Dreyer — New Determination of Constant of Precession. 619 



stars were determined with an interval of about forty years, with first- 

 class instruments, and by eminent observers, and as Nyren took special 

 precautions to clear Bessel's positions of constant errors by comparisons 

 with the Dorpat and Armagh Catalogues, the material thus prepared 

 must be considered very promising for k good result. It was there- 

 fore very surprising that the comparison between Schjellerup's and 

 Bessel's Right Ascensions showed a well-established difference 

 S - JB = - O''lo5, which, after reducing the equinoxes of the two 

 Catalogues respectively to "Wolfers' equinoxes for 1830, and to an 

 equinox for 1862, deduced from Greenwich and Paris observations, 

 did not become smaller than - C^'OGS. This indicated a rather con- 

 siderable correction to the adopted constant. Nyren's final result was, 

 for 1800, 



50"-1882. 



The deviation of this result from Bessel's and Struve's is very 

 much larger than the accuracy of modern observations should lead one 

 to think possible. In my Paper on the personal errors in transit ob- 

 servations (these Proceedmgs, ser. n., vol. ii., p. 518) I have mentioned 

 a fruitless attempt to find out whether either Bessel's or Schjellerup's 

 Right Ascensions show any sign of errors depending on the magnitude 

 of the stars, as such an error, if it existed, might explain the difference 

 between the constants of precession found from bright and from faint 

 stars. But although it turned out to be impossible to find directly 

 whether the personal errors of Bessel and Schjellerup depended to any 

 sensible extent on magnitude, there are many circumstances which 

 must lead us to believe that this was not the case. Argelander has 

 compared Bessel's Eight Ascensions with those of W. Struve (''Posi- 

 tiones Mediae"), and found 8 - B = ~ 0'-043 ; but as the reduction of 

 Struve to Wolfers' Pundamental- System is zero, while Bessel's is 

 - 0'-012, the difference between Wolfers' {i.e. Bessel's Fundamental 

 Catalogue) and Bessel's zones would only appear to be - 0'-031. As 

 to Schjellerup, we possess a comparison between his catalogue and 

 that deduced by Copeland and Borgen from their observations of stars 

 between the equator and - 2° declination ; and as these observers used 

 the same standard system (Nautical Almanac), the result, C and ^- /S' 

 = - 0'-005, leaves scarcely any doubt that Schjellerup's Right Ascen- 

 sions are free from constant errors. 



Nyren's result must therefore be affected by some considerable 

 errors in one of his equinoxes, and a later investigation by himself 

 has, in fact, shown conclusively that the observations of the sun, made 

 in Greenwich (and on which his equinox for 1862 chiefly depended), 

 require a large positive correction to make them agree with the ob- 

 servations made at Pulkowa, Edinburgh, Cambridge, Paris, and 

 "Washington ; and their fair accordance with the Paris observations of 

 1856-59 must be due to the chapter of accidents. But it is evident 

 that this positive correction will do away with the greater part of the 

 large difference between Schjellerup and Bessel, and consequently 



K. I. A. PROC, SEll. II., VOL. III. — SCIENCE. 3 I 



