( 429 ) 



After the first objoct hnd Ix'Cii (ihsci-xcd, the iiistriiiiiciil was 

 carefiillv hi'ouirlil iiil<» the a/.iiimlh of the >ec()ii(l. The (lin'creiiee 

 helweeii the level I'eadiiius in ihe two posiliuns was alwavs smaller 

 than 1 (Inision = 5". 4 and for 'A'.\ of' flic oil ohscrvations smaller 

 than {)^^St. The interxal hctween two eorrcspoiidinu- Iransils of the 

 moon and the star was alwavs less than 2."i nniiiilcs and ^cncralU' 

 nmeh less. 



Finally 1 remark that, exelndijiji; 11)01 Fehr. 25, the \alne of the 

 [»arallaetie angle durin<i- the observations was always enclosed between 

 80° and 110^ or between 250' and 280°. 



o. J)('t('i-niui((t'unis of f'nu<'. Rati's of tin' cliroiiotiii'lcr. Ext*e|)t on 2 

 ni<ihfs (ii)Ol Oct. () and J 3), when I obser\ed ti-ansits of stars over 

 the nieridiaJK the eoi'reetions of my ehronometer to the mean time 

 of ('hiloan,u'o \\ere always deferinine»! by oi)S('i'\ations of altitude 

 For this 1 refer to my }»re\ions |)a|)ei* (( Contributions 1 p. (27(i) o), 

 where I have also gi\en the eorreefions of the ehroniunefer and the 

 daily rates for the period U)00 Oct.— J 1101 -Inly. As, however, 1 

 Iiaxe since been able to correct the observed altitudes for the rather 

 lariic division errors and the flexure of my insfriimeid (('omj). I.e. 

 [). (285) 12), I once more liive in the following»- table the corrections 

 and the rates for the whole period HM)] Jan. — 1902 .Alay. The rates 

 hold for the interval between Ihe date on the line abo\e and that 

 on the same line. In the next column are ,ui\en the mean lempera- 

 tnrcs; those foi- 1902 .Ian. () — Febr. 5 are inter[)olated values because 

 thermometer readings were wanting. (See table [). 430). 



Frtnn these data 1 have derived a formnla foi* the influence of 

 the temperatiux' on th<' rate. As the differences in temperature are 

 small 1 could (jidy delernnnc a linear iidluencc 1 found : 



1901 .Ian. 18- Apr. 2<S -f 0\91 -{-2(^.0 



1902 Febr. 5— May 17 -f 1 . 07 +25.8 



Mean + (). 99 -f 25.9 



1901 .lune 17— Sept. 10 +0. 19 +21.4 

 hence : 



lidluence of 1 + OM 7<S M- 



1)\' means of this temperature coeflicieid 1 have reduced all the 

 observed rates to + 24 and these reduced rales are gixcu in the 



1) For this computatiun Ion the rate in the inteival li'O-* Aitiil oD-Muy 1:2, during 

 which the chronomelci- \va.-< ti anspoiteil to Mayili and l)ack, was used erroneously. 

 If we exclude it, we find for the second .summer rate -(- 1^.05 at 26°.0 and for the 

 temperature coefUcient -fOs.lTI, which dilVers lillle from the value found aiiove. 



