[ I^o ] 



Aaguant mercury^ But in thefe portable barometers with 

 narrow boxes, though the common fcale be adjufted at firft 

 to the true height of the mercury, it will not fhew its true 

 height afterwards when the mercury has rifen above, or fallen 

 below, tliat point or divifion of the fcale w-here it ftood at firft. 

 P'or as the box is narrow, the height of the mercury will vary 

 in the box whenever it varies in the tube, and it is the fum 

 of thefe two variations that gives the true variation which has 

 taken place in the height of the elevated niercury. Now the 

 fcale annexed to the tube only fliews one of thefe variations ; 

 and therefore when the mercury ftands above or below the 

 divifion x)£ the fcale to which its real height was at firft 

 adjufted, we cannot tell, by merely infpecfling the fcale, how 

 many aliquot parts of an inch the height of the mercury has 

 varied, or how much it differs from the height it had when it 

 ftood at that divifion to which it was adjufted at firft. Con- 

 feqtiently when the mercury departs from that divifion the 

 fcale will not fliew its true altitude in inches and aliquot parts 

 of an inch. 



To corredl this error of the fcale, by which the variations 

 in the height of the mercury alway appear lefs than they really 

 are, you propofe that tables fhould be formed which may fhew 

 what additions ought to be made to each partidular variation. 

 This however might be done in an eafier way than by tables 

 prevloufly calculated : For when you have found the pro- 

 portion between the furface of the mercury in the box and 

 that in the tvibe, fay as the furface in the box is to the fum of 



the 



