NEWS zigzag line, which he subsequently elaborated and described in his Presidential Address to the Chemical Society, March 26th, 1902 (see Journ. Chem. Soc., Part 1, vol. lxxxi.) In his final arrangement he spaced the elements on curves which represent major and minor chord vibrations (stationary waves). The inert gases fell on or near to the nodal points. Tchitchèrin, in 1888, working on atomic volumes, proposed the formula At. vol.at. wt. X (2—0'00535 x at. wt. x N). J. Schmidt (Zeit. Anorg. Chem., 1902, xxxi., 146) cal thereto (except to hydrogen) chemically inactive portions a, culates the atomic weights of the elements by assigning and assumes that chemical attraction obeys the universal laws of attraction. b is the attractive portion. Monovalent atoms are represented by a+b, bivalent ones by a+b+b2, and so on. and the calculations are in fairly satisfactory accordance Certain assumptions are made, with weight determinations. A. Marshall points out some curious regularities that appear when the atomic weights are referred to other standards than those of O and H (Chem. Zeit., 1902, xxvi., 663, and CHEMICAL NEWS, 1902, lxxxvi., 88). For example, when those of Cl, Br, Ag, and I (German Committee weights taken at the time), are multiplied by 2.53868, whole numbers 90, 203, 274, and 322 are obtained. By dividing Li, NH4, Na, K, and Rb by 1'004, each division gives respectively 7, 18, 23, 39, and 85 exactly, except a small difference in the case of Na. The same HSn, Ca, and Al; the factor 1'0551 is given for the Zn-V factor gives practically whole numbers for Pb, Cd, V, Pt, series. The author connects these regularities with the phenomena of isomorphism. Vincent finds (Phil. Mag., 1902, iv., 103) "that if a list of all the atomic weights in ascending order of magnitude be taken, and the order in this list be called n, then the nth atomic weight, from n=3 to n = 60, is given by the equation W=(n+2) 121." Iodine is least in agreement by a difference of +3.9. Hughes considers the elements as built up from corpuscles analogous in type to the radicals NH4, CN, CH, and NH (CHEMICAL NEWS, 1903, lxxxviii., 298). Calculations are given. Wetherell takes 4 and multiples of 4, and classifies the elements according to the Periodic System (CHEMICAL NEWS, 1904, xc., 260; see also vol. xc., p. 271). He regards the abnormal properties of beryllium (low specific heat, &c.) as due to the presence of a satellite. In the case of Be, the satellite would be relatively large. Te, he suggests, may have a a relatively small satellite, N is an integer which for Li and Na=8, K=4, Rb-3, will make this clear : and account for the Te - I inversion in the Periodic Table. Minet (Comptes Rendus, 1907, cxliv., 428) assigns to the elements serial numbers, and, by adding six new elements to the series, plots the atomic weights by a curve, x=xI215. Two modifications of the same order are given. Delauney (Comptes Rendus, 1907, cxlv., 1279) finds curious whole-number relations, thus : 2 3'98 D. 0'02 3 7 8 II C = 19 2008 36 40 44 0'32 O'I 32'07 0'07 40'09 Ο ΙΟ 0'09 I'I 48.1 51 51.2 O'I 0'2 ΟΙ 54'93 More examples of this kind are given. From the foregoing it will be observed that the methods employed may be classed as-(1) those involving a mathematical treatment without regard to any particular chemical order or classification; (2) those in which the parallel arrangement or gradation of the elements is accomplished, although the mathematical treatment is somewhat imperfect; and finally (3) those involving the calculation of the weights, some useful classification being arrived at at the same time. While the best general chemical classification of the elements may not accord with their mathematical or geometrical harmonisation, the two ideas are not necessarily opposed. The difficulties are certainly great. But the failure to bring the two methods into perfect accord does not imply that a purely mathematical treatment is not possible. Moreover, a mathematical treatment, if it accords with facts, may lead to new discoveries, and bring to notice other properties than those usually associated with the classification of the elements. NEWS It is with the view of showing a remarkable series of regularities of an apparently exact mathematical nature that the plan below given is suggested. Curve Nos. I. 2. 3. 4. 5. 6. 7. 8. 9. Kr The CI This method is based upon two simple operations. first is to arrange the elements in a regular series, that passes through a zero, according to the empirical equation,(4P)+KW; P is a number of an integral series, o, 1, 2, 3, and K is a constant of somewhat arbitrary selection which, Be (?) Mg Na Fe Cu Br &c., V Zn Co Sr As Si Yt Ge Nb Ru II. 12. 13. 14. 15. 16. 17. 18. Hg Dy The second operation consists in projecting the dif ferences between the calculated and actual atomic weights horizontally along a line on which, by the above equation, the element is placed. Referring to the accompanying chart, it will be seen that the elements are first, as it were, hung on the rungs of a ladder, according to their nearness in weight to the values assigned to the rungs by the equation. The ladder passes through a zero, the rungs being four units apart. It should be here noted that, in order to avoid making the letterpress of the tabular part too wide, the names that fall on the same line are shifted to the nearest blank space, and dotted arrows are introduced to connect them with the lines on which they belong. The final allocation of the elements in the "difference" portion of the chart reveals some chemical regularities, although the true significance of the classification in this respect is not attempted as yet. The alkali metals, except rubidium, fall in the central part of the table, and if the table is folded down the centre on the O-line, so as to superpose the and sides, many like elements are brought into approximate line. It is not my present object to attempt any chemical classification, and I do not believe that the system in its present form lends itself to a periodic arrangement worth considering. Gallium, lutecium, and neoytterbium do not fit into the series as they give differences greater than o'g. Moreover, their weights are not known accurately to the first decimal place. This may be said of a number of the heavy elements which are included. The chart may, of course, be extended to take radium and uranium. It is of importance to note that only those elements having certain numerical values fit into the series, and about the same number fall on the positive as on the negative side. The true zero of the system may be taken on the hydrogen line or a fraction of a unit below it. Those elements that fall on the oʻg line (−) fit also on the o'g line (+), and this suggests that such elements should easily change the sign of their valence. Carbon is an example in this respect. The most interesting part of the table is the elliptical curves which are drawn mechanically, and are symmetrically disposed with reference to the hydrogen line. They are all formed round the common foci (++), and spring from the 4-unit lines (where the latter intersect the +0.8 vertical line) on which the elements fall. The curves occur in groups of 9, and intersect practically all the elements. There are two such groups clearly defined. The third is doubtful, as the elements belonging to it would be of very large magnitude, and no accurate determinations are available for the purpose of establishing the curves. The regular manner in which the double curves intersect the elements will be seen from the table (see next column). In preparing this table, those elements not practically on the curves are moved to the nearest curve. The transpositions are very slight, considering the number of elements involved, and the fact that a number are not accurately known to the first decimal place. To make the Te Rh Ti Cr 2500 table even more symmetrical, I should be moved to the curve No. 14 and Er to the vacant space on curve No. 16. The table predicts an unknown inert gas, designated by Y, which has been looked for. The altered atomic weights of some of the inert gases are those kindly given to me by Sir William Ramsay, these being derived from new density determinations. Otherwise the table (chart) is in accord with the International Weights for this year. The prediction of Y has been arrived at by three more or less independent ways. In the first place, it will be noticed that a circle intersects He, A, and Ne exactly, and also Kr by a correction in the second decimal place. These gases, except He, are also intersected by the elliptical curves, and Xe is only a little out. Their position on the curves were noted to be thus:-A first, Kr second, Ne first, Xe second; therefore Y should be first, i.e., not preceded by an element, to preserve the regularity. Referring to the table above, it will be seen that Y should be on the curve with Sa, which is the second element. Assuming that Kr, Xe, and Y might also be intersected by a circle of the same radius as the one that intersects the others, by a trial method I found the position of Y, which coincides with the elliptical curve where it intersects a 4-unit line. The correction of the weights of the inert gases by the "three " intersections gives the following: Nitrogen and beryllium are the only elements which do The difference between the not fit into the system. observed weight of N and the calculated value for the Be falls away from any series is of the order of II. Wetherell (see citation above) has sugprobable curve. gested that Be and Te carry satellites, which he suggests and the wrong position of Te with respect to I in the would account for the abnormal specific heat of the former periodic classification. I suggest that Be, N, and I carry a satellite each, which by trial I find to be about o'2684. Subtracting this figure from N (14.0064) gives 13.7380, which value falls on the first circle. Similarly, assuming Be to be 9.0874 instead of 9.1, and subtracting o.2684, gives 8.8190, which places Be, minus its satellite, on the first * Ramsay (Roy. Soc. Proc., Ser. A, Aug. 27, 1908, lxxxi., pp. 178 180) says the Periodic Table shows gaps for two or three inert gases of higher atomic weight than Xe, and an exhaustive search has failed to indicate any new gas, which he says may be due to its instability, like the emanations of radium, thorium, and actinium. Their weights should be about 172, 216, and 260. In the same Proceedings, pp. 195209, Moore reports that he has found no traces of a new constituent of the atmosphere upon careful search, and says that if a new element exists, it must, in all probability, be in less quantities than I in 2,560,000,000 parts of air. NEWS Sa curve, and brings the number of elements on that curve It will be seen that there is a gap in the ninth curve, and it may he wrong to assume that the first and ninth curves should carry seven elements, since, by shifting Al to curve No. 2, and assuming that the ninth curve carries six elements, the elements on the first group of curves will be symmetrically disposed. W Pr Nd. Gd. Tu +0'11 +0.07 Ta Os Difference. 0:00 + 0'00 While I have gone into the third and fourth decimal place in some of the calculations, this was not done with a view of assuming such a high degree of accuracy as the figures might imply. Indeed, it is probable that the calculations are only approximate at best, and that some of the curves will require shifting, especially since the weights of the elements are not accurate enough to establish the system on an absolute basis. However, the agreements are very close, as will be seen from the following table : Th. Be ::: H.. - 0'035 +0.03 (a) It is impossible to determine the value of this element in the second and third decimal places by the curve drawn. The agreements between the observed and corrected values are sufficiently close to make the probability of coincidences impossible, when excluding those elements of high atomic weight which are not known with any degree of accuracy. Due allowance should be made for those elements that fall by chance on a curve when it is known that their values are uncertain. In general, the above table may be regarded as good up to cerium, although Dy, Au, and a few others of high magnitude are fairly accurately determined. It should be borne in mind that a change in the fundamental standards (hydrogen, chlorine, nitrogen, and silver) would involve a re-calculation of many elements, and necessitate a slight general readjustment of the curves. The exact position of the foci and also of the starting-points for the curves-i.e., whether on the 4-unit lines where they intersect the +0.8 line, or at some other point of intersection-can hardly be fixed until all the standards are established beyond doubt. The accuracy with which the curves are drawn may be judged by their positions with reference to the machineruled squares, the chart here shown being a reduction from a drawing 56 cm. x 19 cm., the smallest squares being I mm., the next 5, and the largest 10 mm. 7, Doughty Street, London, W.C., February, 1909. Compounds of Gold with Bromine.-Fernand Meyer. -The final product of the action of bromine on gold is pure AuBr3, easily obtained by shaking the reagents together for long periods, or by heating to 100° and then cooling and repeating this operation many times. AuBr3 is a black crystalline substance soluble in bromine and volatile in an atmosphere of bromine at 300°. There appears to be no reason for believing in the existence of the compound AuBr2. At temperatures at which auric bromide dissociates, two compounds, and only two, exist, namely, AuBr3 itself and AuBr.-Comptes Rendus, cxlviii., No. 6. |