74 
ASTRONOMY: C. D. PERRINE 
are distant also. Campbell concludes, 3 I understand, that the rule that 
when one spectrum is considerably fainter than the other, the spectrum of 
the secondary is apparently of a slightly earlier type than the spectrum of 
the primary, applies to these stars, which we have just found to be distant. 
Both of these pieces of evidence are confirmatory of the condition pre- 
viously found. 
I have also examined several well known systems whose colors have been 
observed, viz., a Herculis, rj Geminorum, 7 Delphini, tj Cassiopaeiae, £ Bootis, 
]8 Cephei and 7 Leonis. These stars also show in general the same relation 
to distance. 
Careful consideration of the data seems to show beyond doubt that the 
relation is to distance coupled with low galactic latitude. It cannot well 
depend upon the actual separations of the stars, for the ranges in that respect 
seem to be about the same in all of the groups. It can scarcely depend upon 
the differences of mass of the conponents as indicated by their differences of 
brightness, for a similar reason. There is a difference of absolute magnitude 
between the two groups, but there seems to be no reason for suspecting a 
relation in this case, whatever may be the bearing of such differences in 
others. 
It is not known whether both components of double stars were originally 
of the same spectral type or not. Investigators have generally assumed that 
they were. It is not possible to say, therefore, just what the course of change 
has been. However, we may justly conclude, I think, that the conditions 
are such in these regions as to produce opposite spectral effects in the com- 
ponents. There can be little doubt that the fainter components are in general 
also of smaller mass. The conclusion may be stated, therefore, in the fol- 
lowing form: — The conditions appear to be such that if two stars of unequal 
mass were introduced into the near region the smaller body would move 
more rapidly toward the later stage than the larger one, whereas in the rela- 
tively distant galactic regions the tendency would be for the smaller body to 
become of earlier type more rapidly than the larger one. 
This investigation also shows that the greatest differences in spectral type 
are in general found among the stars whose components show the greatest 
differences in brightness. The differences of brightness in the pairs having 
the same spectra are consistently small for all types. 
The results of this investigation seem to indicate with considerable force 
that some external cause is operating in more or less definite regions of our 
stellar system upon the conditions which produce spectral class. 
The details have been given more fully in a paper which will be published 
in the Astrophysical Journal. 
*IMd. } 6, 1910, (47). 
