( 385 ) 



Although tlic differences between the results of (he separate years 

 seem to be real, I have applied only to the Z\ é derived from obser- 

 vations of (he north and the south limb the constant corrections 

 + 0".l and — 0".l. 



For (he observadous of 1900 — 1902 1 did no( know which 

 limb was obserxed. While, however, in the preceding years the 

 constant errors appeared to be small and in the mean for (he two 

 limbs were found to be 0".0, I thought myself justified in ncglecling 

 them altogether for 1900—1902. 



ÏÏ. In (lie second place the Lö had to be corrected for the errors 

 of longilnde. 



We tind to a sufticient degree of approximation (comp. also Inrestl- 

 (/(ftiou p. 31 — 32 ^) ) that the derivative of the declination relatively 

 to tiie mean longitude is: 



JÖ 



— = (f (1 -f c — c- COS 2X) c'O.s X -\- b cos {X — &) 

 dX 



+ 2 a e cos {2X — .-r) -\- 2 b e cos {2X - :t — 6) 

 where d = .sv';/ 8 = 0.398 



h = co.^ 6 sill I = 0.083 

 c = \ sill' e = 0.040 



For our purpose we may neglect the 3'^ ami the 4"^'' terms ; their 

 siiorl periods permit of their intluence being regarded as fortuitous. 

 Also the 2'^ term has provisionally been neglected, as its influence '■'), 

 may easily be accounted for afterwards. 



So there oidy i-emains the 1*' term, which has been tabulated by 

 Newcomb in his Table XI, and I multiplied it by the total errors 

 of the mean longitude. The errors of the true longitude depending 

 on ƒ/, gix'e rise in (l(f only to terms of very short aud of \ery long 

 period which could be neglected as being without influence on (he 

 resul (s to be derived. 



12. The A d corrected in this way were arranged for each year 

 into 18 groups according to the values of the argument of the 

 latitude ?/, in the same way as it was done for the A a according 

 to (he values of//, and (hen (he sums and the means for each grou}) 

 wove formed. 



I do not give here these annual means, but oidy (he general means 

 derived from the total sums. 



1) hi the formula on p. 32 S e K and 3 e H ougiit to be 2 e K and 2 e if . 



-) This tcini is llie influence of the enoi- in longitude on the latitude and 

 con?e({uently directly infhiences llie determination of' the longitude of the node, 

 but not tlia}, of the inclination. 



26 



Proceedings Royal Acad. Amsterdam. Vol. VI. 



