arbitrary). Certain particles separated out which appeared to be the denser ones, and by the introduction of moisture a remarkably homogeneous deposit was obtained on the wall of the casing. If the vapours of those elements that are monatomic could be subjected to a prolonged treatment in the above manner a partial if not complete separation of the isotopes might be possible, but instead of using moisture to collect the particles they could be condensed on the casing by cooling it. Lead from different radio-active sources has different atomic masses in accordance with the principle of isotopy revealed amongst many radio-atoms in process of change. This need not be adduced as an argument that the elements below lead in atomic mass cannot have isotopic atoms. The process of evolution followed by one of devolution seems to indicate that in the genesis of the elements, according to the well-known pendulum illustration of Crookes, the swing of the pendulum laid down the stable elements up to and possibly including some lead; but the pendulum evidently swung so far as to finally lay down two elements, thorium and uranium, that are too unstable to exist indefinitely; therefore in their disintegration through a long series of changes to lead they represent the final recession of the pendulum, and the end-products of devolution have different masses amongst themselves, and these differ from that of lead produced in the first instance by evolution. Lead, therefore, has atomic masses differing slightly according to the source from which obtained or its isotopic composition; but in the process of evolution from hydrogen to lead there is no reason why two kinds of atoms chemically alike yet differing in mass may not have been produced, though apparently there has been no vast time of separation or other condition whereby such isotopes could have appeared separately in different minerals. In conclusion, the alteration in the conductivity of the elements by the presence of isotopes of differing mass in different proportions would lead to the possibility of converting metals into insulators as one extreme and improving their conductivity as the other. This hardly seems possible from a common sense point of view, though the suggested relation leads to such a conclusion. One is therefore forced to seriously consider whether an equalisation of mass has actually taken place, and whether the variation in conductivity is due to an energy condition which cannot be altered back so as to restore the original masses to the atoms, and that concidences in numerical relations as indicated can never have more than a hypothetical interest at best. At the same time, fractional atomic weights are so simply accounted for by the supposition of whole-number isotopes that one is tempted to indulge in speculations with a view of finding some new road for experimental investigation. Frankland concludes that "the uniformity with whi trans-addition and trans-elimination appear to be the m favoured processes suggests that some common cause is at the root of both." An explanation of the mechanism of trans- and cisreactions can be given in terms of Bohr's theory of the arrangement of the atoms and electrons in the molecule. On Bohr's conception the hydrogen molecule consists of two positively charged hydrogen nuclei with two electrons rotating around the line joining the centres of the nuclei (Lewis, 1919, vol. iii., p. 194). Thus for the carbon linking we can assume the presence of electrons rotating around the lines joining the centres of the carbon nuclei. Diagrams I., II., and III. show a possible arrangement of the electrons for a simple, double, and triple carbon linkings respectively, and afford an explanation of the difficulty of rotation of the two carbon atoms in the case of double and triple linkings. THE NATURE OF THE ETHYLENIC AND ACETYLENIC LINKINGS IN CARBON COMPOUNDS. By W. E. GARNER. FRANKLAND, in his Presidential Address to the Chemical Society (Trans., 1912, ci., 673), has drawn attention to the peculiar nature of the unsaturated linkings in carbon compounds. The appearance or disappearance of the unsaturated linking is accompanied by either trans elimination or trans-addition of the groups leaving or entering the molecule. Michael (Journ. Prakt. Chem., 1895, [ii.], lii., 352) states thas he knows of no exception to the rule of trans-addition at the double bond of halogen acids. Even where the cis-reactions occur they take place much more slowly than the corresponding trans-reactions. The following instances of trans-addition and elimination may be given : The elimination of hydrobromic acid from a bromo acid is illustrated by means of the Diagrams (a), (b), and (c). On the extrusion of the bromine atom from (a) we may assume that it is accompanied by one valency electron and that the other electron remains attached to the upper carbon atom unpaired. This electron can pair with one of the valency electrons of the carbon-carbon linking, giving rise to the arrangement shown in (b). Diagram (b) shows two hydrogen atoms attached to the lower carbon atom, and either of these could be eliminated at the same time as the bromine atom. Also it is seen that an unpaired by a-rays from polonium deposited on copper-foil, but failed to make any comparison of the chemical and electrical relationships (Phys. Zeit., 1910, xi., 273). In a brief notice in the fifth edition of his "Lehrbuch der Elektrochemie" (p. 317) Le Blanc called attention to this omission, and stated that the a-ray effect closely approached the requirements of Faraday's law. Later, Le Blanc (Zeit. Phys. Chem., 1913, lxxxv., 511) published in full his calculations from Bergwitz's data on which his statement was based. electron is attached to one of the valencies of the upper, Bergwitz gave the results of the decomposition of water carbon atom. The probability of this electron pairing with an electron from H(I) is greater than with an electron from H(II). The extrusion of H(I) is accompanied by the formation of a double linking, and trans-elimination of HBr takes place. cis-Elimination could, however, take place, but for this to occur the lower carbon atom must rotate so as to bring H(II) into the right position. Thus H(II) would not be extruded so often as H(I). The final state is given by (c). The intermediate stage (b) is only given for the sake of clearness and is not necessarily an actual step in the elimination of HBr from a bromo acid, for the hydrogen and bromine atoms may be liberated simultaneously. In any case it is evident that the rate of transelimination H(I)Br is greater than that of cis-elimination H(II)Br. The addition of hydrobromic acid, bromine, hydrogen, &c., at the ethylenic or acetylenic linking may be treated in the same manner, and it will be seen that the reaction will be more likely to take place in the trans-position than in the cis-position, on account of the greater rapidity of the former process. Bohr's theory thus supports the van't Hoff conception of directed valency, while it explains the apparent mobility of the carbon valencies. The racemisation of optically active compounds and that particular case of racemisation, viz., the Walden inversion, is thus due to rearrangements of the valency electrons and not necessarily to any change in the directions of the carbon valencies. The number of valency electrons at each linking may be more than two, and their arrangement in the case of the double and triple linkings may be different from those shown in the diagrams, but it seems certain that with any possible arrangement the probability of the formation of trans-compounds will be greater than that of the forma tion of cis-compounds. The Chemical Laboratories, In 1911 the writer determined chemically the amount of ozone formed in oxygen by a-rays, and calculated that one molecule of ozone was formed per two pairs of gaseous ions (Sitzb. Akad. Wiss., Wien., 1911, cxx., 1709; Monatshefte, 1911, xxxii., 295; Am. Chem. Fourn., 1911, xlvii., 397; Le Radium, 1911, ix., 104). This was the first comparison between ionisation and chemical action in a gaseous system, and hence the first instance where ionisation and chemical reaction referred to the same medium in each case. A little later Krueger and Moeller (Phys. Zeit., 1912, xiii., 729) devised an ultra violet absorption method for the determination of ozone in very minute quantities, which Krueger employed to determine the amount of ozone formed by the passage of electrons of high velocity through gaseous oxygen ("Nernst Festschrift," p. 240; Phys. Zeit., 1912, xiii., 1040). Krueger arrived at a conclusion similar to that of the writer, that one pair of ions is involved in the formation of each molecule of ozone. In 1912 the writer (Journ. Phys. Chem., 1912, xvi., 564) collected all the available data on the chemical action of a-particles, and drew from his calculations based on them the general conclusion that ionisation by aparticles and the resulting chemical action are always of the same statistical order (ion for molecule), and may be treated as illustrative of a modified form of Faraday's law. The results and applications of this theory were discussed in a number of papers (Am. Electrochem. Soc., 1913, xxiv., 339; Le Radium, 1914, xi., 108; Zeit. Phys. The experimental data Chem., 1913, lxxxiv., 759). employed by the writer in the comparison of ionisation and chemical action were largely those of Cameron and Ramsay obtained by mixing radium emanation with various gases and following the course of the reaction manometrically (Fourn. Chem. Soc., 1908, xciii., 966). In order to calculate the ionisation from their results a method of calculating the average path of all the aparticles in a given volume was devised (S. C. Lind, loc., cit.) and applied to the data of Cameron and Ramsay, with the result stated in the preceding paragraph. CHEMICAL ACTION PRODUCED BY RADIUM | water, measured by Duane and Scheuer employing EMANATION. I.* THE COMBINATION OF HYDROGEN AND Oxygen. 1. Introduction. IN 1903 Ramsay and Soddy reported quantitative data on the decomposition of water in a solution of radium salt (Proc. Roy. Soc., 1903, lxxii., 204). In 1907 Bragg used their data to make the first comparison between the chemical and ionising effects of a-particles (Phil. Mag., 1907, [6], xiii., 356). Although Bragg calculated that the number of molecules of water decomposed was almost exactly equal to the number of ions that would have been produced in air by the emanation employed, he was apparently not impressed by the equality be found, and regerred to it as a "curious parallelism in numbers." In 1910 Published with permission of the Director of the U.S. Bureau of Mines. From the Journal of the American Chemical Society, xli., No. 4. More recently several chemical reactions under the influence of a-particles have been very carefully studied in the laboratory of Mdme. Curie. The decomposition of emanation in an a-ray capillary tube, showed a close equivalence between ionisation and chemical effect (Le Radium, 1913, X., 33). The combination of hydrogen and oxygen was studied by Scheuer in a mixture of emanation and electrolytic gases in glass spheres, by determining the diminution in pressure in a single measurement after decay of most of the emanation (Comptes Rendus, 1914, clix., 423). Scheuer calculated the ionisation from the Duane and Laborde empirical formula (Comptes Rendus, 1910, cl., 1421), and found, with a good agreement among all his experiments, that about 5.5 molecules of gas recombined for each pair of ions. A small proportion of the molecules recombined to form H2O2, according to Scheuer's analysis; but, approximately, we may express his result as 36 molecules of water formed for one pair of ions. This is a much higher value than 10, the average earlier calculated by the writer from the data of Cameron and Ramsay. The decomposition of hydrogen sulphide was also measured in the Curie laboratory by Wourtzel (Comptes Rendus, 1913, clvii., 929) who found 3'3 molecules decom posed per pair of ions (in air), and calculated the decomposition to be 4'7 times greater than Duane and Scheuer found for water. Wourtzel later reported briefly the results of other reactions, namely, decomposition of ammonia, of nitrous oxide, and of carbon dioxide (Fourn. Russ. Phys. Chem. Soc. Proc.,1915, xlvii., 210, 493, 494). In all these reactions Wourtzel finds the amount of reaction to be in excess of the ionisation. A. Debierne (Ann. Phys., 1914, [9], ii., 97) has been led by the statistical disagreement found by Scheuer and by Wourtzel between ionisation and chemical action to reject the theory of ionisation put forward by the writer as the primary cause, and to substitute one based on the hypothesis that the passage of an a particle through a gas may thermally decompose molecules lying outside the path of its ionising effect. This view of thermal decomposition is not favoured by Wourtzel (loc. cit.) because in some cases he found reactions actually having negative temperature coefficients. The difference between the conclusions drawn by the writer in regard to the rôle played by ionisation and those of Debierne, Scheuer, and Wourtzel, based on the Paris measurements, demonstrate the desirability of further experimental work to determine, first, whether the discrepancy lies in the data themselves, or in their treatment; and, second, whether the higher chemical values, if correct, are too great to be brought into accord with ionisation. In order to settle the first point it appeared advisable to make an exhaustive experimental study of the simplest possible case, such as the combination of hydrogen and oxygen gases, in order to establish thoroughly the laws governing the reaction under various conditions as regards volume of the reacting vessel, pressure of the gases, concentration of emanation, temperature, and variation of the proportions of hydrogen and oxygen. The results of this study are reported in the present paper. In the main the experimental method of Cameron and Ramsay has been used and found well suited for the purpose. The writer was able to profit by the experience of Cameron and Ramsay to improve the manipulative details somewhat, to which attention will be called later. Following the course of the reaction manometrically enables one to study the kinetics of the reaction thoroughly. The reaction between hydrogen and oxygen has been chosen because the products of reaction are continually removed, and the system maintains itself in a constant condition with respect to the composition of the gases being acted upon. The kinetic equation earlier deduced by the writer (loc. cit.) for the data of Cameron and Ramsay has been confirmed over a much wider range than was formerly possible. By varying the size of the spherical reaction bulbs, experimental confirmation has also been obtained of the law of the average path of a-particles as applied to their chemical effect in such vessels. Briefly expressed, all the assumptions previously made by the writer in treating the Cameron and Ramsay data have now been verified by direct experiment, and show that the treatment was in every way justified. The disagreement, however, between the data of Cameron and Ramsay and of Scheuer has been found to be real, and must be decided in favour of Scheuer through a good agreement between his and the new results. The explanation of the discrepancy lies in the quantities of emanation reported by Cameron and Ramsay, which were not measured in loco, but calculated from the amount of radium employed and the time of accumulation, which apparently led to a considerable error through incomplete evolution or collection of the radium emanation. 2. Source of the Radium Emanation. The radium employed as a source of emanation was part of that produced in the co-operative work of the U.S. Bureau of Mines (Bull., 104, 1915, by C. L. Parsons, R. B. Moore, S. C. Lind, and O. C. Schaefer) and the National Radium Institute. 297 8 mgrms. of radium element in the form of bromide, protected by a nine-fold excess of the setting of the mercury at e. By means of a capillary tube, b, connection was made at a with the Duane apparatus. After thorough exhaustion through a, b, and c to e by means of the Gaede mercury pump the purified emanation was introduced into c and sealed off at b. Electrolytically prepared hydrogen and oxygen were collected over mercury in a Ramsay gas pipette (Proc. Roy. Soc., 1905, A, lxxvi., 113; Trans., 1907, xci., 939) sealed on at j, from which the gas could be passed through g and later through f to c, where it mixed with the emanation. This arrangement avoids bringing the emanation into contact with any stopcock grease, which caused Cameron and Ramsay much trouble in their early experiments through the continued generation around the stopcock f of foreign gases which would rise into c and vitiate the manometric results. The gas mixture can be collected below f, either before or after the collection of emanation in c. Only the latter procedure, however, permits of the measurement of the initial pressure (usually negligible) of gas collected with the emanation. The mercury levelling bulb i is connected by means of rubber tubing to the glass apparatus which is provided with a gas trap at h. The reaction can be begun immediately after collecting emanation, or after equilibrium with induced activity. The stopcock f was open only while taking readings. A water-jacket (not shown) was brought about the bulb c to prevent temperature fluctuations while reading. cury level between e and i was determined by mounting the whole apparatus close before a vertical mirrored millimetre glass scale. The difference thus determined was added (algebraically) to barometric pressure after all necessary corrections had been made to reduce the results to standard conditions. (To be continued). RARE EARTHS. II.* By L. M DENNIS and P. A. van der MEULEN. (Concluded from p. 6.) When a drying agent was desirable a mixture of equal THE ELECTROLYSIS OF SOLUTIONS OF THE weights of fused sodium and potassium oxides was introduced into c before assembling the apparatus and fused to the wall with a weak flame. Owing to the low melting. point of the mixture it could be melted in a very thin layer at two or three spots on the inner wall without deformation of the bulb or materially affecting the spherical volume. The volume was determined by calibration with mercury after the mixture was in place. The quantity of emanation actually employed in each experiment was determined by the y-ray method of measurement at any time (usually the next day) after the introduction of the emanation. For this comparison three of the Bureau of Mines standards were employed containing 10:56, 59-26, and 157.3 mgrms. of radiuin element respectively, all of which had been compared with the III. Electrolysis of a Neutral Solution of the Nitrates of the Rare Earths, Using a Diaphragm. THE oxides of the rare earths (that were obtained in the twenty fractions of the electrolysis of the chlorides were united and were boiled with an amount of nitric acid that was not quite sufficient to dissolve them completely. The excess of oxides was removed by filtration, and the neutral nitrate solution, which contained about 700 grms. of the oxides, was diluted to 10 litres. This volume of th solu 130 126 Nd 2 3 4 5 U.S. Bureau of Standards international standard. (The writer wishes to take this opportunity of expressing his indebtedness to Dr. N. E. Dorsey for making these comparisons). This measurement was greatly facilitated by the fact that the whole apparatus was mounted on a single iron stand, and could be readily transported and the bulb brought into any desired position with reference to the electroscope. A correction of o'8 per cent is applied to the 7-ray indication on account of the lag of radium C. (E. E. Rutherford, "Radio-active Substances and their Radiations," Sec. 197, also p. 659). The course of the reaction was followed by determining the pressure at suitable intervals. The difference in mer 7 8 9 10 tion was maintained throughout the electrolysis. A sheet of platinum 5 cm. square was used as anode. In other respects the cell was exactly the same as that used in the electrolysis of the chlorides. The nitric acid that accumulated in the anode cup was removed by allowing distilled water to drop into the cup from a separatory funnel, and keeping the height of liquid in the cup constant by means of a syphon. A current of 5 amperes was used. The voltage varied from 7.5 to 18. The conduct of the electrolysis, the treatment of the fractions, and the deter *From the Journal of the American Chemical Society, xxxvii. No. 9. mination of the atomic weights of the fractions were performed in exactly the same manner as in the fractionation of the chlorides. The fractional electrolysis of this solution of the nitrates, using a diaphragm, was found to precipitate the hydroxides of the earths about four times as rapidly as in the chloride electrolysis. It was possible that the hydrogen set free at the cathode might reduce some of the nitric acid to ammonia, as was found to be the case when no diaphragm was employed. Careful qualitative tests of the solution after the ninth fraction had been removed failed, however, to show the presence of a trace of ammonia. The results of this fractional electrolysis of the nitrates are shown in Fig. 5 and Table V. It will be noted that in this series of electrolyses the coloured earths erbium, holmium, and thulium were present in preponderating amount in the first four fractions and were probably accompanied by considerable amounts of yttrium, as is indicated by the relatively low average atomic weight of the earths in these fractions. The last fractions, however, consist chiefly of yttrium, mixed with small amounts of cerium and neodymium that also were concentrated at this point. The method yields, therefore, in the first fractions a rapid concentration of erbium that is nearly entirely free from the didymium group; and, in the later fractions, yttrium that is quite free from the coloured earths of the erbium group, but is accompanied by any of the didymium group that may have been present in the original solution. It is interesting to note that the later fractions contained detectable amounts of cerium, although qualitative tests of the original solution failed there to disclose the presence of this element. To obtain further indication as to the efficiency of the method in the separation of the earths of the erbium and yttrium groups, the first four fractions given in Table V. were combined, and were converted into a neutral solution of the nitrates. The weight of the oxides in the solution was about 220 grms., the volume of the solution was 2300 cc., and the average atomic weight of the earths was 116 54. A cell measuring 21 cc. in diameter and 20 cm. in height was used for the electrolysis. In other respects the same apparatus was employed as before. Six fractions were obtained by electrolytic precipitation. The liquid remaining after the last fraction had been removed possessed high resistance and gave no precipitate with oxalic acid. The results of this fractional electrolysis are |