[ io6 ] 



change in their dimcnfions from expanfion or contradion might 

 defeat the accuracy of obfervation. It were at all times defirable, 

 with every fpecies of barometer, that the feparate obfervations 

 fhould as nearly as pofliblc be made at the fame temperature 

 of the annexed thermometer. If this circumftance is attended 

 to the fourcc of errors depending on this caufe will entirely 

 vanifh. 



REMARK VI. 



On carrying and ujing the Barometer. 



The only fafe way of carrying this inftrumeht is with the 

 point downwards ; the attached thermometer is to be loofe in 

 its cafe with a quill fcale, and a bit of cork or cotton within the 

 cap for its bulb to reft on ; in this way I have both on horfeback 

 and in carriages conveyed this inftrument fafely for many 

 hundred miles. 



From the conftrudion it is evident that, if accurately made, 

 they will hang truly plumb when inverted and held lightly 

 between the fore-finger and thumb. But this obvious advantage 

 does not preclude many contrivances which might be thought 

 of to hang them in gimmals, or fufpend them in any manner 

 that might be thought more advifable. This and a great many 

 varieties in the conftrudion, which might be adopted without 

 interfering with the fimplicity of the principle, I leave at large 



to 



