214 
VARIABLE STARS 
must be taken to indicate that the incident light is re-emitted without 
much change of quality—that it never really mixes with the ordinary 
radiation L 2 . Take as a typical case T x = 12,000°, T 2 = 6,000°. The high 
frequency radiation of S x will be able to do things in the atmosphere of S 2 
which the latter’s own radiation cannot do ; and it will do them vigorously 
because it finds virgin material. It is for this reason absorbed rapidly in 
the upper atmosphere of S 2 —more rapidly than in the atmosphere of S x 
because the conditions are further from thermodynamical equilibrium. 
By keeping its mechanism of absorption and emission distinct from that 
of the general radiation of S 2 and by the great departure from conditions 
of thermodynamical equilibrium it seems likely that the reflected radiation 
will, to a large extent, preserve its original constitution. 
This seems to be confirmed by spectroscopic observations. In many 
cases the spectrum of the faint component is measured for radial velocity. 
Sometimes it is explicitly stated that the two spectra are nearly the same ; 
more often no remark is made—a silence which is equally significant, 
since a difference of type corresponding to the known values of J X IJ 2 
would scarcely pass without comment. 
We may suggest further that the spectrum of the hemisphere of S 2 
turned towards a component S x at higher temperature is likely to approxi 
mate to the spectrum of S x even when the reflected radiation is relatively 
weak. The reflected radiation has the strategic advantage of attacking 
the material of S 2 from the outside, so that the conditions caused by it 
imprint their spectrum last of all on the radiation flowing out from S 2 , 
and the lines cannot be blurred out by subsequent experiences of the 
radiation. 
These suggestions however require a more detailed consideration than 
I have been able to give, and I am by no means confident that they will 
be found tenable. 
The star TV Cassiopeiae is among those treated in Table 19. It appears 
from Table 28 a that it is well behaved as regards reflection effect and the 
circumstances seem to be favourable for a more intensive treatment. 
Using the most recent data, viz. Plaskett’s orbits (Pub. Bom. Obs. Victoria, 
2 , p. 141) and McDiarmid’s photometric solution ( Princeton Pub. No. 7) 
we have 
M x = 1-74, M 2 = 1-02, R x /a = -279, B 2 /a = -306, 
R x = 2-45 x R q , R 2 = 2-68 x R Q , J x /J 2 = 13-6, 
T x = 10,500, T 2 = 5,200, 
the former temperature being estimated from the observed type B 9 and 
the latter from J x /J 2 . The results are-— 
Bol. Mag. (from iî) Bol. Mag. (from M) Residual 
+ 0-34 + 1-86 + 1-52 
+ 3-18 + 4-58 + 1-40 
51 
5 2