1872.] 



some Stars to wards or from the Earth, 



387 



parisons which necessarily make the numerical estimations given further on 

 less accurate than I could wish. Even when spectroscope C, containing 

 four compound prisms, and a magnifying-power of 16 diameters are used, 

 the amount of the change of refrangibility to be observed appears very 

 small. The probable error of these estimations is therefore large, as a shift 

 corresponding to five miles per second (about T \j of the distance of D 1 to 

 D"), or even a somewhat greater velocity, could not be certainly observed. 

 The difficulty arising from the apparent smallness of the change of refran- 

 gibility is greatly increased by some other circumstances. The star's light 

 is faint when a narrow slit is used, and the lines, except on very fine nights, 

 cannot be steadily seen, in consequence of the movements in our atmosphere. 

 Further, when the slit is narrow, the clock's motion is not uniform enough 

 to keep the spectrum steadily in view ; for these reasons I found it neces- 

 sary to adopt the method of estimation by comparing the shift with a wire 

 of known thickness, or with the interval between a pair of close lines. I 

 found that, under the circumstances, the use of a micrometer would have 

 given the appearance only of greater accuracy. I wish it, therefore, to be 

 understood that I regard the following estimations as provisional only, 

 as I hope, by means of apparatus now being constructed, to be able to get 

 more accurate determinations of the velocity of the motions. 



Sirius. — The comparison of the line at F with the corresponding line of 

 hydrogen was made on several nights from January 18 to March 5. Spec- 

 troscope C and eyepieces 2 and 3 were used. These observations confirm 

 the conclusion arrived at in my former paper, that the star is moving from 

 the earth ; but they ascribe to the star a velocity smaller than that which 

 I then obtained. 



These observations on different days show a change of refrangibility 

 corresponding to a velocity of from 26 miles to 36 miles per second. The 

 part of the earth's orbital motion from the star varied on these days from 10 

 miles to 14 miles per second. We may take, therefore, 18 to 22 miles per 

 second as due to the star. 



The difference of this estimate, which is probably below rather than in 

 excess of the true amount, from that which I formerly made may be due in 

 part or entirely to the less perfect instruments then at my command. At 

 the same time, if Sirius be moving in an elliptic orbit, as suggested by Dr. 

 Peters, that part of the star's proper motion which is in the direction of the 

 visual ray would constantly vary*. 



* Dr. H. Vogel at Bothkamp seems to have repeated my observations on Sirius with 

 the necessary care. He says (Astron. Nachr. JS"o.gl864) : — " Mit der eben beschricbenen 

 Anordnung gelang es Herrn Dr. Lohsc unci mir am 22. Miirz (1871) bei ganz vorziig- 

 licher Luft die Nichtcoincidenz der drei Wasserstofllinien Ha, H /3, und H y, der 



Geissler'scben Kohre mit den entsprechenden Linien des Siriusspectrum zn seben 



mit Beriicksicbtigung der Geschwindigkeit der Erde zur Zeit der Beobachtung berechnct 

 sich die Geschwindigkeit mit welcber sich Sirius von der Erde bewegt zu 10 Meilen in 

 der Secunde, wogegen Procyon sich 13 8 Meilen in der Secuude von unscrer Erde ent- 

 fernen wiirde." 



