4.12 Mr. David Gill [May 29, 



its motion round the sun. But the velocity of the earth's motion in 

 its orbit is known with an accuracy of about one five-hundredth part 

 of its amount, and therefore, within that accuracy, we can allow 

 precisely for its effect on the relative velocity of the earth and 

 a AurigsB. When we have done so we get the annexed results for 

 the velocity of the motion of a Aurigse with respect to the sun. You 

 see by the annexed table how beautifully they agree in the Potsdam 

 results, and how comparatively rough and unreliable are the results 

 obtained by the older method at Greenwich. 



I believe that in a few years, at least in a period of time that 

 one may hope to see, we shall not be content merely to correct our 

 results for the motion of the earth in its orbit only, and so test our 

 observations of motion in the line of sight, but that we shall have 

 arrived at a certainty and precision of working which will permit the 

 process to he reversed, and that we shall be employing the sjjectroscope 

 to determine the velocity of the earth's motion in its orbit, or in other 

 words to determine the fundamental unit of astronomy, the distance 

 of the sun from the earth. 



I will take as another example one recent remarkable spectro- 

 scopic discovery. 



Miss Maury, in examining a number of photographs of stfUer 

 spectra taken at Harvard College, discovered that in the spectrum of 

 P Aurigas certain lines doubled themselves every two days, becoming 

 single in the intermediate days. Accurate Potsdam observations 

 confirmed the conclusion. 



The picture on the screen (Fig. 3) shows the spectrum of ^ AurigaB 

 photographed on November 22 and 25 of last year. In the first the 

 lines are single, in the other every line is doubled. Measures and 

 discussion of a number of these photographs have shown that the 

 doubling of the lines is perfectly accounted for by the supposition 

 of two suns revolving round each other in a period of four days, each 

 moving at a velocity of about 70 miles a second in its orbit. 



When one star is approaching us and the other receding, the lines 

 in the spectrum formed by the light of the first star will be moved 

 towards tl;e blue end of the spectrum, those in the spectrum of the 

 second star towards the red end of the S23ectrum. Then, as the two 

 stars come into the same line with us, their motions become at right 

 angles to the line of sight, and their two spectra, not being affected 

 by motion, will perfectly coincide ; but then, after the stars cross, 

 their spectra again separate in the opposite direction, and so they 

 go on. 



Thus by means of their spectra we are in a position to watch and 

 to measure the relative motions of two objects that we can never see 

 apart ; nay more, we can determine not only their period of revolution 

 but also the velocity of their motions in their orbits. Now, if we 

 know the time that a body takes to complete its revolution, and the 

 velocity at which it moves, clearly we know the dimensions of its 

 orbit, and if we know the dimensions of an orbit we know what attrac- 



