1881.' 



on the Distances of the Stars. 



617 



the error of pointing the telescope to the top edge of a penny-piece 

 instead of to the bottom edge when the pcnny-pieco was fifteen or 

 twenty miles oli". 



Ordinates indicate parallax. 

 Dots indicate observations. 



Still, however, the entire quantity to be measured is so small that 

 the errors, minute as they are, bear a large proportion to the parallax. 

 In this lies the weakness of such work. By sufficiently multiplying 

 the numbers of the observations, and by discussing them with the aid 

 of the method of least squares, considerable confidence may be 

 attached to the results. 



Groomhridge 1830. 



This star has been the subject of much parallax work. It has a 

 proper motion of seven seconds annually. Mr. Huggins or Mr. 

 Christie could perhaps ascertain by the spectroscope what its motion 

 may be in the line of sight. From the theory of probabilities it may 

 not improbably be nine seconds. We shall, however, take it at seven 

 seconds. The parallax has been determined by Struve and by Briinnow. 

 It is very small, being one-tenth of a second. The actual velocity of 

 1830 Groombridge must therefore be 70 radii of the earth's orbit per 

 annum, or 200 miles per second. 



Newcomb has employed this result to throw light on the question 



