252 



POOR 



Table XII. 



Observation minus Computation. 



25. The corrected elements represent the observations con- 

 siderably better than Bauschinger's. The residuals, however, 

 still show the systematic character which was mentioned in 7. 

 The differences in right ascension are all small and the signs 

 are irregular, but the differences in declination are larger and 

 the signs of the terms show a systematic arrangement. With 

 the exception of the last normal all the differences in declina- 

 tion of the first appearance are negative, while in the second ap- 

 pearance, with the exception of the .second, they are all posi- 

 tive. The large difference in declination, -f 2^.49, obtained for 

 December 21, 1890, is completely accounted for by the fact that 

 the normal place rests on but five discordant observations, one of 

 which differs by more than 10" from the mean of the other four. 



Taking the mean by weights of the differences, both in a 

 and d, for each appearance I have, 



