( 210 ) 



ponent, and llio coiislaiil co-ordiiuites of the mean pole, 1 ol)laiiie<l 

 the coni|ni(e(l \alues of ,/■ and // of Table IV. 



T A 15 L E IV. 



If we look al the columns () — C, and eom])are them with the 

 ('oiTes})ondiiig ones in Tables II and III, we see, how nitieh the 

 agreement between obser\ation and eoniputation has been iin|)roved. 

 As mean error we now^ tind ±()".()15(), wliieh is even less than 

 the nneei'tainty (m. e. = ± U".02) of the co-ordinates deduced from 

 observation. 



The calculated polar ciir\'e, obtained by adopting formnla (5) has 

 been represented in the ligiii'e by dashed lines. In Judging the 

 agi'eement it nniy be useful to remind, that the unit of Vioon second 

 of arc, in which (he nnnd)ers of the tables are expressed, re[)i'esents 

 a length of 31 millimeter on the surface of the earth; the scale of 

 the tigure is nearly I -. 77.5. 



Now that we may hold it proved by the foregoing comparisons 

 that the observations of 1904 up to J91J may all be represented by 

 the same elements with the exception of the increased value of c*.^, 

 the (|uestion arises, in how far the sauie elements also satisfy the 

 former observations. Small variations in the yearly ellipse can be 



