680 DR. W. HUGGINS ON THE PHOTOGRAPHIC SPECTRA OE STARS. 
Now the lines H and K agree in position with two strong lines in the spectrum of 
calcium, and are therefore usually believed to be produced by the vapour of that sub¬ 
stance. It was therefore of some interest to ascertain if any of the other strong lines 
of the typical spectrum were concident in position with a pair of strong lines in the 
calcium spectrum at 3736‘5 and 3705 ‘ 5 . The calcium line 3736 • 5 falls nearly half-way 
between e and £. The stellar line 6 is indeed very near the calcium li ne 3705 ' 5 , but 
does not coincide with it, its position being 3707 ' 5 . 
I prefer in the present paper not to enter into any discussion of the physical 
lengths) and made the following table of their first and second differences. Assuming that the irregulari¬ 
ties in the second differences cannot be referred to errors of observation, I think that the accuracy of your 
work gives evidence which must be accepted that the second alternative is the true one, viz., that the 
lines do not lie on but near a definite curve. 
W.—Wave length in air. v .—Wave frequency in air. 
A n .—First difference. A ' 2 n .— Second difference. 
w. 
n . 
An . 
A hi . 
c 
6562-1 
1523-9 
F 
4860-7 
2057-3 
h 7 
4340-1 
2304-1 
134-2 
h 
4101-2 
2438-3 
81-8 
52-4 
H x 
3968-1 
2520-1 
52-2 
29-6 
a 
3887-5 
2572-3 
35-9 
16-3 
z 3 
3834 
2608-2 
26-8 
91 
7 
3795 
2635-05 
19-3 
7-5 
c 
3767-5 
2654-3 
15-6 
3-7 
e 
3745-5 
2669-9 
11T 
4-5 
3730 
2681-0 
Q. 
2-1 
>i 
3717-5 
2690 
& 
7-2 
1-8 
e 
3707-5 
2697-2 
6-2 
l-o 
i 
3799 
2703-4 
This so far goes to show that the typical lines are not consecutive members of one series, but the 
members of one or more series whose positions lie near the curve. This appears to be corroborated by 
finding that H x and the hydrogen line near G are connected harmonically, these rays being exactly the 
35th and 32nd harmonics of a vibration whose fundamental is 
T 
72-003 
(t being the time in which light 
