1917] on The Complexity of the Chemical Elements 129 



cision not hitherto possible. Fortunately I was not deterred by 

 the non possumus, for it looks as if everybody was right ! An 

 explanation of this paradox will later be attempted. In point of 

 fact, there are exceedingly few thorium minerals that do not contain 

 uranium, and since the rate of change of uranium is about 2 ' 6 times 

 that of thorium, one part of uranium is equal as a lead-producer to 

 2 • 6 parts of thorium. Thus Ceylon thorianite, one of the richest 

 of thorium minerals, containing 60 to 70 per cent of ThO.,, may 

 contain 10 to 20 and even 30 per cent of UgO^, and the lead from 

 it may be expected to consist of very similar quantities of the two 

 isotopes, to be in fact very similar to ordinary lead. I know of only 

 one mineral which is suitable for this test. It was discovered at the 

 same time as thorianite, and from the same locality — ^Ceylon thorite, 

 a hydrated silicate containing some 57 per cent of thorium and 1 per 

 cent of uranium only. In the original analysis no lead was recorded, 

 but I found it contained • 4 per cent, which, if it were derived from 

 uranium only, would indicate a very hoary ancestry, comparable, 

 indeed, with the period of average life of uranium itself. On the 

 other hand, if (1) all the lead is of radioactive origin, (2) is stable, 

 and (8) is derived from both constituents, as the generalisation being 

 discussed indicated, this O'-l per cent of lead should consist 95*5 per 

 cent of the thorium isotope and 4 • 5 per cent of the uranium isotope. 

 Thorite thus offered an extremely favourable case for examination. 



In preliminary experiments in conjunction with H. Hyman, in 

 which only a gram or less of the lead was available, the atomic weight 

 was found relatively to ordinary lead to be perceptibly higher, and 

 the difference, rather less than one-half per cent, was of the expected 

 order. 



I was so fortunate as to secure a lot of 30 kilos, of this unique 

 mineral, which was first carefully sorted, piece by piece, from admixed 

 thorianite and doubtful specimens. From the 20 kilos, of first grade 

 thorite, the lead was separated, purified, reduced to metal, and cast 

 in vacuo into a cylinder, and its density determined together 

 with that of a cylinder of common lead similarly purified and 

 prepared. Sir Ernest Rutherford's theory of atomic structure, to be 

 dealt with in the latter part of this discourse, and the whole of our 

 knowledge as to what isotopes were, made it appear probable that 

 their atomic volumes, like their chemical character and spectra, 

 should be identical, and therefore that their density should be 

 proportional to their atomic weight. The thorite lead proved to be 

 0*26 per cent denser than the common lead. Taking the figure 

 207*2 foi'Hhe atomic weight of common lead, the calculated atomic 

 weight of the specimen should be 207 '71:. 



The two specimens of lead were fractionally distilled in vacuo, 

 and a comparison of the atomic weights of the two middle fractions 

 made by a development of one of Stas's methods. The lead was 

 converted into nitrate in a quartz vessel, and then into chloride by a 



Vol. XXII. (Xo. Ill) k 



