14 TRANSIT OF VENUS, 1874. HONOLULU. 



4 feet from the ground, from sunset to sunrise on the nights of 1875, 

 February 6, 7, 8, and 9, and its image bisected with the micrometer every 

 half hour. The hanging level, which was left in position and reversed only 

 occasionally, was read at the same time ; while as many circumpolar stars as 

 possible were observed. 



The results of the four nights' observations showed that the direction of 

 the optic axis moved steadily to the eastward (when the telescope was pointed 

 to the north) from sunset until ll h . 30 mean time. It then remained steady 

 until sunrise. The observations of the mark corrected for level-error, com- 

 pared with the stars, give the following absolute azimuths of the center of 



the mark from the center of the transit instrument : 



// 



1 876, February 6, S Ursas Minoris 9 66 east of north. 



Cephei 5i 7 ' 85 



, , 7,8 Ursas Minoris 6 g5 



Cephei 5i 8'i3 



X Ursae Minoris 7 88 



Polaris S.P 7-11 



, , 8, 8 Ursse Minoris 9 '20 



X Ursse Minoris 8 53 



Polaris S.P 7-67 



, , g, 8 Ursas Minoris 8 " I o 



Cephei 5i 8' 85 



X Ursas Minoris 9*44 



Polaris S.P 8' 88 



The adopted absolute azimuth is 8" -32. 



35. On 19 mornings there 'are observations of Polar Stars not very long 

 before the mark was observed. Assuming that the instrument was steady in 

 azimuth after midnight, each of these gives a determination of the absolute 

 azimuth of the mark. The mean of them is 8"'23. 



36. The azimuth-error adopted in the reductions for a group of stars has 

 been found in the following manner: If the middle time of the group is 

 later than ll h . 30 m mean time, a mean has been taken of the azimuth-errors 

 found from circumpolar stars and the morning observation of the meridian 

 mark. For example, on 1874, November 22, Table III., there are four 

 determinations of the Error of Azimuth after midnight (three by circum- 

 polar stars, one by the meridian-mark), the mean of which has been used 

 in reducing the group of clock stars observed after midnight. When 

 tne middle time of the group falls before ll h . 30 m mean time, it has been 

 assumed that the change of azimuth, denoted by the difference between the 



