160 
ASTRONOMY: W, S. ADAMS 
groups of *giant' and ^dwarf' stars.^ This hypothesis was based pri- 
marily on parallax observations. The absolute magnitudes calculated 
from these parallaxes showed almost a complete absence of stars of bright- 
ness intermediate between exceedingly luminous stars Uke a Orionis, and 
extremely faint stars such as Groom. 34. It has been thought probable 
by some astronomers that this apparent gap is due to the fact that 
parallax determinations have hitherto been restricted almost entirely to a 
few stars of great apparent brightness, and to stars of very large proper 
motion, while the connecting Hnks would probably be found among 
stars of moderate apparent brightness and moderate proper motion. 
The spectroscopic evidence, however, is based upon numerous stars of 
just this character, and the gap still appears to persist. 
These results may be summarized briefly as follows. Two groups of 
M stars are indicated clearly by an examination of the intensities of the 
hydrogen lines: in the first the hydrogen Hues are very strong; in the 
second they are very faint. A computation of the absolute magnitudes 
of these stars on the basis of certain peculiarities in their spectra shows 
the existence of these groups distinctly. Connecting Hnks over a range 
of 7 magnitudes are entirely lacking, and the conclusion seems to be 
unavoidable that among these stars the intensity of the hydrogen lines 
varies with the absolute magnitude. 
The results given for the high and the low luminosity stars may be 
used to furnish an approximate relationship between the intensities of 
the hydrogen lines and absolute magnitude. Thus we have for H^: 
Assuming a linear relationship between intensity and absolute mag- 
nitude we obtain the equation 
This is remarkably similar to the corresponding equation found for 
the Hne X 4216 and given on a preceding page. It seems probable, 
therefore, that in the case of the M stars, at least, the hydrogen lines 
may be used for absolute magnitude determinations in the same way 
as X 4216. 
There is, however, one characteristic of the spectra of these high lu- 
minosity stars which must be taken into consideration when use is made 
of the relative intensities of the hydrogen lines. This is a relationship 
which appears to exist between the intensities of the hydrogen lines and 
the intensities of the bands, the hydrogen lines being stronger when the 
High luminosity stars 
Low luminosity stars. 
Intensity of Efi 
+ 1.9 
-3.0 
M = -1.8 A -I- 4.8 
