[ 53 ] 



grains, it appears that the above accefllon of weight to the co- 

 lumn of air that fupports it can raife it little more than ^ of 

 an inch, inftead of nearly -h, the variation to be accounted for. 

 This caufe, therefore, though not abfolutely inefficient on the 

 fuppofition that the whole mafs of the fuperincumbent column 

 is encreafed by the acceffion of new air in proportion to the 

 condenfation, is notwithftanding inadequate to the effedt pro- 

 duced. 



0/ the Influence of Winds. 



The winds, whofe efficacy in producing the variations of 

 the barometer I am about to difcufs, are thofe which reign in 

 the lower regions of the atmofphere, they being principally 

 referred to by thofe philofophers who have had recourfe to their 

 agency. 



First. Dodor Halley attributes the rife of the barometer over 

 its mean altitude to the accumulation of air over the place of 

 obfervation, which accumulation he attributes to two contrary 

 winds blowing towards that place ; but if this were the caufe of 

 the elevation of the mercury, we Ihould always have a calm 

 when the ^ ftands higheft j for the accumulation ffiould take 

 place only when the two contrary winds blow with equal force, 

 fince if one of them prevails it fhould repel its antagonift, and it 

 is only during the prevalence of neither that the air can be 

 accumulated. Now it is notorious that the greateft mercurial 

 heights are accompanied by an eafterly or northeriy wind, as he- 

 himfelf has obferved. Nor can that equality of barometrical 



heights 



