Of BAROMETRICAL MEASUREMENTS. 89 



troduced. The atmofphere, therefore, is warmed by the earth, 

 from the furface of which a quantity of heat is contirmally 

 flowing off, and afcending through the different ftrata of the 

 air into the regions of vacuity, or of sether. But this afcent, 

 on the whole, is uniform; becaufe there is a certain tempera- 

 ture which, though varied by periodical viciffitudes, remains 

 under every parallel the fame, as to its mean quantity. Eve- 

 ry ftratum, therefore, of the atmofphere, whatever be its height, 

 gives out, at a medium, the fame quantity of heat that it re- 

 ceives j in other words, its mean temperature is conflant, and 

 neither increafes nor decreafes, on the whole. 



3. LET there be three flrata, then, of the atmofphere of the 

 fame thjcknefs x, and contiguous to one another ; fo that, if x 

 be the diflance of the firft from the furface of the earth, that 

 of the fecond may be x + x, and of the third x + zx. Let Z>, b\b\ 

 be the heats of the flrata, and A, A', A", their denfities refpec- 

 tively : Then, fince the quantity of heat, communicated in an 

 inflant from one flratum of a fluid to a contiguous ftratum, 

 mufl be, as the difference of their temperatures, multiplied in- 

 to the denfity of the colder, and divided by the denfity of the 

 warmer, the heat communicated, in an inflant, from the firft 



flratum to the fecond, = (b b'} ; and that communicated by 



the fecond to the third, = (b'b"}, . But, fince the difference 

 of A and A" is indefinitely fmall, as alfo that of A' and A", we 

 have = i, and ^- i ; fo that the heat gained by the mid- 

 dle flratum is = b b', and that loft by it b' b''. Now, thefe 

 two quantities muft be equal, in order that the temperature of 

 the ftratum may remain uniform, that is, h b'~-h' b" ; or, in 

 other words, the heat of the firft ftratum exceeds the heat of 

 the fecond, as much as the heat of the fecond exceeds the heat 

 of the third. Therefore, the heat of the fucceffive ftrata muft 

 decreafe, by equal differences, as we afcend through equal 



M fpaces, 



