ii. 97 General and Physieal Chemistry. Refraction of Light at Corresponding Temperatures W. HERZ (2. p,kysikaE. Chem. 1921 98 175-180).-A theoretical paper in which from the examination of the data for a large number of substances it is shown that the refractive index of all substances has the same value 1.126 a t the critical temperature. The formulze of Gladstone and Dale and of Lorentz and Loreilz yield approxi- mately the same specific refraction values for this temperature. It is also shown that at other comparable temperatures the refrac- five indices approximate to the same value although in these cases considerable differences do occur. Thus a t the boiling point ten liquid substances give mean value of 1.335 for the refractive index the extreme values being 1.348 and 1.320.The Spectrochemistry of Aliphatic Dienes with Conjugated Double Bonds. K. VON AUWERS and H. WESTERMANN (Ber. 1921 54 [B] 2993-2999).-The physical constants of a number of carbinols and the corresponding dienes are recorded. The specific exaltations in the cases of the latter substances with " un- disturbed doubly disturbed and singly disturbed " conjugation are E&fr. +2.0 +1.4 and +0.9 andE8Dbp. +50% +43% and -+%yo respectively. The boiling points of the &enes are de- pressed by branching of the chain but raised by the approximation of the methyl group to the double bond; this regularity is not always very definitely marked. The density of isomeric com- pounds increases relatively greatly when the methyl group is attached to a doubly-bound carbon atom.Similarly the index of refraction increases with similar alteration in structure ; these changes are consonant with those observed with cyclic compounds. The following constants are recorded As-pentene-8-01 b. p. 64"/62 mm. d:795 043382 ni795 1.42558 ngy5 1-42821 nFQ5 1.43502 72"':15 1.44065. As-Hexen-8-01 b. p. 59"/27 mm. djJJ 0.8370 n?j 1043025 nE3 1.43286 n r s 1.43962 n,;' 1,44510. Ab-Heptene-8-01 b. p. 63"/11 mm. d:" 0.8422 n:4 1.43698 nE'4 1.43965 nF4 1.44620 nY.4 1.45176. c-Methyl-Ab-hexen-8-01 b. p. 75"/45 mm. dig 0.8411 nz 1.43539 n$ 1.43788 ng 1-44478 n y 1.45045. [-Methyl-As- hepten-8-01 b. p. 67"/11 mm. dY2 0.8354 nyd 1-43679 nEd 1.43926 nlS2 1.44585 ny2 1.45127. y-Methyl-Afi-octen-&ol b. p. 85"/12 lnn~. d:4'5 0.8402 n:' 144202 ng5 1.44448 nF5 1.45114 nId5 1.45651.&Methyl-Ay-heptene-c-ol b. p. 66"/17 min. dli 0.8525 i L 7 9 1.44525 n::9 144792 nF9 1.45454 nyy 1.46018. 6-Methyl-A~-octen-~-oI b. p. 89"/16 mm. d y 5 0.8495 n:' 1.44576 ng5 1.44838 nF5 1.45490 ny'B 1.46059. A@-Pentadiene b. p. 43" d:"''j 0.6887 di;6 0-685 n26 1.42991 ?g~ 143M3 nFS5 1.44655 "Rt;"j 1.45694 nz!j' 14309. Afi8-Hexadiene b. p. J. F. S. Y B VOL. UXXII. ii. 4ii. 98 ABSTRACTS O F CHEMICAL PAPERS. 80" d:"' 0.7237 d;,7O5 0.720 n;'O5 1.44962 n:'O5 1.45420 n;'O5 1.46651 n;'O5 1.47764 n'," 1.4514. A@-Heptadiene b. p. 107" d.i6'; 0.7341 dit.3 0.731 nz"3 1-45101 n;'' 1.45543 ng'3 1.46695 n;"" 1.47735 1.4534. [-Methyl-A@-heptadiene b. p. 117",d~5'50~7361,d~~'50~733,n~'~ 1.44887 n:.' 1.45302 nF'5 1.46397 ny'5 1-47379 n$ 1.4505.Y-Mefhyl-AflS- octadiene b. p. 149" 0.7515 dii'6 0.751 145427 nS.6 T.45831 nF-6 1.46903 n:'6 1.47812 7~: 1,4553. c-Methyl-Afls-hexadiene b. p. 104" dt:'7 0.745 n 1.4606 (i) dig'' 0.7439 7~:'~ 1.45659 n',"'' 1.46117 nF'7 1.47340 n;:'i 1.48433 (ii) 0.7473 n:'65 1,45589 72g6 1.46037 n;' 1.47262 ny'Os 1.4831 6. P-Methyl-Aay-butadiene b. p. 34" dig'* 0.6826 dii'8 0.682 n:'* 1.41902 n;'* 1.42309 nF's 1.43422 ny'* 1.44405 72; 1.4224. 6-Methyl-Aye-heptadiene b. p. 131" d";"4 0.7598 dii'-? 0.763 n24'4 1.45610 n$'4 1.46003 nFy 1.47125 ny'4 1.48103 n?, 1.4625. 8-kethyl-AyE-octadiene b. p. 150" die" 0.7708 d;;''" 0.771 nz,0'3 1.46206 n::.'" 1.46617 d:"" 1.47724 n20'3 1.48708 n$ 1-4663. py-Dimethyl- Av-butmdiene h. p. 70° d$ 0.7239 d;f" 0.725 nz'l 1.43307 ng'l 1.43703 ny'l 1.44758 n7.l 1.45704 n$ 1.4382.K. VON AUWERS (Ber. 1921 54 [B] 3000-3003).-A reply to K. H. Meyer (A. 1921 i 853). The view that the methyl group causes optical exaltation cannot be mainta,ined. In certain of the examples quoted by Meyer the differences lie within the limits of experimental error whereas in ot,her examples readily polymerised substances for instance acrylic acid and its esters have been chosen and the data obtained are not due to the unimolecular forms. The Molecular Refraction of Aromatic Hydrocarbons and '' Aromatic " Carbon. K. VON AUWERS (Ber. 1921 54 [B] 3188-3194).-1n a recent communication (A. 1921 ii 473) von Steiger has endeavoured to show that the molecular refraction of hydrocarbons can be calculated more accurately from linking refractions than from atomic refractions and he has postulated that all C-C bonds and all C-H bonds in aromatic hydrocarbons are equivalent among themselves but different from the corre- sponding values in the aliphatic series.This mode of calculation appears to the author to be unsound theoretically since the atomic refraction of aromatic hydrogen as deduced from the molecular refraction of benzene diphenyl and naphthalene in these circum- stances is negative whilst also the " linking refraction " for (C-H) falls with decreasing wave-lengths. V O ~ Steiger has illustrated his arguments by citing the cases of benzene diphenyl and naphthalene and the agreement between the observed and calculated values is excellent. Unfortunately the data quot,ed for naphthalene are incorrect and the amended figures show that the presumption of additivity is not fulfilled.Further the data for the three hydrocarbons are not directly comparable since they are obtained under widely-diff ering conditions ; when due allow- ance is made for this factor it is found to be impossible to deduce constant atomic or linking refractions. The molecular refractions H. W. [Substitution Processes]. H. W.GENERA4L AND PHYSICAL CHEMISTRY. ii. 99 of homologues of benzene and other aromatic hydrocarbolls cannot be calculated from von Steiger’s aromatic and aliphatic linlcing refractions since the differences between the observed and calcu- lated values increase more and more with increase in the number of side-chains. The fundamental idea in voii Steiger’s refractometric hypothesis and thermochemical theories (A.1920 ii 355) is the existence of a practically tervalent “ aromatic ” carbon atom of the graphite type which differs from the quadrivalent “ diamond ” carbon atom of the paraffins (cf. Debye and Schemer A. 1917 ii 437). To the author this view of the varying valency appears to be mis- leading. The supposed ter- and quadri-valency of the carbon atom does not as in the case of other elements correspond with different stages of oxidation or different electric charges ; the atom is invariably quadrivalent and the apparent variation is caused by alteration in t’he distribution or compensation of its affinity. H. W. The Mechanism of Continuous Luminous Radiation. J. J ~ C L A U X (Compt. rend. 1921 173 1355-1357).-The author has previously put forward the hypothesis (cf.ibid. 1914 158 1879) that the emission and absorption of the continuous spectrum depend on the establishment or rupture of chemical valencies. Evidence in support of this hypothesis is now given. WILLIAM R. SXYTHE (Astrophys. J . 1921 54 133-139).-Wjfh a discharge passing between gold electrodes the ultra-violet region of the fluorine spectrum was photographed through a fluorite window the gas having been prepared by the electrolysis of potassium hydrogen fluoride and purified by being passed through sodium fluoride and a freezing trap. Ten fluorine lines all in the red were observed arid tabulated with an accuracy of A0.l A. An impurity which was supposed to have the composition CF was observed following a ff are back from the charcoal chamber.The approximate positions of nine heads of bands between A 4829 and A 6525 were observed in this spectrum. ARTHUE ST. C. DUNSTAN and BENJA- MIN A. WOOTEN (Astrophys. J . 1921 54 65-75).-The spectra of strontium barium lithium copper and lead were observed when the vapours of the metals were introduced into a direct current arc by means of an alternating current arc placed below it in the same enclosure. It was shown that by cooling the elec- trodes alternately the intensity of the metallic iines remained as before greater a t the cathode than at the anode when the metal was introduced into the arc symmetrically; neither was the effect due to electrolytic or electrostatic separation. Since the pheno- menon is fully developed in 1/120 second the transference of vapour from one electrode to the other must be of little importance.The light appears to be due chiefly to bombardment of the metallic vapour by electrons from the cathode. Whilst the lines of all W. G. The Spectrum of Fluorine. CHEMICAL ABSTRACTS. Arc-cathode Spectra. 4-2ii. 100 ABSTRACTS OF CHEMICAL PAPERS. metals are stronger a t the cathode the difference decreases with increasing atomic weight. CHEMICAL ABSTRACTS. The Electric Furnace Spectrum of Scandium. ARTHUR S. KING (Astrophys. J. 1921 54 28-44).-The spectrum of a pure sample of scandium oxide was observed in the carbon arc and in the electric furnace through the range A 3015-6559 A. The furnace was operated a t the usual low gas pressure and tem- peratures of 2200° 2250° and 2600" were used.A table gives the classification of 257 lines with respect to their behaviour in the arc and a t various furnace temperatures. Some 25 lines are indicated as being enhanced in the arc; 160 are enhanced in the furnace. The 29 arc-flame lines belong to the latter class some of them showing moderate strength a t all temperatures whilst others increase in intensity more or less rapidly with increasing temperature. These lines are not due to oxidation but the bands on the other hand are quite probably due to the oxide. Lines which appear a t a low temperature and those enhanced in the furnace are weak or lacking in the general solar spectrum but prominent in that of sun-spots. The Zeeman effect for scandium lines is large and apparently uniform in sunspot spectra although laboratory observations have not as yet been made.Scandium oxide fused in the furnace seemed to form a carbide with the graphite of the boat. The glossy black residue became a greyish- brown powder on exposure to the air presumably reabsorbing oxygen. CHEMICAL ABSTRACTS. Wave-length Measurements in Arc Spectra Photographed in the Yellow Red and Infra-red. F. M. WALTERS jun. (Bureau of Standards Sci. Papers 1921 17 No. 411 161-177).- Observations were made with a grating spectrograph of 640 cm. radius and 299 lines' per mm. The photographic plates were sensitised to yellow red or infra-red light with pinacyanol and dicyanin. The wave-lengths of silver aluminium gold bismuth cadmium mercury lead antimony tin and zinc are given in international units and to 0.01 A.The longest wave-length measured is that at 10395 8. in the spectrum of cadmium. The work is compared with previous results of others and discrepancies are accounted for as arising from impurities ghosts or overlapping spectra of higher orders. Extension of the Ultra-violet Spectrum and the Progres- sion with Atomic Number of the Spectra of Light Elements. R. A. MILLIKAN (Proc. Nut. Acad. Xci. 1921 7 289-294).-The results of an investigation of the extreme ultra-violet radiations emitted by the second ring or shell of electrons in the atoms of atomic number 2 to 13 (helium to aluminium) are recorded. The ultra-violet spectrum has been photographed down to 1=136*6 A. in the case of aluminium and to h=149.5 A. in the case of copper.The La lines of aluminium magnesium and sodium are found a t 144.3 b. 232.2 8. and 372.2 b. respectively. Aluminium emits no radia- tions when excited by condensed sparks in a vaeuum between CHEMICAL ABSTRACTS.ii. 101 144.3 8. and 1200 A. where the JI spectrum due to the three outer electrons begins. The chief lines below 2000 8. due to the three outer electrons of aluminium are 1379.7 1384.5 1606.9 1612.0 1671.0 1854.7 and 1862.7 8. Magnesium is like aluminium there is a blank between its La line a t 232.2 8. and the lines due to the two outer electrons which begin a t 1700 8. The lines measured below 2000 A. are 1735-2 1737.9 1751.0 and 1753.7 A. Sodium edits no lines between the L line 376.5 8. and the lines due to the single M electron which have their convergence wave-length a t 2412.63 8.The spectrum due to the six L ring electrons of oxygen begins a t 230 A. and extends with much complexity and strength up to 834.0 8. where the strongest line is found; above this point the lines are few in number and relatively faint. The ratio of the K / L frequency for oxygen is about 35. The strongest oxygen lines are 321.2 374.3 507.8 525.7 554.2 599.6 610.1 616.7 625.2 629-6 644.0 703.1 718.5 and 834.0 8. The spectrum of the four L-ring electrons of carbon begins a t 360.5 8. and extends with much complexity and strength up to 1335.0 8. where the strongest line lies; above this point the lines are widely scattered and rela- tively weak. The ratio of the K / L frequency in carbon is about 30. The spectrum due to the five L-ring electrons of nitrogen is simple; it begins a t 685.6 8.and reaches a maximum at 1085.3 8. The only other strong lines of nitrogen are 685.6 8. 916.2 a. and 991.1 8. The ratio of the K/L frequency is 34-8. In the case of fluorine a strong line appears a t 657.2 A. and a second a t 607.2 A. ; these are the only lines thus far identified as coming from the seven L-ring electrons of fluorine. The spectrum from the three L-ring electrons of boron is very simple. It consists of 676.8 760.0 1624.4 two doublets 2164.2 2166.2 and 2496.9 2497.8 and a single spark line 3451.5 A. The ratio of the frequency K/L is 37. The spectrum of the two L-ring electrons of glucinum begins a t 2175 8. and finishes in its La doublet at 3130.6 and 3131.2 A. Similar experiments with lithium reveal no lines between the shortest wave-lengths measurable by the present method and the familiar series due to its single L-ring electron the Laline of which is a t 6708 pi.and its convergence mave-length a t 2209 A. The progression shown in these optical spectra is simple and very like t,hat exhibited by X-ray spectra J. I?. S. GENERAL AND PHYSICAL CHEMISTRY. The Structure of tho Elements of Mean Atomic Number. A. DAUVILLIER (Conapt. rend. 1921 173 1458-1461).-An exten- sion of previous work (cf. A 1921 ii 669) further results being given for gold and new measurements of the L series for cerium The Structure of the Bismutli Lines. €1. NAGAOKA and Y. SUGIURA (Astrophys. J. 1931,53,339-348).-The light resulting from the bombarding of bismuth by an electronic current was observed by means of a glass Lummer-Gehrcke plate crossed with one of quartz or crossed with an echelon grating.For the line A 4722 the five known positive components were confirmed and and antimony. FV. c.ii. 102 ABSTRACTS OF CHEMICAL PAPERS. seven weak negative satellites were discovered showing the line to have the general structure common to lines of heavy metals. The principal component is probably multiple its constituents varying in relative intensity. The line A 4122 has four strong components. The line A 4308 consists of two principal lines separated by 0.350 A. each being a narrow doublet; there are also two faint satellites. The authors point out some constant frequency differences and suggest that these may be interpreted in terms of the quantum theory.The Zeeman Effect. H. R. WOLTJER (Chem. Weekblad 1921 18 677-682).-A paper written in connexion with the twenty-fifth anniversary (October 31st 1921) of the first announce- ment of Zeeman’s discovery. The importance and subsequent developments of his work are discussed a t some length. Anisotropy of Molecules. C. V. RAMAN (Nature 1922 109 75-76) .-An extension to direct visual observations is described of the method whereby it has been shown photographically (Lord Rayleigh A. 1920 ii 574) that the light scattered by molecules is in general not completely polarised when observed in a direction transverse to the pencil of light traversing the gas. With carbon dioxide the effect is conspicuous and visual determinations of its magnitude have been made.Similar evidence that the molecules of gases are not spherically symmetrical and are anisotropic in their properties is furnished by observations on the polarisation of light of the sky; liquids also show an imperfect polarisation attributable to aniwotropy. H. ZOCHER (2. physikaZ. Chem. 1921 98 293-337).-The double refraction discovered by Diesselhorst Freundlich and Leonhardt (A. 1916 ii 65) with flowing vanadium pentoxide sols has been found in the following sols Soap solution clay suspensions sols of silver cyanate benzo- purpurin “ benzo-brown,” primulin sodium alizarinsulphonate alizarin p-azoxyphenetole p-azoxyanisole anthracene cerasin- orange and aniline-blue. Some observations on the spacial par- tition of the double refraction in agitated sols showed that in the eddies phenomena were ohserved which exhibited many analogies to the behaviour of uniaxial crystals in convergent light.These observat!ions serve as metliods for determining the flowing-double refraction and pleochroisiii and for determining the character of the sols. The rotation of the dark cross in eddies in opposition to the polarisation direction shows the direction of flow does not always need to correspond with the main vibration direction of the light. The coincidence of the dark Cross with the direction of polarisation in old vanadium pentoxide sols and soap solutions is attributed to the bending of the doubly refracting particles. The sign of the double refraction of soap solutions varies with the age and concentration of the sol.The change in the sign of the double refraction is explained as follows Colloidal particles in the form of short rods must arrange themselves with their symmetry axis CHEMICAL ABSTRACTS. S. I. L. A. A. E. Sols with Mon-spherical Particles.GENERAL SND PHYSICAL CIIEI\IISTRY. ii. 103 in the direction of flow particles in the form of leaflets must arrange themselves with the axis of symmetry a t right angles to this direc- tion. If now in the process of ageing the rods pass into leaflets then the optical character of the direction of flow must' a t first he the same as that of the colloidal particles and afterwards of the oppsite sign. Aniline-blue sols prepared by pouring an alcoholic solution of the dye into water show on flowing a strong negative double refraction in the red negative pleochroism in the region orange to green and a positive double refraction in the blue.Anomalous interference bands are produced by this anomalous behaviour. The optical anisotropy produced by a magnetic field has the opposite sign that is the colloidal particles arrange themselves a t right angles to the lines of force. On the other hand they arrange themselves parallel to the electrical lines of force. The negative double refraction of benzopurpurin can be made to disappear by heating. The addition of electrolytes causes the double refraction to reappear on cooling. The coagulum obtained by the addition of an excess of electrolyte gives a doubly refracting sol on peptisa- tion if the ad'dition of the electrolyte was slow or if the original sol was doubly refracting but if the coagulation took place rapidly or if the original sol was isotropic there will be no double refraction. When acidified double refracting red sols may be changed into double refracting blue sols.The strong pleochroism has always a negatlive sign and the particles of the sols are diamagnetic. The double refraction of vanadium pentoxide sols disappears more slowly the older and more concentrated the sol. Very concentrated old vanadium pentoxide sols and concentrated electrolyte-poor benzopurpurin solutions show a persistent double refraction similar to that of the crystalline liquids. The view of Diesselhorst and Freundlich (Zoc. cit.) that the growth of non-spherical particles is not a crystallisation but an aggregation of non-spherical primary particles in parallel layers is confirmed with the addittion that the primary particles may be crystalline and that eventually they may undergo crystalline processes.J. F. S. Recent Advances in Stereochemistry. B. I<. SINGH ( J . Proc. Asiatic Xoc. BengaZ 1921 17 213-230).-A general account of the historical development of stereochemistry and a discussion of the relation between optical activity and chemical constitution. Particular consideration is given to optical activity in homologous series the effect of conjugated nnsaturation and of position isomerism on optical activity. The Walden inversion is also discussed. H. W. Investigations on the Fundamental Law of Photochemistry. Y. LASAREV (2. physikal. Chem. 1921 98 94-97).-A theoretical discussion of work previously published on the bleaching of dyes by light (A.1912 ii 219; Ann. Physik 190'7 [iv] 24 661). The experiments on the bleaching in the presence of oxygen under increased pressures (up to 150 atm.) are particularly considered. The fundamental law for the bleaching of dyes has the formii. 104 ABSTRACTS OF CHEMICAL PAPERS. -dC/dt=cto(l-e-KC')(l-e-"~f~) in which C is the concentration of the dye K a constant proportional to the absorption constant C and K similar values for the substance non-sensitive to light and a the photochemical constant. It is assumed that in the reaction during the absorption of light an electron from the first Bohr orbit springs over to the furthest orbit. It is shown that the total absorption of light is proportional to the absorption of the largest (ionised) atoms.The increase in volume during photo- chemical reactions is explained by the assumption that the light inasmuch as it occasions an increase in the volume of the atoms by effecting the electron spring brings about an increase in the gaseous volume. J. F. S. New Measurements of Precision in the X-Ray Spectrum. M. SIEGBAHN (Compt. rend. 1921 173 1350-1352).-By the use of three spectrographs each adapted to a particular spectral region it was possible to measure a wave-length with an accuracy of 0-01-0.005~0 and by this means the wave-length of the copper line Ka was found to be 1537.36 x 10-11 em. in a vacuum. A more precise instrument is described the readings of which are accurate to within 0.002~0 and with this the value of A for Ka was found to be 1537.302 x em.W. G. Spectrographic Study of the De-intensifying of Barium Platinocyanide in the Villard Effect. A. ZIMMERN and E. SALLES (Compt. rend. 1922 174 80).-Fluorescent screens which have undergone the Villard effect can be regenerated by exposure to diffused light. It is shown that the radiations capable of de- stroying the Villard effect are found in four principal bands of almost equal breadth. The first is entirely in the infra-red the second in the greenish-yellow the third in the blue and the fourth in the ultra-violet. W. G. The Theory of Absorption of X-Rays by Matter and the Principle of Correspondence. LOUIS DE BROGLIE (Compt. rend. 19.21 173 1456-14&3).-An expression for the atomic coefficient of absorption of a substance for a radiation of wave- length A has previously been given (cf. A.1920 ii 208) in which the constant a was defined by the hypothesis that for the possible changes of internal configuration of an atom placed in a system in thermodynamic equilibrium a t the temperature T the probability A of the return of the atom from a configuration of energy fl to a configuration of less energy c2 is proportional to and also to the absolute temperature in such a way that ~412=M(E1-€2)T. A mathematical proof of the truth of this hypothesis and a means of calculating the constant a are given and it is shown that the coefficient of the Bragg-Pierce law is %t universal constant capable of being expressed as a function of the constants of the electron and of the radiation.Calculation of the X-Ray Absorption Frequencies of the Chemical Elements. I and 11. WILLIAM DUANE (Proc. Nut. Acad. Xci. 1921 7 260-267; 267-273).-1n the first paper W. G.GENERAL AND PHYSICAL CHEMISTICY. ii. 105 the author has calculated the K critical absorption frequencies of the elements magnesium sulphur calcium iron selenium molybdenum tin cerium dysprosium tungsten lead and uranium on the basis of the Rutherford-Bohr theory of the structure of atoms and the mechanism of radiation. The assumption is made that the electrons are distributed in circular orbits which do not lie in planes passing through the nucleus of the atom. In order to estimate the forces exerted on an electron in one orbit A due to the electrons in the parallel orbit B it is assumed that they are the same as if the charges are concentrated half at the nearest point to the orbit A and half a t the point furthest from A .A series of values is obtained which are in fair agreement with those observed. I n the second paper it is assumed that the forces are the same as if the electricity of the electrons in the orbit B is uniformly distributed along the orbit. The values calculated on the second assumption are somewhat smaller than those of the first calculations. The Laws of Absorption of X-rays. B. K. RICHTMYER (Physical Rev. 1921 18 13-30).-An experimental study of the absorption of X-rays by water (oxygen) aluminium copper molybdenum silver and lead each over a great part of the range of wave-lengths 0.093 8.to 0.95 d. The longest of these waves is shorter than the critical K absorption wave-length of oxygen aluminium or copper and falls between K and I; for molybdenum silver and lead. Rays from a Coolidge tube were analysed by a spectrometer with a sodium chloride crystal and the intensity was measured by ionisation of methyl bromide. Scattered and fluores- cent radiation from the absorbing screen were suppressed by placing the screen in front & the spectrometer slit nearest the tube. De- tailed tables and curves of the results are given. The rnass- absorption coefficients p / p of these substances is given by p/p= Ph3+(u/p) where h is wave-length and u / p is the mass-scattering coefficient except in the region close t o hR on the short-wave- length side where the equation gives values which are somewhat too large the error being greater as the difference between h and hg is smaller.For each substance P has a constant value for h<hg and another much smaller constant value for h>hK. The mass-scattering coefficient is independent of wave-length and is to a first approximalion the same for all element's. The atomic- absorption coefficient is given by pa=2.29 x 10-27 N413+~. where N is the atomic number and ca the atomic-scattering coefficient for A<hg. No trace of the suggested " J " absorption discon- tinuity was found. At 0.093 A. p / p for water is still much greater than for the hard rays of radium-C. The Luminescence of certain Oxides Sublimed in the Electric Arc. E. L. NICHOLS and D. T. WILBER (Physical Rev.1921 17 707-717).-Thin films of metallic oxides were obtained by subliming the metal or one of its salts from the crater of a direct- current carbon arc on to a metal disk. Oxides prepared by sub- limation are found to respond to excitation by cathode rays. J. F. S. CHEMICAL ABSTRACTS. 4"ii. 106 ABSTRACTS OF CHEMICAL PAPERS. The shift in colour with increasing temperature in luminescence is in general to the violet. The effect of pressure on luminescence is discussed. The most active oxides were those of calcium mag- nesium zinc zirconium silicon and aluminium. CHEMICAL ABSTRACTS. Anode Rays of Glucinum. G. P. THOMSON (Nature 1921 107 395).-The method of positive ray analysis applied to lithium (Aaton A. 1920 ii 344) has been extended to glucinum and found to yield a well-marked parabola corresponding with a single charge and an atomic weight 9.0 &0- 1 (Na= 23).No second parabola was observed which could be ascribed with certainty to glucinum but i t is doubtful if a line of intensity one-tenth that of the first could have been detected. No indieahion was found which would suggest that the atom of glucinum can lose two electrons under the prevailing experimental conditions. Reactions in Penetrating Radium Radiation and in Ultra- violet Radiation Filtered by Quartz Glass. 11. The Hydrogen Peroxide Equilibrium set up in Radium Radiation. ANTON KAILAN (2. phydcal. Chem. 1921 98 474-497; cf. A . 1912 ii 10 522; 1920 ii 576).-Radiation from 0.1 gram of radium metal after passing through 1 mm. of glass into 100 C.C.of water in contact with air produces a solution containing 6 x gram-equivalent of hydrogen peroxide ; in the presence of 0-01N- sulphuric acid the amount of hydrogen peroxide becomes and in 1.0N-acid the amount is 5x10-4 gram-equivalent. I n the last-named solution 6 x 10l2 and in the first 3 x 1012 molecules of hydrogen peroxide are formed per second. These values are considerably less than those obtained nine or ten years ago and reasons for the discrepancy are advanced. A comparison of the values obtained for the velocity of formation of hydrogen peroxide then and now with the rate of formation by other reactions which have been obtained in a similar manner to the present shows that both in aqueous and in non-aqueous solutions when the decom- position of hydrogen peroxide is neglected the number of mole- cules of hydrogen peroxide formed is of the same order as that of the ion pairs absorbed from the radiation by the medium.In ultra-violet light from a quartz glass lamp under similar conditions the reduction of the peroxide concentration by increase in the hydrogen-ion conckntration is less than in the radium rays whilst the decomposition of hydrogen peroxide is more than one hundred times greater than in the radium rays. Stationary Electron Vibrations without Radiation Resist- ance. A . D. FOKKER (Physics 1921 1 107-109).-1t is theoretically shown that a Bohr atom with electrons revolving jn closed paths may be free from radiation resistance without violating the principles of the classical theory.CHAGLES STAEHLING (Compt. rend. 1921 173 1468-1411; cf. A. 1920 ii 5).-In continuation of previous work (Zoc. cit.) it is shown that A. A . E. J. F. S. CHEMICAL ABSTRACTS. The Radioactivity of the Oxides of Uranium.CENEUL AND PHYSICAL CHEMISTRY. ii. 107 the radioactivity of the green oxide which has been restored to its original value commences to diminish as before. The restoration of activity may be brought about by simple calcination without passage through ammonium uranate. During this process of restoration the oxide undergoes loss in weight due for the most part to the removal of moisture. The black oxide which shows little loss in activity also shows little loss in weight or alteration in activity on calcination. The author considers that the green oxide exposed to the air in thin layers undergoes hydration as well as loss in activity the hydration diminishing the superficial density of the uranium atoms.The black oxides obtained by calcination at high temperatures do not hydrate and show little or no decrease in radioactivity. These facts tend to confirm the existence of a definite black oxide which is probably an allotropic modification of the green oxide. W. G. Isotopy of the Radio-elements and Meitner 's Nucleus Model. MAXIMILIAN CAMILLO NEUBURGER (2. physikal. Chem. 1921 99 161-167).-A theoretical paper in which on the basis of Meitner's nuclear hypothesis (A. 1921 ii 293) the author has subdivided isotopes into four groups. These are termed isotopes of the first second third and fourth order. Isotopes of the first order have the same nuclear charge and arrangement of the outside electrons but different nuclear mass total number of nuclear constituents arrangement of the nuclear constituents number of each kind of nuclear constituent and probability of disintegration ; radium and meso-thorium-I illustrate this group.Isotopes of the second order have the same nuclear charge arrangement of the outside electrons nuclear mass and total number of nuclear con- stituents but different numbers of each kind of nuclear constituent arrangement of the nuclear constituents and probability of dis- integration. An example of this class is furnished by ionium and uranium-Y. Isotopes of the third order differ only in the arrangement of the nuclear constituents and the probability of disintegration all other properties being identical ; this is the case with radium-D and actinium-B.Isotopes of the fourth order are identical in all the properbies mentioncd ; no examples of this class are known. J. F. S. Calculation of the Branching Relationships for Dual a-Disintegration and the Meitner Nucleus Model. MAXI- MILIAN CAMILLO NEUBURGER (2. physikal. Chem. 1921 99 168- 171; cf. Meitner A. 1921 ii 293).-A theoretical paper in which i t is shown that the method of calculation adopted by Smekal (A. 1921 ii 149) for determining the branching relationship is so closely dependent on the assumptions made as to the structure of the radioactive nucleus that it cannot possibly be of general use. The branching relationship cannot be calculated by this method since it leads to inaccurate values which lead to results opposed to the experimental data.The agreement found for the branching relationship by this method for the uranium radium thorium 4*-2ii. 108 ABSTRACTS OF CHEMICAL PAPERS. and actinium families with the experimentally determined values for this relationship is accidental. Emissivity of Iron and Copper. K. LUBOWSKY (Elektro- techn. Z. 1921 42 79-8l.).-The emissivity of both bright and blackened iron and copper was determined. Curves are given showing the effect of convection currents obtained from a flat plate of the material first perpendicular to and second in the direction of the air stream. Expressed in watts per sq. cm. per 1" of temperature excess the emissivity varies from 0.75 x to 1-5 x lob3 under the various conditions.CHEMICAL ABSTRACTS. The Effects of Impurities on the Ionisation Potentials Measured in Themionic Valves. L. S. PALMER (Radio Rev. 1921 2 113-125).-The method of investigation used is based on a study of the current-voltage curves measured in hard and soft 3-electrode tubes. The details of the method have been given in a previous paper. Certain bends in these curves are interpreted as being due to the bombardment of the filament by positive ions. The ionisation potentials of helium mercury and argon have thus been determined. There are variations in the values for the ionisation potentials obtained in helium. These are ascribed to impurities and the amount of impurity active accords with the values found by Horton in helium (A.1919 ii 210; 1920 ii 660; 1921; ii 672). This effect of the impurity also explains the low values of the potentials obtained by Stead and Gossling (A. 1920 ii 659). The characteristic of the soft 3-electrode tube can be used to determine the resonance and ionising potentials of the contained gas. The nature of the gas and presence of the impurities can be determined from the critical points of the characteristics. Methods of measuring ionisation potentials are liable to error if impurities are present. Horton's previous work is confirmed and the result has been extended to argon with traces of mercury. The value of the ionisation potential decreases slightly with increase of gas pressure causing changes in the characteristics similar to those for impurities. A number of critical potentials obtained for helium and argon are given.Estimation of the Radium Content of Radioactive Luminous Compounds. E. A. OWEN and WINIFRED E. PAGE (Proc. Physical SOC. 1921 34 27-32).-The absorption of 7-radiation in zinc sulphide has been measured by different methods and its value found to be approximately the same by all. No indica- tion was observed of the occurrence of an abnormal absorption when the radioactive salt was mixed with the sulphide. A table has been drawn up which gives the values of the ratio of the true to the apparent radium content for tubes of different diameters filled with a radioactive luminous compound. This table applies to a compound the apparent density of which is 2.03 grams per C.C. The value of the absorption of y-radiation from radium has been measured for zinc sulphide (OalOl) barium sulphate (0.079) barium chloride (0-loo) barium carbonate (00046) lead monoxide (0*294) J.F. 8. CHEMICAL ABSTRACTS.GENERAL AND PHYSICAL CHEMISTRY. ii. 109 bismuth sesquioxide (0.206) and uranium pentoxide (0.315) a The absorption coefficients are given in brackets in The mass absorption coefficient has been calculated in each case and it is shown that the barium salts have nearly the same value the mean value being 0.053. Electrical Conductivity of Salts and Mixtures of Salts. A. BENRATH and K. DREKOPF (Z. physikal. Chem. 1921 99 57-70).-The specific conductivity of molten mixtures of potass- ium sulphate with sodium sulphate magnesium sulphate potass- ium fluoride and lithium sulphate respectively has been determined Over the whole range of compositions and from the results relation- ships to the diagram of condition are drawn.It is shown that the conductivity isotherm can be deduced only to a small extent from the diagram of condition. It is also shown that the point a t which the eutectic appears is a well defined conductivity point. Pure salts as well as salt compounds change their conductivity with temperature according to the formula log k=a+bT. This formula is not generally applicable to mixtures of salts and mixed crystals; attempts to obtain a theoretical basis for the formula gave the relationship log k”+bT”=const. The formula of Tubandt log K=c-d/T is not generally applicable. For the complete explanation of the phenomena other factors particularly the viscosity must be known and taken account of in the calculation. The Electrical Conductivity of Anhydrides of the Higher Fatty Acids.D. HOLDE and IDA TACKE (Chem. Ztg. 1921 45 1246-1247; cf. A. 1921 i 842).-The values previously given for the molecular conductivity of oleic anhydride and oleic acid in acetone on being adjusted t o a dilution of 50 and 8 respectively gave 2 . 0 1 5 ~ 1 0 ~ and 0*53x104. If it is assumed that these two compounds are only slightly dissociated that is that the molecular conductivity increases ProportionalIy to the square root of the dilution the calculated figure for A for oleic acid becomes 1.33 x lo4 a figure not differing very greatly from that of the anhydride; the difference may be due to impurities in the acetone used. A.R. P. Measurement of Dielectric Constants. J. F. KING and W. A. PATRICK ( J . Amer. Chem. SOC. 1921 43 1835-1843).- A bridge method for the measurement of dielectric constants is described in which an arrangement consisting of an electron tube a condenser and an induction coil connected in a circuit serve as the source of alternating current of symmetrical wave form. Preliminary measurements of the dielectric constant of mixtures of ethyl alcohol with benzene ethyl ether and carbon tetrachloride respectively up to 100% by weight of alcohol are recorded. The dielectric constant increases with each addition of alcohol to the other constituent in a fairly regular manner. Electrical Moments of Carbon Monoxide and Dioxide Molecules. HANS WEIGT (Physikal.Z. 1921 22 643).-The J. F. S. 5. F. S. J. F. S.ii 110 ABSTRACTS OF CHEMICAT PAPERS. dielectric constants of carbon monoxide and carbon dioxide and their dependence on temperature have been determined with the object of ascertaining by means of the Debye theory the electrical moment of the two gases. The values obtained are p,,,=(0*1420&0.0017) x lo-" pc~=(0~l180-l-0~0016) x both of which are in good agreement with 6hat calculated from the Lorentz-Lorenz formula. J. F. S. Application of the Theory of Allotropy to Electromotive Equilibria. IV. A. SMITS (2. phgsilcal. Chem. 1921 98 455- 459 ; cf. A. 1915 ii 217 ; 1916 ii 77 ; 1917 ii 232).-The author offers a corrected interpretation of that published (" Theorie der Allotropie," Leipzig 1921,403405) for the electromotive behaviour of metals as soon as they are attacked by acids.The E.M.F. of M/Z-zinc sulphate dissolved in water O*05AT-sulphuric acid OalN- 0.5N- and l.0N-sulphuric acid and N-potassium sulphate has been determined a t 18" in an atmosphere of nitrogen. The results show that the potential of zinc is not effected by the presence of acid below 0 6 N and from this point upwards the potential is only changed by 2 millivolts. Similar experiments are described with zinc chloride in hydrochloric acid solutions; here it ii shown that a concentration of hydrochloric acid of 0.1N is without influence on the potential but that 0.5N and 1.ON cause a change in the E.M.F. of 0.021 and 0.038 volt respectively. Electromotive Behaviour of Aluminium. A.GUTNTHER- SCHULZE (2. Elelitrochem. 1921 27 579-582).-A reply to Smits in which the author maintains his previous assertions (cf. this vol. ii 20). J. 3'. S. Electromotive Behaviour of Metallic Compounds with Electron Conductivity. G . TRUMPLER (2. physikal. Chem. 1921 99 9-56).-Metallic or mixed conducting binary compounds combined with the metallic component as conductor and immersed in a solution of a salt of the metal which is saturated with the compound have a zero potential when measured against the metallic component in the same solution. The formula put forward previously by Haber for such Cases has been confirmed in the case of cuprous iodide and cuprous lead and silver sulphides. It is also shown that metallic and mixed conducting compounds combined with the metallic component as conductor possess the same potential as an indifferent electrode (platinum) when im- mersed in a solution saturated with these substances and containing the negative component both as ion and in the free condition.A necessary condition in this case is the stability of the compound towards the metallic conductor and the solution. This has been invest)igated in the case of lead cupric cuprous silver and ferrous sulphide against sodium sulphide and polysulphide solutions. If in the last case pure electrolytes are used the free negative component has no influence on the potential against the solution; this is much more dependent on the concentration of the com- ponents in the solution. This has been shown in the cells AgJAgBrJKBr,Br,(sat.) ; Ag]AgI[KI,I,(sat.).In the case of the J. 3'. S.QIEKERAI AND PHYSICAT CHEAIISTBY. ii. 111 mixed conducting silver sulphide in the presence of free sul- phur the potential lies between that of the pure components. This potential is a characteristic of the mixed conductor. The three forms of conductors metallic mixed and electrolytic are characterised by the potential relationship obtained by the above- named measurements. From the different behaviour of the metallic and electrolytic conductors (compounds) with respect to the influence of the negative component on the potential relations as obtained by the above method a fundamental difference of the internal structure is deduced according to which in the pure metallic conducting compounds the space lattice points are occupied by atoms or molecules and not by ions.They appear in opposition to electrolytes to be non-polar. It is shown that pure cuprous sulphide is an electrolyte or a mixed conductor in which the metallic component of the conductivity is very small. The passivity of lead sulphide in solutions of lead salt has been investigated and a process for activating the substance is put forward. The data in the literature on the solubility of lead sulphide and cupric sulphide are corrected. The electro- metric determination of the solubility of lead sulphide is not possible on account of the complete passivity of the lead electrode in sulphide solutions. The solubility product of lead sulphide has been determined approximately from the solubility in hydro- chloric acid and the value [Ph”]*[S”]=5 x obtained.It is shown that cupric sulphide is not stable in sodium sulphide and that in tlhe presence of copper only cuprous sulphide can exist in sodium sulphide solution and consequently the measurement of the copper potential in sodium sulphide solution does not give the solubility product of cupric sulphide (cf. Knox A. 1908 ii 830). The solubility product of cuprous sulphide is found to be [Cu’]2*[S”] =2 x 10-47. S. P. L. Trav. Lab. Carlsberg 1921 14 No. 14 pp. 31).-Biilmann’s quinhydrone electrode (A. 1921 ii 372) gives in the presence of sodium chloride too small a potential (with 3.99N-sodium chloride + 0.OlN-hydrochloric acid 0.6931 volt instead of 0,7044 volt a t 18”). This is due to a lowering of the solubility of quinhydrone and can be overcome by saturating the salt solution with quin- hydrone and with one of its constituents for example with quinol.This and various other questions are discussed theoretically in detail and the theory is verified by experiments. EINAR BIILMANN and HAKON LUND (Ann. Chim. 1921 [ix] 16 321-340).-It has previously been shown (A. 1921 ii 372) that aqueous solutions of quinhydrone may be used for the preparation of reversible electrodes with very constant potential which may serve to determine hydrogen-ion concentrations in solutions which are one-tenth molar. Concor- dant results are not obtained with stronger solutions. It is possibie to construct electrodes however in which the electromotive J. F. S. The ’‘ Salt Error ” of the Quinhydrone Electrode. SURENSEN M.S0RENSEN and K. LINDERSTRBM-LANG (Compt. rend. G. B. The Quinhydrone Electrode.ii. 112 ABSTRACTS 03 CHEMICAL PAPERS reaction is a transformation of one solid body into another solid body. The preparation of such electrodes with benzoquinone- quinhydrone and with quinol-quinhydrone is described. Such electrodes have hydrogenation potentials ~,,=0-7562 and 0.6179 respectively. They may be used for measurements over a wide range of hydrogen-ion concentration Electromotive Force produced by the Relative Displace- ment of an Electrode and an Electrolyte. ST~FAN PROCOPIU (J. Chim. Physique 1921,19,121-134).- It is found that the move- ment of an electrode in a liquid gives rise to an E.N.F. of move- ment which is general for all metallic electrodes.The dimensions of this E.M.F. are determined only by the solution pressure of the metal and the osmotic pressure of the metallic ion in the solution. The effects produced when a series of metals is moved in water nitric acid sulphuric acid salts of the metals and potassium hydroxide have been measured. The phenomenon is explained by the existence of a layer of solution round the electrode of composition different from that of the rest of the solution. This leads to a relationship between the photoelectric E.M.F. and that of movement and makes it possible to find the effect of movement on the electrical resistance of liquids. The Overvoltage of the Mercury Cathode. EDGAR NEWBERY Theory of the Electrolytic Ionic Condition and the Calcu- lation of the Electrolytic Solution Constants and of the Related Quantities from the Chemical Relationships.KARL FREDENHAGEN (2. physikal. Chem. 1921 98 38-69).-A theoretical paper in which it is shown that neither Arrhenius’s electrolytic dissociation theory nor Nernst’s theory of solution tension furnishes any information on the nature of the ionic con- dition or the nature of ionic dissociation. Neither of these theories can be directly deduced from the facts of inorganic chemistry and they do not bring these facts into relationship with one another The author has put forward an hypothesis of the iodc condition and the electrolytic solution constants which removes the above- named defects and shows how ordinary thermal dissociation passes into electrolytic dissociation. The hypothesis also shows how the solution constants of the elements and the partition and dissociation constants of chemical compounds may be calculated from the chemical relationships of the compounds and elements and from a factor which expresses the condition of the solvent.The hypothesis permits a qualitative deduction of the electrical and dielectrical behaviour of conductors of the second class. W. G. J. F. S. (T. 1922,121,7-17). J. F. S. An Electrolytic Current Intensification Effect a New Electrolytic Displacement Effect and the Connexion between Electrolysis and the Emission of Electrons in a Vacuum. 11. D. REICHINSTEIN and 3’. KLEMENT (2. physikal. Chem. 1921 99 275-289 ; cf A. 1921 ii 729).-A continuation of work previouslyUENERAL AND PI3YSIOAL CHEMISTRY.ii. 113 published. (Zoc. cit.). It is shown that in a closed circuit by means of superimposed alternating current a stronger direct current can be produced the work from which can be greater than that of the alternating current used to generate it. The action of a high frequency Ourrent on an electrolytic cell is not a specific high frequency action; qualitatively it is the same as that of a low frequency current. A11 passivity phenomena both anodic and cathodic show on superposing an alternating current on the direct current a decrease of the direct current polarisation and thereby allow themselves to be used as detectors or current intensifiers. Using the commutator method an intensification of two hundred times has been obtained and the experiments indicate that this may be still further increased. A new experiment with a highly evacuated electron tube is described which indicates that after removal of the polarisation it is possible to generate cathode rays with low potentials and cold electrodes.11. J. N. BR~NSTED (K. Danske Videnskab. Selskab. Math.-fys. Medd. 1920 3 1-21 ; cf. A. 1920 ii 78).-By a modification of the original electrometric method consisting in the use of a streaming electrode E.M.P. determinations have been carried out with one electrolyte dissolved in a concentrated solution of another. Cells solvent (n-cl) solvent (n-c,) AgC1,Ag and Of the ' Ag'AgC1 1 chloride ( c l ) I chloride (cz) I (n-c2) Ag were employed. Support is afforded to the theory that the simple gas laws are valid for ions present; in salt solutions the concentrations of which are large in comparison with that of the ions concerned.Conditions under which the law fails to apply are discussed and the results of Loomis Essex and Meacham (A. 1917 ii 353) that slight changes in the activity coefficient occur with varying concentrations are confirmed. CHEMICAL ABSTRACTS. Newer Investigations on the Anomaly of the Strong Electrolytes. LUDWIG EBERT (Jahrb. Radioaktiv. Elektronik 1921 18 134-196).-An alphabetical bibliography of the work on strong electrolytes is given. The whole subject of the behaviour of strong electrolytes is discussed under the headings (i) chief anomalies and (ii) the hypothesis of complete dissociation of the strong electrolytes and the activity of electrical forces between the ions.J. F. S . Transport Numbers of Sulphuric Acid by the Concentra- tion Cell Method. ALFRED L. FERGIUSON and WESLEY G. FRANCE ( J . Amer. Chem. SOG. 1921 43 2150-2160).-A method is described for the determination of the transport numbers of a uni- bivalent electrolyte by measurement of the potentials of con- centration cells. The transport number of the anion of sulphufic acid for concentrations between 0.1M and 0.01M has been measured and found to be 0.1868&0.0007 a t 25". This value is compared J. F. S. Applicability of the Gas Laws to Strong Electrolytes. I I silver salt (cl) silver salt (cz) solvent (n-cl)ii. 114 ABSTRACTS OF CHEMICAL PAPERS. with the values obtained by other investigators. It is shown that the dissociation values determined from freezing-point data are more satisfactory for calculating the potentials of concentration cells than those obtained from conductivity data.A correction to the formula for the potential of a concentration cell has been developed which takes into account the undissociated part of the acid. It is shown that the concentration cell method is entirely satisfactory for the determination of the transport numbers of sulphuric acid. J. F. S. Influence of Gelatin on the Transport Numbers of Sulphuric Acid. ALFRED L. FERGUSON and WESLEY G. FRANCE ( J . Amer. Chem. Soc. 1921 43 2161-2171).-The effect of gelatin on the transport numbers of gelatin has been investigated. It is shown that the addition of sulphuric acid up to 20% increases the transport number of the anion of sulphuric acid (0-1-0-01M) from 0.187 in the absence of gelatin to 0,685 in the presence of 20% gelatin.The effective concentration of O-lM and 0.01M sulphuric acid solution is reduced by the addition of gelatin. The conductivities of sulphuric acid are reduced by the addition of gelatin. An hypothesis is put forward to account for the action of gelatin in the presence of electrolytes. Ek?ctro-QsmOsis. A. 13. W. ATEN (Chem. Weelcblad 1921 18 690-692) .-The theory of electro-osmosis of liquids through porous solids is discussed and the volume passing in unit time shown to be directly proportional to the current density the dielectric constant of the liquid and the potential difference between the liquid and the porous material and inversely proportional to the viscosity and conductivity of the liquid.The various attempted technical applications of the phenomenon and of kataphoresis (which term is restricted to the motion of solid particles suspended in a liquid under the influence of the electric field) are described. s. I. L. J. F. S. Some Electro-osmotic Experiments with do Haen's Membrane Filters. P. H. PRAUSNITZ (Kolloid Z. 1921 29 293-309) .-The electro-osmosis of water solutions of ammonia sulphuric acid ammonium sulphate sodium sulphate potassium ferricyanide and aluminium sulphate through de Haen filters for varying size of pores has been investigated a t 25" and under a pressure of 43 em. of water. It is shown that the de Haen filters are particularly well adapted to the study of the influcnce of electrolytes on electro-endosmosis.Filters of medium-sized 1:ores (Nos. 20 and 320) when 5 em. thick allow 100-500 C.C. of water to pass per hour under the pressure used in the experiments. Dis- tilled water ( A = 6 x lo+) travels toward the cathode through the mem- brane under a current of 100 volts and about 0.1 ampere at the rate of 1-2 litres per hour. The addition of sulphuric acid to the anode compartment or any addition of thorium chloride or aluminium sulphate reduces the electro-endosmosis to values below thzt of water and in cases may produce even a weak anodic endosmosis. InGENERATI ANT) PRYSICAT CIIIEMISTRY. ii. 115 other cases the addition of electrolytes brings about an increased cathodic endosmosis in which the amount of water transported varies with the time the current is flowing and the nature and concentration of the electrolyte. It is shown that in general with an electrical conductivity of 2 0 0 4 0 0 x 10+ and a consump- tion of energy of 0.01-0.06 K WHllitre of transported liquid 8-16 litres/hour of liquid can be made to pass through a dia- phragm of 100 cm.2.It is not always necessary that the electrolytc should pass through the membrane as such; particularly in the case of ammonia it is shown that the addition of the base to the cathode compartment can bring about a very considerable endos- mosis of the water which is to be attributed to the wandering in the opposite direction of those ions which are responsible for the discharging of the membrane. The transport of the SO,-ion through a de Haen membrane from negative to positive compart- ments brings about a considerable passage of water from the positive to the negative compartment.In every case there is an optimum electrolyte concentration and if this is exceeded anomalies occur which are found to be connected closely wit'h the electrical conductivity. The discharge of the membrane is held to be. due in all probability to the adsorption of hydroxyl ions for all processes which effect a reduction of the hydroxyl-ion concentration also effect a reduction of the electro-endosmosis (cf. D.R.-P. 333575 1919). Attainment of Constant High Temperatures. E. MOSER (2. angew. Chem. 1921 34 625).-The apparatus or object to be maintained a t constant temperature is immersed in paraffin con- tained in a cylindrical glass vessel surrounded by an outer glass vessel of similar shape but enlarged to a spherical form a t the bottom to accommodate a liquid of suitable boiling point which is kept in ebullition so that its vapour surrounds the inner vessel.The two vessels are sealed together round the top except where a tube serves to connect the vapour jacket with a reflux condenser. The lower end of the condenser tube projects downwards for some distance into the vapour jacket and the vapour enters it fhrough a lateral orifice. The end of the tube is drawn out to a jet and disposed vertically above the boiling liquid. By this arrangement the condensed liquid is warmed again before it falls back into the jacket and is prevented from falling on the hot walls of the vessels. Liquids of high boiling point such as ethyl benzoate b.p. 212-,5" The Expansions of some Refractory Materials at High Temperatures. B. BOGITCH (Compt. rend. 1921 173 1358- 1360) .-The expansion curves for bauxite clay silica chromite and magnesia over the temperature range 0" to 1600" are given. The smallest expansion was shown by the bauxite which is suitable for use in furnaces subject to sudden changes in temperature Silica showed the most irregular expansion ; it expanded rapidly up to 600" and then only very slowly and above 1000" showed a slight contraction. The curve shows two singular points a t 210" J. F. S. have been successfully used in the outer vessel. J . H. L.ii. 116 ABSTRACTS OB CHEMICAL PAPERS. and 570" respectively which are the transformation points of cristobalite and quartz. Furnaces constructed with such bricks must only be heated up very slowly the temperature rise being less than 50" per hour.The curve for the clay bricks is intermediate between the curves for bauxite and silica. Magnesia and chromite showed the biggest expansions and such bricks are only suitable for the construction of furnaces in continuous work. Coefficient of Expansion of Molten Salts. W. HERZ (2. physikal. Chem. 1921 98 98-102).-1n previous papers (A. 1914 ii 25 245; 1921 ii 381) it has been shown that Mendeldev's equa- tion is true for the expansion of normal organic liquids. In the present paper the equation has been applied to the molten inorganic substances sodium molybdate lithium nitrate sodium nitrate rubidium bromide czsium nitrate potassium tungstate czsium iodide potassium molybdate sodium tungstate rubidium fluoride caesium sulphate and thallous nitrate.The data employed are taken from the work of Jaeger (A. 1918 ii 33). The densities calculated by the Mendelgev formula agree in most cases to about 1 per 1800 with the experimental values. The only cases where divergence is shown are thallous nitrate and sodium tungstate and here it is not much greater than in the other cases. The density values calculated by the Mendeldev equation are compared with those calculated by Jaeger (Zoc. cit.) by a single constant formula for caesium iodide and rubidium fluoride and by a two constant formula for caesium sulphate and the value of the single constant Mendelgev formula demonstrated.General Methods for the Determination of the True Mole- cular Constitution of Pure Substances and their Mixtures. MARIO BASTO WAGNER (2. physikal. Chem. 1921 98 81-93).-A theoretical paper in which the author differentiates between the thermodynamics of mixtures treated in connexion with their independent components and the thermodynamics of mixtures treated in connexion with their true components. The advantages and drawbacks of both methods of treatment are explained and the nature of the results obtainable is pointed out in each case. The various properties which may be used in the determination of constitution are enumerated and their application is indicated. These include volume heat change vapour pressure solubility freezing point boiling point specific heat compressibility thermal coefficient of expansion and osmotic pressure.The properties dielectric constant viscosity refraction and surface tension do not permit conclusions being drawn which may be used in the second method of treatment. Thermodynamics of Mixtures. IX. MARIO BASTO WAGNER (2. phy?ilcal. Chem. 1921 98 151-156; cf. A. 1921 ii 375).-A theoretical paper in which the theory of the heat of mixtures is developed on the basis of the thermodynamics of mixtures treated in connexion with the true components (cf. preceding abstract). J. F. S. W. G. J. F. S. J. F. S.GENERAL AND PIXYSICAL CHEMISTRY. ii. 117 Theory of Equations of State. 11. MARIO BASTO WAGNER (2. physikal. Chem. 1921 98 24G251; cf. preceding abstract).- The hypothesis developed previously is extended.The present paper deals with the theory of the heat of vaporisation and the internal pressure. J. F. S. Equation of Condition of Nitrogen at Small Pressures and Low Temperatures. R. BARTELS and A. EUCKEN (2. physikal. Chem. 1921 98 70-79).-The equation of condition of nitrogen has been measured by means of a constant pressure gas thermometer between the temperatures 90" and 76" Abs. and between the pressures 0.5 and 1-0 atm. The results are repre- sented by means of an empirical formula for the corkection quantity B' of the form v=RT/~-RTB'. This formula has been used to calculate the molecular heat of nitrogen for the ideal gas con- ditions and the value 6.873 obtained for Cp. This value is 0.07 cal. or 1 yo lower than that (6.944 cal.) demanded by the classical theory whilst the Berthelot equation of condition leads to a value 0.2 cal.lower than the classical value for the temperature 92" Abs. A comparison of the reduced B' values for a range of temperature shows that the curves for hydrogen and nitrogen cannot be brought into coincidence but that at low temperatures they diverge. J. F. S. Equation of Condition. II. Principal Equation of Con- dition. 111. Principal Equation of Condition and the Equation of Condition for Individual Substances. IV. Compressibility Equations of Liquids. A. WOHL (2. physikal. Chem. 1921 99 207-225 226-233 234-241; cf. A. 1914 ii 430).-A series of theoretical papers in the first of which the author further develops the equation of condition previously published (loc. cit.).The various relationships brought out by this equation are considered and compared with the equations of van der Waals and Kamerlingh Onnes. I n the second paper equations of condition are developed and considered for helium hydrogen water and ethyl alcohol. The third paper deals with compressi- bility equations for liquids particularly highly compressed sub- stances. The substances ethyl ether carbon dioxide and hydrogen are considered. J. F. S. Extension of Ebullioscopy and its Application to Binary Mixtures. EBNST BECKMANN and OTTO LIESCIIE (2. physikal. Chem. 1921 98 438454).-The authors have combined the Raoult equation P=Po[N/(N+n)] with Dalton's law P=B[N'/ (N'+n')] to form the equation Po[N/(N+n)]=BIN'/(N'+n')] and shown that this equation may be used generally in the ebullioscopy of binary mixtures and may be extended to meet complicated cases since the determination of the absolute boiling point fixes the value of Fo with the help of vapour pressure tables or formula.The introduction of the boiling point difference A and the ebullio- scopic constant E leads t o a special ebullioscopic equationii. 118 ABSTRACTS OF CIIEMlCAL PAPERS. A=E(n-n') which may only be used for small A values. If in either of the above equations n or n' is given the value zero the two limiting cases are obtained in which the second component (represented by small letters) is non-volatile or is not soluble in the first component. J. F. S. A Vapour Pressure Formula with a General Integration Constant. F. A. HENGLEIN (2. physikal.Chem. 1921,98,1-13).- A theoretical paper in which the constants of a previously pub- lished rule (A. 1920 ii 732) €or the comparison of the vapour pressure curves of two substances log T,=a log T,+b have been brought into relationship wit'h the Trouton constants. From this relationship a vapour pressure formula has been deduced which contains two specific constants and has the form log p(atm.) = - k/T,'+ 4.6222. The formula holds in comparison with other two constant vapour pressure formulq over greater temperature ranges and is characterised by the fact that the integration constant is the same for all substances. The index n is shown to be equal to ~ / ( v - - v ' ) p . K where K has the same value €or all substances and consequently A/(v-d)p may be calculated from the above formula; on the other hand n may be determined from A/(v-d)p or AIRT.To calculate the actual formula two vapour pressure determinations are neces- sary or the knowledge of one vapour pressure point and the molecular latent heat of vaporisation a t the same pressure. The formula is shown t o be equally true for sublimation pressures and it may also be applied to decomposition pressures. Theory of Binary Mixtures. VII. Mixtures of Ethyl Ether and Bromoform. F. DOLEZALEK and M. SCHULZE (2. physikul. Chem. 1921 98 395-429; cf. A. 1920 ii 32).-The vapour pressure of a series of mixtures of ethyl ether and bromo- form has been determined a t 25" and 75" ; the contraction on mixing the coefficient of expansion and the density of the mixtures have been determined a t 25".The results show that on mixing the components a mutual partial complex formation occurs. The resulting compound consists of one molecule of ether and one molecule of bromoform; it occurs in the solution in a dissociated condition and on heating it decomposes. Bromoform itself is strongly associated but with decreasing concentration the associa- tion decreases. The vapour pressure was determined by a new static method. A theory is developed whereby the molecular constitution may be calculated from the vapour pressure measure- ments. From two measured vapour pressures the combination constant K and the association constant k were determined and from these the molecular constitution of the mixture was calculated for 25". The results show that a t 25" bromoform consists of 65% of bimolecules and that in the mixture of 1 mol.of ether and 1 mol. of bromoform 23% of the eyui-molecular compound is present. The vapour pressure for ail mixtures has been calculated from the molecular composition and found to be in good agreement with the experimental values. A marked coritraction occurs on mixing J. F. S.GENERAL AND PHYSICAL CHEBLISTRY. ii. 11.9 ether and bromoform the contraction curve is strongly depressed on account of the association of the bromoform. The contraction has been calculated on the basis of the theory and found to agree with the observed contraction. Prom the coefficient of expansion the contraction a t 15" 20" and 30" has been calculated. From the contraction a t 25" the density of the mixtures was calculated and the results show that the measurement of volume change by the differential method is the most exact for the measurement of easily volatile liquid mixtures.Vapour Pressure of Binary Mixtures. GERHARDT C. SCHMIDT (2. physilcal. Chem. 1921 99 71-86).-A new simple static method of measuring the vapour pressure of liquids is described. It is shown that the dynamic method of measuring the vapour pressure of mixtures gives only moderate results and then only when the experiments are made under definite conditions and it is inferior in every way to the new static method. The vapour pressures of the following mixtures benzene-ethyl ether benzene-chloroform benzene-methyl acetate methyl acetate- ethyl acetate benzene-carbon disulphide benzene-toluene toluene- carbon disulphide ethyl alcohol-methyl alcohol benzene-methyl alcohol chloroform-carbon disulphide and acetone-chloroform have been measured up to a pressure of 4 atmospheres.By plotting the molecular compositions as abscissz and the vapour pressures as ordinates it is found that a t all temperatures (20-100") mixtures of ethyl alcohol and methyl alcohol give a straight line. Mixtures of benzene and ether and of benzene and chloroform give approximately straight lines up to 30" but a t higher temperature curves which a.re convex to the abscissa axis and the higher the temperature the greater the convexity. Mixtures of acetone and chloroform a t 0" give a straight line and a t higher temperatures curves convex to the abscissa axis which become more convex as the temperature is raised.Mixtures of benzene and toluene ethyl acetate and methyl acetate and benzene and methyl acetate at low temperatures give straight lines but a t higher temperatures the curves are concave to the abscissa axis and the concavity becomes greater as the temperature is raised. Mixtures of benzene and carbon disulphide toluene and carbon disulphide chloroform and carbon disulphide and benzene and methyl alcohol give at 0" curves which are concave to the abscissa axis and the higher the temperature the greater the concavity this being particularly noticeable in the case of benzene- methyl alcohol. From the fact that the higher the temperature the greater the concavity or convexity the conclusion can be drawn that for every liquid mixture there is a temperature a t which the relationship between the vapour pressure and the molecular com- position may be represented by a straight line generally speaking this occurs at low temperatures.It is shown tha.t a series of deduc- tions from Dolezalek's theory of binary solutions which refer to the influence of temperature on the course of the isotherms are not confirmed by the experiments. Consequently the theory cannot be maintained. J. F. S. J. F. S.ii. 120 ABSTRACTS OF CHEMICAL PAPERS. Intersecting Vapour Pressure Curves and Deductions therefrom. CARL VON RECHENBERG (2. physikal. Chem. 1921 99 87-104).-A large number of pairs of substances are recorded for which the vapour pressure curves intersect. From a comparison of the intersecting vapour pressure curves the following conclusions are drawn.All liquid and solid substances consist of large molecular complexes of undetermined variable composition of different stability and present as many classes of compounds as they have different molecular types. Every individual com- pound behaves in a characteristic manner. The division of liquids into associated and non-associated groups from surface tension measurements is incorrect. All compounds are associated but the so-called associated compounds are more loosely combined. The sub-division only separates the most divergent compounds but does not take account of the intermediate and transition com- pounds. A table is drawn up of the various aggregates according t o the increase of vapour pressure with increase of temperature.The table is very similar to Rothmund's solubility table from which it follows that vapour formation and solubility are parallel phenomena and must be due to the same cause. Vaporisation of Solutions of Liquid Pairs possessing Intersecting Vapour Pressure Curves. CARL VON REGHENBERG (2. physikal. Chm. 1921 99 105-115 ; cf. preceding abstract).- The Bancroft rule that liquids with intersecting vapour pressure curves form solutions with a characteristic point has been con- fimed when notice is taken of the fact that the property of the characteristic solution only obtains inside a limited pressure region. A large number of homogeneous mixtures with intersecting vapour pressure curves are tabulated with the minimum boiling point and the concentration of the characteristic fraction a t 760 mm.J. F. S. J. F. S. Distillation and Rectification. L. GAY (Chim. et I d . 1921 6 567-578 ; cf. this vol. ii 85).-In order to obtain the maximum efficiency from a column ceteris paribus the higher the tem- perature of the initial mixture the lower should be the level of its introduction into the distillation apparatus. This level should be such that an initial liquid or gaseous mixture meets a liquid or gaseous mixture of approximately the same composition in the column. If the liquid mixture is a t its boiling point or rather if the gaseous mixture is a t its condensation point then the com- position of the two mixtures is identical. The determination of the minimum intensity of heat and the minimum number of com- partments of the column necessary for an initial gaseous mixture has been carried out in a similar manner to that used for an initial liquid mixture (Zoc.cit.) and applied to the separation of practically pure water and alcohol from any gaseous mixt,ure of water and alcohol at 95". The determination of the minimum intensity of heat necessary t o remove a trace of impurity from a liquid has also been carried out. F. M. R.GENERAC AND PHYSICAL CHEMISTRY. ii. 121 Heat of Formation of Silver Iodide. HUGH STOTT TAYLOR and WILLIAM THEODORE ANDERSON jun. (J. Amer. Chem. Sec. 1921 43 2014-2017).-Eraune and Koref's calorimetric measure- ment of the heat of formation of silvcr iodide (A. 1914 ii 536) using AT-potassium cyanide instead of 3N-solutions as solvent for silver iodide have been repeated.The result 15,100 cal. per mol. of silver iodide has been obfained which confirms Braune and Koref's value. J. 3'. S. Relationship between Heat of Solution and Heat of Fusion of Organic Substances. GEORG GEHLHOBF (2. physikal. Chem. 1921,98,252-259) .-The heat of solution of naphthalene diphenyl diphenylamine azobenzene a-naphthylamine p-naphthylamine urethane paraffin (wax) nitronaphthalene nitrophenol m- dinitro - benzene phenol menthol thymol pyrogallol resorcino1 citric acid and chloral hydrate has been determined in 1-3% solutions in the solvents water benzene ethyl ether alcohol carbon disulphide acetone amyl acetate aniline chloroform and acetic acid. The results show that for a series of substances particularly hydro- carbons aromatic amines urethane and azobenzene the heat of solution in various solvents is practically identical with the heat of fusion.Organic substances containing the hydroxyl group give very different values for the heat of solution and the heat of fusion whilst substances containing the nitro-group take up an intermediate position between the other substances. The heat of solution of the sodium potassium and ammonium haloids has been deter- mined in aqueous solution and the results are discussed in con- nexion with the structures and space lattices of these salts. J. P. S. Heats of Solution and of Transformation of the Acido- and Aquo-cobalt Pentammines. ARTHUR B. I A M B and JOHN P. SIMMONS ( J . Amer. Chem. SOC. l921,43,2188-2199).-A method of calorimetry has been developed in which only observations 011 a silver coulometer and of time and electrical resistance are required.It is shown that the same cobalt sulphide CopS3 is produced by the action of sodium sulphide on chloro- and aquo-pentammine- cobalt chloride in aqueous solution. The solubility of chloro- pentami~inecobalt chloride in water a t O" 30° and 58" has been determined and the values 2.244 6.047 and 10.01 respectively have been obtained in grams per litre. The heat of solution has been calculated from these data and the vdue 13,440 cal. obtained which is in fair agreement with the experimental value. 12,400 cal. The heats of solution of a number of acido- and aquo-pentammine- cok)alt salts in water have been measured. The following values are found a t 25" in cals.chloro-chloride -12,420 ; aquo-chloride -6,460 ; bromo-bromide - 12,660 ; aquo-bromide -9,220 ; nitrato-nitrate - 14,900 ; and aquo-nitrate - 15,300. The heat of reaction of solutions of pentammine saks with solutions of sodium sulphide has been measured at 25" and that of the solid salts with sodium sulphide solutions calculated. Theii. 122 ABSTRACTS OF CHEMICAL PAPERS. following values in cals. a t 25" are obtained chloro-chloride 15,070 ; aquo-chloride 12,930 ; bromo-bromide 13,290 ; aquo- bromide 11,600 ; nitrato-nitrate 12,340 and aquo-nitrate 11,860. From these results the heat of transformation of the acido-salt to the aquo-salt in each case has been calculated and the values chloride 2,140 cal. ; bromide 1,690 cal. and nitrate 480 cal.have been obtained. J. F. S. Heat of Wetting of Silica Gel. W. A. PATRICK and F. V. GRIM ( J . Amer. Chem. Xoc. 1921 43 2144-2150).-The heat change occasioned by wetting silica gels with water ethyl alcohol benzene carbon tetrachloride and aniline respectively has been determined a t 25". The quantity of liquid used in each experiment was 50-07 C.C. and the amount of gel 2-3-2.5 grams. The follow- ing values in cals./gram of gel were obtained water 19.22 ; alcohol 22.63 ; benzene 11.13 ; carbon tetrachloride 8.42 and aniline 17.54. The heats of wetting have been discussed and found to be in keeping with the changes of surface energy. J. F. S. Viscosity of Gelatin Sols. ROBERT H. BOGUE ( J . Amer. Chem. Xoc. 1921 43 1764-1773).-A number of experiments have been effected with gelatin sols to determine the relation between viscosity and concentration.The data obtained have been applied to Hatschek's formula for the viscosity of emulsoids and it is shown that the value of A'IA representing the volume occupied per unit weight of dispersed phase is not a constant with varying concentration but rises regularly to a maximum and thereafter regularly diminishes with increasing concentration. This behaviour is not,iced also in the case of other colloids. A tentative explanation is presented based on the effect which increas- ing concentrations of dispersed phase will have on the surface tension of the dispersion medium. Assuming the equilibrium surface tension solvation potential an empirical relation is developed which with a fair degree of accuracy defines the equi- librium.At very high concentrations a reversal of phase probably occurs a t which no expressions yet developed adequately represent the relations involved. A high degree of solvation is shown to be indicated by a high coefficient of viscosity that is solvation and viscosity are parallel functions. Isoelectric gelatin at a hydrogen-ion concentration 2x10-5 is shown to have the lowest viscosity and the lowest degree of solvation gelatin chloride a t a hydrogen-ion concentration 3.1 x the highest and calcium gelatinate a t a hydrogen-ion concentration 2.5 x is intermediate. If an excess of acid is allowed to remain in the gelatin solution even although the acid be of very low concentration the viscosity and degree of solvation will be reduced.A New Viscostalagmometer for the Estimation of Surface Tension and Viscosity of Liquids of Very Different Limpidity. I. TRAUBE (Biochem. Z. 1921 120 106-l07).-The essential feature consists in the adaptation to the ordinary form of stalagmo- ,J. F. 8.QENERAL AND PHYSICAL CHE3TISTRY. ii. 123 meter of five interchangeable ground-in mouth-pieces of differ- ing capillary bore thus allowing measurements to be performed with the same apparatus on liquids of a great range of fluidities. H. K. Ionisation and Surface Activity of Aqueous Solutions of Aliphatic Acids. WILH. WINDISCII and PHIL. OSSWALD (2. physikal. Chem. 1921 99 172-188).-The connexion between ionisation and surface activity of organic acids has been investigated. It is shown that of N/lO-acid or alkali when added to an aqueous solution acts only as a diluent that is the surface activity is independent of the hydrogen-ion concentration. Experiments have been made to determine the concentration of acids of different strengths such as hydrochloric formic and acetic acids which must be added to solutions of nonoic acid so that the same size of drops may be obtained and also the concentrations of these acids which will produce the same colour with methyl-orange solution.It is shown that free nonoic acid in aqueous solution is 10% dissociated and therefore the addition of acid increases it,s surface activity. The simultaneous presence of sodium chloride is without action on the activity. With continued addition of acid the size of the drops sinks to a constant minimum.A solution of sodium nonoate exhibits a small activity because through hydrolysis a small quantity of acid exists in the free condition. By the addition of a small quantity of alkali the free molecules are ionised and the size of the drop sinks to that of water. An addition of hydrochloric acid produces fresh acid molecules and the size of the drop sinks to a minimum which is reached when all the aliphatic acid ions have disappeared. Consequently it follows that the surface activity of aqueous solutions of aliphatic acids is due to ionisation processes and the possibility of calculating the size of the drops due to the action of various additions is pointed out. J. F. S. High Pressure due to Adsorption and the Density and Volume Relations of Charcoal.WrrmAni D. HARKINS and D. T. EWING ( J . Amer. Chem. SOC. 1921 43 1787-1802).-The present paper furnishes evidence in favour of the hypothesis pre- viously expressed (A. 1921 ii 87) by the authors which states that the liquids which penetrate into charcoal are compressed by a force due to molecuIar attraction which acts as a pressure of many thousand atmospheres (20,000 or more). The present paper shows that it is the liquid in the micro-pores and not the macro- pores which is compressed and consequently the pressure mould probably be much greater than the figure mentioned. Kot only charcoal but other porous substances and fine powders should exert this compressive effect but to a much smaller extent. A number of experiments are described in which the compression is demonstrated by volume changes but the values measured have not exceeded 72 atlms.Ethyl ether which is much more compressible than water occupies a volume in charcoal which isii. 124 ABSTRACTS OF CHEMICAL PAPERS about lOyo less than that occupied by the amount of water which outside the charcoal is equal in volume to the ether. It is probable that the water in the micro-pores is compressed by about 25% or more whilst the ctlicr is compressed by 40%. The liquids in tho macro-pores of the diameter 1 * 2 ~ l O - ~ em. are not compressed sufficiently to produce a noticeable effect on the volume. The true volume relations in 1 C.C. of a characteristic coconut shell charcoal are 0.28 C.C. micro-pores 0.18 C.C. macro-pores 0.54 C.C. carbon which give a density of 1.60 for the carbon.The density of the lumps of such a charcoal is 0.866. Charcoals which are inactive as adsorbents of gases do not exert a compressive effect on the adsorbed liquids of a sufficiently high magnitude to be very evident although there appears to be a slight effect of this nature. The densities of the carbon in the two inactive wood charcoals investigated are 1.65 and 1.50. When coconut-shell charcoals alone are considered it is found that the lower the apparent density in an organic liquid the less is the adsorptive action on vapours and it is indicated that this relation may be expected to hold better the more compressible the liquid which is adsorbed so that either ethyl ether or pentane or another highly compressible liquid should be used in such tests.Simple thermodynamic equa- tions are given for the heat of immersion or adsorption of a plane surface. Whilst there is probably no definite area of surface inside a lump of charcoal a definition for an apparent area may be given. The one chosen here is that the apparent area with respect t o the heat of immersion is the area of the plane surface of carbon which will develop the same amount of heat on immer- sion as is equal to that developed by the immersion of 1 gram of the charcoal in the same liquid. Since the film in the charcoal is probably a number of molecules thick this apparent area is prob- ably larger than corresponds with the carbon surface. This method indicates that the area of 1 gram of charcoal is 120 sq. metres.The magnitude of the heat of immersion of liquids on mercury is 3.3 x 10-6 cal. for octane sncl 3.25 x cal. for isobutyl alcohol. J. F. S. The Origin of the Potential Differences Responsible for Anomalous Osmosis. ~ACQUES LOEB ( J . Gen. Physiol. 1921 4 213-226).-1f a collodion bag coated on the inside with protein (gelatin) separate two identical acid solutioiis and a neutral salt be added to the inside the rate of diffusion inwards through the membrane is a minimum a t the isoelectric point of the protein increases with the acidity to a maximum and then decreases. The potential difference bettween the two solutions varies similarly. In the absence of protein these characteristic variations are not obtained. The anomalous variation of the potential difference is apparently n consequence of the Donnan equilibrium between solution and protein and this potential difference causes electrical endosmosis and hence the irregular rates of diffusion.The potential difference is also influenced by the diffusion potential but this exists although no protein is present. W 0. K.GENERAL AND PEYSICAL CXlEMISTRY. ii. 125 Penetration of Electrolytes into Gels. 111. Influence of the Concentration of the Gel on the Coefficient of Diffusion of Sodium Chloride. WALTER STILES and GILBERT SMITHSON ADAIR (Biochem. J. 1921,15,620-626).-The coefficient of diffusion from normal solution into gelatin and agar gels was measured by the indicator method previously described (A. 1920 ii 235). With increasing concentration of gel the coefficient of diffusion at first decreases rapidly according to an exponential relation but with concentrations above 2% the curve connecting coefficient of diff u- sion and concentration is approximately a straight line.It is shown how this result can be harmonised with the relation between diffusion and viscosity found by Einstein Sutherland and von Smolukowski. The diffusion coefficients found when extrapolated for diffusion in water are about 7.5% higher than those obtained from the results of Oholm (A. 1905 ii 147). This is attributed to the fact that in Oholm's experiments the salt concentration refers to ionised+un-ionised salt whereas the authors measure only the rate of penetration of the ions which diffuse about twice as fast as the un-ionised salt. The resistance which even a stiff jelly offers to salt diffusion is quite small; the rate is reduced by less than 2y0 for each gram of gelatin added to 100 C.C.when the concentration exceeds 2%. G. B. Penetration of Electrolytes into Gels. IV. Diffasion of Sulphates. WALTER STILES (Biochem. J. 1921 15 629-635) .- In 0.5% agar ammonium potassium sodium and magnesium sulphates diffuse almost as fast as in pure water. In 10% gelatin the rate is decreased by about 25%. The diffusion coefficient of copper sulphate in agar is considerably greater than in water and in ISYO gelatin less than in water but still greater than the values for other sulphatea indicate. The abnormal results of von Fiirth and Bubanovid (A. 1920 ii 94) have not been confirmed. The comparatively small uptake of sulphates by plant tissue is not to be explained on the basis of retarded diffusion through a gel.A General Theory of Solutions of Electrolytes. JNANENDRA CHANDRA GHOSH (2. physikal. Chem. 1921 98 211-238).-A theoretical paper in which the author develops a new theory of electrolyte solutions on the basis (i) that in solutions of strong electrolytes only ions exist and (ii) in solutions of weak electrolytes or those of medium strength there exists an equilibrium between a non-polar form (non-electrolyte) and a polar form (which behaves as a strong electrolyte). This hypothesis is examined in connexion with the data obtained in recent years on electrical conductivity activity coefficients conductivity in non-aqueous solvents influence of salts on solubility relations absorption of light by salt solutions and the relation between osmotic pressure and activity coefficients.Theory of Strong Electrolytes. H. KALLMB" (2. physihl. Chern. 1921 98 433437).-A number of criticisms of GhosPs G. B. J. l?. S.ii. 126 SBSTRACTS O F CHEMICAL PAPERS. hypothesis of solutions of electrolytes (T. 1918 113 449 627 707 and preceding abstract) are put forward. GEORGE PUCHER and WILLIAM 91. DEIXN ( J . Arne;.. . CY~enz. Xoc. 1921 43 1753- 1758; cf. A. 1917 ii 445).-The solubility of forty-five organic substances has been determiiied a t 20-25" in ethyl alcohol quinoline and eyuimolecular mixtures of the two solvents. In some cases the solubility in the mixed solvent lies either above or below that in the individual solvents; this is most marked in the case of carbamide " saccharin," p-nitrophenol acetanilide resor- cinol piperine cinnamic acid and chloral hydrate and in all these cases compound formation is indicated.The solubility of carb- amide " saccharin," brucine and o-nitrophenol in a series of mixtures of alcohol a i d quinoline has been determined a t 2 5 O and the solubility G€ lactose a t 1" and 25" and raffinose at 25" in a series of mixtures of water and pyridine has also been determined. It is shown that the solubility of carbamide decreases quite regularly as the concentration of quinoline increases but a t 60% quinoline a break occurs and thc compound of one molecule of quinoline and three molecules of carbamide is formed m. p. 154". Compound formation is indicated in the case of " saccha,rin," brucine lactose and raffinose but no definite compound was isolated. Compound Formation and Viscosity in Solutions of the Types Acid-Ester Acid-Ketone and Acid-Acid.JAMES KENDALL and ELIZABETH BRAKELEY ( J . Anzer. Chem. SOC. 1921 43 1826-1834; cf. Kendall and Gross this vol. ii 32 33).- The viscosity-composition curves a t 25" have been determined for the systems trichloroacetic acid-ethyl acetate acet'ic acid- ethyl acetate trichloroacetic acid-ethyl benzoate acetic acid- ethyl benzoate trichloroacetic acid-acetone acetic acid-acetone trichloroacetic acid-acetophenone acetic acid-acetophenone and trichloroacetic acid-acetic acid. The results obtained indicate extensive compound formation in solutions of the esters and ketones with a strong acid (trichloroacetic). The same substances with a weak acid (acetic) give viscosity curves which are much less abnormal although compound formation is still evident.The curve for the system acetic acid-trichloroacetic acid exhibits con- siderable compound formation and an examination of the results of previous workers on systems of the types acid-acid and acid-base shows that this behaviour is general except when the components are of similar acidic strength. The rules formulated in previous papers relating to the extent of compound formation with diversity in character of the components are confirmed by the present work. Compound Formation and Conductivity in Systems of the Types Formic Acid-Metal Formate and Sulphuric Acid- Metal Sulphate. JAMES KENDALL HOWARD ADLER and ARTHUR W.DAVIDSON ( J . Amer. Chem. Soc. 1921,43 1846-1853; bf . preceding abstract).-Electrical conductivity determinations J. F. S. Solubilities in Mixtures 01 Two Solvents. J. F. S . J. F. S.GENERAL AND PHYSICAL CHEMISTRY. ii. 127 have been made of solutions of various concentrations of barium magnesium and lead formates in anhydrous formic acid at 25" aiid of solutions of sodium barium calcium magnesium zinc lead and silver sulphat,es in anhydrous sulphuric acid. I n this case the solutions contained 0.04 equivalent of solid sulphate. Comparing the results for the formates with those obtained for other formates in formic acid by Schlesinger (&4. 1919 ii 91; 1920 ii 72 73) it is shown that a definite parallelism exists between the degree of ionisation and the extent of compound formation.I n the case of the sulphates however the uncertainty of the correction to be applied for thc solvent is so great as to obscure the results. Vari- ations in compound formation solubility and ionisation in systems of the general type HX-RX are all more extreme the weaker the acidic radicle X . Hydroxides in aqueous solution differ more widely than formates in formic acid and these are more diverse than sulphates in sdphuric acid solution. The Photography of Opaque Crystals. 119. FRANQOIS and CH. LORMAND (Bull. SOC. chim. 1921 [iv] 29 1056-1059).-The application of the apparatus previously described for the photo- graphy of transparent crystals (cf. A. 1921 ii 626) to the photo- graphy of opaque crystals is given in detail.Stereoscopic Photography of Crydals. M. PRANQON and CH. LORMAND (Bull. SOC. chim. 1921 [iv] 29 1059-1063).-For this purpose a microscope the tube of which can be inclined suc- cessively t o the left and right is used. It is fitted with the apparatus previously described for t'he photomicrography of crystals (cf. A 1921 ii 626 and preceding abstract). A photograph is taken with the tube inclined to the Ieft and then one with it inclined equally to the right. The examination of the two photographs in a stereo- scope gives a stereoscopic picture of the crystals in relief. Ionic Properties and Crystallo-chemical Relationships. I. The Properties of the Ions which appear in Crystals. H. GRIMM (2. physikal. Chem. 1921 98 353-394).-,4 theoretical paper in which it is shown that the electrostatic action of the ions in crystals is due to the following ionic properties (i) the charge of the ions (ii) the radius of the ions (iii) the structure of the ions particularly the surface and number of exterior electrons.The ions are divided into five principal groups depending on the number of exterior electrons. (i) Ions of the helium type (2 exterior electrons) (ii) ions of the Ne A Kr Xe and emanation type (8 ex- terior electrons) (iii) ions of the Cu+ Ag+ Au+ type (probably with 18 exterior electrons) (iv) ions of the Te+ and PbT+ type (v) ions of the type Mn++ Fe++ CO++ Ni++ (transition ions). It is shown from many compounds and elements that the expression $A-$Ke>.$xe-+Kr > +Kr- $A holds for many physical properties where $ is the numerical value of the property and the index is that of the inactive gas the structure of which comes nearest to that of the ion concerned.This inequality is due to an analogous inequality of the ionic radii. The ionic radii of 0- S- Se- Te-; AIg++ Cat+ Sr++ Ba++ have been calculated and that J. F. S. W. G. W. G.ii. 128 ABSTRACTS OF CHEMICAL PAPERS. of Cs+ has been approximated. The series order of the other ionic radii has been deduced from the molecular volumes of iso- morphous compounds in which the interchangeable ions have the same number of exterior electrons. Ions with different structures and different exterior sheaths show a similar electrostatic action when the difference in the ionic radius compensates the other properties. J. F.S. Structure of Organic Crystals. SIR W. H. BRAW (Proc. Physical Soc. 1921 34 33-50).-The author has determined the X-ray spectra of naphthalene a-naphthylamine benzoic acid acenaphthene p-naphthol and a-naphthol by the powder method. To interpret the results he has made the assumption that the benzene and naphthalene rings are actual structures having definite size and form and that they are built as a whole into the organic substances in which they occur. It is shown from crystallographic and other data that this hypothesis is an extremely probable one and it leads to results which are in excellent agreement with crystallographic data. It is shown that the unit cell of naphthalene contains two molecules and has a mass of 213 A.U. where 1 &U.= 10-24 grams the linear dimensions of the cell are a=8-34 b=6.05 and c=8.69 B.U.Two molecules of anthracene occupy the unit cell and this has the linear dimensions a=8.7 b=6*1 and c = l l . 6 A.U. The unit cell of acenaphthene contains four molecules and has linear dimensions 8.32 14.15 and 7.26 A.U. respectively whilst a-naphthol has a unit cell which also contains four molecules and has linear dimensions a=13*1 b=4*9 and c=13.4 A.U. ; the same number of molecules occupy the unit cell of p-naphthol and here the linear dimensions are a=5.85 b=4.28 and c=8.7 A.U. The unit cell of a-naphthylamine also contains four molecules and has the dimensions a=8.62 b=14.08 and c=7*04 A.U. Benzoic acid has a wide spacing between the planes the unit cell contains four molecules and has the dimensions a=5.44 b=5.18 and c=21*8 B.U.I n a note the author stated " It is convenient in this work to extend the Angstrom system of units so that an A.U. of area is cm.2 of volume ~111.~ and of mass 10-24 grams." Romtgen Spectroscopic Investigation of Organic Com- pounds. K. BECKER and W. JANCKE (2 physikab. Ghem. 1921,99,242-266,267-.274).-1n the first part an account is given of the X-ray examination of the structure of a number of organic substances. The substances were examined in the form of compressed pastilles by a method which is a combination of that due to Bragg and Debye-Scherrer. In the case of indigotin it is found that the unit cell is an hexagonal parallelopiped with a rhombus as base of edge 19.55 A.U. and height 11.80 A.U. This contains 12 indigotin molecules.The figures refer to indigotin which has been subjected to a pressure of 4000-5000 atoms. Indi- gotin which has not been so compressed has a unit crystal cell with the Linear dimensions a=b=20-20 A.U. c= 12.15 B.U. which means that there is a volume contraction of 9.S% the base de- creases 3.2% and the height %9y0. Carbamide has a unit cell J. F. S. I. and 11.GENEIAAL AND PHYSICAL CHERIISTRY. ii. 129 of dimensions a=b=8.75 A.U. c=7-24 A.U. and contains 8 mole- cules. The unit cell of succinic acid contains 2 molecules and has dimensions a=5.00 A.U. b=8.20 A.U. c=5-40 A.U. p=94.7O0. Compression reduces all values the c value relatively most the values being volume 10*3y0 a 2.2y0 b 4-8% and c 5.4%. Measure- ments are recorded for quinol quinone anthraquinone anthracene naphthalene phenanthrene resorcinol phthalic anhydride phthalic acid cinnamic acid (trans) p-phenylpropionic acid azobenzene hydrazobenzene lithium oxalate maleic acid succinic acid maleic acid d- and I-tartaric acids pentaerythritol a-methylglucoside and acetylenedicarboxylic acid.The results show that the number of molecules in the unit crystal cell is always greater than one and is not the same for nearly related substances such as maleic acid and succinic acid. I n the second paper the results of X-ray examin- ation of a large number of aliphatic acids are given. The results are arranged in four groups. I. Acids with an odd number of carbon atoms formic propionic Valerie heptoic nonoic undecoic. 11. Acids with an even number of carbon atoms acetic butyric octoic lauric myristic palmitic stearic (also aa’- distearin). 111.Isomeric acids isobutyric isovaleric trimethylacetic. IV. Unsaturated acids crotonic undecenoic oleic and elaidic. I n many cases the acids were measured as their lithium salts. It is shown in the case of the even-number saturated acids apart from the first two members that they possess an hexagon symmetry with 72 molecules in the unit crystal or possibly rhombohedra1 with 24 molecules; the uneven-number acids are tetragonal with 24 or 48 molecules in the unit cell. There is no fundamental difference between the free acids and the lithium salts. J. F. S. Mixtures of Anisotropic Liquids and the Identity of Grandjean’s Stratified Liquids with Liquids of the Azoxy- phenetole Type.G. FRIEDEL and L. ROYER (Compt. rend. 1921 173 1320-1322; cf. Grandjean A. 1921 ii 91).-In the case of mixtures of substauces capable of furnishing anisotropic liquids the two phases the conic and thread-like (liquides a coniques et liquides A fils) are perfectly distinct and separated from one another by an absolute discontinuity. Mixtures of Grandjean’s stratified liquids with other anisotropic liquids always show an absolute discontinuity between the stratified liquid and the conic phase. On the other hand these mixtures do not show any dis- continuity between the thread-like phase and the stratified liquid of Grandjean. It is suggested therefore that Grandjean’s stratified liquids are only a special aspect of the thread-like liquids of the type of azoxyphenet ole.W. G. Some Fundamental Conceptions of Colloidal Chemistry. RICHARD ZSIGNONDY (2. physikaI. Chem. 1921 98 14-37).- A theoretical paper in which the nature of matter in the col- loidal condition is considered. The views put forward by Niigeli (“ Garung,” Munich 1879) on the nature of micell= are considered and it is shown that in general these views are in keeping with VOL. CXXII. ii. 5ii. 130 ABSTRACTS OF CHEMICAL PBPERS. experimental data. The conceptions to be generally adopted are expressed as follows by micellae in its broadest sense is to be understood a molecular complex of the dispersed material which is not permeated by the dispersion medium. I n the narrower sense micellae are crystalline ultramicroscopic particles of the dispersed material.Micellar complexes are formed by the association of micellq and are ultramicroscopic particles generally but in isolated cases may be microscopic. According to the form of the association micellar complexes form on the one hand microscopic or ultramicroscopic particles which are permeated by the dispersion medium or on the other ultramicroscopic or microscopic swelling crystals. J. F. S. An Interesting Colloid Gel. Ross AIKEN GORTNER and WALTER F. HOFFMAN ( J . Amer. Chem. SOC. 1921 43 2199- 2202).-Dibenzoylcystine is insoluble in water and in the crystalline state has no hydrophilic properties but it can give rise to rigid gels which do not contain more than 0.2% of dibenzoyl-I-cystine. The 0.2% gel is prepared by dissolving 0.2 gram of dibenzoyl-Z-cystine in 5 C.C.of 95% alcohol. The solution is heated and water added to make the solution up to 100 C.C. After cooling for two to three hours .the solution sets to a gel comparable with that of a 5% gelatin gel. The gel is transparent but after several days opaque nuclei are formed due to crystallisation which is accompanied by syneresis and after several weeks most of the dibenzoyl-I-cystine has separated in definite crystals. On applying strong suction the liquid may be drained from the gel thus showing the coarse dis- persion. The gel has a fibrillar structure. J. I?. S. Coagulation of Colloidal Solutions of Arsenious Sulphide by Electrolytes. E. F. BURTON and E. D. MACINNES ( J . Physical Chem. 1921 25 517-525).-The coagulation of solutions of arsenic sulphide sols of concentrations 0.027 gram per C.C. to 0-00337 gram per C.C.by means of potassium lithium magnesium barium aluminium and zirconium chlorides lanthanum sulphate and cerium nitrate of various concentrations has been investigated. It is shown in the case of aluminium chloride that the volume of the electrolyte required for coagulation of a given amount of the dis- perse phase varies inversely as the concentration of the electrolyte solution provided the final concentration of the colloid is kept constant. For univalent ions it is shown that the concentration of ion necessary to produce coagulation increases with decreasing concentration of the colloid whilst for bivalent ions the concen- tration of ion required to produce coagulation is almost constant and independent of the concentration of the colloid.With tervalent ions the concentration of ion required to produce coagulation varies almost direetly with the concentration of the colloid whilst quadri- valent ions necessary for coagulation decrease much more rapidly than the concentration of the colloid. These results are in keeping with earlier results of Burton and Bishop (A. 1921 ii 176). It is also shown that the coagulation curve for barium chloride showsGENERAL AND PHYSICAL CHEMISTRY. ii. 132 a tendency to resemble the curve for tervalent ions as given by aluminium chloride whilst the curve for the quadrivalent zirconium ion differs from that of the ceric ion which is similar to that for tervalent ions. J. 3'. S. Acidity and Basicity. RUDOLF KELLER (2.physikal. Chem. 1921 98 338-351).-A theoretical paper in which the terms acidity and basicity are considered. It is shown for the determin- ation of the acidity of a combination solute-solvent or the sense of the electric charge not only the inner chemical constitution of the dissolved substance is to be considered but also the chemical constitution of the solvent which may in the case of substances which are nearly amphoteric be determinative. Further the difference in the dielectric constants of the solvent and solute must be considered ; this quantity makes its influence noticeable in the case of amphoteric non-ionised and associated solutes. The so-called basic dyes in aqueous solution usually migrate toward the anode that is they are acid according to the ruling terminology; this is particularly so in the case of the slightly dispersed and non- dissociated colour colloids; the so-called acid dyes are for the greater part basic or amphoteric.Certain acid dyes such as picric acid Martius-yellow aurantia are really acid that is negative and wander to the anode. The capillary rule of Fichter-Sahlbom is generally correctly expressed in the literature but incorrectly defined since it places the dyes in the incorrect position in the electro-polarity series. All the experimental capillarity experi- ments show that immediately on dipping the paper strips the neutral and positive dyes rise with the solvent. I n this correct form the rule holds not only for colloids but also for ion- dispersoids. Gobbelsroder (" Kapillaranalyse," Dresden 1910) has actually found the strongest acids below in the anode region.The strongest acids and bases such as hydrochloric and sulphuric acids and potassium hydroxide in suitable solvents migrate in directions which are opposed to their nature. Concentration cells may be .built up in which the strongest acids behave as bases. J. F. S. Homogeneity and Dispersity. WA. OSTWALD (2. physikal. Chem. 1921 99 155-159).-A theoretical paper in which the conception specific surface is investigated on the basis of the phase rule. J. F. S. Reactions in Heterogeneous Systems. The Systems K2CO3JBaSO4-K2SO4~BaCO3 ; K,CO CaC,04-K2C,04~CaC03 SALLINGER (2. physikal. Chem. 1921 98 103-150).-The systems named in the title have been experimentally investigated. By the addition by barium chloride solution to an excess of a mixture of potassium carbonate and potassium sulphate solution the relation in which barium divides itself between the carbonate and the sulphate ions has been approximately determined. This ratio [CO,l] [SOa"] is termed the preoipitation equilibrium.I n the and M,Cr04~AgI0,-KI03~AgCr0,. I . RAMANN and H. 5 - 2ii. 132 ABSTRACTS OF CHEMICAL PAPERS. case of the addition of silver nitrate solution to an excess of the mixed solutions of potassium iodate and chromate the original equilibrium partition of the silver takes place according to the anion ratio [IO,’] [Cr04”/2] as in the previous case but the two equilibria differ to an extraordinary degree in stability. Whilst the first can be approximately fixed by the rapid addition of alcohol the second can only be fixed by a special experimental procedurc.I n the system xK2CO,+yK,C,O,+xCaC1,(x <x+y) the precipi- tation equilibrium could not be fixed by the addition of alcohol. The precipitation equilibrium in general is labile inasmuch as immediately after the precipitation a reaction is set up which brings about a change in the relationship of the salts in the solid phase and leads to the formation of one of the salts (solution equilibrium). The solution equilibrium has been determined for various concentrations for the system K C0,I BaS04-K,S041BaC03 ; for the system K,CO,I CaC,04-K2C,0~CaC03 a t various total potassium salt concentrations and various temperatures and for the system K,Cr0,(AgI03-I~I0,~Ag,Cr0 at various total salt con- centrations and one temperature (20’).I n the system K,CO,I BaS0,-K,SO,IBaCO the partition relationship K,CO,JK,SO shows a marked dependence on the total salt concentration. No measurements were made on the temperature influence on this ratio. I n the system K,C0,~CaC,04-K,C,0,~CaC0 the partition relationship K,CO,IK,C,O is but slightly influenced by the total salt concentration but very strongly influenced by temperature. The ratio K,CQ31K,C20 is shown to be a strict linear function of the temperature. The data obtained for the systems investigated and those obtained by Knupffer (A. 1898 ii 420) for the system KCNSITlCl-KClITlCNS and by Noyes and Kohr (A. 1903 ii 201) for the system KOHIAgCl-KClIAgOH have been examined by means of the formula put forward by Nernst and Noyes (A 1891 ii 142).J. F. S. Studies on Chemical AEnity. XIII. Reaction M n i t y in Systems of Solid Salts. J. N. BRONSTED (2. physikal. Chem. 1921 98 239-243).-1n earlier papers (A. 1920 ii 290 298) a method was described whereby the reaction affinity between very soluble solid salts might be determined. This method does not yield the required results when applied to sparingly soluble salts. In the present paper a method is described for the determination of the reaction affinity in systems of sparingly soluble solid salts. The method has been applied to the reaction [CoCl(NH,),]Br,+ [ COB~(NH,)~]C~ -+ [CoCl(NH,) ,]Cl,+ [ CoBr (NH,) ,]Br2 and consists in determining the solubility of the four cobalt complex salts in mixtures of (i) potassium formate and chloride solution (ii) potass- ium formate and bromide (iii) sodium chlorate and chloride and (iv) sodium chlorate and potassium bromide where the total alkali concentration is 0.5N in each case.Then from the solubility products of the four cobalt complex salts by means of the gas lawa the reaction affinity may be calculated. The results of the experiments lead to the value d,>=621 cal. for T=273’ a valueOFlNERAL AND PHYSICAL CHEMISTRY. ii. 133 which is probably accurate to 1%. The system is characterised by the fact that the lighter ions are combined with t,he lighter ions and the heavier ions with the heavier. J. F. S. The Velocity of Decomposition of High Explosives in a Vacuum. 111. Mercuric Fulminate. R,OBERT CROSBIE FARMER (T.1922 121 174-187). Sensitiveness of Very Sensitive Explosives. J. EGGERT (2. Elektrochem. 1921 27 547-558).-The sensitiveness of a number of explosives including the iodides of nitrogen has been experimentally investigated. The results show that the inter- molecular decomposition of nitrogen iodide takes place according to the equation 8NH,NI,=5N,+GNH,I+91 whether the decom- position takes place in the dark or in light or is occasioned by detonation. Nitrogen iodide is not sensitive to shock and many effects which were previously attributed to detonation are now shown to be secondary mechanical effects. The sensitiveness of nitrogen iodide and silver amide is not materially changed by lowering the temperature to - 190". Gradual isothermal increase of pressure to 5000 afm.brings about a decomposition in 70% of the cases of nitrogen iodide or silver amide but all othcr explosives examined are unchanged by this treatment. An attempt is made to find an explanation of Ohe sensitiveness of explosives ; since earlier experi- ments have shown that the mechanical effects are not to be explained as secondary temperature effects there only remains the possibility that local increases of pressure occur with all mechanical influences and these are t o be regarded as causing the detonation. This view is supported by the present experiments on the behaviour of the sensitiveness with increase of pressure. A consideration of the processes occurring during the decomposition of explosives from the point of view of Herzfeld's theory of velocity of chemical reaction leads to a possible explanation of some of the properties of these Substances.I n this connexion the primary generally very simple unimolecular decomposition of the explosive which depends on the persistence of the molecule alone is to be differ- entiated from the accompanying at. least bimolecular reaction due to the collision between the products of reaction of the primary process. If in the primary reaction products are formed which possess more energy and a greater persistence than the original substances or if following on the primary process other processes which are subjected to several negative accelerations occur then there will be no marked tendency to detonation. But if on the other hand the primary reaction leads to the stable end products of the reaction or if these products undergo other reactions with no retardations or a t most only one retardation then the explosive is one which has a high sensitiveness and a great tendency to detonation.Pressure is the factor which removes the retardation t o the reactions since it causes the atoms molecules and atomic groupings which have affinities for one anobher to come closer together. J. I?. S.ii. 134 ABSTRACTS OF CHEMICAL PAPERS. The Autoracemisation of Potassium Chromioxalate. ERIC KEIGHTLEY RIDEAL and WILLIAM THOMAS (T. 1922 121 196- 202). Neutral Salt Action at Higher Salt Concentrations ; the Velocity of Hydrolysis of Ethyl Acetate and the Hydrogen-ion Activity of the Catalyst. GOSTA AKERLOF (2. physilcal. Chena. 1921 98 260-292).-The velocity of hydrolysis of ethyl acetate by 0~~5N-solutions of hydrochloric sulphuric and nitric acid a t 25" has been determined in the presence of various concentrations of the chlorides of ammonium potassium sodium magnesium calcium and barium the nitrates of ammonium potassium sodium magnesium and calcium and the sulphates of ammonium potassium sodium and magnesium.The hydrogen- ion activity of the catalyst in the various reaction mixtures has been determined by means of E.H.P. measurements. An attempt is made to ascertain the cause of the neutral salt action on the velocity of hydrolysis a t constant acid concentration by combining the velocity measurements with the hydrogen-ion activity. A relationship between these two quantities is furnished by t'he formula K=Ca .f(n);/& in which K is the reaction velocity n the acid titre and a the hydrogen-ion activity.The formula holds for all concentrations of catalyst and is not influenced by the nature or concentration of the added salt as far as the present experiments go. This formula is probably also true for the inversion of sucrose and allied reactions. The foregoing shows that the hypothesis o€ the catalytic activity of the non-ionised molecule is incorrect. The salt molecules have no activity of their own but they are the cause of the change of activity of the catalyst. Exactly how the neutral salt changes the activity of the catalyst is not known with certainty but the most probable hypothesis is tlhat the water sheath of the hydrogen ion is changed by the addition of neutral salts.The concentration of the hydrogen ion remains constant whilst the activity changes. According to this view the hydration numbers do not represent definite quantities but the ions control all the water molecules in the solution. The various neutral salt actions are controlled by the various forces with which the different ions attract the water molecules. J. F. S. Velocity of Hydrolysis of Acetals. ANTON SKRABAL and ANTON SCHIFFRER (2. physikal. Ch,em. 1921 99 290-313).-The velocity of hydrolysis of methylene dimethyl ether methylene diethyl ether ethylidene diethyl ether methylene diacetate and ethylidene diacetate has been investigated. The reactions were carried out with hydrochloric acid of various concentrations (0.001-0.3N) and also in alkaline solution.It is shown that the measured velocity constants represent the velocity of hydrolysis of the first ether group; the removal of the second group is much more rapid and is regarded as a water hydrolysis. The velocity constants vary very much for the different casm thus for the acid hydrolysis the value of Ic is 70,000 for methylene diethylGENERAL AND PHYSICAL CHEMISTRY. ii. 135; ether whilst for ethylidene diacetate it is 0.00650 whilst in the case of alkaline hydrolysis the value of E is 1,100 for methylene diacetate and 96 for ethylidene diacetate. In the case of the hydrolysis of esters it is shown that the velocity depends largely on the nature of the acid but very little on the nature of the alcohol. These rules are established from the published data of a number of aut h ors .Velocity of Hydrolysis of Sucrose. R. H. CLARK (J. Arner. Chem. Xoc. 1921 43 1759-1764).-The velocity of hydrolysis of sucrose by hydrochloric hydrobromic sulphuric or nitric acid has been determined a t 25". The experiments were effected in the presence of acids of varying concentration and in the presence of potassium nitrate and dextrose. It is shown that the addition of a neutral substance to keep the concentration of the water constant while varying the concentration of the acid has no appreciable effect in producing a numerical proportionality between the quantity of hydrogen ions present and the inversion velocity of sucrose. The increase in the velocity of hydrolysis of sucrose produced by a strong acid in the presence of a neutral salt of that acid over the velocity produced by the acid alone can only t o a small extent be attributed to the fact that the volume of the reaction is reduced by the presence of the salt.On the assumption that both the dissociated and undissociated forms of an acid are catalytically act'ive in sucrose inversion the values of Ki for the hydrogen ion from hydrochloric hydrobromic and nitric acids have the same value (Ki=0.234) ; the corresponding value for sulphuric acid is considerably lower (Ki=O-144). Affinity Dimensions of Weak Acids and Bases in Alcoholic Solution and the Alcoholysis of their Salts. HEINRICH GOLDSCHMIDT CARL GORBITZ HAAEON HOUGEN and KRZSTIAN PAHLE (2. physilcal. Chem. 1921 99 ll&l54).-The theory of the alcoholysis of salts of weak acids with bases of the aniline type in anhydrous alcohol and in alcohol containing water is developed.A method of determining the degree of alcoholysis based on electrical conductivity measurements is described. The electrical conduc- tivity of a series of sodium aniline and ammonium salts in pure and in aqueous alcohol has been determined. The affinity dimen- sion of several organic acids in alcohol solution has been deter- mined. The alcoholytic constant of eleven aniline salts has been measured and from the results the affinity constant Kg corre- sponding with the equilibrium (Aniline x H')/Aniline ion has been calculated. The same quantity has been determined for the three toluidines and mono- and di-methyl anilines. The influence of water and the alcoholytic constants has been investigated.J. F. S. J. F. S. J. F. S. The Rate of Solution of Iron in Dilute Sulphuric Acid both when Stationary and under Rotation. JOHN ALBERT NEWTON FRIEND and JOHN HORACE DENNETT (T. 1922 121 4 1 4 4 ) .ii. 136 ABSTRACTS OF CHEMICAL PAPERS. Theory of Catalysis in Homogeneous Gas Reactions. KARL F. HERZFELD (Z. physilcal. Chem. 1921,9€?,161-174).-A theoretical paper in which an attempt is made to answer the question " Why does a reaction proceed more rapidly through an intermediate product than by the direct action? " It is shown that since a t equal concentration the slowest of all the partial reactions is deter- minative of the velocity of the whole process this slowest reaction must proceed more rapidly than the original reaction when catalysis occurs.After a short review of the previously published theoretical work on gas reactions the author considers such reactions in a way which assumes the existence of a few free atoms. Then for catalysis i t is necessary that a t not too small concentrations 01 the catalyst the heat of activation of the intermediate product (that is practically its heat of formation from atoms) is markedly smaller than that of the starting material ; in the best case about one-half. This gives then an upper limit for the acceleration which at not too high concentrations may be reached by a given catalyst. Examples on which to test the hypothesis are not known. J. P. S. Catalytic Oxidation of Carbon Monoxide. T. K. ROGERS C. S.PIGGOT W. H. BAHLKE and J. M. JENNINGS ( J . Amer. Chem. Xoc. 1921 43 1973-1982).-The present work was undertaken with the object of finding a catalyst which will completely oxidise carbon monoxide a t ordinary temperatures. A number of such catalysts have been prepared which cause rapid and complete oxidation. The chief of the catalysts obtained is a specially pre- pared manganese dioxide on which is precipitated the oxide of silver or copper or both. The silver or copper is best precipitated as the carbonate and subsequently hydrolysed to the hydroxide. The presence of wat,er vapour limits the life of these catalysts. The decomposition temperature of silver oxide when simultaneously precipitated with calcium hydroxide is considerably lower than that of silver oxide alone.J. 3'. S. Catalytic Oxidation of Carbon Monoxide at Ordinary Temperatures. DAVID R. MERRILL and CHARLES C. SCALIONE ( J . Amer. Chern. Soc. 1921,43,1982-2002).-A number of catalysts for use in protective masks against carbon monoxide are described. The most successful are mixtures containing manganese dioxide and basic copper carbonate and manganese dioxide copper oxide cobaltic oxide and silver oxide respectively. These mixtures oxidise carbon monoxide rapidly and at ordinary temperatures and when protected by a drying agent such as calcium chloride are suitable for use in protective masks against carbon monoxide of all concentrations below those in which an oxygen helmet would be necessary. The factors influencing activity such as the con- stituents of the mixtures the conditions of precipitation the mechanical treatment of the precipitate and method of drying are discussed and a description is given of the efficiency of the mixtures under various conditions. J.F. S.GENERAL rWD PHYSICAL CHEMISTRY. ii. 137 Oxidation Catalysis. 11. L. KARCZAG (Bioclzern. Z. 1921 119 16-22).-Experiments with dyes at high dilutions show that the ferrous ion catalyses oxidation with hydrogen peroxide much faster than the ferric ion. The order of mixing is of influence the controlling factor being whether the ferrous ion is fist oxidised to ferric or not. A. D. FOKKER (Arch. NLerZand 1921 5 [iii~] 193-242).-A mathe- matical paper in which it is shown that the application of the electronic theory to the mutual actions of the electrons inside the atom gives results which are in accordance with experimental observations.J. F. S. MAURICE COPISAROW ( J . Amer. Chem. Soc. 1921 43 1870-1888).-Allotropy is defined as the capacity of an element to exist in forms differing in the mode of their intra- molecular linking. It is therefore a function of the valency but it does not imply isomerism or polymerism since allotropes need not contain the same or a multiple number of atoms in their respec- tive molecules. Thenumber of possible forms in which an element can exist on this basis is indicated for the elements of all valencies. It is theoretically possible for an element t o have in certain cases more than one molecular form corresponding with each mode of linking. Regarding allotropes as the most chemically and physically distinct forms of an element i t follows that several molecular forms each containing a different number of atoms but all having the power of free rotation will differ among themselves t o a less extent than when compared with a molecular structure of the same element in which all atoms are rigidly fixed.Thus it follows that valency and the saturation or fixation of atoms and not the actual number of atoms play the predominant part in the deter- mination of allotropes. In this light allotropy becomes the capacity of an element to exist in forms differing in the mode of their intra- molecular linking. Molecular forms differing in the distribution of their intramolecular linking and in thc number of atoms but all belonging to one type of linking can be termed allotropoids.The allotropy of carbon phosphorus iron and nickel are coilsidered from the point of view of the hypothesis. Types of Valency. IRVING LANGMUIR (Science 1921 54 59-67; cf. ibid. 1921 53 290; J . Ind. Eng. Chem. 1920 12 386 and A. 1919 ii 328; also Rydberg A. 1915 ii 94).-Recog- nition of three distinct types of valency (a) positive valency (the number of electrons an atom can give up) ( 6 ) negative valency (the number of electrons an atom can take up) and (c) co-valency (the number of pairs of electrons an atom can share with its neigh- bours) as well as the numerical values for most of the elements of each type are shown by a simpler method than that hitherto employed to be derived from certain postulates regarding the struc- ture of atoms and the relationships between the different types of valency are thereby further clarified.On the basis of the H. K. Electronic Theory of the Interior of the Atom. Theory of Allotropy. J. F'. S. 5'ii. 138 ABSTRACTS OF CHEMICAL PAPERS. Rutherford atom only three postulates consistent with those previously proposed are necessary (a) The electrons in atoms tend to surround the nucleus in successive layers containing 2 8 8 18 18 and 32 electrons respectively (the extra electrons if any remaining in the outside layer as an incomplete layer or “ sheath ”). ( b ) Two atoms may be coupled together by one or more duplets (stable pairs of electrons forming the first complete layers) held in common by the completed sheaths of the atoms. A group of neutral atoms interacting by transfer of electrons gives rise to the conception of electrovalency including positive and negative valency differing only in algebraic sign whilst interaction by the sharing of duplets corresponds with co-valency.It is shown that the sum of the electrovalencies and co-valencies for all the atoms in any complete compound is zero a complete compound being defined as one in which all the atoms possess complete layers of electrons. (c) The residual charge on each atom and on each group of atoms tends to a minimum. The term “ residual charge ” is employed to express the total charge of an atom or aggregate of proximate atoms regardless of sign. It is pointed out however that the first and third postulates are often in conflict and that in such cases the tendency of the first may prevail against that of the third.Chemical compouirds are theref ore classified according to the types of valency exhibited by their atoms as follows (1) complete compounds including (a) compounds without co-valency such as certain simple metallic salts volatile halogen compounds silicates and most minerals and ( b ) compounds without electropositive atoms such as organic compounds ; (2) incomplete compounds including metallic substances and compounds such as zinc oxide ferroso-f erric oxide lead sulphide and cupric oxide which contain both electropositive and electronegative atoms ; (3) exceptional cases including nitrogen carbon monoxide nitric oxide and the cyanogen radicle the structure of which is not accounted for by this theory.These substances may have a single octet structure or possibly a triple bond structure. Other exceptions are boron hydride and certain other compounds f orrning double molecules. A. A. E. The Expression of the Octet Theory of Valence in Structural Formulae. GRANVILLE A. PERKINS (Philippine J . Xci. 1921,19 1-22).-A system for writing structural formulae is developed based on the octet theory of valency as presented by Langmuir. Electronegative valence the physical interpretation of which is vacancy for one more electron in the outer shell of the atom is represented by a line in the usual way. Electropositive valence on the other hand is variable and its maximum is represented by the total number of electrons in the shell. It cannot be repre- sented by “unsatisfied bonds” as it never causes direct union between atoms and is expressed as follows for example Na’ CaII. The neutral atoms having electronegative valence may satisfy that valence by acquiring an electron that is becomingGENERAL AND PHYSICAL CHEMISTRY. ii. 139 a negative ion; this electron may be obtained from the shell of an atom showing positive valence tendencies and it is then left as a positive ion. These two oppositely charged ions may then form a stable compound by electrostatic attraction. Such a union is termed a salt-forming union and is expressed thus Naf . . . . . Cl- or Ca*+ 0-. On the other hand two atoms may be held together by the shell-completing forces of both atoms acting on one or more pairs of electrons which are then shared between them. Each shell thus fills one or more electron vacancies by sharing electrons ‘and thus “ satisfies ’’ one or more “ bonds.’’ This is expressed in the usual way 13-€I H-0-H etc. and is termed a direct union. All valency relations cannot however be expressed in terms of either of the above unions and a special feature of the system is the introduction of the conception of a “borrowing direct union,” for cases in which both the shared electrons are supplied by the same atom. In this case the borrow- ing atom fills two vacancies in its shell and the lending atom neither gains nor loses electrons; but the union is essentially polar since the originally neutral borrowing atom has acquired at least an interest in two negative electrons and therefore becomes a negative pole. This relationship is expressed. thus :>0 oc0,O-S a0 indicating that the 0 atom is bound to S or to 0 by virtue of having its two electron vacancies filled by borrowing an interest in two of the shell electrons of the other oxygen atom in hydrogen peroxide an$ in the sulphur atom of sulphur trioxide. The mode of application of the system to all the known elements is indicated in the original paper. G . P. M. CLAUDE W. BOURLET and WALTER THOMAS (Chem. News 123 336).-The thermos flask can be utilised advantageously in the laboratory in numerous instances where it is desired to maintain substances above the normal temperature as for example in reactions between liquids or liquids and solids a t elevated temperatures such as hot oxidation by permanganate and other similar reactions which are not strongly exothermic. Oil and water emulsions and similar mixtures are conveniently ‘‘ split ” by warming and keeping over- night in a thermos flask and flocculent or colloidal precipitates can be dealt with in a similar manner. In carrying out certain colour tests for example the Halphen reaction and reactions in media which gelatinise or become viscous a t ordinary temperatures a water-bath can with great advantage be replaced by the vacuum flask. 8 0 The Thermos Flask in the Chemical Laboratory. G. F. M. Water Pump. JOHANNES WETZEL (Chem. Ztg. 1921 45 1122).-A water pump of improved efficiency is described; par- ticular attention is directed to the relative diameters of the injector tube and outlet tube and to the central position of the jet as regards the outlet tube. w. P. s. 5”-2