474 Scientific Intelligence. 



value the difference of the factors for east and west observations 

 is found to be a maximum. And taking the average double par- 

 allax factor to be 3 -50 the last column gives the extreme correc- 

 tion that would have to be applied to the parallax, in which for 

 A/3 is to be taken its value for a star of the same degree of red- 

 ness. 



tan z cos (p — q) 



Decl. 





a 





E—W An 



y 



East 



West 



3-50 A/? 



0° 



30-3° 



-0-59 



+ 0-59 



—0-34 A/5 



+ 20 



52-3 



-0-94 



+ 0-94 



-0-54 



+ 40 



65-0 



-1-08 



+ 1-08 



-062 



+ 60 



74-4 



-1-14 



+ 1-14 



-0-65 



+ 80 



77'5 



-1-15 



+ 1-15 



-0-66 



Thus considering that the color effect for a most highly colored 

 star is not likely to be greater than 0*66 X 0*046" = 0*03* we feel 

 justified in concluding that any vitiation in our parallax results, 

 due to this cause, is presumably well within their probable error." 



Appreciating the rigorous methods employed in the work we 

 can accept with confidence the inferences drawn by the authors 

 from the analysis of the parallaxes obtained. 



These are best summarized by again quoting from the text : 



" The present investigation furnishing, as it seems to us, a homo- 

 geneous series of parallax values, we have thought it of interest 

 to make a number of classifications of the results independent of 

 the material hitherto obtained by other observers. While we 

 cannot claim any very great precision for the individual results, 

 determined as these are from such a comparatively limited num- 

 ber of observations, yet we do attach considerable weight to the 

 mean values of the various groups we have formed, and consider 

 these values of unquestionable significance. 



These classifications are comprised in the five following tables. 

 Table I gives an arrangement in order of the size of the proper- 

 motion of the star under consideration ; Table II, a similar one 

 in order of the star's magnitude ; Table III, in order of the size 

 of the parallax ; Table IV, in order of Right-Ascension, sub- 

 divided into two classes north and south of Declinations + 30°, 

 and finally Table V, according to the spectral type and classes as 

 given in the Draper Catalogue." 



Of the results shown in these tables the following are the most 

 interesting : 



Table I reveals, as might be expected, a distinct interdepend- 

 ence of parallax and proper motion, especially in the group with 

 proper motions exceeding 1". The data of Table II point to a 

 small but undoubted relation between parallax and magnitude. 

 Table III emphasizes the connection between proper-motion and 

 parallax, none of the negative parallaxes corresponding to proper- 

 motions greater than 1 , and a marked progression being very 

 apparent. 



