( 434 ) 



4''' and 5^'' colunins coiila'm the coi-reeliojis of L /. rpsiiliinu from 

 llic n(l()|)le(l correclioiis of the i-iL>iit ascension and the (leclirialioiiof llie 

 moon, tlie (5^'' contains the corrected values of Z. L and the 7''^ (he mean 

 results for L L from the observatioJis of each dav. The hist coiiinni 

 shows the derivatives of the hm^'itnde rehitivelv to the corrections 

 of the chronometer. The dei'ixatives rehitivelv to the other eh'ments 

 are omitted to sa\'e room. I shall oidv ui\(^ hereafter their mean 

 \alnes for the ohserxations of each lind). (See table p. 435). 



Nou' the (|iiestion arises >\hat wei.uhts must he assiuned to these 

 results and how thev w HI he hest cond)ined. 



First it miisi he hoi-ne in mind that in HM)! Jauuarv the transits 

 A\ei'e ohserxed o\ei' 2 threads onlv and ni all the follo\x"in_^' ohser- 

 \alions oxer 7, e\ce|»t in \\\<) cases o\er (5 thi'eads. Tiierefore W3 

 liave assi<iiied the \\ei,uht Yj *** ^''^^ ^'I'^^f mentioned observations. 

 Secondlv we must pav attention to special uid'axom-able circum- 

 staives dui-inu' the observatiojis, and then we lind thai for that 

 of 1902 Jaji. 24 the star was onlv faintiv \ isiblc Ihrou^ii the 

 iia/\- atmosphere, and that for the lii-st obser\ation of 1 902 March 25 

 an uncei'taintv [u-evails about one of the level-readin^us. To these 

 observations also the xveiii'ht '/., was assigned 'j. 



Finally we had to pay lieed to nnfavonral)le geometrical cojiditions 

 for some obser\ations, xvhich might highly increase the ijdiuence of 

 some reductioJi elements, ami also to special cases of uncertainty 

 in ojie of those elemejits. I liaxe already mentioned that on 1901 

 F'ebr. 25 the observation had been made \\ith very unt;n()ni'al)le 

 relatixc positions of the moon and the star. The values of the 

 ])arallactic angle were for the moon ami the stai- 124 and 117 

 respectively, while iu the other cases the most unfavourable \-alne 

 was 10(S' and the greatest diUei-ence betxNcen the \alues foi' the 

 moon and the star amounted to o . We now see that owing to this 

 the derivative relati\ndy to the chronometer-correction is excejttionally 

 large, while also those relatixely to the latitude of the station and 

 to the moon's declination ha\-e the largest value heiv. 



A special uncertainty on certain days can only be ex[)ecte(l 

 in the chronometer-correction. A measin-e for this uncertainly is 

 given 1\\ the interval between the observatio]i of the moon and 

 the nearest timc-detei-mination, and il appears that this intcrxal 

 for almost ail the obsei-vations lies between and 3 days, bui 

 for 1901 Febi-. 25 amounts to (5 and for 1901 Alaivli 3 to 12 

 da\s. Another liuie-delernnuation had slill been made on Febr. 2<S 



1) This has been i.-ikeii into nccuiinl in tlie derivnlinti of llir incaii resnll r<i|- 

 March 25 in the table above. 



