162 
ASTRONOMY: W. S. ADAMS 
in brightness by nearly 8 magnitudes, and also in the case of the N and 
the R type stars of the Harvard classification. The differences, accord- 
ingly, appear to be fundamental in nature, and associated with the 
intrinsic brightness of the stars of the several types. They indicate a 
lower temperature in the absorbing gases constituting the atmospheres 
of the fainter stars, and are analogous in many respects to those observed 
in the spectrum of sun-spots. 
The division of the M type stars into two well-defined classes of high 
and low luminosity stars raises the question at once whether a corre- 
sponding separation may be found among other types of spectra. From 
his discussion of parallax observations Russell concludes that such a 
separation does exist among the K stars. The spectroscopic evidence 
tends to support the existence of such a division at least for the K5-KQ 
stars. This evidence is of just the same character as that in the case 
of the M type stars, and is of two kinds. First, the hydrogen lines 
have an abnormally high intensity in the very luminous stars, and there 
is an absence of intermediate values of the intensity between these and 
the low values characteristic of the fainter K5-K9 stars. Second, com- 
putations of absolute magnitude indicate the existence of two mean mag- 
nitudes, one high and the other low, about which the values for the in- 
dividual stars showed a marked tendency to gather. This effect is not 
so well defined as for the M stars, but still very clear. It may perhaps 
