68 On Fiiiiilitj/, and 



The relative places of portions of vapour, or of permanent 

 gas, are determined, then, by this reciprocal pressure, as well 

 as by that which gravitation causes. The aggregate, or possible 

 aggregate, must be considered as terminated one way, either by 

 the earth's centre, or, perhaps, by the surface of its solid 

 matter ; and the other way, by the variable confines of restrained 

 gas, and unconfined elastic fluid. 



Between those extremes must be placed gaseous particles, in 

 all gradations, from that where repulsion barely exceeds attrac- 

 tion, to that in which its excess above particular attraction barely 

 falls short of gravity. The first are lowermost, confined on one 

 side by solid or liquid matter, and compressed on another by the 

 weight of superincumbent gas or vapour. The others are upper- 

 most, upheld by the repulsion of the subjacent gas, and restrain- 

 ed only by the almost evanescent excess of gravitation, above the 

 difference between repulsion and attraction, and by no resistance 

 of repulsive force and gravity of particles situated beyond them. 



According to this conception of the aerial sphere, an atmo- 

 spheric column is an inverted cone, the apex of which reaches the 

 earth's centre ; or, considering the column with reference to the 

 earth's solid surface only, it is the frustum of a cone, the summit 

 of which touches the circumference of the earth, and the base 

 reaches the circumference of an orb, the site of equipoise of 

 gravitation and elasticity. Reckoning from the summit of the 

 frustum, and taking it to consist of a given quantity of particles, 

 or atoms, which are nearly in equilibrium of particular attrac- 

 tion and repulsion ; every successive stratum, or single layer, 

 consisting of a like number of particles, is deeper, as the com- 

 pression of gravitating contiguous particles is less ; and as gra- 

 vitation decreases : the mutual attraction of homogeneous par- 

 ticles being lessened with the consequent increase of reciprocal 

 distance. 



The density, then, of a gravitating and elastic fluid, the elasti- 

 city of which grows with the decrease of gravitation consequent 

 on increase of elevation, is determined by two causes ; taking 

 repulsion as constant, viz., direct gravitation, or weight of the 

 fluid; and mediate gravitation, or superincumbent pressure; 

 and, admitting the gravitating tendency of atmospheric air to 



