590 



SCIENCE 



[N. S. Vol. XLI. No. 1060 



positive sodium, and, in general, the transi- 

 tion per saltum from period to period had 

 been discussed by Reynolds and others. It 

 needed explanation and was impossible 

 mathematically except by passing through 

 zero or infinity. Some, as Sedgwick and 

 de Boisbaudran, seem to have predicted 

 such transition elements, and when argon 

 was discovered it was not difficult for 

 Julius Thomsen and de Boisbaudran to ar- 

 range an entire zero group with approxi- 

 mate atomic weights three years before 

 Eamsay's brilliant discovery of the other 

 inactive gases. 



There are other anomalies in the system 

 which are difficult to explain with the ac- 

 cepted tabulation. Such, for instance, is 

 the existence of the rare earths, now some 

 sixteen in number, so closely alike chemic- 

 ally and so different from other chemical 

 individuals. The more they are studied, 

 the less possible does it seem to fit them in 

 any vacant places in the table. Meyer has 

 recently suggested that they may form a 

 miniature periodic system in themselves 

 reproducing the relations of the main sys- 

 tem. But a more serious breakdown in the 

 supposed fundamental principle of the 

 system comes in the relative position of 

 such elements as argon and potassium, co- 

 balt and nickel, tellurium and iodine. 

 After most exhaustive investigation of their 

 atomic weights it has become evident that 

 these can not be used in deciding the rela- 

 tive order and at the same time have these 

 elements fall into the proper grouping with 

 those elements chemically most nearly re- 

 lated to them. So the order of the atomic 

 weights has been tacitly abandoned and the 

 superior determining power of the chem- 

 ical characteristics acknowledged. This 

 can only mean that the mass of the atom 

 is not the sole, nor indeed the chief, deter- 

 mining variable, and it would seem that the 

 search for the latter can only be ended by 



the solution of the problem as to the nature 

 of the atom itself. 



Certain clews to this have undoubtedly 

 been in our hands for a long time, but 

 their leading was not clear and the thought 

 of them baffling. Such, for instance, were 

 the facts that by taking an atom of nitrogen 

 and four of hydrogen a grouping of atoms 

 was obtained which behaved in general as 

 an atom and was the analogue of potas- 

 sium. Or, again, carbon and nitrogen give 

 us an analogue of chlorine — and so with 

 compound radicals in general. But while 

 both building and tearing down again were 

 easy, they seemed to throw no light on how 

 those we could not tear down were once 

 built up. 



Still another thought-inspiring fact which 

 would seem to have important bearing on 

 the nature of the atom and hence the mean- 

 ing of the periodic system is the ease with 

 which certain elements by a change of 

 valence change their chemical character 

 and form distinctive series of salts as if 

 they had been transformed into different 

 elements. This causes some confusion and 

 what would ordinarily be called forcing in 

 the present tabulation of the system, and 

 it will be recalled that Mendeleeff, in his 

 earlier tables, actually placed certain of the 

 metals, as copper and mercury, in two 

 different groups, assigning each two differ- 

 ent places. Signs are seen in the work of 

 Barbieri and others of a tendency to place 

 certain of the elements in two or more 

 different groups according to valence. 



I believe that one should keep in mind 

 the idea involved in Patterson-Muir's defi- 

 nition of an element as a collection or group 

 of properties. Thus there are weight, 

 electro-chemical nature, affinity, valence 

 and other properties by which we recognize 

 it and differentiate it from other elements 

 and to which our knowledge of it is neces- 

 sarily limited. There is a more or less defi- 



