Group of the Pleiades. 827 



headed lltiVr. The number., in this column are the direct differ- 

 ences of the two triangulations executed by Elkw at rale; and it 

 is quite plain from tbem that tbe photographic observations agree 

 with those of the heliometer, about as well as do tbe latter inter ne. 

 The part of* tho table relating to deelinations furnishes equally satis- 

 factory results. We may therefore conclude that the present photo- 

 graphic observations are entitled to be taken into consideration, if 

 it be desired to make a study of proper motion-;, or to form a defini- 

 tive catalogue of the Pleiades. 



It has Dot appeared desirable to compare directly the photographic 

 and belipmeter values of tbe distances and position angles. For the 

 use ofadifferent star as tbe origin of measurement if) tbe twoca les 

 would lead to differences of considerable magnitude, and might tend 

 to make the results of a comparison misleading. Moreover, tbe 

 comparison of right ascensions and declinations given here .-hows 



that the rather large systematic error- in position angle found by 



Elkin,* upon examining Gould's results from tbe earlier Ruthee- 



1 1 .ico photographs, do not appear in the present case. In fact, there 

 is no very apparent reason why tbe method of orientation adopted 

 by Rutherfurd should lead to erroneous results. 



The probable errors of the distances and position angles for the 

 various stars are given in tbe table of Section VI I. : the correspond- 

 ing one:-: for the resulting right ascensions and declinations may 



easily be computed/ If we- let: 



i\ and r v be the probable errors of distance and position angle, the 



latter being taken in arc of a great circle, 

 then we have for b' — b : 



Probable error of 5' — 5 = |/C08 a 7t r} -f sin 2 * r p a 

 and for COS h (V — a): 



Probable error of COS A (a — a) = y sin' 7t r, 2 -f- cos 2 ft r p 2 . 

 It does not seem worth while to make a table of probable errors 



computed in this way for the several stars. But it will be of interest 

 to set down the mean values. Tbe average probable error, accord- 

 ing to the table in Section VII. may be taken as: 



rfc o".o4 for the distances, 



=b o".o6 for the position angles. 



The average value of sin 2 n 18 \ t and that of COS 3 rt also \\ SO 

 * Astronomical Journal, No. 197, p. 33. 



