274 



SCIENCE. 



[N. S. Vol. IX. No. 217. 



validity of the theory of gases and accept 

 the logical deductions than to deny the 

 truth of the present theories. The only 

 reason for not admitting the correctness of 

 these atomic weights is that that of argon is 

 greater than that of potassium, but this is 

 no severer attack upon the validity of the 

 periodic law than the accepted position of 

 iodin after, instead of before, tellurium. As 

 a matter of fact, all the more recent deter- 

 minations of the atomic weight of tellurium 

 give the figure 127.6, while that of iodin re- 

 mains unchanged at 127. 



Since these new elements form no com- 

 pounds, it is not possible to decide the 

 question by purely chemical methods. Were 

 it only possible for us to prepare a single 

 volatile compound of helium or of argon our 

 problem would be solved. In spite of many 

 attempts, I have not been able to confirm 

 Berthelot's results with benzene or carbon 

 bisulfid. I have, however, offered to place 

 a liter of argon at the disposal of my dis- 

 tinguished colleague, that he may repeat his 

 experiments on a larger scale. No one can 

 doubt that it is exceedingly desirable that 

 the question of these atomic weights should 

 be finally decided and that by chemical 

 methods. 



In order that the subject may not depend 

 wholly on physical theories, I have con- 

 sidered it from another standpoint. If we 

 assume, as from countless chemical facts we 

 are fully justified in doing, that the periodic 

 law is true, then, giving helium the atomic 

 weight 2 and argon 20, there is no possible 

 place for an element of their mean atomic 

 weight ; for, unless we absolutely overturn 

 the accepted views, there is no vacancy in 

 the table for such an element. This ap- 

 pears from the following portion of the 

 table : 



H = l He = 2(?) Li=7 Gl = 9-2 B = ll C = 12 

 N = 14 = 16 F = 19 A =20(?) 



It is true there is space enough between 

 He = 2 and Li = 7, but it is highly im- 



probable that an element belonging to the 

 argon-series could have so low an atomic 

 weight. The difference between adjacent 

 members of the same group of elements is 

 generally from 16 to 18 units, but here such 

 a difference is wholly excluded. If, on the 

 other hand, we assume He = 4 and A = 40, 

 it would be, in my opinion, by no means im- 

 probable that such an element could exist 

 whose atomic weight would be somewhere 

 about 16 units greater than that of helium, 

 and consequently 20 units less than that of 

 argon. The discovery of such an element 

 would be, therefore, not only a proof of the 

 correctness of 40 as the atomic weight of 

 argon, but also a confirmation of the pres- 

 ent views regarding the significance of the 

 specific heats of gases for their molecular 

 weight. 



A glance at the periodic table will make 

 these considerations clear, for in the latter 

 case we have the following series : 



Li = 7 Gl=9.2 B = ll C = 12 N=14 

 Na = 23 Mg = 24.3 Al = 27 Si = 28 P = 31 



He = 4 

 0=16 F = 19 (?)=20 

 S = 32 Cl = 35.5 A = 40 



Shortly after the discovery of helium I 

 began the search for this suspected element 

 of atomic weight of about 20, at first in con- 

 nection with Dr. Collie, my former assistant, 

 and later with my present assistant, Dr. 

 Travers. 



At first it appeared not improbable that 

 this element might be found in those 

 uranium minerals from which helium had 

 been obtained. We did not, however, con- 

 fine ourselves to these minerals, but tested 

 all available minerals either by heating in 

 a vacuum or by fusion with sodium bisul- 

 fate. In many of these minerals helium 

 was found ; in many, on the other hand, 

 only traces of hydrocarbons and hydrogen. 

 One mineral only, malakon, gave suf&cient 

 argon to be recognizable by the spectro- 

 scope ; the others which contained helium 



