ARC SPECTRUM OF ELECTROLYTIC IROK 
985 
rid of many of the lines due to impurities in the carbon. Observations were also 
made with iron poles 15 millims. in diameter, but although these did not melt so 
readily, the results obtained were not deemed satisfactory. Finally iron was 
volatilised on carbon poles. Thalen used the best Swedish iron in his investigation, 
but found that impurities were always present in it, and also in his carbon poles; for 
the spectrum of the arc always exhibited lines which were known to be due to 
calcium, manganese, barium, titanium, lithium, sodium, and other substances. In 
order to distinguish between lines due to foreign substances and those really due to 
iron, the spectra of suspected impurities were separately examined. Lines common 
to all or any of the elements observed and to the spectrum of iron on carbon poles, 
were assigned to the one in whose spectrum they were most intense. The origin of 
many of the foreign lines was known from the work of previous investigations, and 
It was therefore often only necessary to make exact determinations of wave-length to 
decide whether such lines did or did not coincide with lines attributed to iron. 
As to the success of this method of eliminating impurities Thalen remarks :— 
“ Malgre tous les soins que j’ai pris, il est pourtant bien probable que quelques lines 
des raies attribuees au fer doivent etre rejetees de ma liste comme appartenant a des 
corps etrangers. Neanmoins, apres avoir examind en somme cinq fois le spectre du 
fer, je suis porte a croire que je peux enoncer comme resultat de ma recherche 
pr^cedente que le nombre des raies du fer obtenu dans le spectre visible monte reelle- 
raent au moins a 1200, et que ces raies coincident toutes avec des raies sombres du 
spectre solaire, Je ne doute pas qu’on ne puisse encore augmenter beaucoup ce 
nombre, au fur et a mesure qu’on augmente I’intensite du courant, c’est a dire en se 
servant de machines dynamo-electriques plus puissantes que la mienne.” 
II. The Peesent Work, 
Necessity of the Research. 
Observations of the variations undergone by the spectrum of a single element 
subjected to changes of temperature, led me to make an investigation of the sjaectra 
of different strata of the sun’s atmosphere. The considerations which made me hope 
for help in this quarter were stated as follows :—“ Whatever be the chemical nature 
of this atmosphere, it will certainly be hotter at the bottom—that is, near the photo¬ 
sphere—than higher up. Hence, if temperature plays any jiart in moulding the 
conditions by which changes in the resulting spectrum are brought about, the 
spectrum of the atmosphere close to the photosphere will be different from that of any 
higher region, and, therefore, from the general spectrum of the sun, which practically 
gives us the summation of all the absorptions of all the regions from the top of the 
atmosphere to the bottom, 
“ Now, as a matter of fact, we have the opportunity, when we observe the spectrum 
6 K 
MDCCCXCIY. 
A. 
