1904.] on Spectroscopic Studies of Astrophysical Problems. 443 



to this, and, contrast failing, there remains only greater intensity to 

 account for the temporary strength of the lines. 



To me, therefore, the more probable interpretation of the chart 

 is : that shortly after the chief minimum there is a more or less 

 sudden evolution of heated gases, which cool down during the rest of 

 the period. The decline of the bright lines is in strict harmony with 

 this fall of temperature, allowance being made for an apparent 

 revival after the second spectral conjunction during which they are 

 hidden by the masking effect of the dark lines. The periodic rise of 

 temperature is quick and its fall is slow, as might be expected from 

 a tidal disturbance at the near approach on an elliptical orbit. 



Secondly. The highest temperature, brought out by tidal action, 

 may be expected shortly after the near approach of the stars ; and the 

 whole-day-delay in the manifestation of the bright lines is possibly 

 too long. But the greatest heat and greatest light need not go hand 

 in hand together. Heat may have done its work, and the light display 

 may be delayed until the masking effect of the superposition has been 

 removed by the separation of the bright and dark lines, which is 

 always the condition of this bright evolution. 



Thirdly. At this first appearance of separated bright and dark 

 lines, the relative displacement is already well advanced as early as 

 33 hours after conjunction at the minimum. It is so far advanced that 

 it may already have reached its maximum, for the measured width of 

 the bright part differs so little from the subsequent measures in the 

 series that it cannot be asserted confidently that there is any greater. 

 And this early indication of the maximum velocity in the sight-line 

 is in keeping with the quick swing back on an elliptical orbit, and 

 impossible in the slower rounding of the circle. 



Fourthly. To consider a hitherto neglected peculiarity in the be- 

 haviour of the line marked H^, you will see on the spectrogram at six 

 hours after the chief minimum that this lino has anticipated the other 

 lines of the spectrum in appearing with its one-sided bright line. It 

 is quite 24 hours earlier ; and it has lost its more refracted bright part 

 witliout any gain of width on the opposite side ; the more refracted 

 part seems to have simply vanished. I can only think of two 

 possible explanations. Either the two bright lines, as seen on the 

 previous date 0-5, are parts of one broad hydrogen line which has 

 contracted asymmetrically, so far as to show nothing on the more 

 refractive side ; or they are separate lines, and one of them has 

 vanished. Of these suppositions the first has against it serious 

 dif&culties, which must be passed by here as requiring too long a 

 discussion, and may be omitted without loss : for the periodic en- 

 larging and thinning of the line cannot well be accounted for 

 otherwise than as an effect of disturbances set up at the near a23i)roaches 

 of two stars, a state of things which implies elliptical motion, and is 

 the condition wanted for the simpler explanation of the one of two 

 lines disappearing. 



Supposing, therefore, that at the principal minimum there are 



Vol. XVII. (No. 98.) 2 h 



