RADIOACTIVITY .*Nuclear Constitution of Atoms.THE nuclear theory of the atom is based on the form of the trajec-tory of the a-particle in passing through the atom, which in turnis deduced from the deviation suffered by the a-particle in itspassage.' The fact that the overwhelming majority of thea-particles pursue practically rectilinear trajectories, whilst afew of them are deviated more or less abruptly, led to thewell-known conwption of the atom as a system of sparsely dis-tributed single electrons occupying the atomic volume, equal innumber 60 the atomic number of the atom, with a concentratedcountervailing positive charge, equal in magnitude to the corn-bined charge of the electrons, a t the centre of the atom, and con-stituting a nucleus of dimensions excessively small relatively tothe atomic volume.The inferenoe that this nucleus contains allbut some 0.05 to 0.02 per cent. of the mass and weight of the atomfollows from the known mass of the contained negative electrons,and is in general accord with the electrical theory of mass.According to this, the mass of an electric charge is proportionalto the square of the charge and inversely proportional to itsdiameter. To account for its mass on this view, the diameter ofthe nucleus of the uranium atom would be 4 x l O - l 5 cm., or 1150thof the diameter of the single negative eleatran, if it consisted ofpure positive electricity. That the nucleus is not a pure positivecharge, but contains negative electrons, the net charge being posi-tive and equal to the atomic number, is shown by the emission of@-rays from the radio-elements and by the mode of formation ofisotopes in radioactive changes. Hence the view is not free frominconsistencies.Impact of a-Particles on Heavy A toms.-Great improvementshave been made in the comparison of the experimental results ofscattering with the mathematical theory.The magnitude of the* This Report covers the years 191 9 and 1920.Ann. Reports, 1913, 10, 272 ; compare also R. Seeliger, Jnhrb. Radioaktiv.It'ektronik;, 1919, 16, 19 ; A., ii, 145.21218 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.nuolear charge for platinum, silver, and copper has been evaluatedaccurately by a determination of the ratio of a-particles scatteredover a solid angle between 2 2 O and 36-5O.This ratio is propor-tional to the square of the atomic number and a quantity depend-ing on the velocity and known physical and instrumental constants.The minuteness of the ratio makes a direct determination, bycomparing the number of a-particles scattered with those in theoriginal beam, difficult. The difficulty was ingeniously overcome,however, by introducing a notched rotating disk into the path ofthe a-particles, when oounting the direct beam, and so reducingthem in a known ratio a t will to a number comparable with thenumber scattered. I n this case one obtains intermittent gusts ofscintillations with any desired interval between, conditions whichare very favourable t o counting, and actually enable the numberper second capable of being counted accurately to be five timesas great as without the device. This is apart from the completecontrol over this number by varying the relative size of the notch.The experimental values for the three metals named, 77.4, 46.3,29.3, are, in eaoh case, within the known probable error of theaccepted values of the atomic numbers, 78, 47, and 29.This,incidentally, is an important confirmation of the correctness ofthe absolute magnitude of the atomic numbers, and shows that thePeriodic Table contains no unsuspected vacant places.With the same arrangements, i t was possible to verify accuratelythe inverse-square law of force over the region in which scatteringof the a-partiole occurs for the platinum atom. The number ofparticles scattered, other conditions being constant, depends onthe initial velocity of the a-particle raised to the power 4/(1 - p ) ,when the force around the nucleus deviating the partiole variesas l / r g , r being the distance.The experimental value of p foundwas between 1-97 and 2-03, a variation from the inverse-square lawwithin the counting error of 4 per cent. The actual least distanceof approach to the nucleus was between 7 and 14(x10-12 cm.)for high and low velocity a-particles respectively. From otherexperiments in this field and in that bf the wave-lengths of theE-series of X-ray spectra, it follows that the inverse-square lawholds over a range between 3 and 100( x 10-12 cm.), and that therecan be no electrons in this region in the case of a heavy atom likeplatinum.These are fundamental oonclusions.2Impact of a-Particles on Light A toms.-Turning now to impactsof a-particles with the nuclei of light atoms, where the nucleusstruck is violently repelled,3 and itself constitutes a new type ofJ. Chadwick, Phil. Mag., 1920, [vi], 40, 734.3 Ann. Reports, 1914, 11, 274; 1916, 13, 261RADIOACTrVITY . 219radiant particle, such as the H-particle resulting from the passageof a-rays through hydrogen, most striking results have beenachieved. Here the a-particle approaches within 0-25( x 10-12 cm.)of the hydrogen nucleus, and the results point to rapid changesand possibly to variations of the direction of the field of formwithin the distance 0-35. The inverse-square law no longer holds.Only one in lo9 of the hydrogen atoms penetrated by the a-particleis repelled with a velocity sufficient to enable it to be detectedbeyond the range of the a-rays, or one H-particle is producedby a hundred thousand a-particles passing through 1 cm.ofhydrogen gas a t N.T.P. This number, though minute, is fromten to thirty times that to be expected if the inverse-square lawheld. The absorption of these H-particles over their range, whichis four times that oT the a-particle producing them, is reminiscentof the absorption of a-particles themselves, and is totally differentfrom what is theoretically to be expeoted. With H-particlesgenerated by long-range a-particles, over a range equivalent to22 cm. of air, there is practically no diminution of the number ofH-particles, whilst between this and the end of the range, 28 cm.,there occurs a gradually increasing diminution.With H-particlesgenerated by short-range a-particles, the theoretical curve is morenearly approached. The H-particles in the first case appear to beprojected in the same direotion as that in which the a-particle istravelling, or within a few degrees of it, all a t the same velocity.It is clear that to this case, where a very intimate approach ofthe helium and hydrogen nuclei occurs, special considerationsThe identity of these H-particles with hydrogen was proved bya determination of the deviation suffered in electromagnetic andelectrostatic fields. The value of elm found, 104, is in perfectaccord with that of the hydrogen ion, 9570 (E.M.U.).I t svelooity, namely, 1.6 times that of the a-particle generating it, isin perfect agreement with the maximum value calculated for ahead-on ” collision. The charge is positive in sign, and nonegatively charged particles were observed .5With regard to light gases other than hydrogen, helium givesno particles differing in range from the generating a-particle.From this it is inferred that singly charged atoms of helium, theestimated range of which would be four times that of the a-particle,are not formed. But oxygen, nitrogen, air, and carbon dioxide allgave scintillations of similar brightness over a range of 2 cm. ofL. R. Zoeb, ibid., 38, 533; A., ii, 145.aPPb*4(Sir) E. Rutherford, Phil.Mag., 1919, [vi], 37. 537: A., 1919, ii, 256(Sir) E. Rutherford, ibid., 562 ; -4., 1919, ii, 258220 AKNWAL REPORTS ON THE PROGRESS OF CHEMISTRY.air beyon’d that of the range of the generating a-particles. Thenumber was of the same order as those obtained in hydrogen gas.The range, 9 cm. of air, was only one-third as great, and thebrightness of the scintillations, a t a distance equivalent to 7.5 cm.of air, was equal to that of an a-particle a t 1 cm. from the endof its range, instead of a t 0.5 cm., as for the H-particle. Theoriginal presumption,6 that these short-range particles were dueto atoms of oxygen and nitrogen carrying unit charge, has nowbeen shown to be a t fault.7 It has been found possible to deter-mine their nature by special arrangements for the examination oftheir deflexion in a magnetic field, allowing the use of wide slits.Instead of the particles from oxygen being less deflected than thegenerating a-particles, as should be the case if they were singlycharged atoms of oxygen, they were more deflected, which excludesthe possibility that they oan be oxygen atoms, either singly ordoubly charged.A mass intermediate between 1 and 4 and adouble charge were indicated. The deviation suffered wasestimated to be 5 per cent. less than that suffered by H-particlesin a direct comparison, and the conclusion was drawn that theyconsist of doubly charged positive particles of mass about 3 witha velocity 1.2 times that of the generating a-particle. There wasno noticeable difference between oxygen and nitrogen, so far asthese short-range particles are conaerned.Both appear to yielda new particle of mass 3, differing from the ‘(H3” of positive-raymethods of gas analysis in that it carries two units instead of oneof positive electricity, and therefore is presumably an isotope ofhelium.Nitrogen, however, differed sharply from oxygen in giving, inaddition to these new particles, a very much smaller number (aboutone-twelfth) of H-particles. The range of these is slightly greaterthan of those obtained from hydrogen, but their identity wasproved by direct comparison of the electromagnetic deviation in theapparatus above referred to.* It is estimated that, to produce1 cubic millimetre of hydrogen by this means, the total a-radiationof 2 kilograms of radium acting for a year would be required.So far as can be seen, artificial disintegrations of atoms bycollision with the a-particle appear t o be endothermic.Theparticle of mass 3 appears to escape with rather more than theenergy of the a-particle striking the nucleus of the oxygen ornitrogen atom. Even neglecting the kinetic energy of the residueof the nucleus and of the a-particle after the collision, the dis-(Sir) E. Rutherford, Phil. Mag., 1919, [vi], 37, 571 ; A., 1919, ii, 259.7 Ibid., Bakerian Lecture, Proc. Roy. SOC., 1920, [A], 97, 374 ; A., ii, 541.* Ibid., Zoc. cit., and Phil. Mag., 1919, [vi], 37, 581; A . , 1919, ii, 260RADIOACTIVITY. 221integration, as in the case of the radio-elements themselves, mustbe accompanied by the liberation of energy.On the view thatthe actual energy required to bring about the disintegration issmall, and that the energy of the a-particle is mainly expended’against the strong repulsive field, in getting near enough to thenuoleus to affect it, electrons or /3-rays, which would move up tothe nucleus in an attracting field, may be able to bring aboutsimilar changes. This raises anew the whole question, so fre-quently discussed in these Reports,g as to the origin of the helium,found by some and not by other investigators, after passage of theelectric discharge through gases in vacuum tubes and in old X-raybulbs and the like. The latest experimental contribution gavenegative results, in so far as the discharge through carefullypurified hydrogen is concerned.In no case was helium or neondetected .lo Naturally, these highly significant results have pro-duced a flood of speculation as to the mnstitution of the atomicnucleus, which does not yet call for consideration here.Is0 t opes.Our knowledge of the heterogeneity of common elements hasbeen notably advanced, during the period under review, beyondthat recorded in the Reports four and seven years ago,ll by theperfection of the positive-ray method of gas analysis and its appli-cation to the detection of heterogeneity, if it exists, in somenineteen non-radioactive elements.12 The methods depend on thesame general principles as those which sufficed to detect thepresence of meta-neon, of atomio mass 22, in atmospheric neon in1913, but the electromagnetic and electrostatic deviating fields arerearranged in such a way as to secure an effect precisely analogousto focussing in optics.The trajectories of the positive ions in aslightly divergent beam are brought to a focus in a plane contain-ing the photographic plate. All those for which the value of themass divided by the charge is the same are brought to the samepoint in the plane, those with greater and less values, respectively,being on either side. The oomplex pencil is resolved into a “massspectrum” in every respect analogous to a light spectrum pro-duced by a prism or grating. The terms “first-order and second-s Ann. Reports, 1914, 11, 45, 289.lo A. Piutti and E. Cardoso, Gazzettu, 1920, 50, i, 5 ; A., ii, 312.l1 Ann.Reports, 1916, 13, 245 ; 1913, 10, 265.l2 F. W. Aston, Nature, 1919,104, 334, 393 ; 1920,105, 8, 547 ; 106, 468 ;Phil. Mag., 1919, [vi], 39, 449, 611; 4, 628; A., ii, 277, 344, 718222 ANNUAL REPORTS ON THE PROGRESS O F CHEMISTRY.order mass-spectrum” are used to denote spectra produced by ionssingly and doubly charged respectively. The existence of ions withmore than one unit of charge introduces a complication, but fortu-nately these are experimental peculiarities which enable the twoorders usually to be distinguished without uncertainty. The relativemass of the ion causing any line in the spectrum can so be evaluatedto an accuracy of one part in a thousand, and the atomic massdetermined to a degree of accuracy comparable with that attainedin the best determinations of the atomic weight by chemical means.Incidentally, the complete agreement between the two in manycases affords much the most important evidence of the constancybetween mass and weight for different elements.This question hasbeen much canvassed of recent years.Of the nineteen elements so far examined, ten prove to be homo-geneous and nine t o be heterogeneous and composed of more thanone isotope with different atomic masses. The following table,taken from the author’s last communication to Nature, gives theresults.TABLE OF ELEMENTSAtomicElement. number.Hydrogen- ............Helium ...............Boron ..................Carbon ...............Nitrogen ............Fluorine ...............Neon ..................Silicon ...............Phosphorus .........Sulphur ...............Chlorine ...............Argon ..................Arsenic ...............Bromine ...............Iodine ..................Xenon ...............Oxygen ...............Krypton ...............12567891014151617183335365354Mercury ...............80Atomicweight.1-0083.9910.9012.0014.0116-0019-0020.2028.3031.0432.0635.4639.8874.9679-9282.92126.92130.32200.60AND ISOTOPES.Minimum Masses of isotopes,number ofisotopes. intensity .1 1.0081 42 11, 101 121 141 161 192 28, 29, (30)1 31in order of their2 20, 22, (21)1 322 35, 37, (39)2 40, 361 752 79, 8161 1275 (7)84, 86, 82, 83, 80, 78129, 132, 131, 234, 136,(128, 130 ?)(6) (197-200), 202, 204(Numbers in brackets are provisional only.!Apart from a possible uncertainty, already alluded to, as tothe order of spectrum to whioh any line belongs, the photographspublished reveal the great power and certainty of the new method.Unfortunately, only non-metallic elements have so far beenincluded.The difficulties in the way of examining metallicelements by this means have not yet been overcomeRADIOACTIVlTY. 223In every case, except hydrogen, the atomic mass of each homo-geneous component proves to be an exact integer, in terms of thatof oxygen as 16, within the error of measurement already stated.For hydrogen, however, the chemical value, 1.008, is exactly con-firmed and its homogeneity proved.Hydrogen, of course, is anexception to every generalisation concerning the chemical elements,and its simple structure, consisting probably of a single positivecharge as nucleus and a single electron as satellite, is a sufficientreason for its uniqueness. I f the hydrogen nucleus is theelementary positive constituent of the nuclei of other atoms, anumber of electrons, equal to the difference between the atomicweight and atomic number, must be present also. Thus, if thenucleus of uranium is made up of 238 hydrogen nuclei, there mustbe in the nucleus 238-92=146 electrons. The close packing ofthese positive and negative constituents may account for the differ-ence of mass, 1.9 units, between the mass of the constituents andthat of the resulting atom, that is, essentially to the difference inthe atomic weights on the basis H = 1 and O= 16.l3The integral value of the atomic weights then points to anatomic constitution of secondary units, such as helium nuclei,packed sufficiently openly not to influence their mutual masses, thewhole of the packing effect being due to the close packing withinthese secondary units.Isotopes of Lead.Atomic Weight of Leud of Radioactive Origin.-Fuller detailsof the atomic weight determination of the lead from Norwegianthorite, which gave 207.9, the highest yet found, have been pub-lished, together with those found for lead from three Ceylonthorianites, partioulars of which follow : 14Per cent.Th. Per cent. U. Per cent. Pb. At. wt.I. ......... 68.9 11.8 2.3 207-2111. ......... 62.7 20-2 3.1 206.90111. ......... 57.0 26.8 3.5 206.83Lead separated from samarskite, containing 12.2 per cent. ofU30, and 1.03 per cent. of Tho,, gave the value 206.30.15 Leadfrom a Japanese source, of possible, though doubtful, radioactiveorigin, gave the value 207.13, which does not differ appreciablyfrom that of common lead.16l3 Compare Ann. Reports, 1916, 13, 253.l4 Compare Ann. Reporter, 1918, 15, 201 ; 0. Hhigschmid, Zeitsch.l5 A. L. Davis, J . Physicul Chem., 1918, 22, 631 ; A., 1919, ii, 107.l6 T. W. Richards and J. Sameshima, J . Amer. Chem. SOC., 1920, M, 928 ;Elektrochem., 1919, 25, 91 ; A., 1919, ii, 285.A., ii, 434224 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Melting Point .-Two determinations show that, within the errorof measurement, the melting point of lead of radioactive origin isidentical with that of common lead.I n 0 ~ , 1 7 the lead, comparedwith common lead, had the atomic weight 206-57. A constantin-manganin couple was used, and the melting points were found tobe identical to 0 - 5 O , the experimental error. I n the other,l8thermocouples of copper-nickel were employed, the single couplebeing capable of reading hundredths, and the multiple couplethousandths, of a degree.The lead compared in this case was from an Australian radio-active mineral of atomic weight 206.6. Neither specimen wasspectroscopically pure, the common lead being the less pure, butprobably the impurities did not exceed 0-005 per cent.Theyshowed slight diff erenoes of behaviour. The super-cooling wasgreater for the purer sample, and its freezing-point-time curvewas more horizontal. The radioactive lead had the higher meltingpoint by 0*05O, but part, if not all, of this difference is to beascribed to its greater purity. The thermo-electric power, electricconductivity, and change of the latter with temperature andpressure, were for each sample the same. These negative resultsthus have now decided between opposing theoretical views beforediscussed . l gSpectrum.-The minute difference, 0.0043 A., in the wavelengthof the line 4058 A., already reported, has been confirmed.%Ordinary lead, lead from Joaohimsthal pitchblende of undeter-mined atomic weight, and lead from Ceylon thorite of atomic weight207.77, were compared.The method consisted in photographingthe respective interference fringes, produced by a Fabry and Perotttalon, the source of light being an arc between an alloy ofcadmium with the lead and a button of tungsten in a vacuum.Important sources of error present in the first series of experi-ments,21 which gave a negative result, were eliminated by reducingall observations t o a selected cadmium fringe as standard, whichregisters any variation due to a change of temperature or to thewandering of the souroe of light. These causes affect the standardfringe equally with the fringe under examination, and are soeliminated.The wave-length for the pitchblende lead was foundto be 0*0050 A.*0*0007 A. greater than that for ordinary lead,17 M. Lembert, Zeitsch. Ebektrochem., 1920, 26, 59 ; A., ii, 216.18 T. W. Richards and N. F. Hall, J . Amer. Chem. SOC., 1920, 42, 1550;l9 Ann. Reports, 1916, 13, 252.2o Compare Ann. Reports, 1918, 15, 2 0 4 ; T. R. Merton, Proc. Roy. Soc.,-4., ii, 622.1920, [A], 96, 388 ; A., ii, 140. a1 Ann. Reports, 1916, 13, 248RADIOACTIVITY. 225which, in tum, was 0.0022 A.+0-0008 A . greater than that forthorite lead. Also, a difference was found for the wave-length ofthe line 5350 A. of thallium when ordinary thallium and thethallium contained in pitchblende residues were compared, theformer being the greater by 0.0053 A.kO.001 A.I n this case,owing to the thallium not having been separated from the residues,the result cannot be entirely depended on, for the displacementof lines, by the presence of impurities, in the arc spectrum, thoughrare, is not entirely unknown. But it indicates a presumptionthat the thallium in pitchblende is of radioactive origin anddifferent in atomic weight from ordinary thallium.I n an interesting discussion of the spectra of isotopes,zz it ispointed out that the differencm in the case of lead, although onlyof the order of a millionth of the wave-length, are one hundredtimes greater than are to be expected from the Bohr theory, ascorrected to take into account the displacement of the centre ofmass o f the vibrating system with a change of the mass of thenucleus. They are enormously greater than can be ascribed to anypurely gravitational effect of the mass of the nucleus on theelectron.The result indicates the existence of a force, due to themass of the nucleus, on the eleotronic system of the atom nothitherto known. In the original experiments, in which a 25 cm.grating was used and the spectrum photographed in the sixthorder, the line was shifted, not broadened, to a position correspond-ing with the mean atomic weight of the lead, although a broaden-ing, if not an actual resolution, into two or more lines correspond-ing with the separate isotopes present, in these circumstances,although not in the subsequent ttalon experiments, is apparentlyto be expected. This minute difference of wave-length of the linesin the spectrum is the only difference in the physico-chemicalproperties of isotopes, apart from atomio mass, so far substantiated.Separation and Properties of Isotopes.It cannot yet be considered proved beyond doubt that any actualanalytical separation of the components of a mixture of isotopeshas been effected.Systematic fractionation of atmospheric neonby the use of cold charcoal failed to effect any separation. Evenfractional diffusion through pipe-clay has not, so far, given con-sistently positive results.’3 The theoretical question of the possi-22 W. D. Harkins and L. Aronberg, J . Arner. Chern. Soc., 1920, 42, 1328;pa F. A. Lindemann and F. W. Aston, Phil. Mag., 1919, [vi], 37, 5 2 3 ;REP.---voL. xvrr. 1A., ii, 841.A . , 1919, ii, 209226 ANNUAL REPORTS ON THE PROGRESS OB CHEMISTRY.bility of separation by various means has been much dis0ussed.aMethods, such as fractional diffusion, centrifugal separation, andthermal diffusion, which depend on diff ereiices of niolecular mass,if not those, such as vaporisation and chemical fractionation, oughttheoretically to be effective. The thermal diffusion method,depending on the maintenance of two intercommunicating vesselsa t widely different temperatures, which produces a oondition ofequilibrium, in which excess of the heavier constituent is presentin the colder vessel, and centrifuging! both appear promising fromthe theoretical point of view.Preliminary announcements of the partial separation of theisotopes of chlorine, mercury, and iodine ( !) have been made.I nthe first case,25 indications of a separation of hydrogen chloride byfractional diffusion into a heavier and lighter fraction have beenannounced, but no definite experimental data are given. In themse of mercury,2G evaporation a t low pressure is stated to give acondensate less in density than the residual mercury. Each frac-tion was redistilled before the density was taken, and the differencein the pyknometer determinations amounted to 5 parts in 100,000,the error of measurement being less than one part in a million.Iodine, the most recent of the elements to be submitted to positiveray analysis, and found, unlike chlorine and bromine, to be homo-geneous, has, from speculative reasoning, been ascribed five isotopes.Fractional diffusion gave products with atomic weight varying from128.22 upward, the mean being 2.04 per cent.above the acceptedvalue .27Very interesting new results have been obtained along the linesof the use of radioactive isotopes of common metals to indicatewhat is occurring t o the latter in chemical operations. Thus i thas been shown that a free exchange of the metallic atom amongthe competing acid radicles occurs for ionised, but not for non-ionised, compounds, The general method was to mix solutions oftwo different compounds of lead in equimolecular proportions, theone compound only being " activated " by presence of thorium-B,which is isotopic with lead, and to determine the activity of thelead in the less soluble compound crystallising out.When activelead nitrate and inactive lead chloride are dissolved in molecularzp F. A. Lindemann, Phil. Mag., 1919, [vi?, 38, 173; S. Chapman, ibid.,182 ; A., 1919, ii, 390.W. D. Harking, Nature, 1920, 105, 230 ; Science, 1920, 51, 289.E. Kohlweiler, Zeitsch. physikal. Chem., 1920, %, 513; 95, 95; A.,26 J. N. Brijnsted and G. von Hevesy, Nature, 1920, 106, 144.ii, 610, 615RADIOACTIVITY. 227proportion in boiling pyridine, the lead1 in the lead chloride crystal-lising out is half as active as the lead in the original lead nitrate,but when an active lead salt is so mixed with an organic compoundof lead, such as lead tetraphenyl or diphenyl nitrate, in suitablesolvents, no interchange of lead occurs, and the active lead saltretains its original activity.This const'itutes something like adirect proof of the ionic dissociation theory and of the currentviews as to the difference between the nature of chemical union inelectrolytes and non-electrolytes. When the acetates of quadri-valent activated lead and of bivalent inactive lead are mixed inglacial acetic acid, the activity of the first compound, after crystal-lising out from the mixture, is reduced to one-half. This indicates,since the two lead ions differ only by two eleotrons, a free inter-change of electrons between them and a dynamic equilibriumbetween ions and electrons, and between free electrons and theelectrodes in electrolysis.28Isotopes have been used to determine the velocity of diffusionof molecules among themselves. The rate of diffusion is dependenton the molecular diameter, and not on the mass, so that a radio-active isotope diffusing among the inactive moleoules of the sametype of element affords the means for investigating experimentallythis otherwise insoluble problem.The case has been tried withmolten lead. A t the bottom of a narrow, vertical tube was placeda layer of lead rendered active by the presence of thorium-B, andabove it a layer three times the height of common lead. Thewhole was kept a t 340° for several days. After cooling, thecylinder was cut up into four equal lengths, each melted andhammered into foil, and the concentration of thorium-B in eachdetermined by a-ray measurements.Values for the diffusioncoefficient between 1.77 and 2.54 per sq. cm. per day, with a meanof 2.22 in seventeen experiments, were obtained. This oorrespondswith a diameter of the lead molecule between 0.78 and1.16( x 10-8 cm.), according to the formulae used to connect thetwo quantities. The value found, when reduced to a temperatureof 1 8 O and for a fluid of the viscosity of water, becomes 2.13.Since the value for lead ions diffusing in aqueous solutions is 0.68,this indicates that the molecular diameter in the case of metalliclead is only a third of that in the case of the ion, and shows thatthe latter is probably hydrated.29a* G. von Hevesy and L. Zechmeister, Ber., 1920,53, [B], 410 ; A , , ii, 278 ;a0 J. Gr6h and G.von Hevesy, Awn. Physik, 1920, [iv], 63, 85; d.,Zeitsch. Elektrochem., 1920, 26, 161 ; A., ii, 345.ii, 739.1 228 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Researches analogous to those reported have led t o the detectionand isolation of volatile hydrides of lead and tin.30A notable rival has been developed t o the view described in thelast Report31 of an atom in which the electrons are supposed t orevolve in orbits around the nucleus, with special assumptions asregards the radiation of energy in quanta rather than continuously.I n this atom, the electrons surrounding the nucleus are supposedto occupy, or oscillate about, certain fixed positions in the struc-ture. This fixed electron type of atom has been found to possessmany advantages in chemistry and physics, notably in acmuntingfor the Periodic Law, the various categories of chemical compounds,ionised and un-ionised, and the arrangement of the atoms in thecrystal space-lattice as determined by X-ray methods.It may besaid to draw its underlying postulates from facts in these fieldsrather than from any purely mathematical or fundamental reason-ing.32 The chief idea is that, in the outermost shell of electronssurrounding the nucleus, the electrons tend to form an octet andto owupy the corners of a cube. I n the outermost shell all eightcorners are occupied in those atomic structures corresponding withthe zero family of chemically inert gases. The chemical activityof other elements is due to some of the corners being not occupiedby electrons, whereby two or more atoms tend chemically to“combine.” The oombination may be of two kinds.Either theatoms with only a few of the corners occupied by electrons, thatis, of those elements in the first families of the Periodic Table, losetheir electrons altogether, forming positive ions, such as Na’, Mg”,Al”’, to the atoms which have all but a few of the corners occupied,that is, of those elements in the last families of the Periodic Table,with the formation of the negative ions, such as Cl’, S”, or moreoften to groups of these atoms. This way of regarding ionisedcompounds was, of course, arrived a t long before this theory wasproposed, but it emphasises the completely separated existence ofthe two ions forming the moleoule, even in the solid state, whichis supported by the character of the space lattices of the crystals3 0 Arzn.Reports, 1918, 15, 225 ; 1916, 13, 266 ; F. Paneth and K. Furth,Ber., 1919, 52, [B], 2020; F. Paneth and 0. Norring, ibid., 1920, 53, [B],1693 ; A., ii, 41, 758.3l Arm. Reports, 1918, 15, 206.32 G. N. Lewis, J. Amer. Chem. SOC., 1916, 38, 762; A., 1916, ii, 310;I. Langmuir, Proc. Nut. Acad. Sci., 1919, 5, 252; J . Amer. Chem. ~ o c . , 1919,41, 868, 1543; 1920, 42, 274; A., 1919, ii, 328, 506; 1920, ii, 243; Science,1920, 51, 605 ; A., ii, 656 ; W. Kossel, Zeitsch. Physik, 1920, 1, 395 ; A.,ii, 681RADIOACTIVITY. 229of salts. The forces a t work are the opposed charges on the ionswhich act statistically, n sodium ions requiring the simultaneouspresence of n chlorine ions, rather than each sodium atom beingattached to one chlorine atom, as in the formula Na-C1.I n the second kind of combination, namely, that correspondingwith definite atomic linkings, such as are regarded to exist betweenthe atoms of the molecule in organic compounds and in non-electrolytes generally, the theory is mdre original.Different atomsso rigidly linked together are regarded as sharing electroiis inpairs. Two electrons held in common by two atoms constitute theordinary single bond. Fourelectrons in common correspond with a double bond. The cubesare attached face to faoe. A corner to corner attachment of twocubes, which the single bond most closely suggests, is not con-sidered to occur a t all.The sharing of a pair of electrons by twoatoms is regarded as the single unit of valency.I n addition to this type of definite linkage, two others arepostulated. The hydrogen nucleus is capable of sharing a pair ofelectrons, its own and one derived from another atom, either anatom of itself, as in the hydrogen molecule, or ail atom containingail uncompleted octet. Thus in water the oxygen nucleus is atthe clentre of a cube of electroiis, two pairs of which, a t two oppositepairs of contiguous corners, being shared with two externalhydrogen nuclei. Pairs so held are supposed to be drawn closertogether, distorting the cube. I n this way, the tetrahedralcharacter of the carbon atom is accounted for. The uncombinedatom of carbon, if i t existed free, which, of course, never occurs,would have four of the eight corners of the cube oocupied withelectrons. If symmetrically distributed, these would occupy thecorners of a regular tetrahedron. When it shares these in pairswith electrons of other radicles or atoms in compounds, the draw-ing together of each pair shared preserves the tetrahedral characterof the arrangement in the combined atom.The facts of stereo-chemistry require free rotation to be possible about a single bond,and not about a double bond, whereas, unless further assumptionsare made, such as that the pair are drawn together to one pointor supposed to rotate round one another, free rotation would notbe a possibility for a single linkage on this theory.The existenceof triple bonds again, whiclh is possible on a tetrahedral, isimpossible on a citbic atom. if only partly deformed to a tetra-hedron.The second type of combination postulated is rather surprisingin that a pair of nuclei are supposed to be contained in a singleciihe. or octet, i n such combinations as t,he nitrogen molecule,The cubes are attached edge t o edge230 ANNUAL REPORTS ON THE PROQRESS OF OHEMISTRY.carbon monoxide, hydrocyanic acid, and nitric oxide. If this iscorrect, suah compounds would represent, as it were, structureshalf-way between those typical of atoms and molecules respectively.Although something might be said for such a structure representingthe properties of nitrogen, one would scarcely have expected it tobe capable of representing also such an extremely active gas asnitric oxide.Into this theory of valency, which so far seems to be confinedmainly to the lighter elements in the earlier part of the PeriodicTable, it is unneoessary further to enter here.Of more topicalsignificance is the way in which the atomic numbers of the PeriodicLaw are accounted for. The atom is regarded as made up of con-centric shells of electrons of relative diameter 1, 2, 3, 4, and relativearea 1, 4, 9, 16. Each electron is regarded as occupying the samesuperficial area, to whatever shell i t belongs. The inert gases arethe elements formwhich the outer shell contains its full oomplementof electrons. Helium, of atomic number 2, has two electrons atthe poles of the first shell.The line joining them and passingthrough the nucleus is regarded as the polar axis of the atom.The plane passing through the equator divides the shell into twohemispheres. There are no electrons in the equatorial plane of anyatom. I n the outer shells, concentric with the first, they are dis-tributed according to the symmetlry of a tetragonal crystal. Foursecondary planes of symmetry, at 45O with each other, pass throughthe polar axis. The second completed shell, being four times thearea of the first, contains eight eleotrons, occupying the eightcorners of a cube. This is the neon atom. The atomic numberis 10, and i t contains eight electrons in the second shell-four ineach hemisphere above and below the equatorial plane-and twoin the first, or helium, shell.Every shell, other than the inner-most, after getting filled up with electrons once, is filled up twice,and the next inert gas is argon, atomic number 18, containingsixteen electrons in its second shell and two electrons in its first.I n the next, krypton, of atomic number 36, the third shell con-tains eighteen eleotrons, two distributed a t the poles and the othersixteen symmetrically with regard to the polar axis and the sixteenunderlying electrons of the second shell. By filling the third shellagain we get xenon, of atomic number 54. The fourth shell con-tains thirty-two electrons, and the next inert gas must have anatomic number 86. This is the correct atomic number for theemanations of ,the radioactive elements.Unfortunately, thePeriodic Table comes to an end before this ingenious theory canbe further tested. That, however, the table should prooeed touranium, which possesses complete chemical analogy to tungsteRADIOACTIVITY. 23 1and molybdenum, instead of t o a second lot of rare earth elements,after radium and thorium, raises the doubt whether i t just doesnot come to an end in time for the theory. Undoubtedly, how-ever, it is an achievement, even by such arbitrary assumptions, tohave accounted for the actual sequence of elements in the tablea t all.The theory has found general support in the explanation of thearrangements of atoms in crystals as elucidated by X-raymethods.33 It is possible to assign t o each atom in the space-latticea definite approximate diameter, and to regard the crystal as builtof spheres of these diameters closely packed.When the atomicvolumes corresponding with these diameters are plotted againstatomic weights, a curve, in every respect analogous t o LotharRfeyer’s atomic volume curve, is obtained, but applicable to thecompounds of the elements. Then i t is found that two electro-negative elements are situated close together, and are assignedsmall diameters when, ’according to the above theory, they shareelectrons ; but the electropositive elements, which exist as separatedions and do not share electrons with their neighbours, are situateda t a distance from them, and appear to have large diameters.From crystal data, the diameter of the electronic shells correspond-ing with neon, argon, krypton, and xenon are put at 1.30, 2.05,2-35, and 2-70 Angstrom units respectively.The theory, beingdefinite and easily visualised, if arbitrary, will doubtless justifyitself in drawing attention to the many different types of chemicalinteraction, which hitherto have been too liable to be confusedtogether and forced into a mould to fit just the one type of inter-action which the ordinary valency-bond theory suffices to explain.It is not yet possible t o bridge the gap between this idea and thatof the rotating electron atom, which has grown up largely 1roiiithe study of the wave-lengths of the characteristic X-rays theni-selves. Undoubtedly each type has its advantages, but forohemistry the fixed electron type seems easily to hold the field.I n the light of these advances, an experiment showing that thea-radiation from different faces of a large crystal of uraniurunitrate was, within the error of experiment, of the same intensity.seems to show that the a-particles are shot out from trhe nucleusduring disintegration, without, relation to the orientation of t h eatomic axis, for it may be regarded as at least highly probablethat in the crystal space-lattice the atoms have their axes orientatedin a regular manner 34** W.L. Braggy Royd Institution Evening Lect,ure, May 28th, 1920 ;a4 T. R . Merton, Phil. Maq., 1919. [vil, 28. 4 6 3 : A . . 1919, ii, 453Phil. Mag., 1920, [vi], 40, 169; A . , ii, 637232 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Ifiyh- f re y I I e 12 c?/ S p e c t ra of t h e Ele m e n t s .35Work in this field proceeds apace, though without breaking muchfresh ground.Precision measurements, to a degree of acouracyone-hundredfold greater than previously, have been carried out forthe lines in the K series of a number of elements from chlorine tocopper, and the results compared with the various mathematicalforniulae .proposed. The spectrum of tungsten, for which the X,L , and iif series can all be studied, has also been examined, and aspectrograph constructed t o bridge the gap between these two seriesof precision meas~rements.3~ The Jf series has been further in-vestigated, and extended from uranium as far as dysprosium.37New measurements of the absorption bands of thulium,neoytterbium, and luteciuin in the Ii series have been made.38An examination of the.L absorptioii spectrum of a pure radiumchloride solution gave two lines, 0.802 and 0.670 A . , in agreementwith the atomic iiuinber 88 assigned by the Periodio Table.39 Bythe use of a reinforcing screen of calcium tungstate, the fl, line ofthe A- spectrum of tungsten (0.1844 A.) has been shown t o be adoublet separated by about 0.0007 A.40 In ail examination of theX-ray absorption spectrum of phosphorus, differences of wave-lengthwere observed for different forms. The wave-length 5.767 A. wasfound for the black phosphorus of Bridgeman, and 5.750 A. forphosphoric acid and its ammonium salt, whilst red phosphorusshows a double limit, corresponding with each of the two wave-lengths given. This is believed to be the first case noticed of thechemical state .of an element affecting its X-ray spe~trum.~'Arrangements have been described for the examination of thespace-lattices of powdered materials, by which it has been shownthat thorium and nickel in powder form have face-centred cubicallattices, and niagnesium a lattice ooniposed of two interpenetratingsimple hexagonal lattices.42X-Rays have also been used to determine the size and structureof the particles of organic and inorganic colloids.Gold and silver35 Compare Ann. Reports, 1916, 13, 257.36 M. Siegbahn, PhiE. Mag., 1919, [vi], 37, 601 ; 38, 639 (and with A. U.TAde), 647 ; A., 1919, ii, 261, 498, 489 ; E. Hjalmar, Zaitsrh.Physik, 1920,1, 439 ; A . , ii, 665.37 W. Stenstrom, Atzti.. Physik, 1818, [iv], 57, 347 ; A., 1919, ii, 90.38 M. de Broglie, Compt. rand., 1920, 170, 725 ; A . , ii, 277.3 y Ihid., 1019, 168, 854; 169, 134; A . , 1919, ii, 207, 358.41 J. Bergengren, +bit?., 1920, 171, 624 ; A . , ii, 654.42 H. Bohlin, L41?j/. T'l/?/sik, 1920, [iv'), 61, 421 ; A . , ii, 214 ; compareIbitl., 1920, 170, 1053 ; A., ii, 344.A. W. Hull, J . d v t r r . C'ittm. 8 o c . , 1919, 41, 1 1 6 8 ; .4.. 1919, i i . 470RADIOACTIVITY. 233in the colloid fomi possess the same bpace-lattices as in largecrystals, even when the particles are too small to be visible underthe ultramicroscope. I n old silicio acid and stannic acid gels,traces of crystalline structure can be detected, but not in thetypical organic colloids, such as albumin, gelatin, cellulose, starch,and the like.43It, has been pointed out that it is a necessary consequence ofthe modern views of crystal structure that, in certain cases, thechemical composition of the crystal must depend on its size.Thusiron pyrites, with a space-lattice consisting of an atom of sulphurwithin a cube, four alternate corners of which are occupied by ironatoms, instead of the formula FeS, must possess a compositiongiven by Fe(n+l$z2n::, where n is the number of elementary cubesin the crystal. I f n is 50, the particle would still be visible bythe aid of the ultramicrosoope, and its composition would be givenby FeS,.,,,.44a-Rays.The Geiger-Suttall Relation.-The logarithmic connexioiibetween the period of average life of an atom and the range ofthe a-ray expelled from i t during disintegration, and the theoryo€ the cause of atomic disintegration to which it has led, have beenthe subject of closer examination.On this the0ry,~5 the instabilityof the atom is supposed to result from the simultaneous conjuiic-tion of a large number, N , of separate particles, moving independ-ently of one another within the atomic nucleus, in a certain favour-able relation. The chance of disintegration depends on somethinglike the one-hundred-and-sixtieth power of the velocity of thea-particle expelled, and suoh a law can scarcely be explained exceptas an expression of the probability of the fortuitous occurrence ofa very large number of independent events.The actual value first deduced from the periods and ranges forAT was about 80, a nuniber of the same order as the atomic number.The relation between range and period on this theory becomeslog A =Na + &?V log R ,where R is the range and a is, approximately a t least, a constant.From new data on the ranges of some of the members of thevarious disintegration series, values for L7\T, 81, 77, and 71, havebeen assigned for the uranium-radium, thorium, and actiniumseries respectively.46 Clearly these must be of the nature of mean43 P.Scherrer, Nachr. Ges. Wiss. Gtittingen, 1918, 96 ; A . , 1919, ii, 274.44 A. Quartaroli, Gnzzetta, 1920, 50, ii, 6 0 ; A., ii, 609.45 Ann. Reports, 1916, 13, 257.46 S.Bloyer, V. F. Hess, and F. Paneth, Sitzungsber. Akad. Wiss. Wien,1914, 123, 2n, 1459; S. ? r l e y ~ , ibid., 1916, 125, Za, 201.I234 ANNUAL REPORTS ON THE PROQRESS OF CHEMISTRY.values, since the number of independently moving particles in thenucleus must diminish by unity with each a- or &particle expelled,that is, with each step in the disintegration series. The experi-mental numbers given for the uranium-radium and the thoriumseries agree with the mean values to be expected if the atomicnuclei are practically composed of helium nuclei and bindingelectrons. Thus for thorium, with atomic weight 232 and atomicnumber 90, 58 helium nuclei and 116-90=26 binding electronswould constitute a system consisting of eighty-four independentunits.The value of N for thorium itself would be one less, sincethe probability relation holds between one of the particles and therest of the nucleus. The mean value for the series between thoriumand thorium-C, considering N to be reduced by unity for eaoh a-and for each &particle expelled, would be 78 instead of the valuefound, 77. For the uranium atom, with atomic weight 238 andatomic number 92, two hydrogen nuclei a t least must be postulated.I f there are only two, there must be 59 helium nuclei and118+2-92=28 binding electrons, making a total of 89 inde-pendent particles. The mean of the number, diminished by one,is, for the series uranium to polonium, 81 or 82, again in excellentagreement with the experimental value.If, now, the correct values for N are introduced for eachmember of the series, and the value of the constant a calculated,i t is found that a is truly constant for the middle members ofeach series, but is markedly, although not greatly, different forthe first and last members of each ~eries.4~ The actinium series isscarcely yet worth consideration here, as, in absence of all experi-mental evidence as to the atomic weights of its members, the valuesto be assigned to N must be a matter for speoulation.I n addition,i t obeys the logarithmic relation only very imperfectly. Thedifficult question as to the cause of the disintegration of the atomin radioactive changes seems at least to be progressing towards asatisfactory and highly suggestive answer.I n other papers, Bohr’s principle of angular momentum has beenapplied t o the internal economy of the nucleus, and the conclusionreached that the motions of the particles remaining in the nucleusare not affected by €he successive steps in the atomio disintegration.The radius of the orbit of revolution of the a-particle in thenucleus before expulsion has been calculated, and found to diminishby steps with each successive disintegration.**A collection of papers has appeared on the counting and photo-47 G.Kirsch, Physikal. Zeitsch., 1920, %, 452 ; A., ii, 577.48 H. T. Wolff, ibid., 175, 393 ; A., ii, 366, 578RADIOACTIVITY. 235graphic registration of a-particles. The eleotrometer method,using high potential gradients just below the sparking potential,whereby the ionisation is enormously magnified by collision, hasbeen the one employed.As the most suitable gas for filling thecounting chamber, a mixture of 54 per cent. of carbon dioxideand 46 per cent. of air was used. It was found that a mixtureof carbon dioxide and air, with the former in excess, respondedonly to the a-rays, and not to the p- and y-rays.A new determination of the number of a-particles expelled persecond per gram of radium (element), free from disintegrationproducts, gave 272( k0.02) x 1010. In arriving a t this value,80,000 a-particles were counted. This is about, 4 per cent. abovethe previously accepted value, even after oorrection in terms ofthe International Standard.49I n a special research, it was found that 1.5 milligrams persq.cm. of mica, of density 2.87, correspond, in stopping powertowards a-rays, with 1 cm. of air a t 760 mm. and 1 5 O . 6 0The individual intervals between the emission of u-particles bypolonium have been systematically studied by photographicregistration methods for the case of 10,000 emissions. The require-ment of the theory of probability was very exactly verified. Thefraction of the total number of intervals of duration greater thanT is E - @ ’ ~ , where 8 is the mean interval. For very short intervalsthe law is departed from, owing to the emission of a-particles withintervals between them too small t o be distinguished by the meansemployed. From the results obtained, the number of such‘‘ doublets ” could be very exactly evaluated.51An effeot, analogous to the ‘‘ spluttering” of metals under theaction of the cathode rays in discharge tubes, has been observedwith a-rays for the noble metals nickel and aluminium, but notfor such metals as copper, the surface of which is easily oxidisedby the action of the atmosphere.Another effect due to a-rays wasobserved with polonium, electrolytically deposited on metal foil.The a-particles emitted .at grazing incidence appear to “ knock off”the polonium in aggregates of several molecules a t a time, causingan effeot analogous to the volatilisation of the polonium, which iscalled “aggregate recoil.” The effect is very much more pro-nounced in a vacuum than in the atmosphere. It is greatest withfreshly deposited preparations, and diminishes with their ags.*II Ann.Reports, 1914, 11, 274.bo R. W. Lawson and V. F. Hem, &kung.&er. dkad. W&. Wkn, 1918,121,61 (Mrne) 116. Curie, J . Phyrr. Radium, 1920, [viJ, I, 12 ; -4., ii, 727.3a, 406, 461, 636, 699, 943.I* 236 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.The most regular results are obtained with palladium and platiiiuuifoils electrolytically saturated with oxygen. Saturating withhydrogen diminishes the effects and makes them irregular .52y-Rays.In a continuation of the researches fully described in the lastRep0rt,52~ two types of secondary y-rays, referred to as S; and S,,have been found to be associated with the two primary componentsof the y-rays of radium, designated as K , and K,.The first isof the nature of a scattered primary, possessing the same coefficientof absorption as K,, and distributed with deoreasing intensitywith increasing angle of scattering. None is detectable a t anangle of 90° or beyond, or, in other words, this secondary radiationis confined to emergence. For this type, the scattering power ofdifferent atoms is proportional to their atomic number. The typeS2 is, in general, different in penetrating power from K,, and isscattered over an angle of 180°, constituting an incident, as wellas emergent, radiation. For light atoms the scat’tering is pro-portional to the atomic number, but for heavy atoms to the squareof this number.53 The absorption of divergent beams of y-rays hasalso been studied, with the view of throwing light on the reasonwhy y-rays, although complex and scattered, so nearly obey thetheoretical law of absorption to be expected for a homogeneous,non-scattered beam.54Methods have been developed for ‘‘ oounting ” y-rays analogousto those referred to under “a-Rays.” The effect of a y- or B-rayis, in general, twenty to twenty-five times less than that of ana-ray, but with a sufficiently sensitive counting arrangement theymay be counted with ease and certainty.Some special precau-tions are taken, on account of the high potential necessary torender the response very sensitive, but otherwise the arrangementsare very much as for the a-rays. The gas used in the countingchamber is air, drawn from the free atmosphere and stored oversulphuric acid until any emanation initially present has had timeto decay, and filled into the chamber through cotton wool andphosphorio oxide.The y-ray acts by liberating a high-velocity &particle from themetal walls of the counting chamber, and the same methods are52 R.W. Lawson, Sitzungsber. Akad. Wiss. Wien 1918, 127, Za, 1315;1919, 128, 2a, 795.52a Ann. Reports, 1918, 15, 211.53 K. W. F. Kohlrausch, Sitzungsber. Akad. Wiss. Wien, 1919, 128, 2a, 853.Ann. Reports, 1918,15,213 : M. Blau, Sitzunpber. Akad. Wiss. Wien, 1918.127, 2a, 1253RADIOACTIVITY. 237equally applicable for /3- as for y-rays. It was found that thenumber of y-rays given per atom from radium-B and radium-C,respectively, were practically the same. The total number ofy-rays from both together is, in terms of the number of a-particlesfroin radiund', between two and 0ne.55C'lzemical Actions of the Rays of Radimm.The reactions prooeeding in common gases, when mixed withradium emanation, and due to a-rays, have been the subject of twoexhaustive investigations, chiefly to ascertain whether the facts arein agreement with the theory that the reactions obey a form ofFaraday's law, that the molecules formed in the reaction are equalin number to the pairs of ions formed from the rays in the gas.56I n one research, four gases, hydrogen sulphide, ammonia, nitrousoxide, and carbon dioxide, were studied.Other conditions beingthe same, the amount of decomposition is proportional to theamount of emanation present.The decomposition increases as thesize of the reaction vessel is increased, to a limit correspondingwith the state in which praotically the whole of the energy of thea-rays is spent in traversing the gas molecules. For hydrogensulphide, the thermal coefficient of the velocity of reaction is prac-tically zero from -180° to 1 8 O , and above this, to 220O; is slightlynegative. For nitrous oxide, the reaction proceeds probably intwo directions, for the most part with the formation of nitrogenand oxygen, but also with the formation of nitric oxide andnitrogen. The accumulation of nitrogen peroxide as a result ofthe second reaction retards the reaction. Here, again, changes oftemperature producie but a slight effect on the velocity of reaction,the coefficient being negative below and positive above 1 8 O .Forammonia, the coefficient is positive and considerable up t o 220O.Carbon dioxide was found t o undergo oiily a very slight decom-position, and the rapid change recorded by other investigators isattributed to the effect of mercury and phosphorus in the vessel.For these reactions, Faraday's law was found not to apply. Theratio between the number of molecules produced and the numberof pairs of ions formed by the a-rays exceeds unity in cases whereno catalytic acition is involved.57I n the other research, the combination of hydrogen arid oxygenwas re-studied. Here it was found that 3-92 molecules of waterresult for every pair of ions formed in the gas. It appears that" V.F. Hess and R. W. Lnwson, Sitzungsber. Akad. Wiss. Wien, 1916,125,2% 285, 585, 661. '' E. Wourtzel, I;e Radium, 1919, 11, 2139, 332; A., ii, 214.GG Ann. RrpoTts, 1912, 9, 323238 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.in all the earlier results of Ralusay and Cameron, which gave sup-port to the ionisation theory, the amount of emanation used hadbeen much overestimated. It had been calculated from the timeof accumulation and quantity of radium, which, since the develop-ment of exact methods of measuring the emanation by the y-rays,is known to be quite untrustworthy. The velocity of reaction wasfound to be proportional to the quantity of emanation and in-dependent of the temperature. It was increased by inorease ofoxygen above the stoicheiometrical proportion, the velocity ofreaction continuing to rise as this excess increases with the reaetion, and diminished by increase of hydrogen, the velocity con-tinuing to fall as the excess of hydrogen increases.This is to beexpected from the ionisation theory, since the relative ionisationsin oxygen and hydrogen are as 1-09 to 0.24, that of air being unity.In very small vessels, particularly at low pressures, the velooityof reaction is abnormally high. This is ascribed to the atom ofradium-A, recoiling from the atom of emanation, bringing aboutthe combination in the same way as an a-particle. Under themost favourable conditions for magnifying this recoil effect rela-tively to that produced by the a-rays, i t may exceed the latter sixor seven times.The relative effect produced is in agreement withthe data as to the magnitude of the ionisation produced by recoilatoms.With the single exceptioii of hydrogeii and chlorine, where thechemical action may be several thousand times as great as theionisation theory requires, i t is claimed that there is a generalstatistical agreement between the iiuiiiber of ions and the numberof molecules produced for a large number of reactions. The twonumbers are not the same, but they correspond within a inultiplierof a few units only in either direction. This is true for reaotionsproduced by the cathode rays and @rays, as well as those resultingfrom the action of a-rays and recoil atoms. The ratio of four toone, in the present case, between the numbers of the moleculesand ions can be explaiiierl by ionic possibilities, without recourseto other theories.6*In the reverse reaction, the dercmposition of liquid water bythe a-particle into hydrogen and oxygen, about one molecule ofwater is decomposed per pair of ions formed.In practice, thereoombination of hydrogen and oxygen under t.he action of theemanation proceeds almost t o completion a t constant volume,because the water condenses to droplets, and so is removed for themost par! from the action of the a - r n y ~ . ~ ~5 3 S. C. Link J . Amer. Chem. SOC., 1919, 41, 531, 551: A., 1919, ii, 210.6 9 Idem, Tram. Amer. Etectroch-mn. Soe., 1918, 34, 211RADIOACTlVITY. 239The long series of parallel experiments on the action of thepenetrating rays of radium and of ultra-violet radiation from aquartz mercury lamp on organic substances has been continued.The substances studied comprise a mixture of maleic and fumaricacids, solutions of formic and benzoic acids and of carbamide, dryand wet toluene, chloroform, and carbon tetrachloride.Theeflects of the two kinds of irradiation are, in general, similar, theultra-violet light being usually almost incomparably the morerapid. The results bear out the general view that these agentshave a shattering effect on almost all molecules, followed bynumerous secondary reactions among the products.GOThe thermoluminescence and decolorisation of glass which hasbeen exposed to the rays of radium, on heating, have been shownto be independent of one another.For freshly exposed glass,thermoluminescenaa starts on heating below looo, and a t 200° forspecimens exposed some years previously. Decolorisation does not,however, occur until the temperature of 500° is reached.61 Thecolorations and thermoluminescence produced in a great variety ofminerals have been examined. The fluorspars, by reason of thealmost bewildering variety of colour changes they undergo and thebrilliance of the thermoluminescence produced, are among themost interesting.@ I n this connexion, the variety of fluorsparfrom Wolsenberg, Bavaria, locally called " Stinkfluss," deservesspecial mention. On being crushed, it emits a peculiar odour, whichthose who have made a careful study of the mineral assert is with-out doubt due to free fluorine.Radium rays easily reproduce thenatural dark blue colour in the mineral after the colour has beendischarged by heating, but do not restore its odoriferous quality.63Studies of Radioactive Minerals.Age of Thorium Minerals.--In a careful review of the difficultquestions connected with the age of thorium minerals, both theisotopes of lead derived from thorium are regarded as stable,and the age of the mineral, A , is deduced from the formulaA = Pb x 7,9?0 million years,U + 0-384 Thwhere Pb, U, and Th are the percentages of these elements in the6o A. Kailan, Zeitsch. ph?~&al. Chem., 1920, 95, 215 ; A., ii, 576 ; &tzungs-ber. Akad. Wias. Wkn, 1919. 128, 2a. 831 ; 1917, 126, 2n, 741.61 5. C. Lind, J .Phys&ccrl Chem., 1920, %, 437 : A., ii, 576.E. Newbery and H. Lunton, Mem. Manchpster Phil. S'oc, 1918, 62,No. 10; A., 1919, ii, 130.B3 F. Henrich, Zeitsch. angew. Chem., 1920, 33, 5, 13, 20 ; A., ii, 216240 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.mineral. The conclusion is reached that in (1) the Middle-Devonian formation a t Brevig, Norway, in the Precambrian form-ations of (2) Arendal-Gegend, Norway, and (3) Moss, Norway, and(4) in the thorianite-bearing pegmatites of Ceylon, all thoseminerals with less than three times as much thorium as uraniumgive quite concordant ages for the formations, namely, (1) 300,(2) 1300, (3) 950, (4) 500 million years, respectively. These mustbe regarded as true primary minerals; but all those for which theTh/U ratio is greater than 3 give smaller ages, and must beregarded as secondary minerals derived from the primary byvarious processes of change in which the content of thorium hasbeen enriched.I n the first class are to be found, in (1)eudidymite, eucolite, zircon, pyrochlor, and biotite, in (2) clsveite,in (3) broeggerite, and in (4) pitchblende and thorianite, whilst inthe second class are, in (1) freyalith, thorite, and orangeite, in(2) uranothorite and orangeite, in (3) uranothorite of Rittero,and in (4) thorite. The atomic weights of the lead from the threevarieties of thorianite, already given, conform well to the viewthat they are primary constituents of the pegmatite, which has anage between 400 and 500 million ~ e a r s .6 ~The ratio of thorium to uranium in a number of minerals hasbeen determined by radioactive methods. I n Morogoro pitch-blende there is 0.5 per cent. of thorium and 74.5 per cent. ofuranium ; in pitchblende of St. Joaohimsthal, per gram of uranium,4.68 x 10-5 grams of thorium, making, with the estimated1-96 x 10-5 grams of ionium, a total 6-64 x 10-5 grams of thoriumisotopes. A monazite sand of unstated origin, containing 7.23 percent. of thorium, was found to contain 0.087 per cent. ofuranium.65 I n another estimation, monazite sand from Brazil wasfound to coiitain 0.8, and from India Oa102 ( x 10-9 gram ofradium per gram). These correspond with 0.235 and 0.03 percent. of uranium respectively, and, on a thorium content of 4 and9 per cent., mesothorium preparatioiis obtained from them wouldowe 28 and 2.1 per cent., respectively, of their initial y-aotivity tor a d ium .66In a nlonograph on broggerite, which contained 63-66 percent.of uranium, 6-6.5 per cent. of thorium, 9.5-10 per cent.of lead, and 0-7-1.5 per cent. of rare earths, hydrofluoric acidG 4 R. W. Lawson, Sitzungsber. Akad. Wiss. Wien, 1917, 126, 2n, 721;G5 S. Mttyer, ibid., 1919, 128, 2a, 897 ; A . , ii, 548.06 J. E. Underwood and H. Schlundt, Trans. Amer. Electrochem. SOC.,(Tn the abstract, lo-' gram should read loA9A . , ii, 149.1918, M, 203; A., ii, 146.gram.RADlOACTIVITY. 241was found to be the best precipitant for thoriuin in presence ofuranium. As is well known, the niethod of separation based onthe solubility of uranium nitrate in ether or acetone is useless.The Pb/U ratio in this mineral is essentially constant a t 0.12 to0.13, which corresponds well with the age of about a thousandmillion years, already given, for the pegmatite dykes in the granitesof Moss, Norway, from which it is obtained.67T h e I;runizc?n-Radizcn Rcrtio .--This important ratio has beenredetermined for a carefully seleoted Colorado pitchblende.Theuranium was estimated analytically, and the radium by theemanation method against specially made absolute standards ofradium. These were prepared, by dilution to 1.5 x 10-9 grams ofradium per c.c., from a radiuin chloride of 100 per cent. purity,measured against. the International Standard by y-rays. A milliontimes the quantity of barium was added to the diluted standardto protect, the minute amount of radiuin from precipitation.Theresult gave 3.4 x 10-7 grams of radium as the quantity in equil-ibrium with 1 gram of uranium. This was the original “Ruther-ford-Boltwood ” value, but it was subsequently corrected t o3.23 x 1 0 - 7 on the Internatioiial Standard. Much independent)work has shown that the uncorrected value was substantiallycorrect, and it is very satisfactory t’o have had this oonclusioiiconfirmed so convincingly.68Relative a-Activities of FULII iu?n, t m t l Rndium. --Many pointsremain to be cleared up with regard to the relative a-activities ofradium and uranium minerals. A new determination has sub-stantially confirmed the original determinations.Taking theactivity contributed by uranium (U-I and U-77) as unity, thetotal activity of the mineral is now found to be 4.73 instead of4.69, and the part due to radium itself as 0.488 in place of 0.45.I f , however, the radium were produced from the uranium in adirect line without branching, its a-ray activity, calculated fromthe law that the ioiiisation is proportional to the two-thirds powerof the range of the a-particles, should be 0.57. The value found,0.488, can only be accounted for if the actinium braiich claimseither 26 per cent. of the atonis disintegrating if it starts fromuranium-I, or 14 per cent. if i t starts from uranium-11. Fromthe proportion of the total activity contributed by the actinium6 7 E.Cleditsch, Archiv for Mathematill: og A-atitrvideiLsTcnb, Christianiu,1919,36, Nr. 1 ; compare A. Fleck, T., 1913,103, 384.6 8 S. C. Lind and L. D. Roberts, J . Arne?.. Chem. Soc., 1920, 42, 1170 ;A., ii, 463 ; compare Becker and Jannasch, Jahrb. Radioaktiv. Elektronik,1915, 12, 1 ; F. Sod+- and (Miss) A. F. R. Hitchens, Phil. Mag., 1915, [vi].30, 218 ; A., 1915, ii, 726 ; E. Gleditscb, Zoc. cit242 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.series, estimated from both the new and the old measurements as0.26 to 0.28 in terms of that of the uraniums as unity, a branch-ing factor of 8 per cent. for the actinium series has hitherto beenaccepted, but work about to be considered has reduced this to3 per cent. or less, so that a real inconsistency between the experi-mental data and our theoretical interpretation exists, the clearingup of which might throw much light on the branching of theseries.69The Uranium-Actinium Ratio.-In a study of the pitchblendesof Morogoro and St.Joachimsthal, broeggerite from Norway, andtwo thorianites from Ceylon, representing extremes of Th/U ratio,the constancy of proportionality between radium, and thereforeuranium, and actinium has been confirmed. Since the thorium-uranium ratio varied between the limits of 6x10-5 and 6, theindependence of actinium and thorium, and the genetic connexionbetween actinium and uranium, follow.70The Uranium-X-Uranium-P Ratio.-For uranium derived fromthe same materials, the constancy of proportionality in the ratesof production of uranium-X and -P was established, and the geneticconnexion extended to the supposed first member of the actiniumbranch series, uranium-P.In this work, periods of average lifefor uranium-X, of 34-37 days, and for uranium-P of 35.53 hours,were found. The former is slightly, and the latter considerably,less than the previously accepted values, namely, 35.5 days and52.8 hours.The method adopted for separating from uranium theuranium-X, and -P, both being isotopes of thorium, consisted inneutralising the strong uranium solution with sodium hydroxide(not potassium hydroxide or ammonia) and adding a small quantityof a solution of a cerous salt and hydrofluoric acid. The cerousfluoride carries down with it the thorium isotopes, and is redis-solved in hydrochloric acid.A milligram of dissolved zirconiumis added, and the solution is precipitated with a solution of sodiumhydrogen hypophosphate, NaHB03,3H,0, according to the methodof Koss. This precipitates the thorium isotopes with the zirconiumand' leaves the cerous salt dissolved. The relative activity ofuranium-X, and -P proved to be independent of the source of theuranium, and from i t a branching factor for the actinium seriesof a t most 4.2 per cent. was ded~ced.~1an Ann. Reports, 1909, 6, 259 ; J. H. L. Johnstone and B. R. Boltwood,Phil. Mag., 1920, [Vi], #, 50 ; A., ii, 523.70 S. Meyer 8nd V. I?. Hess, 8itzung8ber. Akad. Wiss. Wkm, 1919, 128, 2a,909 ; A., ii, 658.71 G. Kirsch, ibid., 1920, 129, 2a, 309; compare M.Ross, Chem. Zeit.,1912, 36, 686 ; A,, 1912, ii, 809RADIOACTIVITY. 243Parent of A ctznium-Details have been published of the separ-ation of the direct parent of actinium, proto-actinium or eka-tantalum, from pitchblende residues.72 By prolonged and repeatedtreatment with nitric acid, the other radioactive constituents,including radium, may be almost completely removed. A littletantalum oxide is then added, and the material extracted withhydrofluoric and sulphuric acids. The addition of a few milli-grams of thorium and lead nitrates a t this stage serves to keeptraces of ionium, uranium-S, and radio-lead in the insoluble form.The filtrate is evaporated to dryness, which leaves the tantalumand proto-actinium in an insoluble form, from which impurities,such as iron, zirconia, and the like, may be removed by boilingwith aqua regia.So far, all attempts to separate proto-actiniumfrom tantalum have failed.By various elaborations of this method, the whole of the proto-actinium from pitchblende containing a known amount of uraniumwas carefully separated, and its a-activity measured. If thebranching factor of the actinium series were 8 per aent., the pre-paration should have an a-activity equal to 4 per cent. of thatof the uranium in the mineral. As a mean of six determinations,i t was found that the branching factor was only 3 per cent. Thecompleteness of the separation and the avoidance of loss duringthe chemical treatment were proved by carefully chosen tests witha previously prepared and measured preparation.Period of A ctinium.-From these proto-actinium preparations, anew and independent determination of the life of actinium wasarrived a t in the following manner.From the known ranges ofthe six a-rays of the actinium series, including that of proto-act'inium, i t follows that the initial a-activity of proto-actiniummust be in the ratio 1 :5.74 to the a-activity of the substance afterequilibrium with the five a-ray-giving members of the actiniumseries has been attained. The proportionate increase of thea-activity over periods from 400 t o 600 days corresponds with ahalf-period for actinium of 20 years, or to a period of average lifeof 28.8 years. This is in good agreement with the further resultsobtained by observations on the decay of activity of actinium itself.so that the period of actinium may now be regarded as known withreasonable certainty.7372 Ann.Reports, 1918, 15, 195.'s 0. Hahn and L. Meitner, Ber., 1919, [B], 52, 1813; '4., ii, 147, 148 ;PhyeikaE. Zeitsch., 1919, 20, 529 ; d., 1919, ii. 209244 ANNUAL REPORTS ON THE PROURESS OF CHEMISTRY.Various.Radioactivity of Rubtdiun~.-An exainination of the very feeblypenetrating B-aotivity of rubidium compounds has confirmed theview that it is an atomic property of rubidium, and is unaffectedby chemical purification or treatment. The rays are somewhatinore penetrating than those of uranium-X1,[pA, =347 - 308(cm.)-],as compared with 463 for the B-rays of uranium-S, and 312 forthose of radium].Their velocity is estimated as 0.6 that of light.The activity is feeble. A surface covered with 0.025 gram ofrubidium sulphate per sq. cm. possesses the same activity as onecovered with’ 0.00033 gram of uranium oxide per sq. cm. (totalP-rays) . Eliminating the penetrating &rays of uranium-X2, andextrapolating to a film of zero thickness, so correcting for absorp-tion, i t is estimated that the speoific activity of the elementrubidium is one-fifteenth of that of uranium, due to the change ofuranium-X,. The product of the change, if the normal law isfollowed, should be an isotope of strontium. It is suggested thatthe search for calcium, strontium, and cesium, respectively, inminerals containing potassium, rubidium, and czsium, and thedetermination of their atomic weight, if foulid, might throwfurther light on the radioactivity of the alkali metals.i4Changes in the Radioactivity of the Oxides of r/mnium.-Sonieresults in this field incompatible with the present theory of radio-activity have been recorded.The a-activity of various prepar-ations of oxide of uranium showed a diminution over a term ofyears from 1 to 31 per cent. The greater decreases occurred withthe green oxide, prepared by the gentle ignition of ammoniumuranate, and the smaller with the black oxide, obtained by strongignition of the nitrate, preparations of intermediate colour show-ing intermediate behaviour. I n a preparation the a-activity ofwhich had fallen from an initial value 5.95 to 4.64, the initialactivity was restored by solution in nitric acid and ignition.Itis unfortunate, perhaps, that no changes of weight of the prepar-ations were looked for, for such results might be due to the possible,although hitherto unobserved, gain of oxygen or moisture by thefeebly ignited green oxides from the atmosphere. On the otherhand, from the impurities separated from the uranium, increasesin a-activity from 7 to 93 per cent. were observed, the rate ofincrease in one case corresponding with a period of 1.1 months75’* 0. Hdm and RI. Rothenbach, Physikal. Zeitsch , 1919, 20, 194 ; -4..1919, ii, 312. 75 C. Stsehling, Con~pt. rend., 1919, 169, 1036; A . , ii, 5R AD I0 A CTI VIT Y . 245Period of ZO?LZUTIL.--T~~ minute growth of radium froin largeamounts of carefully purified uranium, already recorded, has sinoeproceeded regularly according to the square of the time, and adefinite estimate of the period of ionium can now be deduced -fromthe measurements.This is the same as that already provisionallycalculated. Actually, the product of the periods of average lifeof ioiiiuni and radium alone is involved, and this, to an uhcertaintyof a t most 5 per cent., is 237,500,000 years. The period of ioniumis thus 100,000 years if that of radium is 2375 years. The actualgrowth of radium from 3 kilograms of uranium (element) in10 years has been 2 x 10-10 gram.76Fractional Crystallisation of Rczdium and ilfe.sothorium fromBarium.-The theory and practice of the enrichment of radiumand inesothorium from barium in the fractional crystallisation ofthe chloride and bromide has been the subject of two communica-tions. The enrichment factor, K , is defined as the ratio of theactive material in the crystals to that in the original material inthe solution.As regards the chloride, K varies from 1-65 for anacidity 0.25~1-, with 44 per cent. of the salt crystallising, to 1.49for an acidity 2X, with 58.3 per cent. crystallising. The condi-tion chosen for study was O*5riT-acidity, with 50 per cent. crystal-lising, for which K is 1.62. For the bromide, the value of R fellfrom 2.60 for O.25iT-acidity and 30 per cent. crystallising, to 2.45for 1.ON and 38.2 per cent. crystallising. The condition chosenfor further study was O.33N-acidity and 33.3 per cent.crystal-lising, for which K is 2-5. The enrichment factor is independentof the relative concentration of radium or radium and meso-thorium in the solution. From the second communication itappears that, so long as the same fraction of crystals separate, itis independent of acidity. As already known, the bromide offersadvantages over the chloride in speed of separation, especially inthe earlier stages of the separation. Some evidence was obtainedof t3he formation of a compound, RaBr2,2BaBr,,6H20, as the finalproduct of the fractionation in a weakly acid solution, correspond-ing with 39 per cent. of anhydrous radium bromide, which wouldexplain the advantage of the chloride over the bromide in thelater stages of the purification.On the other hand, in verystrongly acid solutions, above 2W, and very small concentration ofthe radium, below 10-7, the process is actually reversed, and moreof the radium remains in the mother liquor than separates outwith the crystals.The most favourable method of carrying on the fractionation76 Ann. Reports, 1916, 13, 249; F. Soddy, Phil. Mag., 1919, [vi], 38, 483;A., 1919, ii, 443246 ANNUAL REPORTS ON THE PROGRESS OB CHEMISTRY.in practice is by a system in which the crystals and mother liquorsmove one step in opposite direations in the series a t each crystal-lisation, except for the fractions enriched above the initial con-centration, the mother liquors from which move two steps. Insuch a system carried out continuously, representing the initialconcentration as unity, the series runs as shown, the figuresrepresenting the concentrations in each unit of the system: 77Mother liquor +-- + +- + +--- +--f------ +--+ 0.0016 0.008 0.04 0.20 1.0 2.3 5.0 12.3 27 +Crystals + ++ -++ -+- + -+ + +In another study of mesothorium-radium-barium bromides, theactivities were determined by y-ray methods.The value of K wasfound to vary from 2.57 with 24.3 per cent. crystallising to 1'44with 69 per cent. arystallising.78Solubility of Radium Emanation.-Two series of determinationsof the solubility of radium emanation in organic solvents, for themost part, have appeared. The solubility in these is much greaterin general than in aqueous liquids, and increases as the hydro-carbon character of the solvent predominates over the aqueous,rising steadily, for example, in a series of aliphatic alcohols oracids.79 A new determination of the coefficient of diffusion of theradium emanation in water a t 14O gave the value 0.82 cm.perday, which corresponds with a molecular diameter of 1.85 A.80The value deduced from the space-lattiae of crystals was abouttwo-thirds of this.y-A ctivity of Thorium-D.-The conclusion that in a thoriummineral 36.3 per cent. of the y-radiation is due to mesothorium-2and 62.7 per cent. to thorium-l)*l has been confirmed by a com-parison of the y-activities of quantities of the two products inequilibrium with the same quantity of thorium. The y-activitydue to thorium-l) was found t o be 1.81 times as great as that dueto mesothorium-2.Since only 35 per cent. of the thorium atomsdisintegrating produce thorium-D, it follows that, atom for atom,thorium-D gives 5.17 times as much y-radiation as mesothorium-2.From these data, a table of the changes of the y-activity of a puremesothorium preparation with time has been constructed. I f the7 7 C. E. Scholl, J . A ~ T . Chem. Soc., 1920. 42, 889; A., ii, 408.7 8 J. L. Nierman, J. Physical Chem., 1920. 24, 192 ; A., ii, 408.'4 Alfred Schultze, Physikot. Zeitach., 1020, 95, 257 : G.80 E. Ramstedt, Medd. R. Vetewkapsakad. Nobel Inat., 1919, 5, No. 6 ;A., ii, 577 ;Hofbauer, Sitzungsber. Akad. W~RU. Wkn, 1914, 123, 2a, 2001.A., ii, 72. 81 Ann. Repork, 1918,15, 220RADIOA(JTIVITY . 24 7initial activity is unity, in three years it is 1.62, and in ten yearsunity again.8Natural Radioactivity.Bmss.-The improvements in the methods of recording thepassage of individual a-particles have been applied to theexcessively feeble natural radioactivity of common materials, andhave thrown light on the important question whether this is whollydue to known radioactive impurities or is in any part a specificactivity.A statistical examination of the a-particles emitted froma hollow brass sphere showed that a large number of the a-particlespossessed a very short range, shorter than that of any of the knownradio-elements. The rate of emission was one a-particle per sq.cm. of surface in 9'25 hours. The range of this new type wasestimated a t 1.8 cm. of air.By the Geiger-Nuttall relation, thisoorresponds with a period of life 1-5 x 106 times that of uranium.Hence the inference is formed that the a-particles are derived froman actual disintegration of the metal, either copper or zinc, withthis excessively long period of 1016 years. From copper an isotopeof cobalt, and from zinc an isotope of nickel, would result in ana-ray change. It is sad, however, that such elaborate andimportant physical experiments should be conducted on such amaterial as-brass !Rocks.-A survey of the radioaotivity of the rocks encounteredin the piercing of the Loetschberg Tunnel, which runs fromKandersteg to Goppenstein, in the Bernese Oberland, showedunusual similarity of composition along the length of the tunnel.This agrees with the fact that no abnormal temperature gradient,such as was encountered in the St.Gothard Tunnel, was experi-enced. The average of all the rocks (eighty-two specimens) was2*2( x 10-12 grams of radium per gram). The rocks a t theKandersteg end are Jurassic limestones, in the centre Gasterngranite, and, a t the southern end, crystalline schists of all classes.The granites were somewhat lower in radium content, and theaalcareous and schistose rocks somewhat higher, than the averagefor these c1asses.aThe rocks of the Kolar gold field, on the Mysore plateau,southern India, consist of schists of very uniform character, whichcontain as little radium as, and are probably older than, any rocksknown. The temperature gradient in the mines is abnormally84 H.N. McCoy and G. H. Carfledge, J . Amer. Chern. Soc., 1919, 41, 42g8 J. H. J. Poole, Phil. Mag., 1920, [vi], 40, 466 ; A., ii, 667.A., 2919, ii, 89248 ANNUAT, REPORTS ON THE PROGRESS OF CHEMJSTRY.low. The average radiuiii content was 0.19 x 10-12 gralll pergram.84Spring TTTatc.rs.--A survey of sixty Canadian mineral springs,and, later, of six hot springs a t Banff, Alberta, disclosed onlymoderate activities. The latter are the most active in Canada, andpossess an emanation content of from 2 to 6( x 10-10 curie perlitre). For the escaping gases, higher values, up to 24, wereobtained .s5 The springs of Colorado are exceptionally active, theemanation mntent for ninety-five exceeding 10, and for five 100.The most active, 305, is surpassed by few European springs.86The sulphur-soda springs of Bagn5res-de-Luchon have beenfound to be the most radioactive in France, and, apart from suchwaters as actually originate in uranium mines, to be exceeded inactivity by less than a dozen in the whole world.The group ofsome twenty-four springs possessed an emanation content between4 and 415, five being above 240, and higher than any other Frenchsprings. 87The principal spring at Bagiioles de l’Orne showed, over a monthof observations, variations in emanation content from 2 to 15.Previous observations had given much higher values, 24 in 1907and 113 in 1904. These variations have been traced to the rain-fall. After rain, at an interval varying from six to thirteen days,the springs in this neighbourhood were found to reach a maximuniemanation content, the greater the heavier the rainfall. Thisshows that the activity of the spring is derived from surface watersperoolating through radioactive strata and inixing with the deepspring water .88 Two papers dealing with the practical techniqueof such measurements have appeared .@The OcPmz.--The rate a t which radium is supplied to the oceanby rivers and the denudation of the land cannot maintain thequantity present. It follows that there must be in sea-wateruranium and ionium in equilibrium proportion to the radium.Taking the mean radium content of sea-water as 1.2 x 10-15 gramper c.c., the uranium must be 4 x 10-6 gram per litre of sea-water,or 0.1 milligram per kilogram of sea-salt.This is about one-tenthof the estimated content of gold in sea-water. On this view, noa4 H. E. Watson and G. Pal, J . Ind. Tnst. Sci., 1914, 1, 39 ; A., ii, 278.85 J. Satterly and R. T. Elworthy, Trans. Roy. SOC. Canada, 1917-1918,[iii], 11, 17, 27 ; 12, 153 : A., 1919, ii, 41, 72, 312.8 6 0. C. Lester, Amer. J . Sci., 1918, [iv], 40, 621 ; A., 1919, ii, 6.8 i A. Lepape, Compt. rend., 1920, 171, 731 ; A., ii, 727.** P. Loisel, ibid., 1919, 169, 791 ; 1920, 171, 858; A., 1919, ii, 489 ;0. Niirnberger, Physikal. Zeitsch., 1920, 21, 198; A., ii, 345; H1920, ii, 727.Greinacher, ibid., 270 ; A . , ii, 463RADIOACTIVlTY. 249great variation of radiuni coiitent with depth is to be anticipatedin the ocean. It is calculated that the ooean and the land mustbe approximately equal factors in maintaining the amount ofenlanation in the atmosphere, the smaller contribution of the oceanper unit of surface being oounterbalanced by its greater area.90The A tmosphere.-At Innsbruck, as the result of forty-nineobservatioiis of the emanation content of the atmosphere by thecharcoal method, values were obtained between 1110 and 43, witha mean of 433( x 10-15 curie per litre), whioh are considerably abovethose found in other 10calities.~~It will be recalled that in 1912, during balloon ascents, anincrease in the penetrating radiation of the atmosphere wasrecorded, which became considerable a t the height of 3000 metres.92Now kite experiments, carried out a t the aeronautical observatoryof Lindenberg, Prussia, have shown that the active deposit on awire under the influence of the earth's field is much greater a theights between 1000 and 2500 metres than a t the surface. Astudy of the variations of this from July 29th to December 2nd,191 3 (ninety-eight observations), showed that strong increasesoccurred with the fall of the barometer. Since, a t this height, thesupposed explanation of the influence of the fall of the barometricpressure in facilitating the escape of emanation from the surfacesoil fails, the changes of activity and pressure are regarded asoriginating in a common cause. A clear parallelism was found t oexist, between the changes of the activity and what is termed the'' potential temperature " of the layer of the atmosphere between1000 and 2500 metres. By this term is meant the temperaturewhich the air would assume if adiabatically brought to normalpressure. Presumably the changes of this function are independentof internal ineteorological influences, and are a measure of theexternal solar influences. However that may be, temperaturechanges are supposed to be the cause of the pressure changes, andthemselves to originate from a inass radiation from certain limitedzones of the solar surface of eniaiiation particles into the upperatmosphere, which owes t o this its chief source of heat. What-ever the explanation may prove to be, the study of the radio-activity of the upper atmosphere is clearly likely to lead t oimportant advances.93 FREDERICK SODDT.IT. F. Hess, Sitztmgsber. ;1kad. Wiss. Wie)t, 1918, 127, ?a, 1297.91 R. Zlatorovic, Wien. Anz., 1920, 75 ; A., ii, 657.'' -4vn. Reports, 1912, 9, 327 ; 1913, 10, 288.y3 H. Bongards, Ph,ysikal. Zeitsch., 1920, 21. 141 ; A., ii, "7