xi.-xii., Nos, 20-21.