VARIABLE STARS 
213 
For 6 stars the agreement of L r is excellent; 4 give large positive re 
siduals and 1 a negative residual. It is difficult to base any conclusion 
on the table without examining the strength of the determination for each 
star in detail. The probable errors refer only to L r (obs.); but L r (calc.) is 
also subject to uncertainty arising in the determination of R 2 /a • I n some 
cases the circumstances of the eclipse are more favourable for a determina 
tion of the elements than in others. We discuss below the question whether 
the calculated and observed values ought to agree without further cor 
rection ; tentatively we are inclined to think that they should agree nearly, 
and we must hope that the 5 outstanding residuals will be reduced when 
improved data are obtained. 
Table 28 a. 
Reflection Effect in Eclipsing Variables. 
No. 
Star 
T yP e 
RJa 
RJa 
Lx 
L r 
(calc.) 
L r 
(obs.) 
C-0 
P.E. 
1 
SZ Her. 
4-3 
•318 
•331 
•800 
•078 
•030 
+ -048 
— 
2 
U Cep. 
A 
13-5 
•191 
•308 
•838 
•069 
•043 
+ -026 
± -009 
3 
TV Cas. 
B 9 
6-9 
•275 
•302 
•849 
•068 
•074 
- -006 
± -008 
4 
RZ Cas. 
A 
12-2 
•253 
•288 
•902 
•064 
•063 
+ -001 
± -008 
5 
RT Per. 
— 
4-8 
•309 
•269 
•863 
•053 
•022 
+ -031 
± -009 
6 
Z Dra. 
— 
12-8 
•238 
•263 
•911 
•053 
•040 
+ -013 
± -006 
7 
RS Vul. 
B 8 
7-0 
•201 
•262 
•804 
•048 
•078 
- -030 
— 
8 
R CMa. 
F 
13-2 
•245 
•236 
•934 
•046 
•015 
+ -031 
— 
9 
ß Per. 
B 8 
11-6 
•210 
•239 
•895 
•043 
•045 
- -002 
— 
10 
Y Cam. 
A-F 
20-0 
•236 
•225 
•955 
•041 
•041 
•000 
— 
11 
RV Oph. 
A 
12-2 
•125 
•200 
•825 
•027 
•018 
+ -009 
± -006 
In this comparison we have made no discrimination between heat 
reflection and light reflection. If a residual such as that in No. 5 is not 
due to observational error, it signifies that the secondary in reflecting 
the heat of the primary reduces its luminous efficiency to less than 50 per 
cent. But in typical eclipsing variables we expect an increase rather than 
a decrease of luminous efficiency. The primary is usually of type B or A 
so that the original efficiency corresponds to a temperature above 10,000° ; 
it is reflected from a star of lower surface brightness corresponding (as 
judged from JflJJ to a temperature near 6000°. Not unless the tempera 
ture of re-emission is below 4500° is there a loss of luminous efficiency. 
Moreover, even if the temperature is below 4500° there is a compensating 
gain, since the original radiation L 2 is now emitted at a higher temperature. 
We should expect the residuals C-0 to be in general negative as in No. 7 ; 
the puzzling thing about No. 7 is that it is the solitary exception and not 
the general rule. 
I think the fact that we have only one important negative residual