( 423 ) 



The latter would liave required a s|)ecial iiiveslii>atioii of those 

 observations, wliieh 1 <lid not intend. As lo tli<> former point too 

 little eertainty sctMned lo (»\ist. For alllioiigli the ohlicjiiily of the 

 eelij)tie adopted liiially \)y Nkwcomr has for the 2'"' |iarl of llie 

 19''^ cenlnry values wliieh ai-e larger hy 0". 2 ()".!] thaji those 

 from Hansen, yel ihe value derixed direetly by Nkwcomb from the 

 observations of ihe suji oidy agrees well with Hansen for this period. 

 We have : 



1860 1900 



23^27' 23=27' 



Hansen 26"74 8"03 



Newcomb 27.00 8.26 



Newcomb Sun 26.81 7.97 



()n closer eonsidei-alion, however, we see that we t)ughl not to use 

 the absolutely most pi-obable value of the oblicpiity, but that it is 

 best to adopt for each sei'ies of ol)sei-vations of the moon the value 

 wliieh is derived from the observations of the sun whieh have been 

 made at the same time and with the same instrument and have been 

 reduced in the same way. For it may l)e easily seen that in this 

 way we eliminate for the gi'eater part the ei-rors in the refraction, 

 the fle.xure of Ihe instrument etc. Some systematic errors, namely 

 those depending on the conditions of the ol)servalions which are 

 dilfei'eut for the sun and for Ihe moon will be retained undiminished 

 in the results, yet I hold Ihe course |)roposed here for the best, 

 uidess elaborate researches on the errors of the oi)servat ions have been 

 made. In this way, at any rate we may gel some insight into the 

 iidluence which the values adopted for the reduction elements of the 

 observations have on the elements derived for the jilanc of the 

 moons orbit. 



20. For Ihe jjcriods discussed l)y Xkwco.mh and l)y myself we 

 must therefore find the values of the obli(|uily of the ecli[»lic which 

 folloxv directly from the observations as iIh^ are gixen in the Annals 

 of Greenwich and of Washington. 



For (Treenwich I could use the resulting values given for each 

 year in the (rreenirich ohsci-rittioiis ajid derive from them Ihe correc- 

 tions of Hansen's obli(|uity. For 1895 — 1902 these corrections \vere 

 coml)ined into one mean, i»ut the |)eriod 1862 — 1874 wa^ divided 

 into two [)arts: 1862 — 67 and 1868 — 74, because in the year 1868 

 Stone's refraction constant had come into use, which is smaller by 

 about O-S^/o than that of Bessel. There are also dilFerences between 



