34 Hinrichs on Dark Lines in the Spectra of the Elements. 
I. The mutual distances of the different lines in ue separate 
at are multiples of the smallest distance in such g 
roceed further, we must remember that Kirchhoff himself 
a Ghewibaice that he did not map all the lines he saw; besides, 
we know that many lines have been discovered which he did 
not see. Now the intensity of a line is certainly a less import- 
ant element than its place; hence it may be that some lines are 
mapped while corresponding ones are not. Thereby we are per- 
mitted to group several lines into one group which we will call 
a all combination, since it consists entirely of actually ob- 
served lin It would, on account of the above, likewise be 
proper to iat hypothetical lines, thus dividing a given inter- 
val into a wirtual combination ; but since such divisions sii iven 
by the numbers representing the interval, and since we desire 
to keep this part of our article entirely free "from any hypotiens. 
it being our purpose to show just what the given facts mean— 
we will not make any further use of such virtual combinations; 
what is here said will be snfficient to make the experimenter 
look to those places of the spectrum which are pointed out by 
such combinations, 
e six groups of the enginn aes show the following 
remarkable physical combinations 
Group I, interval: observed 1 : 2 4. 
Group Il, range too great, pelt i oprees the ratio of 1 to 2. 
Group Ul, observed 1:3:2: 
physical 4 :2:8, or 2:1:4 
Group IV, observed 4 : sake 3332.9.5:3 
physical 4 : ee epee ee 
#4? € 
or 
Group V, observed 3: 6:1 
Group VI, observed 2:2:1:4:1 
physical oe ie Soa | 
or en 5 
These numbers show 
IL. The intervals in the different groups may be expressed in very 
- simple numbers, as 1, 2, 3 
Thus there is a very great harmony between the intervals of 
the same group. If this vicheeateeing is to be admitted as a physical 
‘fact, it must alg t the whole spectrum of an ele- 
— And if our riotion ve a physical group, as containing 
rrespondi me lines, is correct, these p ‘en groups must have 
ding intervals throughout t 
e find from the preced- 
pe me 
