306 Proi W. D. Harkins on the 



some of the smaller values of this function in 1917 (J. Am. Chem. Soc. 

 xxxix. p. 859 (1917)), while he gave plots of this function in a paper 

 published in October 1920 {ibid. xlii. p. 1975 (1920)). The first plot of 

 this sort was published by Durrant in 1917 (ibid, xxxix. pp. 621-7 

 (1917)). He used for plotting the Harkins- Wilson equation for atomic 

 weights developed in 1915. Attention may be called to the fact that 

 the writer has already published the figure 5 of this paper, which may 

 be considered as the fundamental plot upon which the classification into 

 isotopic numbers is based, in October 1920, so the inclusion of this 

 figure in the present paper is for the purpose of showing its remarkable 

 feature, that all five variables, P, M, N, n, and N/P, are represented as 

 constant values by straight lines. In the plot printed at that time 

 only M and/", where f=n/2, were indicated. The/was used at that time 

 in place of the n for the simple reason that it gives the excess in the 

 number of nuclear negative electrons over half the number of positive 

 electrons of fundamental importance in connexion with the stability 

 relation?, and also the number of beta or cementing electrons in the 

 thorium disintegration series. This note is added, not in order to claim 

 priority, since Professor Masson evidently knew nothing of the work of the 

 writer, but simply to indicate that the fact that this paper bears a later 

 •date than that presented by him, does not show that these ideas have 

 been developed later by the writer. That the system of classification is 

 a natural one, is made evident by its independent development by two 

 different workers. The other systems of classification presented here, 

 <and the stability relations, which are the main topics of the present 

 paper, were not mentioned by Prof. Masson, 



Experimental Correlations (Laws) Related to the 

 Stability and Constitution of Atom Nuclei. 



One of the features of the periodic classification of the 

 elements, as developed by de Chancourtois, Newlands, 

 Mendelejeff, and Meyer, was to arrange the elements in an 

 ordinal series, the order in the series being determined by 

 the chemical and physical properties. While the atomic 

 weight was taken as the value used for fixing the preliminary 

 position of the element in the series, it is to be noted that 

 the final position was determined, not by the atomic weight 

 alone, but largely or even entirely by the other properties. 

 This was recognized by Ry d berg, who suggested that the 

 term " ordinal number " of the element be used to designate 

 the numerical index which fixes the order of the element in 

 the series. Moseley was able to index the elements according 

 to the frequencies of the K lines in the characteristic X-ray 

 ■spectra, and found that the order thus revealed was exactly 

 that given by the periodic arrangement. This element 

 number, which will be represented by M, improperly 

 designated as the atomic number, has the same value as the 

 ordinal number of Rydberg. M is thus shown to be experi- 

 mentally determined by two different methods. 



