558 BIBLIOGRAPHY OF HERSCHEL's WRITINGS. 



Herschel, W.: Synopsis of the Writings of— Continued. 



A. D. Yol. p. 



1796 HG 167 A second layer of stars will be more extensive, etc., etc. 



108 The hypothesis of an equality and an equal distribution of stars is too 



far from being strictly true to be laid down as an unerring guide. 

 163 The stars of the 1st and 2d class, scrupulously examined, prove that 

 we must admit them to he either of different sizes or jilaced at dif- 

 ferent distances. Both varieties undoubtedly take place. 

 168 Thtis it appears that in the classification of stars into magnitudes 

 ° there is either no natural standard or none that can he satisfiictory. 



, 169 If any dependence could be placed uj)on the method of magnitudes, 



it would follow that certain of Flamsteed's stars had undergone a 

 change in their lustre. [Examples given.] 

 170 Flamsteed did not compare the stars to each other, but referred each 

 of them separately to its own imaginary standard. 

 A short inquiry into the confidence to be given to the method of mag- 

 nitudes may be of considerable use. 



170 In Flamsteed's observations an error of 1™ in the brighter classes and 



li™ in the fainter would hardly deserve attention. 



171 In comparing the ohservations of different astronomers larger errors 



may he expected. 



172 Example. From Flamsteed's and Lacaille's ohservaticms of (i 



Leonis 'we may conclude that this star is now less brilliant than for- 

 merly. 



173 I place each star, instead of giving its magnitude, into a short series 



[sequence] constructed upon the order of brightness of the nearest 

 proper stars. 



174 The Greek letters now affixed to the names of stars do not point out 



their order of brightue-ss, except for the few brightest stars of eacli 

 constellation. [Examples.] 



175 A doubt may arise whether any succession of brightness might he 



argued from the very first, second, or third letters of the alphabet, 

 when we find them now arranged thus, as /3 aCassiopece, /3 aCancri, 

 y ft Aquilce, etc. 



177 Lalande, Figott, and Goodericke have used the method I propose 



in special cases. 



178 Simple as my method is in principle, it is very laborious in its pro- 



gress. I began to use it 14 years ago. 

 178 My first design was to draw each whole constellation into one series. 

 Accordingly I began, July 16, 1781, to arrange the stars in Ophiuchus, 

 thus : a, /i, d, Z, ?A "^^ Xi ^- The defect of this arrangement was 

 that we do not always have a proper connection of the steps of the 

 series; the intervals being too great in some cases, too small in 

 others. 



178 To get over these difficulties I marked the stars by degrees, three in 



a magnitude 1', 1", V" ; 2', 2", 2'", etc., as "May 12, 1783; order of 

 the stars in Bocitis "al', s 2", 77 2'", r i^ ^ '^', etc." 



179 Difficulties with this plan. 



180 Other methods tried. 



181 The method finally adopted explained. [The method of sequences.] 



183 Difficulties in carrying out the method stated. 



184 These observations are of importance, as will appear when we remem- 



ber the great number of alterations of stars which have certainly 

 happened within two centuries. 



185 Who would not wish to know the permanency of the lustre of our 



Sun ? If it be allowed to admit the similarity of stars with our sun, 



