136 DETERMINAflON of the LATITUDE 



•which Is fubdivided into eight equal parts ; and the vernier 

 gives one thirty-fecond part of a fubdivifion, or 13", 18. A 

 micrometer fcrew is attached to the vernier, which ferves to re- 

 gulate the motion of the index, and by which, the excefs in fe- 

 conds above the next lefs divifion of the vernier is fhown. 



Each zenith diftance was read off, at leatt, three times, both 

 from tlie ninety and ninety-fix arcs, and the means of each were 

 taken. Thefe ferved as a check on each other ; however, the 

 zenith diftance, as given by the ninety-fix arc only, is ufed 

 for obvious reafons. The ninety-fix arc was found to be about 

 12" lefs than go° ; and the error of the line of collimation at 

 the vertical radius was about a fecond and a half, fubtradive. 



As the tranfit inflrument and quadrant were placed in adja- 

 cent rooms, it was therefore in my power to obferve both the 

 fun's tranfit and zenith diftance the fame day ; however, the paf- 

 fage of the fun's weft limb over the fifth wire, and that of the 

 eaft limb over the firft wire, were by this means loft. Hence, 

 alfo, the zenith diftance of one limb only of the fun could be 

 obferved ; and the true zenith diftance will be aflfedled by the 

 error of the fun's femidiameter, as given in the Nautical Alma- 

 nac, and by the irradiation, which according to M. Du SEjoifR, 

 exceeds three feconds. 



The middle wire in the telefcope of the quadrant fubtended 

 an angle of no lefs than 20", 6 ; therefore, as it was fcarce pof- 

 fible to bring the fun's limb exadlly to the middle of the wire, 

 I conftantly made the lower edge of the wire a tangent to the 

 fun's apparent lower limb. The zenith diftances in the follow- 

 ing table are the differences between thofe obferved and the fe- 

 midiameter of the wire, the tenths of a fecond being negledled. 



The fifth column of the table contains the error of tlie line 



of collimation, combined with that of the ninety-fix arc, taking 



it for granted that this arc is accurately divided. In column 



fixth is the fun's femidiameter, from the Nautical Almanac, to the 



1 neareit 



