I. — Astronomic Observations. 



Referring the daily values of the absolute correction to the mean date of 

 the values of the last double series we get : 



31st July (civil) (p.m.), t.. = 10'' 26"' 48'. ..K; - + 3" 33"' S'^- 

 which represents the absolute correction of the chronometer on Greenwich time 

 before starting on the return journej' from Fort Portal to Entebbe. On 

 reaching the latter place it was found impossible to get a new telegraphic 

 comparison with Mombasa, as on the outward journey. Hence proceeded to 

 the determination of the absolute correction of the chronometer, using for 

 Entebbe the longitude 2'' 9™ 47' East Greenwich given us by the competent 

 local authority. The result was : 



16th August.— Obs. No. 110 ; /, = 10" 29"" 7' ... K, = + S*- 33-" 29"- 9 



16th „ ,, 111 ; ,, = 10 31 13 „ = 28-5 



17th „ „ 112; „= 2 43 .5.5 „ = 25-2 



17th „ „ 113; „- 2 46 1 „ = 23-8 



AVe see (( /'cw/i that in this interval of little over 16 hours the movement 

 of the chronometer indicates a strong variation, such as had never occurred 

 during the whole journey. Instead of taking the mean of these values, it was 

 thought expedient to use the restdts alone of the two series of 16th August 

 observed immediately after the arrival at Enteb1>e. From these we get : 



16th August (civil), (p.m.) ; lO^' .-iO'" 10'. ..Kj = + 3'' 3.3'" 29'- 2 



so that the daily correction of the ehronnmeter to lie used in the interval from 

 31st July to 15th August was : 



Ki = + p-510. 



The question now was to see what degree of confidence might l)e placed 

 in the daily corrections which had so far been obtained. From the fact tliat 

 the longitude of a few points was determined lioth on the outward and the 

 return journey, we were offered a means of control which, if it stood alone, 

 woidd not be absolutely safe, since it was alwaj-s possible that the erroi'.s liy 

 which the accepted daily corrections might be afteeted might lie such, in 

 magnitude and sign (plus or minus), as to lead to longittidinal results apparently 

 concordant though really very incorrect. As, however, there were several 

 points determined under such conditions, so that in some cases we could 

 ascertain the degree of concordance in the longitudinal residts, from this might 

 be inferred Vioth the practical value of the daily corrections that had been 

 adopted, and the measure of confidence that might be placed in the positions 

 obtained from the astronomic observations. 



317 



