( 419 ) 

 l<S52.n .V 



l8(j<s.ri 



181)8.5 

 wlinice 



y = 806=.9 + Ï'XM (/ — 187(>.0) 

 The annual variation of .V arcordiiiu- lo llio now loi-ninla is 

 |)ra('(ically e(|ual lo the \alue found befoi'e. 'i'lie "values of O — ( 

 \\ hirh are joined to those of A' in the tai)le alxtve show, however, 

 that the svsteniatic discordances are still ^'reat and the outstandiiifi: 

 errors of the three noi'uial values are even somewhat greater than 

 before. 



In the second place T have detei-niined anew the coefiicieni of 

 the inequality. 1 have n(tt done this l)y dei-iving directly values for 

 each year from the corrected h and /• and then taking the mean 

 of the separate values, as by doing so 1 should have obtained a too 

 large coeflicient. Hut I ha\e represented the corrected h and k by the 

 formulae : 



h' ^ — c( sin N 

 k' = -j- « cos N 



assuming for .V the computed values. 

 In this ^vay I derived 



from the //' a =z -^ J ".23 



from the /' = + 1 -34 



from the two cond)ined « =: -|- J .28. 



Hence the fornndae for //' and /' become 



h' = — 1".28 sin [307" + 19°.4 {t — 1876.0)] 

 /•' = 4- 1".28 ros [307° + 19^4 (t - 187(3.0)1 



while from the theoretical term II, the -lovian Evection, there would 

 follow : 



hjj — — 0".88 sin [329' -f 20°. 68 (t — 1876.0)j 

 ku — 4- 0".88 cos [329" + 20°.68 {t — 1876.0)]. 

 There still remains a considerable diffeience bet^veen the empirical 

 and the theoretical values; for 1902 the difference between the 

 arguments amounts to as much as 57^. Therefore we cannot but 

 conclude that still other inequalities join their intluence to that of 

 the Jovian Evection. 



17. The expressions obtained for k and /• according to the two 

 I)receding sections are therefore : 



h =: h,; + h' -f liiii 



k == ^v -f h' -f km -f- k[y 



28* 



