i88o.] Spectrum Analysis. 739 
uninfluenced by the nature of the latter substance, unless it 
were supposed that the compound was decomposed in the 
flame, the reverse of which opinion Prof. Kirchhoff at least 
seemed openly to favour. Yet that such is really the case 
seems to be now established beyond doubt. But then it is 
evident also that, in the case of a compound undecomposable 
by heat, the spedtra furnished by MM. Kirchhoff and Bunsen, 
and which really relate but to the uncombined metals, would 
be no guide whatever to the detection of the metallic element 
which it contains. In the next place, an even more im- 
portant remark than the above (which, indeed, has to be 
pradtically considered principally in reference to the matter 
which later on will be treated of, but is of less import in 
ordinary chemical analysis) is suggested by the circumstance 
that it had nowise been proved that, of the fifty-six other 
elements then known, the spedtrum of none coincided, either 
totally or in part, with any one of the six spedtra which 
alone had been mapped and examined; and, in fadt, M. 
Kirchhoff — who had already found tnat the influence of 
temperature is far from being so totally immaterial as at first 
had been supposed — adtually met at a later period with ex- 
amples in which lines in the spedtra of two different sub- 
stances coincided. What some would call apradtical answer 
to these objections, which would certainly have been urged 
from numerous quarters had the subjedt been mooted by 
some fortunate, or rather unfortunate, novice, an answer 
which, however, is not the least to the point, was furnished 
by the discovery, on the part of Prof. Bunsen, of two new 
metals whose existence had been first revealed to him by 
additional lines in the speftra of known metals, together 
with which the former occurred, as otherwise unsuspected 
impurities. A few more such discoveries have since been 
made which, indeed, prove beyond cavil the value of the 
method in question, praaically at least, although not in 
theory ; for, besides the difficulties which. we have already 
pointed out, some other observations have since been brought 
forward which, whilst they may possibly open a wide and 
fruitful field to inquiry for the future, for the present throw 
doubt on the reliableness of the optical test as an absolute 
sign of chemical nature. Not to mention certain observa- 
tions by Dr. Robinson, of Armagh, which bear in the same 
diredtion, it has been alleged by M. Pliicker that an ele- 
mentary gas, such as nitrogen, may possess three spedtra of 
altogether varying charadter, depending upon “ three allo- 
tropic states ; and, on the other hand, an English chemist 
who may be taken to express the opinions of MM. Bunsen 
