[ 78 ] 



altitudes of the celeftial bodies are meafured : for want of fuch 

 a level on land, they who make obfervations v/ith an Hadley's 

 quadrant, which is not fixed and adjufted by a plummet or fpirit- 

 level, are obliged to take thefe altitudes by meafuring the an- 

 gular diftance between the obje£l and the image of the fame 

 feen on the fmooth and level furface of a ftagnant fluid, as water, 

 quickfilver, &c. ; and when the objed is very remote, that an- 

 gular diftance will be double its elevation above the real horizon. 

 But as fuch obfervations muft be made in the open air, and the 

 leaft breath of wind will rufile the furface of the fluid, it is 

 neceflTary to cover it with two glafs planes joined at their upper , 

 edges and ereded like the roof of a houfe over the veffel con- 

 taining the fluid : each of thefe planes muft have its oppofite 

 furfaces exadly parallel, which is never the cafe with plates of 

 looking-glafs, and the proper planes with the requifite apparatus 

 are in few places to be procured, and liable to accidental injury. 

 Inftead of this fluid artificial horizon I have ufed the following 

 •one, which can any where be eafily made, and which I found 

 to anfwer well, as being accurate and portable, lefs liable to ac- 

 cidents in travelling than the former, and difcovering in the adl 

 of obfervatjon its own errors of adjuftment. I conftruded it in 

 the two forms hereafter defcribed, the firft of which is repre- 

 fented in a perfpedlive view in fig. i, and in a fedion through 

 the middle in fig. 2 : the fecond form appears in fig. 8, in a like 

 fedion through the common axis of its parts. 



It 



