( 417 ) 



ïleiK-e, l)v soIxiiiLi- ilic 10 (M|iiali(nis In least s(|iiar('s, wc dci-ixc: 



LI, — J- o".(»2 <v.s' (2 .T 4- w r— :> k) 



Lk — o".uo x//, (2 .T -f- 8 K— 5 7:-') 

 wliile ac't'ordiiiii- (o llicorv llic iwo (•(tcriiciciits oiiulil lo l>c -|- (r.(><S. 



Thus it seems that the term 1 is not eoiiliniicd l)\ the ohsei-xa- 

 tioiis ns(Ml lieiv. I shall show later that a somewhat modiliefl eom- 

 piilatioii leads to the same result. It Diay l»e that in the years 

 considered Ikmv another ine(|Malitv has neutralized its ellect. 



In the second [tlaec we shall trv to liml w hal mav he dei-i\ed IVom 

 \\\(' (). — (\ I about tii(^ term III. 1 lia\e therefore arran.u'cd the 

 Z_// and Li,- aeeor(linu to tli(^ \alues of 2 rr .'^ ,1 -[- 7° and found the 

 follow inu' jiormal xalues \x hieli. liowever, do not cover a full i'e\oin- 

 tion of the aruument. 



I lia\e re|M-eseuted these values bv the e.xpressions : 

 Lh = a + b cos (2 jr 8 ,/ + 7°) 

 Lk — a + // xln (2 .T -3 ,/ + 7^) 

 and l)\- sohinu' the (M|natio]is hy least squares I lound : 



\> — — 0".55 

 // = — 0".40 

 or from the Lji and LI- combined: 



h — h' — — 0".40 

 w hile the theoretical \alue is - ()".o2. 



The empirical determinalioji of tiiese coefiici(Mits is still ^ cit un- 

 certain. In so far as the observations haxe a c(»nclusi\e force, we 

 may say that they conlirm the inequality III. 



28 

 Proceedings Royal Acad. Amsterdam. Vol. VI. 



