i9is.] OBSERVATION OF ECLIPSING VARIABLES. 57 



they both shine iindimmed. Either one star is redder than the 

 other, or the eclipsed star is redder toward the hmb than at the 

 center. 



^ Draconis has been an equally interesting surprise. The aver- 

 age period is quite a little longer than was supposed. Some time 

 ago it shortened up with great rapidity. The top and bottom of 

 this sharp decline are well determined by observations from quite 

 a variety of sources. The two sine terms combined in this case 

 have periods of 7,200 and 2,880 eclipse-periods and nearly the same 

 coefficients as in RT Persei — ten and four minutes respectively. 

 The prolateness and eccentricity of 2 Draconis are about the same 

 as those determined for RT Persei. 



The secondary minimum of 2 Draconis is only 0.06 magnitude — 

 half as deep as that of RT Persei; the observations of the secondary 

 minimum are few and were all taken within a brief interval, and 

 they furnish very little evidence. This star must also be kept under 

 frequent observation. When observing this star with the 23-inch 

 it requires a determined effort to see it at all when in the middle 

 of its deep eclipse. 



The greater depth of the photographic eclipse comes out very 

 nicely in the case of s Draconis. As the eclipse increases, the light 

 from the star becomes redder and redder. At deepest phase more 

 than half the light is known to come from the fainter star. It is ap- 

 parently much redder than its far brighter companion, a fact which 

 is doubtless to be expected. 



Lastly, in reducing the observations of RV Ophiuchi I found it 



necessary to predict the minima with a sine term of about 1,600 



periods, and of small amplitude. The photometric history of RV 



Ophiuchi is short and incomplete compared with that oi RT Persei 



and s Draconis, and no conclusions can safely be drawn from this 



result. 



Princeton University, 

 April, 191 5. 



