26 



Relations of Spectra, etc., to Atomic Mass. [Nov. 7, 



in the density curve above aluminium, and the curve joining the 

 metals above is nearly a parabola. In the melting point curve, boron, 

 aluminium, and gallium are nearly on a straight line ; the curve then 

 takes a totally different form. This break corresponds to the break in 

 the diffuse series of the spectra, so that the same fundamental causes 

 appear to act in the production of these series, and in determining the 

 densities and melting points of the elements. It will be an interesting 

 study to trace those properties of the elements which vary regularly 

 from aluminium to thallium in the way that the sharp series vary. 



Silicon, Germanium, Tin, and Lead. — The connecting lines between 

 these elements are peculiar. There is a suggestion that the density of 

 germanium is too low : the diagram indicates that it should be between 

 5*8 and 5*9 instead of being 5*47. We observe, also, that the melting 

 point is remarkable. If this followed the order of the other cases 

 studied it would be below the melting point of tin. It is possible, 

 therefore, that germanium is a peculiar element, and worthy of more 

 complete investigation. From carbon the melting point falls with 

 increase in the atomic weight until tin is reached ; and it is only then 

 the change of direction takes place, one stage later than in the other 

 groups. 



Diagram (c). — Three of the long connecting lines approach to straight 

 lines in this diagram : those in both density and melting point diagrams* 

 for zinc, cadmium, and mercury, and that in the density diagram for 

 gallium, indium, and thallium. The first of these three lines is almost 

 perfectly straight, and the densities may be calculated from an equation 

 containing the square of the atomic mass : Density = 6*38 + 1803 W 2 

 x 10~ 7 . This formula gives zinc and mercury correctly, and cadmium. 

 0-01 too high. 



Conclusions. 



It has been usual for investigators to rest satisfied when the pro- 

 perties of the elements were shown to be "a periodic function of the 

 atomic mass." Diagrams drawn, by the method employed in this 

 paper, of each of the properties of the elements will show in what 

 degree the properties vary with the atomic mass, and will make it 

 easier to establish the exact quantitative relations. 



The work and results now presented indicate that the properties of 

 the elements are fundamentally due to the structure of the atoms, as 

 revealed by their spectra, rather than to the quantity of matter in them. 

 It seems, for instance, inconceivable that the transition from calcium 

 to strontium proceeded through the intermediate elements, when we 

 consider that the strontium molecules must have a similar structure to- 

 those of calcium. This structure is so simple that the fundamental 

 (Bunsen flame) spectrum of each of these elements consists of a single 

 line. The anomalies, according to Mendeleeff's law, in the atomic 



