xviii THE CLASSIFICATION OF THE ELEMENTS 475 



(c) the resemblance between beryllium and aluminium 

 was similar to that which Newlands had detected between 

 lithium and magnesium and that which is found between 

 boron and silicon, thus : 



Li Be B 



Mg A. Si 



The correctness of the smaller atomic weight was proved 

 finally by determinations of the vapour-density of beryllium 

 chloride, which gave the molecular weight as 817 (Be = 9, 

 C1 2 = 71) between 686 and 812 (Nilson and Pettersson, 

 Comptes rendus, 1884, 98, 988-990). 



2. Indium, Equivalent 38-27. Atomic weight 3 x 38-27 = 

 114-8. This element, discovered by spectrum analysis 1 in 

 the Freiburg zinc ores, by Reich and Richter in 1863, was 

 regarded by Newlands as divalent, but was placed correctly 

 by Lothar Meyer, with the help of his atomic volume curve. 

 Mendele'eff (Joe. cit. p. 178) confirmed the higher atomic 

 weight for the element by measuring its specific heat, which 

 he found to be 0*055 (agreeing closely with a number 0*057 

 given by Bunsen), whence the approximate atomic weight 

 equals 6*3 -"-0-055 = 115. 



3. Uranium. Equivalent 59-6. Atomic weight 59*6 x 6 

 = 238*5. This element was regarded by Newlands as 

 divalent, at. wt. 60 x 2 = 120. Lothar Meyer showed that it 

 could not be placed on the atomic volume curve at this 

 point, and suggested Ur = 60 x 3 = 180. Mendeleeff (loc. tit. 

 pp. 178-184), after tracing its analogies with chromium, 

 molybdenum and tungsten, assigned to it the atomic weight 

 60x4 = 240. Its recently-discovered radioactivity has fully 

 justified the view that uranium should be placed at the end 



1 Compare thallium, discovered in pyrites by Crookes (1861) and 

 gallium (p. 477), discovered in zinc blende from the Pyrenees by 

 Boisbaudran (1875). 



