Appendix B 



Fort Portal. 



Of this the mean for the doul)le series required by the above-mentioned 

 inversion of the glass roof of the artificial horizon is : 



2 1 st July, 4M 5"^ ... K^ = + 3^ 1 5"^ 1 6^- 



21st „ 19 54 „ = + 17-0 



22nd „ 19 54 „ = + 19-5 



23rd „ 20 7 „ = + 22 • 4 



Referring these values to the date coinciding with the first of them, and 

 noting the hour indicated l)y the chronometer, we get : 



21st July (civil)— (p.m.) : /,. = 10'' 57"' 49^..Kl = 3l' 15'" 15«-8 



as the mean on Greenwich mean time. 



From the determinations of the time at Bujongolo on the 10th, 11th, 

 and 12th July (astronomic dates), the results of which have l)een given at 

 p. 312, we obtain the absolute correction of Xo. 1 chronometer. Referring all 

 the values to the date of the last, and passing from the absolute correction 

 of the chronometer to the absolute correction on Greenwich, we get : 



13th July (civil)— (a.m.) : /, = 2'' 51»' 6^..K, = 3'' 15"' 28*-4 



from which in the interval between the 13th and 21st July the daily correction 

 of the chronometer is found to be : 



Kj = - P-521. 



As already stated at p. 315, on the return journey the chronometers at 

 Fort Portal varied considerably through lack of control, so that, before leaving 

 this place, we proceeded to the determination of their correction by means of 

 six series of altitudes, the results being: 



27th July. —Obs. Xo. 82 ; ^. = 10*^ 20"" 1 P ... K,= + 3" 32'" 57«- 6 



27th „ „ 83; ,,=-10 23 „ = 51-0 



28th „ „ 84; „= 2 21 „ = 59*5 



28th „ „ 85 ; „ = 2 23 8 „ = 59 • 9 



31st „ „ 87; ,,-10 25 44 „ = 64-8 



31st „ „ 88; ,, = 10 27 53 „ = 64-7 



The disagreement of the .second .series induced us to abandon it, the 

 influence of some error in the observations being obvious. 



316 



