IN CONNEXION WITH THE SPECTRUM OF THE SUN. 
655 
In the maps the spectrum of the pure vapour is shown below, the spectra of the 
mixtures above, the highest spectrum being that of the mechanical mixture containing 
the smallest percentage. 
The series mapped are as follows : — 
Sn + Cd percentages of Cd 10, 5, 1, 0T5 
Pb + Zn „ Zn 10, 5, 1, 0T 
Pb + Mg „ Mg 10, 1, 0T, 0-01 
I may remark that these substances were used in consequence of their low fluid tem- 
peratures and of the consequent ease with which the mixtures could be made with the 
arrangements at my disposal. 
An inspection of the maps will render any long verbal statement of the results 
obtained superfluous. I may, however, remark that although we have here the germs 
of a quantitative spectrum-analysis, the germs only are present, because from the exist- 
ence of several “ critical points”*, and great variations due to causes which have not yet 
been investigated, the results thus obtained are not sufficiently close or constant for 
practical application. 
Further researches, however, have shown me that, by discarding altogether this method 
of eliminating lines from the spectrum and adopting a different procedure, it is pos- 
sible by means of spectrum-analysis to make quantitative determinations of certain 
mixtures with almost if not quite as great accuracy as that generally obtained by the 
methods in use. 
As these researches are still incomplete, I shall defer an account of them and the 
process adopted till a future occasion. 
Application of these observations to the Solar Spectrum. 
Let it be assumed for a moment that subsequent investigations may justify us in 
applying these considerations to the composition of the reversing layer in the sun’s 
atmosphere, and suppose it to be composed of a mixture of zinc and lead vapours. 
The percentage of zinc indicated by the number of lines actually known to be reversed 
in the solar spectrum would certainly be between one and five per cent. 
Similarly, the percentage of sodium would be 100 per cent., of magnesium 100 per 
cent, at the maximum sun-spot period, a little less at other times ; the percentage of 
iron always something less than 100 per cent., and so on. Adding these percentages 
together, we get an apparently impossible percentage composition of the reversing layer. 
But here another set of considerations comes in. In each case the percentage composition 
has been assumed to be determined by the lines omitted from the spectrum observed at 
atmospheric pressure — a pressure accidental to this planet. 
* For instance, in the cadmium experiments hut little difference occurs between 10 per cent, and 5 per cent. 
In the case of zinc between 10 per cent, and 5 per cent, only a very small change occurs. In the case of mag- 
nesium only two lines are lost between 100 and 10 per cent. 
4 T 
MDCCCLXXIII. 
