These experiments apparently suggested a small decrease of weight as a consequence of chemical processes. On repeating them, however, and making allowance for the different amounts of water absorbed on the surface of the vessel at the beginning and end of the experiment, Landolt found in 1908 (Zeit. physik. Chem. 64, p. 581) that the variations in weight are equally positive and negative, and he concluded that there was no change in weight, at least to the extent of 1 part in 10,000,000.
There is still another question regarding the numerical values of the atomic weights, namely: Are there relations between the numbers belonging to the several elements? Richter had arranged his combining The periodic arrangement. weights according to their magnitude, and endeavoured to prove that they form a certain mathematical series. He also explained the incompleteness of his series by assuming that certain acids or bases requisite to the filling up of the gaps in the series, were not yet known. He even had the satisfaction that in his time a new base was discovered, which fitted rather well into one of his gaps; but when it turned out afterwards that this new base was only calcium phosphate, this way of reasoning fell into discredit and was resumed only at a much later date.
To obtain a correct table of atomic weights the second question already mentioned, viz. how to select the correct value in the case of multiple proportions, had to be answered. Berzelius was constantly on the look-out for means to distinguish the true atomic weights from their multiples or sub-multiples, but he could not find an unmistakable test. The whole question fell into a terrible disorder, until in the middle of the 19th century S. Cannizzaro showed that by taking together all partial evidences one could get a system of atomic weights consistent in itself and fitting the exigencies of chemical systematics. Then a startling discovery was made by the same method which Richter had tried in vain, by arranging all atomic weights in one series according to their numerical values.
The Periodic Law.—The history of this discovery is rather long. As early as 1817 J.W. Döbereiner of Jena drew attention to the fact that the combining weight of strontium lies midway between those of calcium and barium, and some years later he showed that such “triads” occurred in other cases too. L. Gmelin tried to apply this idea to all elements, but he realized that in many cases more than three elements had to be grouped together. While Ernst Lenssen applied the idea of triads to the whole table of chemical elements, but without any important result, the other idea of grouping more than three elements into series according to their combining weights proved more successful. It was the concept of homologous series just developed in organic chemistry which influenced such considerations. First Max von Pettenkofer in 1850 and then J.B.A. Dumas in 1851 undertook to show that such a series of similar elements could be formed, having nearly constant differences between their combining weights. It is true that this idea in all its simplicity did not hold good extensively enough; so J.P. Cooke and Dumas tried more complicated types of numerical series, but only with a temporary success.
The idea of arranging all elements in a single series in the order of the magnitude of their combining weights, the germ of which is to be found already in J.B. Richter’s work, appears first in 1860 in some tables published by Lothar Meyer for his lectures. Independently, A.E.B. de Chancourtois in 1862, J.A.R. Newlands in 1863, and D.I. Mendeléeff in 1869, developed the same idea with the same result, namely, that it is possible to divide this series of all the elements into a certain number of very similar parts. In their papers, which appeared in the same year, 1869, Lothar Meyer and Mendeléeff gave to all these trials the shape now generally adopted. They succeeded in proving beyond all doubt that this series was of a periodic character, and could be cut into shorter pieces of similar construction. Here again gaps were present to be filled up by elements to be discovered, and Mendeléeff, who did this, predicted from the general regularity of his table the properties of such unknown elements. In this case fate was more kind than with Richter, and science had the satisfaction of seeing these predictions turn out to be true.
The following table contains this periodic arrangement of the elements according to their atomic weight. By cutting the whole series into pieces of eight elements (or more in several cases) and arranging these one below another in the alternating way shown in the table, one finds similar elements placed in vertical series whose properties change gradually and with some regularity according to their place in the table. Not only the properties of the uncombined elements obey this rule, but also almost all properties of similar compounds of the elements.
| He 4.0 | Li 7.03 | Be 9.1 | B 11.0 | C 12.00 | N 14.01 | O 16.00 | F 19.0 | .. | .. | .. |
| Ne 20 | Na 23.00 | Mg 24.32 | Al 27.1 | Si 28.4 | P 31.0 | S 32.06 | Cl 35.45 | .. | .. | .. |
| Ar 39.9 | K 39.15 | Ca 40.1 | Sc 44.1 | Ti 48.1 | V 51.2 | Cr 52.0 | Mn 55.0 | Fe 55.9, | Ni 58.7, | Co 59.0 |
| .. | Cu 63.6 | In 65.4 | Ga 70 | Ge 72.5 | As 75.0 | Se 79.2 | Br 79.96 | .. | .. | .. |
| Kr 83.0 | Rb 85.5 | Sr 87.6 | Y 89.0 | Zr 90.6 | Cb(Nb) 94 | Mo 96.0 | .. | Ru 101.7, | Rh 103.0, | Pd 106.5 |
| .. | Ag 107.93 | Cd 112.4 | In 115 | Sn 119.0 | Sb 120.2 | Te 127.6 | I 126.97 | .. | .. | .. |
| Xe 130.7 | Cs 132.9 | Ba 137.4 | La 138.9 | Ce &c. 140 | Ta 181 | W 184 | .. | Os 191, | Ir 193.0, | Pt 194.8 |
| .. | Au 197.2 | Hg 200.0 | Tl 204.1 | Pb 206.9 | Bi 208.0 | .. | .. | .. | .. | .. |
| .. | .. | Ra 225 | .. | Th 232.5 | .. | U 238.5 | .. | .. | .. | .. |
But upon closer investigation it must be confessed that these regularities can be called only rules, and not laws. In the first line one would expect that the steps in the values of the atomic weights should be regular, but it is not so. There are even cases when it is necessary to invert the order of the atomic weights to satisfy the chemical necessities. Thus argon has a larger number than potassium, but must precede it to fit into its proper place. The same is true of tellurium and iodine. It looks as if the real elements were scattered somewhat haphazard on a regular table, or as if some independent factor were active to disturb an existing regularity. It may be that the new facts mentioned above will lead also to an explanation of these irregularities; at present we must recognize them and not try to explain them away. Such considerations have to be kept in mind especially in regard to the very numerous attempts to express the series of combining weights in a mathematical form. In several cases rather surprising agreements were found, but never without exception. It looks as if some very important factor regulating the whole matter is still unknown, and before this has been elucidated no satisfactory treatment of the matter is possible. It seems therefore premature to enter into the details of these speculations.
In recent times not only our belief in the absolute exactness of the law of the conservation of weight has been shaken, but also our belief in the law of the conservation of the elements. The wonderful substance radium, whose Transmutation of elements. existence has made us to revise quite a number of old and established views, seems to be a fulfilment of the old problem of the alchemists. It is true that by its help lead is not changed into gold, but radium not only changes itself into another element, helium (Ramsay), but seems also to cause other elements to change. Work in this line is of present day origin only and we do not know what new laws will be found to regulate these most unexpected reactions (see [Radioactivity]). But we realize once more that no law can be regarded as free from criticism and limitation; in the whole realm of exact sciences there is no such thing as the Absolute.