8S4. Dr. J. W. Nicholson on a Structural 



Their absence from the solar spectrum is significant, and 

 perhaps an indication that in this case the process of devolu- 

 tion has not gone far. I£ they are radioactive products, the 

 fact that their groups are so effective in the constructions 

 for other elements has an important significance. If an 

 atom of caesium, for example, gave off the group He 3 , the 

 remainder might disintegrate into krypton and xenon, 

 without any possible detection by experiment. 



Sodium, Copper, Silver, Gold. — These metals occupy the 

 other sub-^roup in column 2 of the table, and are not closely 

 related chemically, even in respect to valency. Sodium 

 presents much more similarity to the alkali metals, but is 

 devoid of some of their characteristic properties — it does not, 

 for example, form alums. In the alkali group, lithium 

 itself is somewhat exceptional, especially as regards its 

 spectrum, as Hicks* has shown recently. The formula as^ 

 cribed to it is not quite similar to those of the allied metals, 

 as it only contains two of the primary substances. 



Similar remarks apply to sodium in the present group. 

 It is chemically unlike copper, silver, and gold, and it has 

 been found impossible to obtain a formula quite in line with 

 those of the other metals, nor conversely, to endow them 

 with formula? similar to that of sodium. It seems likely 

 that sodium is the only stable representative of a whole 

 group of elements. The formulae suggested for the present 

 elements are : — 



Sodium, {Nu 2 (Pf H) 3 }H 2 Nu 2 . PfH 2 = 23*008 (23*0), 



Copper, 2{He. t (PfH) 3 }Pf 2 .(PfH) 2 = 6:-V57 (63-57), 



Silver, 4{Nu 4 (PfH) 3 }2{NTu 2 (P£H) 3 }He 3 . (Pf H) 2 = 10789 (107-88) 



Gold, 8{He 2 (PfH) 3 ]2{He 2 Nu 2 (PfH) 3 }Nu 2 (PrK) 2 = 197-3 (197-2), 



where the formula for sodium has a pronounced similarity 

 to that of potassium in the previous section. 



Beryllium, Calcium, Strontium^ Barium, Radium. — The 

 spectrum of beryllium is abnormal in the same manner as 

 that of lithium, and is not of the same type as those of its 

 associated members of the table. This peculiarity has been 

 pointed out to me by Prof. W. M. Hicks, who has not yet 

 published his investigations of these series. We may expect 

 io find a similar divergence in its formula, and this occurs. 



* Phil Trans. A. 1909. 



