sodium hydroxide, a result which does not appear to be due to any modification of the -CN group, nor to conversion into ẞ-hydroxybutyronitrile. It is not possible to state whether the isomerisation of sodium vinylacetate (Bruylants, A., 1924, i, 1053) in sodium hydroxide solution results in the formation of a single variety of sodium crotonate. The spectra of cyclopropanecarboxylic acid and its nitrile show a much more intense absorption than is indicated by their formulæ. When the intensities of absorption are compared, vinylacetic acid occupies a position intermediate between cyclopropanecarboxylic acid and the crotonic acids, whilst the position of the double linkings in the various acids exercises a considerable influence on the nature of the curves.

The nitrile of cyclopropanecarboxylic acid (Bruylants and Stassens, A., 1923, i, 213) in hexane, water, and N/20-sodium hydroxide gave spectra showing a band with a strong inflexion at 2400 Å., e=1. The nitrile of methylacrylic acid, b. p. 91.3 91.5°/763 mm., in hexane solution showed a moderately marked inflexion at 2400, e=2.7. Vinylacetonitrile (A., 1922, i, 817) in hexane and in water showed an inflexion at 2350, e-1.75. In alkaline solution, the spectrum varies with the concentration and with the interval between preparing the solution and taking the photograph. With this interval at a minimum, the spectrum of a solution of N/100-nitrile in N/10000-sodium hydroxide is very similar to those in water and in hexane. The displacement of the spectrum and increased absorption, which reach a maximum after 15 days, using the above solution, is much more rapid in more concentrated alkaline solution, and the maximum effect is obtained after 10 days, whilst if N/20-sodium hydroxide is used the change is immediate. The curves obtained using varying concentrations show the isomerisation of vinylacetonitrile into crotononitrile. Crotononitrile, b. p. 108°, in the three solvents hexane, water, and N/20-sodium hydroxide, gives a continuous spectrum. Crotononitrile, b. p. 121°, on the other hand, gives a spectrum showing a marked incurvation in the region of high concentrations. B-Hydroxybutyronitrile, b. p. 102°/10-5 mm. (Henri, A., 1899, i, 182) in aqueous solution shows a slight inflexion at 2650, e=0.87. The spectra of alkaline solutions vary with the concentration, that of N-nitrile in N/20-sodium hydroxide being almost identical with that of an aqueous solution. cyclopropanecarboxylic acid (Bruylants and Stassens, loc. cit.) in hexane solution shows an inflexion at 2500, e=35. Vinylacetic acid (A., 1924, i, 1053) in hexane, water, and sodium hydroxide shows weak inflexions at 2500, e-14 in hexane, and at 2500, e=17.5 in alkali. On keeping the alkaline solution, a progressive displacement to the red and an increased intensity of absorption and more marked inflexions are observed; these reach a maximum after 16 days, at 2604, e=30. Crotonic acid in hexane solution shows a weak inflexion at 2450, e=175; in aqueous and alkaline solutions, it is a little more pronounced, at 2620, e-42. Similarly, isocrotonic acid, m. p. 15.5°, shows a weak inflexion at 2450, e=359. The curves for B-chlorocrotonic acid, m. p. 94°, and ẞ-chloroisocrotonic acid, m. p. 61°, are also given. J. S. H. DAVIES.

Correlation of absorption spectra with ionisation in violuric acid. R. A. MORTON and A. H. TIPPING (J.C.S., 1925, 127, 2514-2517).-Aqueous solutions of violuric acid containing known amounts of alkali or acid or salt have been examined spectrographically and p values have been determined by means of the quinhydrone electrode. The colour band at 5380 Å. is common to the violurate ion and to the undissociated sodium salt, but not to the undissociated acid, which is therefore colourless. In the presence of varying quantities of mineral acid, the intensity of the absorption is quantitatively proportional to the violurate-ion concentration. Violuric acid in aqueous solution shows two bands in the ultra-violet, one with a maximum at 3120, e=3000, and the other at 2490, e=10,300. The former is due to the violurate ion and persists in aqueous solutions of the neutral salt, e (max.) becoming 13,300. The latter is due to the undissociated molecule, and it alone persists in alcoholic solution e (max.) becoming 13,400. The sum of the extinctions of the two bands observed in the aqueous solution of the acid is almost equal to that observed both for the undissociated acid (13,400) and for the completely dissociated salt (13,300). E. E. WALKER.

Absorption spectra and lactam-lactim tautomerism. R. A. MORTON and E. ROGERS (J.C.S., 1925, 127, 2698-2701). The work of Hartley and Dobbie on the absorption spectra of isatin, carbostyril, o-hydroxycarbanil, and their O- and N-ethers has been repeated, giving special attention to the position of the maxima of the absorption curves. The absorption spectra of the O- and N-ethers of isatin and o-hydroxycarbanil resemble each other so closely that no conclusion can be drawn with regard to the constitution of the parent substance. On the other hand, Hartley and Dobbie's observations that the absorption spectrum of carbostyril is very similar to that of its N-ether and different from that of its O-ether is confirmed. The similarity between the absorption spectra of phloroglucinol and its trimethyl ether is also confirmed. The following wavelengths of absorption band maxima in Å. are recorded: isatin: 2430, 2950, 4130; 4-methylisatin (N-ether): 2465, 3000, 4195; methylisatin (O-ether): 2447, 2965, 4140; 5-iodoisatin (red form): 2500 and 4250; carbostyril: 2690 and 3270; N-methyl ether : 2705 and 3280; O-methyl ether: 3222 and 3085; O-ethyl ether: 3226 and 3085; o-hydroxycarbanil 2736, N-ether 2738, O-ether 2735; phloroglucinol (in alcohol) 2665, trimethyl ether 2646.

E. E. WALKER.

Infra-red absorption spectrum of molten naphthalene. F. K. BELL (J. Amer. Chem. Soc., 1925, 47, 2811-2816).-The infra-red absorption spectrum of molten naphthalene is in satisfactory agreement with that reported by Coblentz (Carnegie Inst. Publ., 1905, 35, 127) for naphthalene in carbon tetrachloride solution, and shows marked correspondence with the infra-red absorption spectrum of benzene in the region of the shorter wave-lengths, up to 5.5 μ, beyond which the dissimilarity of the two spectra increases. A simple electrically-heated

cell is described for maintaining substances in the molten condition when their absorption spectra are being plotted. F. G. WILLSON.

Influence of pH on the ultra-violet absorption spectra of certain cyclic compounds. W. STENSTRÖM and M. REINHARD (J. Physical Chem., 1925, 29, 1477—1481).-The absorption bands of a mixture of certain amino-acids in water show a shift towards longer wave-lengths when the mixture is made alkaline. Other cyclic compounds have been examined for a shift in absorption bands with a change in hydrogen-ion concentration. The changes in wavelength with på have been followed at a selected value of the extinction coefficient of the substance concerned. Curves for phenol, resorcinol, and tyrosine in water are given. It is found that the bands in the ultraviolet, between 2200 and 3600 Å., depend on hydrogen-ion concentration for aqueous solutions of the following compounds: phenol, tyrosine, resorcinol, p-hydroxybenzoic acid, salicylic acid, and p-hydroxybenzaldehyde. No shift occurs with benzoic acid, phenylalanine, or tryptophan. The relation between PH and structure of the band is that the latter will move towards the red end of the spectrum and increase in intensity when a certain alkalinity is reached. The presence of an hydroxyl group in the benzene ring of the compound seems essential. In the case of tyrosine, the shift is explained as a change from a curve characteristic of the molecule to one characteristic of the compound ionised at the carboxyl group. A curve showing the shift for blood-serum is also given.

L. S. THEOBALD.

Line fluorescence of cadmium vapour. W. KAPUSCINSKI (Nature, 1925, 116, 863-864).-Fluorescence was first observed at about 350°, when the pressure of the cadmium vapour was 0.2 mm.; it rapidly increases with rise of temperature, changing in colour from bluish-green (with intense emission of the arc triplet 5086-06, 4800-09, 4678-37 Å.) to blue. The lines 3261-2 and 2288-8 were observed, and the lines 2748-7 and 2573.1 Å. may also be present. The effect of using an aluminium, iron, lead, copper, or mercury spark instead of a cadmium spark as the source of light is described. A. A. ELDRIDGE.

Dissociation and fluorescence of iodine vapour. E. G. DYMOND (Z. Physik, 1925, 34, 553-561; cf. Franck, A., 1925, ii, 1077).—The absorption bands of iodine were examined which converge at 4995 A., thereafter extending as a continuous absorption of diminishing intensity towards the violet. When of diminishing intensity towards the violet. When iodine vapour was excited by light of wave-length longer than 4995 Å., fluorescence was observed, but ceased suddenly at this wave-length. Hence, up to this point the iodine could radiate the energy it received; beyond it, this ceased to be possible, and

it is concluded that the molecule must have dissociated into a normal atom and an excited atom. The amount of energy left in the excited atom would be represented by 1 volt; this corresponds with the calculated value 0.9 volt for the difference of the terms 2p2-2p1. No evidence of dissociation could be observed by means of a quartz-fibre manometer. Three new series of resonance spectra are described,

excited respectively by lines in the cadmium, copper, and sodium spectra. E. B. LUDLAM.

sec.

Indirectly excited fluorescence spectra. S. LORIA (Physical Rev., 1925, [ii], 26, 573–584).—In the fluorescence spectrum of thallium excited by collision with activated mercury atoms, every line expected theoretically was identified; the life of the 2p, state for thallium atoms is of the order of 106 In confirmation of the results of Donat, enhancement of the lines was observed in the presence of argon or nitrogen, the effect being neutralised by a trace of oxygen. The results indicate that the metastable mercury atom may survive many collisions with normal argon or nitrogen molecules, but easily gives up its energy when colliding with normal thallium or mercury atoms, probably more easily to the former than to the latter. A. A. ELDRIDGE.

Phosphorescent flame of carbon disulphide. H. B. DIXON and W. F. HIGGINS (Mem. Manchester Phil. Soc., 1924-25, 69, 19-23).-A mixture of air and carbon disulphide ignites when heated to 156°, and the addition of carbon disulphide to any combustible gas greatly reduces the ignition point. In general, the substitution of oxygen for air lowers the ignition temperature by about 25°. Methane containing 50% of carbon disulphide ignites rapidly in air at 195°, and in oxygen at 173°. With only 20% of disulphide, it ignites rapidly at 191° in oxygen, but not below 485° in air. Above 200°, luminescent

clouds are obtained in the last mixture. This phenomenon has been investigated. The luminous cloud begins to form at about 180° and a reddish-brown deposit, carbon monosulphide, separates from it. Traces of coal-gas, or of acetylene or ethylene, destroy the luminosity and prevent the formation of the monosulphide. The luminous cloud may be obtained as a large, cool, phosphorescent flame easily passing through metallic gauzes which prevent the passage of hot flames. The "poisoning" gases must be mixed with the gases before combustion in order to produce an effect. Steam, carbon dioxide, hydrochloric acid, and carbon tetrachloride vapour act as inert gases. Sulphur dioxide, although not an active poison like ethylene, has a marked effect, as also have the vapours of ethyl alcohol and ethyl ether. The effect of the paraffins increases with the mol. wt. It is suggested that the following reaction takes place: CS2+02-CS+SO2. The carbon monosulphide molecules coalesce with each other and with disulphide molecules, the latter then being more readily attacked, leaving further monosulphide molecules joined to the original nuclei. Thus the latter increase in size and condensing power until the chemical action at their surface is sufficient to give luminosity. The "poisoning" gases condense preferentially on the monosulphide molecules and prevent the attachment of disulphide. The same apparatus can be used to show the analogous phosphorescent flame of ether vapour. M. S. BURR.

Mechanics of the perturbed molecule. L. MENSING (Z. Physik, 1925, 34, 602-610).-For diatomic molecules of the type in which the electrons of one atom do not also describe an orbit round the

other, the moment of the electrons is in the direction of the line joining the nuclei; for one model in two dimensions the moment is perpendicular to this line, but for like atoms this model is unstable and for unlike atoms the stability depends on the ratio of moments of the electrons. An oblique moment is not in accord with quantum theory. E. B. LUDLAM.

Glaser's experiments and the orientation of molecules in a magnetic field. G. BREIT (J. Washington Acad. Sci., 1925, 15, 429-434).-Mathematical.

Molecular refractivity of some simple salts. K. F. HERZFELD and K. L. WOLF (Ann. Physik, 1925, [iv], 78, 195-203; cf. A., 1925, ii, 1119).-Previous measurements of the optical dispersion of sodium and potassium chlorides are used as a basis for the calculation of the molecular refractivity of some simple salts, the ions of which are electronically of the type of the inert gas atoms. The calculation involves the electron numbers and the characteristic frequency of the ions. The former are taken as equal to the atomic number of the corresponding inert gas atoms. The latter are obtained from the limiting value of the continuous absorption spectrum of the ion, by adding, in the case of an anion, or subtracting, in the case of a cation, the Coulomb portion of the lattice energy to the electron affinity of an anion, or to the work of ionisation of a cation. The wave-length of the resonance line, which occurs in the second term of the molecular refractivity expression, is about 1-6 times that of the characteristic wave-length. The energy of ionisation of the ions Lit, Rb+, and Cs is assumed to be 1.85 times the ionisation energy of the corresponding inert gas atom, the factor being obtained as the ratio of the experimental values of the ionisation energies of Na+ and of K+ to those of neon and argon. The same ratio is assumed to hold between the ionisation energies of the bivalent ions Mg++ and Ca++ and those of the univalent ions Na+ and K+. The molecular refractivities are calculated of the alkali halides, of the oxides of beryllium, magnesium, and calcium, and of the halogen acids. Where experimental values are available for comparison, the agreement is satisfactory, except for salts with large anions and small cations, for which the theoretical refractivities are too small. F. G. TRYHORN.

Dispersion of carbon disulphide in the ultraviolet. G. BRUHAT and M. PAUTHENIER (J. Phys. Radium, 1925, [vi], 6, 287-294).-See A., 1925, ii, 478.

Effects of chemical combination on the energy of the intra-atomic levels. R. BRUNETTI (Atti R. Acad. Lincei, 1925, [vi], 2, 323-328).-Theoretical. A possible electronic mechanism of the formation of polar compounds is suggested with a view to account for the displacement of the high-frequency lines of an element as a result of chemical combination. The formation of a polar compound may be regarded as occurring in two stages, first, the transformation

of the atoms into ions by addition or subtraction of electrons, and, secondly, the aggregation of the ions into a crystal lattice. The former stage may involve, in addition to a change in the number of electrons in the outer sheaths of the atoms, a change in the number of electrons in the internal shell, consequent on a change in volume of the interior of the atom. It is supposed that the effect of these changes will be to alter the effective atomic number of the atom. This latter will be influenced also by the complex action of attractive and repulsive forces between the ions in the lattice. Empirical terms are suggested for these effects, and expressions derived for the work of introducing an ion into a crystal lattice, and for the variation of the sheath number in the transition of an element from the free to the combined state. It is shown that the variations in the molecular sheath number will depend on the sign and magnitude of the electrostatic potentials existing at the lattice points in the crystal, and that the molecular sheath number may be calculated from a knowledge of the lattice potential. The difference in energy of intra-atomic levels as a result of these changes in the number of electrons in the various levels will result in corresponding differences between the shortwave spectra of an atom in the free and combined states, respectively. F. G. TRYHORN.

Theoretical calculations of physical properties of certain crystals. J. E. LENNARD-JONES and P. A. TAYLOR (Proc. Roy. Soc., 1925, A, 109, 476— 508; cf. A., 1925, ii, 16, 91, 253).-Theoretical. Methods are given of obtaining the forces between certain neon-like and argon-like ions, and these are successfully used to account for the observed distances in certain crystals of the rock-salt type, and for the prediction of other crystal constants yet unmeasured, such as those of calcium chloride. Compressibility, elasticity, and other constants of these crystals are also calculated. The conclusion is that the repulsive force of neon-like ions in crystals is represented by an inverse 11th power law, and that for argon-like atoms by an inverse 9th power law. The latter law does not hold at large distances, when an inverse 15th power law fits the experimental results better.

S. BARRATT.

Forces between atoms and ions. J. E. LENNARD-JONES (Proc. Roy. Soc., 1925, A, 109, 584597; cf. preceding abstract).-From the interatomic distance and compressibility of crystals of rubidium bromide, and the refractivities of krypton and of the ions of bromine and rubidium, the force-field of krypton is calculated. Similarly, the force constants of the xenon group are deduced from measurements of cæsium iodide, and from the fields of krypton and xenon those of the associated ions are derived. By previous methods (loc. cit.), the fields between any two of a series of neon-like, argon-like, kryptonlike, and xenon-like ions are then deduced. The law of force between both atoms and ions is of the form r", and λ and n are calculated. The results are utilised to find the interatomic distances of 32 crystals, including 16 alkali halides. In the latter case, the calculated values are found to lie, with one excep

tion, within 1% or 2% of the observed distances. Compressibilities and crystal energies of the alkali halides are also calculated, moderate agreement being found with the observed values.

L. L. BIRCUMSHAW.

Law of force and size of diatomic molecules as determined by their band spectra. R. T. BIRGE (Nature, 1925, 116, 783-784).-Assuming for the law of force the function F-k1(r-ro)+ kg(r—ro)2+kg(r—ro)3 etc., where r and ro are the nuclear separation and equilibrium nuclear distance, respectively, and for the rotational energy of a nonvibrating molecule the usual function Em-Bm2+ Dm+Fm6 etc., the values of k may be expressed as explicit functions of B, D, F, etc., and evaluations of k may be made with varying accuracy. The underlying theory is confirmed by measurements of the CN band at 3883 Å., whereby k1=16-12× 105 dynecm.1, and k2=-6-18×1014 dyne-cm. 2, whereas from vibrational energy data k1=16-00×105 and k2-5.44×1014. Similarly, values of k, have been obtained for the initial and final states of the CN, CuH, N2, and N2+ molecules, the discrepancy being in all cases within the limits of error. The explanation of the alternation of intensity of successive lines in band series of such molecules as N2 and N2+ is A. A. ELDRIDGE.

discussed.

2

Determination of Avogadro's number by measurements of the birefringence of solutions of dialysed iron. L. TIERI (Atti R. Accad. Lincei, 1925, [vi], 2, 331-334; cf. ibid., 1923, [v], 32, 155).Measurements of the magnetic birefringence at 23°, and in a field of 4000 gauss, have been used to determine the number of particles in different strata of preparations of dialysed iron which had been allowed to remain for 90 days in order to attain statistical equilibrium. Three preparations showing positive and one showing negative birefringence were examined. Two of the former gave a distribution of particles in satisfactory accord with Laplace's law. The density of the granules was determined by means of a pyknometer and their average radius by the application of Stokes' law to measurements of their rate of fall in a long capillary tube. From these data, the values From these data, the values 66×1022 and 50x 1022 were obtained for Avogadro's number. F. G. TRYHORN.

Eka-cæsium and eka-iodine. F. H. LORING

and J. G. F. DRUCE (Chem. News, 1925, 131, 321).The lines on the X-ray photograph ascribed to ekacæsium and eka-iodine were obtained during the examination of material from pyrolusite after the bulk of the rhenium had been removed and not from

the examination of crude rhenium oxide as previously reported (A., 1925, ii, 1124). The plate also showed a line at 0.693 Å., which corresponds with the La line of element 93. A. R. POWELL.

Eka-cæsium. II. F. H. LORING (Chem. News, 1925, 131, 371; cf. A., 1925, ii, 1124, and preceding abstract). Further X-ray evidence is given in support of the author's claim to have discovered element 87. A. R. POWELL.

Search for element 93. II. Examination of crude dvi-manganese (rhenium). III. Foreshadowing elements 75, 85, 87, and 93. F. H. LORING and J. G. F. DRUCE (Chem. News, 1925, 131, 337-341).-Crude rhenium chloride has been isolated from pyrolusite and from "pure" manganese salts by removing the manganese with ammonium sulphide in the presence of ammonia and ammonium chloride, evaporating the solution to dryness, removing calcium with ammonium oxalate in acetic acid solution, expelling the ammonium salts by heating, dissolving the residue in hydrochloric acid, and concentrating the solution until short, feathery crystals deposited on cooling. The chloride solution gives a white precipitate with ammonia which slowly darkens on exposure to the air. Except for the absence of a precipitate with ammonium sulphide, rhenium gives similar reactions to manganese. A further discussion of the lines observed on a film obtained by exposing the crude oxide obtained to X-rays on a copper anticathode is given. (Cf. A., 1925, ii, 1124.)

A. R. POWELL.

Examination of nickel catalysts with X-rays. G. L. CLARK, W. C. ASBURY, and R. M. WICK (J. Amer. Chem. Soc., 1925, 47, 2661-2671).-Powderdiffraction photographs of various nickel catalysts have been taken with a new type of X-ray spectrograph. Catalysts prepared by reduction with carbon, alcohol, ethyl acetate, and hydrogen, and having widely different activities, gave identical lines for nickel, d100=3.536 A.; that prepared by reduction with sodium hypophosphite in solution gave no definite lines and was apparently colloidal. The intensities of the lines, as measured by a photodensitometer, were approximately the same for the various catalysts, and agreed with the values calculated for a face-centred lattice. A rough parallelism exists between decreasing line width and increasing catalytic activity. Nickel monoxide (simple cubic) gave d100-4.16 Å., the dioxide and nickelo-nickelic oxide gave no lines, and the sesquioxide gave only the lines of the monoxide, showing it to be a mixture of this with the dioxide.

A. GEAKE.

Model gratings to illustrate the diffraction of X-rays by crystals. W. L. BRAGG (Mem. Manchester Phil. Soc., 1924-25, 69, 35-38).-Gratings of varying types of complexity have been obtained with the aid of a photographic plate ruled with 400 lines to the inch. These were reproduced on a second plate by putting the two in contact and exposing to light. By then moving the original plate a fraction of the distance between the lines and exposing once more, repeating this as often as desired, gratings with plexity, could be produced. Gratings reproducing 400 groups of lines to the inch, of any required comthe same effects as the crystal gratings, formed by the planes of rock-salt or diamond, for example, have thus been obtained. M. S. BURR.

X-Ray interference in mixed crystals. M. VON LAUE (Ann. Physik, 1925, [iv], 78, 167-176).— Tammann's contention (A., 1925, ii, 20) that X-ray crystal photographs should be very sensitive to slight deviations from the normal distributions

of the atoms in the lattices of mixed crystals, is disproved by a mathematical investigation of the positions and intensities of the interference points resulting from a purely random distribution of two types of atom constituting a simple cubic lattice. The case is also considered in which, as a result of the conditions of preparation, an atom in a mixed crystal is surrounded by more atoms of its own kind than would correspond with a uniform distribution of the constituents of the mixture. F. G. TRYHORN.

Crystal structure of carbon dioxide. J. C. MCLENNAN and J. O. WILHELM (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 51-56).-The crystal structure of carbon dioxide solidified by liquid air has been investigated by means of an X-ray method. Carbon dioxide crystallises in a cubic system with 4 mols. in each unit cell and a lattice distance of

5-76 Å. The distance between carbon and oxygen atoms is 1-25 Å.; the moment of inertia is 8.2x 10-39 g. cm.2 J. S. CARTER.

Structure of a- and B-quartz. (SIR) W. H. BRAGG and R. E. GIBBS (Proc. Roy. Soc., 1925, A, 109, 405-427).-The only elements of symmetry possessed by an ordinary (a-) trigonal crystal of quartz are a trigonal and three digonal axes. It is impossible to determine the crystal structure completely by consideration of these symmetry restrictions and of the distances and angles between the various reflecting planes in the crystal for X-rays. The problem could be solved by studying, in addition, the intensities of the X-ray reflexions, if the principles governing these intensities were sufficiently understood. Actually more indirect methods have to be applied. Chemical considerations have been employed, but with unsatisfactory results. The method here proposed utilises the fact that trigonal (x-) quartz passes into hexagonal (B-) quartz at 575°. The symmetry of the hexagonal form is much higher, and the details of its structure can be fixed with ease. There is evidence that the structure of the crystal is but slightly changed during the transformation from the a- to the B-form, and hence the structure of the B-form gives a powerful clue to the structure of the x-form. A structural model of the a-form, deduced in this manner, accounts for the piezo-electrical properties of the crystals, and for the varieties of twinning which are recorded.

The electrical conductivity of quartz is hundreds of times greater along the crystal axis than across it. This fact may be correlated with the spacings of the atoms in the crystal, as the structure suggested has more open and direct passages along than across

the axis.

S. BARRATT.

Crystal structure of barytes, celestine, and anglesite. R. W. JAMES and W. A. WOOD (Proc. Roy. Soc., 1925, A, 109, 598-620).-Assuming the sulphate ion to be a tetrahedron of oxygen atoms surrounding a central sulphur atom, the distance S-O being 1.5 Å. (cf. Bradley, A., 1925, ii, 638), it has been found possible to assign structures to the three isomorphous crystals investigated which are practically identical, and account completely for the observed intensities of the X-ray spectra

given by them. The structure is based on a simple orthorhombic lattice containing 4 mols. to the unit cell (from the specific gravity and mol. wt. of the crystal), and the space group is Vs. The parameters fixing the relative positions of the different kinds of atoms were determined by X-ray intensity

measurements.

For barium sulphate, the dimensions of the unit cell are: a=8-85, b=5.43, c=7.13 Å. and the axial ratio is 1-6304: 1:1-3136. For strontium sulphate, a=8-36, b=5-36, c=6-84 A., and the axial ratio is 1.5580: 1:1-2800. For lead sulphate, a=8.45, b=5-38, c=6-93 Å., and the axial ratio is 1-57041: 1-2894. The structure appears to be one in which each positive metallic ion is surrounded as closely and as uniformly as possible by negative oxygen ions, perfect freedom of adjustment of the together in fours in the sulphate groups. ions being hampered, because they are attached

In calculating the structure factors, the figures for the diffracting power of ions at different angles, calculated by Hartree (A., 1925, ii, 735), were employed with slight modifications. A comparison of the absolute intensity of reflexion actually observed with that calculated on the classical theory showed that the observed intensity is always lower than, although of the same order of magnitude as, the calculated value. A possible reason for the structure which is common to a number of crystals of the type XRO, is discussed. Observations on potassium permanganate and potassium perchlorate confirm the view that the structures are essentially similar to those of the sulphates under investigation.

L. L. BIRCUMSHAW.

Structure of barium sulphate. R. W. JAMES and W. A. WOOD (Mem. Manchester Phil. Soc., 1924-1925, 69, 39-51).-See preceding abstract.

Crystalline structure of hexachlorobenzene and hexabromobenzene. W. G. PLUMMER (Phil. Mag., 1925, [vi], 50, 1214-1220).-Hexachlorobenzene and hexabromobenzene have been examined by the powder method of X-ray analysis. Photographs were taken with a Shearer X-ray tube fitted with a copper anticathode, the B-lines of copper being excluded by a nickel filter. The observed spacings are tabulated, and agree with the theoretical spacings assuming 2 mols. per cell. The dimensions of the cell for hexachlorobenzene are a=8.10, b=3.86, c= 16-68 Å. These results were later confirmed on a perfect crystal of hexachlorobenzene. Based on these measurements a suggestion is made of the probable arrangement of the atoms in the molecule. The measurements of hexabromobenzene were not so successful, but they prove the general correctness of the structure assumed. A. B. MANNING.

Isomorphism of molybdates of the rare-earth metals with those of calcium, strontium, barium, and lead. F. ZAMBONINI and R. G. LEVI (Atti R. Accad. Lincei, 1925, [vi], 2, 303-305; cf. A., 1925, ii, 1133).-Tabulated values are given showing that the reflecting planes are essentially the same in crystals of the molybdates of calcium, strontium,