136 Hartiey—An Investigation of the Connexion between 
elements, and it does not appear to be consistent with the usual conception of the 
molecule of those substances which contain only one atom; but it becomes quite 
intelligible if the view of the composite nature of the atoms be entertained, for 
the band-spectra belong to the elementary atoms, and may be considered as being 
caused by the vibrations, not only of the atoms themselves, but also of the 
component parts within the atom. The facts adduced with regard to the spectrum 
of silver and of copper, I believe must be regarded as a proof of this. 
I conclude, therefore, that the banded flame-spectra of lead, antimony, bismuth, tin, 
zinc, cadmium, copper, silver, and gold are the spectra of the atoms. 
What relation, then, have the lines of the elements to the bands? It must be 
admitted that the lines are also the spectra of the atoms, and it may be inferred 
therefrom that component parts of the atoms are in a state of dissociation. ‘There 
is, however, an alternative view, which is based upon a somewhat different interpre- 
tation of the facts. It is unquestionable that the are and spark lines are caused by 
the operation of a largely increased amount of energy over that available in the 
flame, and by assuming that this excess is in part added to the material atom, it 
is conceivable that the components of the atom are thrown into a different mode 
of vibration, which increases their radiant energy, and so renders smaller quantities 
of matter recognizable, either by the brilliancy of their spectra, or by the intensity 
of their chemical action on a photographic plate. The acceptance of this view so 
far would naturally lead one to expect that in the periodic system different groups 
of elements, according to their chemical properties, would acquire different 
quantities of energy, and individual elements belonging to the same group would 
show gradational differences in this respect. ‘The evidence of this would be that 
different groups would exhibit line-spectra with greater or less facility; and 
elements of the same group, under the influence of the same source of energy, 
would differ in the number of lines of their spectra, or in the proportion of 
lines to bands. Now, these conditions are exactly in accordance with 
observed phenomena, in forty at least out of forty-four elements. To facilitate 
reference to them, the characteristic flame-spectra of the elements have been 
tabulated, but for particulars regarding the different spectra, the reader must 
consult ‘‘ Flame-Spectra at High Temperatures” * and the ‘Banded Flame- 
Spectra of Metals.” + 
* Hartley, Phil. Trans. 1894, vol. 185, Part. I., pp. 161-212; Parts II. and III., pp. 1029-1091. 
+ Hartley and Ramage, Scient. Proc. Roy. Dublin Soe., 1901, vol. vii., pp. 339-352. 
