446 The Bev. Walter Sidgreaves [Jau. 22, 



The only answer I can give is by the same question on another star- 

 spectrum, the spectrum of o Ceti. You see on the screen those two 

 brilliant lines, Ha and H^ : What has become of He, H^, and H^ ? 

 The latter could not aj)pear in the photograph, but it has been 

 searched for in vain as a feature of the visual spectrum. 



I have now put before you all the parts of my spectroscopic study 

 bearing upon the interesting problem of the form of the orbital 

 motion of /? Lyrae. Tlie velocity of this motion is another problem, 

 the interest of which is enhanced by the difficulties attending its 

 evaluation. It will have been already observed that an eclipsing 

 pair of stars must have their motions at some part of their orbit, very 

 nearly in our line of sight. And, if the orbit is circular, the maxi- 

 mum measured velocity, on the Doppler principle of line displace- 

 ment, must be the uniform velocity of the motion, and must be found 

 at the times when the stars are to us in quadrature ; whereas, in an 

 elliptical orbit the maximum velocity must occur nearer the ej^och of 

 pericentre. If, then, we could measure the displacement-sejjarations 

 of the bright and dark companion lines, and in the process fix also 

 the epochs of greatest and zero separations, we should be able to 

 decide the question between circular and elliptical motion, as well as 

 quote the relative velocity of the pair of stars, independently of any 

 motion of the system in space. But here we meet with the obstacle 

 already alluded to, which throws doubt on all direct measures of the 

 relative displacements of broad lines, viz., the unknown positions of 

 their spectral centres — the positions in which the lines would be 

 found when reduced to the thinness characteristic of the more perfect 

 gaseous state of the origin. 



An indirect method of solving the problem was suggested by the 

 tabulated widths of the components of the composite band at H^. For 

 simplicity of expression we may transfer all the relative motion to one 

 of the pair of stars about the other at rest. The method may then be 

 described as the measurement of the arithmetical sum of the opposite 

 displacements of the same spectral line at two opposite epochs, in 

 place of the relative displacement of the two lines at the same epoch. 

 The method was possible with the H^^ line, because its red side 

 bright edge was never absent, never masked by the dark line. The 

 measures could be taken from the centre of the dark line to the 

 edge of the bright line. And, supposing the bright line star to be 

 the moving one, the measures would be greater when the motion was 

 from us ; and half the difference of the greatest and least measures 

 would give the oscillation amplitude of the bright line, and con- 

 sequently the maximum relative velocity of the stars. The epochs 

 of the greatest and least measures would determine the i^oints on the 

 orbit of greatest positive and negative velocities, and those of the 

 mean measure would settle the points of spectral conjunction. 



The method supposes constancy of width of the lines ; and this 

 may be incorrect physically, or apparently falsified by atmospheric 

 conditions. But both these variations might be expected to be 



