April 24, 1896.] 



SCIENCE. 



619 



as well as on good nights. ISTumerous ob- 

 servations have been made on miscella- 

 neous stars, with and without the screens, 

 to determine at any time the personal equa- 

 tion depending upon the apparent bright- 

 ness of a star ; but these have not yet been 

 reduced. 



The values of the parallaxes resulting 

 from the present solutions are given in 

 Table I. The average number of observa- 

 tions entering into each value is 35. All 

 the stars have been solved in the regular 

 manner except the last two, which pre- 

 sented a peculiar case to be explained in 

 "the following. Of these 85 Pecjasl was re- 

 duced with its second comparison star only 

 with an inappreciable parallax as the result. 



In Table II. are presented all the pre- 

 vious determinations of parallax that I 

 have found for the stars of Table I., exclud- 

 ing some older and much more uncertain 

 values. The several columns are suffi- 

 ■ciently explained by the headings except 

 the third, and here the letters given denote 

 different methods of observing, as follows : 



H. By the heliometer. 



J/j. By the filar micrometer attached to 

 the equatorial telescope and from measures 

 of distance and position angle combined ; M,, 

 from distance alone; il£, from position angle 

 alone; M^, from differences of declination. 



Z. By measures of the zenith distances of 

 the pai-allax star alone, in the meridian. 



R. By observations of right ascension in 

 the ordinary manner. 



P. By measurement of photographs. 



T. By differences of meridian transits em- 

 ploying special comparison stars. 



In the case of « Lyrce the letter c, in the 

 fifth column, indicates that the measures 

 were made from the companion star. The 

 value given for this star from Peters is the 

 only absolute parallax in the table. For ,a 

 Cassiopeue and « Lyrce I have included my 

 own results, assigned the several independ- 

 ent values different weights, somewhat ar- 



bitrarilj', and combined them all into one 

 mean value given in the table. 



Name or Stae. 



ju Cassiopeias 



Lalande 16290 ... 

 Lalande 15565. . . . 

 Lalande 18115, pr 

 5 Ursffi Majoris... 



20 Crateris 



V Serpentis 



1? Herculis 



Lalande 30694 



70 Ophluohi 



aLyree (Vega).... 



Lalande 47019 



85 Pegasi 



As regards the apparent uncertaintj' of 

 results, the present method cannot take 

 rank with the best work done with the 

 heliometer and the filar micrometer, or per- 

 haps with that done by the aid of pho- 

 tograph}'. As shown by Dr. Kapteyn's re- 

 fined determinations, however, this method 

 seems singularly free from systematic error, 

 and its trustworthiness maj^ be higher than 

 that assigned by its accidental error alone. 

 In the present series a material reduction of 

 the apparent uncertainty of any single 

 night's observation of a given star would re- 

 sult fi-om diminishiuff the weights of the 



