94 CAUSES 'which affeft the ACCURACY 



ture 55, or the comprefling force, increafes more flowly than 

 the denfity ; fo that, if the comprefling force be doubled, the 

 denfity will exceed the double by about a tenth part, &c. The 

 law of this variation is exprefled with tolerable exactnefs, by 

 fuppofing, that if D be the denfity of the air, and F the force 



comprefling it, then D F , being a very fmall fraction, 

 nearly .0015. 



10. IT muft be acknowledged, that new experiments are ne- 

 ceflary to afcertain the law of this inequality \vith precifion. 



But as the formula DF is very general, and might be 

 rendered ftill more fo, withoxit affecting the method of integra- 

 tion that is to be employed, the refult of that integration may 

 be ufeful when our phyfical knowledge becomes more accurate. 

 In the mean time, it may not be improper to remark, that the 

 precife knowledge of the law which connects the comprefling 

 force with the denfity of elaflic fluids, is an object well deferving 

 the attention of natural philofophers. The determination of that 

 law may go far to decide the queflion, whether the particles 

 of fuch fluids are in contact or not ; that is, whether the elafti- 

 city of each particle be a force that extends beyond the nearefl 

 particles, Uke the forces of magnetifm and gravitation ; or one 

 which, like that of a fpring, extends only to the bodies which 

 are next it. It is an enquiry, therefore, of no lefs importance 

 in general phyfics than in that particular fubject which we have 

 here undertaken to examine. 



11. THERE is one other correction to be applied to the height 

 of a mountain, as it is ufually found from obfervations of the 

 barometer. This arifes from the diminution of gravity, whe- 

 ther we afcend or defcend from the furface of the earth. The 

 effect of that diminution is to produce a twofold error ; be- 

 caufe, on the fuppofition of uniform gravity, the weight of each 

 particle of air is computed too great, and the weight of the co- 

 lumn of mercury in the barometer, that is not on the furface, is 



alfo 



