Atomic Weights of analogous Elements. 317 



Th^'^e numerical relations are of three kinds. The atomic 

 weights of analogous elements may be the same ; or may be in 

 multiple proportion ; or may differ by certain increments. 

 '' Of the first class we remark the strictly analogous metals, — 

 chromium 26*7, manganese 27"6, iron 28, cobalt 29'5, and 

 nickel 29*6. Then a double group of the platinum ore metals : — 

 palladium 53*3, rhodium, 52*2, and ruthenium 52*2; and also 

 platinum 98-7, iridium 99, and osmium 99*6. We are tempted 

 to add to this — mercury 100. Again, in the mineral cerite we 

 find together — cerium 47, lanthanium 47, and didymium 50. It 

 has been remarked, not only that the metals of each of these 

 groups have similar properties and weights, but that they are 

 found associated together in nature. The question has often 

 been put, — Would more accurate determinations show these 

 atomic weights to be not nearly but exactly the same ? It may 

 be doubted. Yet it ought to be remembered that these numbers 

 are the actual results of experiment, and are not controled by 

 any theory, as is always the case with organic compounds. 



As to the second class of numerical relations among atomic 

 weights, namely, multiple proportions, Who has failed to remark 

 that the platinum group has double the atomic weight of the 

 palladium group, and that gold 197 is again the double of pla- 

 tinum ? These two pairs have frequently been noticed — boron 

 10-9, and silicon 21*3; oxygen 8, and sulphur 16. We now 

 come to a large group, those metals whose oxides principally 

 affect an acid character, being also insoluble in water. The 

 highest of these in weight is tantalum 184; half 184 is 92 — 

 the equivalent of tungsten; half 92 is 46 — the equivalent of 

 molybdenum ; and half 46 is 23 — ^just below the recognized 

 equivalent of titanium 25*. Three times 23 is 69: now 68*6 

 is the equivalent of vanadium, being intermediate between tung- 

 sten and molybdenum. Tin has certain claims to be grouped 

 along with the same elements ; its equivalent is 58 : now 57*5 

 would be two and a half times 23, and intermediate between 

 molybdenum and vanadium. Taking 11*5 as the basis number 

 of this series, we have — 



23*6 according to Mosander. 



