1904.] on Spectroscopic Studies of Astrophjsical Problems. 439 



the spectral changes, notwithstanding the long time-interval between 

 some of the consecutive j)eriodic dates on the chart. 



Secondly. The bright lines of hydrogen and helium lie on the 

 less refractive sides of their dark companions during the first half of 

 the period, and on the opposite sides in the second half, while about 

 the two epochs of minimum-light the pairs of lines are superposed. 

 And, so far, the chart is a confirmation, in a general way, of 

 Pickering's explanation. Neglecting details, we might say that at 

 the two minima of light, when the stars are in conjunction, with 

 partial eclipses, they are crossing our line of sight, and there is no 

 displacement of the spectral lines. "Midway between the two eclipses 

 the stars are travelling along our sight-line in opposite directions, 

 one receding, the other approaching ; the bright-line-star receding in 

 the first half and approaching in the second half of the cycle of changes. 

 Orbital motion in a plane through our line of sight, or only a little 

 inclined to it, is thus indicated by the general aspect of the chart ; 

 and the circular orbit is certainly most in accord with the light 

 variations. 



Thirdly. There is a general decline of the bright lines through- 

 out the light-period, and they are nearly extinct at the end. There 

 is here an obvious suggestion of higher temperature at the beginning, 

 and a progressive cooling during the rest of the cycle — a condition of 

 things not easily explained on any other supposition than that of a 

 tidal disturbance at the periodic returns of a near approach of two 

 stars. 



Collating these three general views of the chart, I consider myself 

 entitled by the first, to treat the spectrographic chart as, in general 

 characters, representative of any single cycle of spectral changes ; 

 and by the second, to assert that these changes must ultimately be 

 attributed to orbital motion. The third is a suggestion, strong in 

 itself, but inconsistent with a circular orbit. And it remains to seek 

 in the details of the chart either confirmation or refutation of 

 elliptical motion. 



To begin with the greater eclipse, we note that just before, at, 

 and just after the principal minimum, the line marked H^ appears as 

 a weak broad bright band divided centrally by a fine dark line, and 

 the other lines are single and dark. There is no displacement. 

 The bright and dark lines are superposed centrally, and the motion 

 of the stars must be across our line of sight. This is the first of the 

 two epochs of spectral conjunction, and it coincides with the stellar 

 eclipsing conjunction. Then comes the outburst of bright lines, all 

 on one side of their dark companions, to your right on the screen, 

 which is the red end of the spectrum ; and they remain there until 

 well after the second minimum, as seen on the 26 successive periodic 

 dates, including the date of 6-20, which is 9 hours after the second 

 minimum. It is not until 17 hours after this minimum that we 

 again find the spectrum as it was at the first eclipse, viz., a broad H^ 

 bright line divided by a fine dark line, and the other hydrogen and 



