56 DUGAN— RESULTS FROM THE [April 24, 



periods. The revolution of the Hne of apsides in the shorter period 

 would cause the time of secondary minimum to oscillate back and 

 forth to an extreme of ten minutes before and after the midway 

 point between successive primaries. 



The evidence is both scarce and somewhat uncertain. The en- 

 tire secondary eclipse amounts to a drop of but little over a tenth of 

 a magnitude. An isolated photographic estimate of brightness is 

 of little value here. Even from a continuous series of photometric 

 observations, it is difficult to fix on the time of mid-eclipse within 

 several minutes. My own observations, weighted according to the 

 apparent certainty with which they determine the time of mid-eclipse, 

 indicate rather strongly the shift in the time of secondary eclipse in 

 the shorter of the two periods. The few observations by Wendell 

 do not furnish any very strong evidence for or against this result. 

 In no case were his observations carried through both branches of 

 the eclipse, and consequently they do not determine the time of the 

 eclipse with much accuracy. What disagreement there is, is in the 

 same direction as in the observations of primary eclipse — for some 

 reason the Wendell times of eclipse are nearly all earlier than mine. 

 The evidence in hand at present points to a revolution of the line of 

 apsides of RT Persei. I hope to observe occasional primary and 

 secondary minima of this star during the next few years. 



As a bi-product, I have determined the photographic curve from 

 the Harvard observations and compared it with the visual curve of 

 primary minimum. According to the observations, the eclipses in 

 the two regions of the spectrum are of the same duration and the 

 same depth, but the curves follow different paths. This is too 

 strange a result to be taken very seriously, considering the paucity 

 of the photographic material. I have, however, one minimum which 

 I observed through a blue color screen made for the purpose. 

 Through this filter the minimum was observed about 0.15 magni- 

 tude deeper than without the filter. So I conclude that the obser- 

 vations at the bottom of the photographic curve, which are few in 

 number, are to be disregarded and the curve extended to the greater 

 depth given by the color-screen observations. The difference in the 

 character of the two curves indicates that when one star is in great 

 part covered up, the light of the system is more reddish than when 



