•26 REPORT — 1891. 



the great problems of the constitution of the heavens, must remain more 

 or less imperfectly known. Now the spectroscope has placed in our 

 hands this power, which, though so essential, appeared almost in the 

 nature of things to lie for ever beyond our grasp ; it enables us to measure 

 directly, and under favourable circumstances to within a mile per second, 

 or even less, the speed of approach or of recession of a heavenly body. 

 This method of observation has the great advantage for the astronomer 

 of being independent of the distance of the moving body, and is 

 therefore as applicable and as certain in the case of a body on the 

 extreme confines of the visible universe, so long as it is bright enough, 

 as in the case of a neighbouring planet. 



Doppler had suggested as far back as 1841 that the same principle, on 

 ■which he had shown that a sound should become sharper or flatter if 

 there were an approach or a recession between the ear and the source 

 of the sound, would apjjly equally to light; and he went on to say that 

 the difference of colour of some of the binary stars might be produced in 

 this way by their motions. Doppler was right in that the principle is 

 true in the case of light, but he was wrong in the particular conclusion 

 which he drew from it. Even if we suppose a star to be moving with a 

 sufficiently enormous velocity to alter sensibly its colour to the eye, no 

 such change would actually be seen, for the reason that the store of 

 invisible light beyond both limits of the visible spectrum, the blue and 

 the red, would be drawn upon, and light- waves invisible to us would be 

 exalted or degraded so as to take the place of those raised or lowered in 

 the visible region, and the colour of the star would remain unchanged. 

 About eight years later Fizeau pointed out the importance of considering 

 the individual wave-lengths of which white light is composed. It is, 

 indeed, Doppler's principle which underlies the early determination of 

 the velocity of light by Roemer; but this method, in its converse form, 

 can scarcely be regarded as of practical value for the motions in the line 

 of sight of binary stars. As soon, however, as we had learned to 

 recognise the lines of known substances in the spectra of the heavenly 

 bodies, Doppler's principle became applicable as the basis of a new 

 and most fruitful method of investigation. The measurement of the 

 small shift of the celestial lines from their true positions, as shown 

 by the same lines in the spectrum of a terrestrial substance, gives to 

 us the means of ascertaining directly in miles per second the speed 

 of approach or of recession of the heavenly body from which the light 

 has come. 



An account of the first application of this method of research to 

 the stars, which was made in my observatory in 1868, was given by Sir 

 Gabriel Stokes from this chair at the meeting at Exeter in 1869. The 

 stellar motions determined by me were shortly after confirmed by Pro- 

 fessor Vogel in the case of Sirius, and in the case of other stars by Mr. 

 •Christie, now Astronomer Royal, at Greenwich ; but, necessarily, in con- 



