6 Harriey & Ramacn— Wave-lengths of Principal Lines in Spectrum of Gallium, &c. 
special and peculiar treatment—we should expect to find that aluminous compounds, 
from whatever source, and aluminium would furnish the gallium lines, which is 
exactly in accordance with our experience. There can be no doubt, from the 
relative intensities of the lines, from their wave-lengths, from the association of 
gallium with aluminium and with iron, that the solar lines, 4172-211 and 4033-112, 
have their origin from gallium contained in the sun, which is present in small 
proportion when compared with iron, and that the solar spectrum, in so far as the 
proportion of gallium to iron is concerned, may be considered to be fairly imitated 
by the arc spectrum of blast-furnace iron containing 5;45;th of its weight of 
gallium, since, if the more volatile metal were in any considerable proportion, the 
gallium lines would be broader and overlap those of iron with wave-lengths, 
4172°296 and 4033-224. 
This research brings to light the fact that, where coincidences are few in 
number, the mere coincidences of lines in the spectra of terrestrial matter with 
reversed lines in the solar spectrum is not equivalent to a proof of the existence 
of the elements in the sun or other heavenly bodies, even when the most powerful 
instruments are employed for resolving the lines. Professor Rowland’s tables of 
solar spectrum wave-lengths show not only how nearly lines of different elements 
may coincide, but how there are some actual coincidences, as for instance of 
nickel, with iron lines. Lines may also overlap. Generally speaking, two lines 
of the same wave-length, belonging to different elements, differ in this respect, 
that one is strong, and the other weak, or perhaps not so strong. 
Examples are familiar to us, and may be cited. For instance, two lines of 
rubidium are very frequently obscured by two of iron; the strong line of 
rubidium corresponds with the weak line of iron, and vice versa.* If therefore the 
two lines appear of the same intensity, we know that rubidium is present; and 
if the order of their intensity is the reverse of that of the iron lines, we know 
then that the proportion of rubidium is larger than in the former case. Of 
course, the presence of iron is determined by other lines than these two which 
coincide with the rubidium lines. The greater mass of a substance may have the 
effect of obscuring or extinguishing some of the lines in the spectrum of another 
element less easily volatilised. On the other hand, the greater mass of a less 
easily vapourised substance may also obscure the lines of one more volatile which 
are in close proximity. 
In conclusion, we tender our sincere thanks to Dr. Agener for the aid so 
cordially given us in arene the photographs from which these measurements 
have been made. 
* Wave-lengths of the rubidium lines, 4215°72 and 4201:98. ; wave-lengths of the iron lines, 4216-28 
and 4202°15. 
