408 Mr. David Gill [May, 29, 



of its use for the accurate measurement of the velocity of the motions 

 of stars in the line of sight.* 



It is beyond the province of this lecture to enter into history, but 

 it is impossible not to refer to the fact that the chief impulse to 

 astronomical work in this direction was given by Dr. Huggins, our 

 chairman to-night — nay, more, except for the early contributions of 

 Fraunhofer to the subject, Dr. Huggins certainly is the father of 

 sidereal spectroscopy, and that not in one but in every branch of it. 

 He has devised the means, pointed the way, and, whilst in many 

 branches of the work he still continues to lead the way, he has of 

 necessity left the development of other branches to other hands. 



From an astrometer's point of view the most important advance 

 that has been made in spectroscopy of recent years is the sudden 

 development of jirecision in the measures of star motion in the line 

 of sight. The method remained for fifteen or sixteen years quite 

 undeveloped from the condition in which it left the hands of 

 Dr. Huggins, and certainly no progress in the accuracy attained by 

 Dr. Huggins was made till the matter was taken up by Dr. Vogel at 

 Potsdam. At a single step Dr. Vogel has raised the precision of the 

 work from that of observations in the days of Ptolemy to that of the 

 days of Bradley — from the days of the old sights and pinnules to 

 the days of telescopes. Therefore I take a Potsdam observation as 

 the best type of a modern spectroscopic observation for description, 

 especially as I have recently visited Dr. Vogel at Potsdam, and he 

 has kindly given me a photograj)h of his spectroscope, as well as of 

 some of the work done with it. 



A photograph of the Potsdam spectroscope attached to the equa- 

 torial is now on the screen. [Description.] 



The method of observation consists simply in inserting a small 

 photographic plate in the dark slide, directing the telescope to the 

 star, and keeping the image of the star continuously on the slit during 

 an exposure of about an hour ; and this is w'hat is obtained on develop- 

 ment of the picture. 



If the star remained perfectly at rest between the jaws of the slit 

 the spectrum would be represented by a single thread of light, and of 

 course no lines would be visible upon such a thread ; but the observer 

 intentionally causes the star image to travel a little along the slit 

 during the time of exj)osure, and so a spectrum of sensible width is 

 obtained. (Fig. 1.) 



You will remark how beautifully sharp are the faint lines in this 

 spectrum. Those who have tried to observe the spectrum of Sirius 

 in the ordinary way know that many of these fine lines cannot be seen 

 or measured with certainty. The reason is that on account of 

 irregularities in atmospheric refraction, the image of a star in the 



* The older metliorls enabled us to measure motions at riglit angles to the 

 line of sight, but till the spectroscope came we could not measure motions in 

 tlie line ot" sight. 



