580 Meeting of the British Association. [Oct., 



their constitution, and the rectification of astronomical speculations 

 previously entertained. 



We are accustomed to apply to the stars the epithet fixed. 

 Defining as fixity, invariabiHty of position as estimated with 

 reference to tlie stars as a whole, and comparing the position of any 

 individual star with those of the stars in its neighbourhood, we 

 find that some of the stars exhibit "proper motions" — show, that is, 

 a progressive change of angular position as seen from the earth, or 

 rather as they would be seen from the sun, which we may take for 

 the mean annual place of the earth. This indicates hnear motion 

 in a direction transverse to the line joining the sun with the star. 

 But how shall we determine whether any particular star is ap- 

 proaching to or receding from our sun ? It is clear that astronomy 

 alone is powerless to aid us here, since such a motion would be 

 unaccompanied by change of angular position. Here the science of 

 optics comes to our aid in a remarkable manner. Suppose that we 

 were in possession of a source of hght capable of exciting vibrations 

 of a definite period, corresponding, therefore, to light of a definite 

 refrangibility. Then, if the source of light and the observer were 

 receding from or approaching to each other with a velocity which 

 was not insensibly small compared with the velocity of light, an 

 appreciable lowering or elevation of refrangibility would be pro- 

 duced, which would be capable of detection by means of a spec- 

 troscope of high dispersive power. The velocity of hght is so 

 enormous — about 185,000 miles per second — that it can readily be 

 imagined that any motion which we can experimentally produce in 

 a source of light is as rest in comparison. But the earth in its 

 orbit round the sun moves at the rate of about 18 miles per second; 

 and in the motions of stars approaching to or receding fi'om our sun, 

 we might expect to meet with velocities comj^arable with this. The 

 orbital velocity of the earth is, it is true, only about the one ten- 

 thousandth part of the velocity of light. Still the efi'ect of such a 

 velocity on the refrangibility of light, which admits of being easily 

 calculated, proves not to be so insensibly small as to elude all chance 

 of detection, provided only the observations are conducted with 

 extreme delicacy. 



But what evidence can we ever obtain, even if an examination 

 of the light of the stars should present us with rays of definite 

 refrangibility, of the existence in those remote bodies of ponderable 

 matter vibrating in known periods not identical with those corres- 

 ponding to the refi'angibilities of the definite ra)-s which wc observe ? 

 The answer to this question will involve a reference to the splendid 

 researches of Professor Khchbofi", which led him to make a careful 

 comparison of the places of the dark lines of the solar spectrum 

 with those of bright lines produced by the incandescent gas or 

 vapour of known elements; and the coincidences were in many 



