1878.] Analysis in connexion with the Spectrum of the Sun. 167 



broken up in the sun, and quite broken up in the discharge from a 

 large coil and jar, into another or others with lines in the violet. 



I said " another " or " others," because I was not then able to deter- 

 mine whether the last-named lines proceeded from the same or different 

 molecules ; and I added that it was possible we might have to wait for 

 photographs of the spectra of the brighter stars before this point could 

 be determined. 



I also remarked that this result enabled us to fix with very consider- 

 able accuracy the electric dissociating conditions which are equivalent 

 to that degree of dissociation at present at work in the sun. 



In fig. 3 I have collected several spectra copied from photographs, in 

 order that the line of argument may be grasped. 



First we see what happens to the non-dissociated and the dissociated 

 chloride. Next we have the lines with a weak voltaic arc, the single 

 line to the right (W. L. 4226*3) is much thicker than the two lines 

 (W. L. 3933 and 3968) to the left, and reverses itself. 



We have next calcium exposed to a current of higher tension. It 

 will be seen that here the three lines are almost equally thick, and all 

 reverse themselves. 



Now it will be recollected that in the case of known compounds the 

 band structure of the true compounds is reduced as dissociation works 

 its way, and the spectrum of each constituent element makes its 

 appearance. If in 3 we take the wide line as representing the banded 

 spectrum of the compound, and the thinner ones as representing the 

 longest elemental lines making their appearance as the result of partial 

 dissociation, we have, by hypothesis, an element behaving like a com- 

 pound. 



If the hypothesis be true, we ought to be able not only to obtain 

 with lower temperatures a still greater preponderance of the single 

 line, as we do ; but with higher temperatures a still greater preponder- 

 ance of the double ones, as we do. 



I tested this in the following manner: employing photography, 

 because the visibility of the more refrangible lines is small, and 

 because a permanent record of an experiment, free as it must be from 

 all bias, is a very precious thing. 



Induced currents of electricity were employed in order that all the 

 photographic results might be comparable. 



To represent the lowest temperature, I used a small induction coil 

 and a Leyden jar only just large enough to secure the requisite amount 

 of photographic effect. To represent the highest, I used the largest 

 coil and jar at my disposal. The spark was then taken between two 

 aluminium electrodes, the lower one cup-shaped, and charged with a 

 salt of calcium. 



In the figure I give exact copies of the results obtained. It will be 

 seen that with the lowest temperature only the single line (2) and 



n 2 



