the Si met It re of the Earth. 59 



be governed by the same law with other matter (that is, inverse- 

 ly, as the squares of the distances,) the densities of the atmo- 

 spheric strata are in geometric progression, answering to dis- 

 .tances in harmonic progression; or, which is the same thing, 

 answering to the reciprocals of an arithmetic progression. 



Applying these speculations to the consideration of the 

 structure of the earth, it may be convenient, in the first instance, 

 to imagine a globe having the same mean temperature of its 

 mass, but composed of water exclusively. The density of its 

 liquid superficies may be assumed to be the same with that of 

 the surface of the mundane ocean ; and, above that level, an 

 atmosphere of watery vapour, decreasing in density upwards, 

 must be supposed. Below that level the density of the liquid 

 increases with the depth, by the augmented force of gravity 

 acting directly, and also mediately through superincumbent 

 pressure. 



Ingenious experiments, yet unpublished, shew that water is 

 more compressible than it has been supposed to be ; and the 

 increase of density, as it has been measured under a pressure 

 equal to that of a very long column of water, has been deter- 

 mined at a greater quantity than was to be expected from previ- 

 ous imperfect trials. Without anticipating the publication of 

 those interesting experiments, it is enough for the present pur- 

 pose to state, that at a certain assignable depth the density of 

 water, and its specific gravity, more than doubled or tripled, 

 would appear to be greater than of those solids, which are 

 known to us as most abundant in the crust of the earth, so far 

 as we are acquainted with it. 



It would follow, then, that water, at very great depth, would 

 be capable of floating bodies, which at its surface sink by their 

 superior weight, provided that density and specific gravity in- 

 crease much less rapidly in the solid than in the liquid, under 

 corresponding degrees of compression. 



Let a sphere be now supposed, having the same mean tempe- 

 rature of its mass, and exclusively composed of gas ; for instance, 

 atmospheric air. It can be conceived that this gas, more com- 

 pressible than a liquid, may, in obedience to the power of gra- 

 vity acting directly, and likewise medialciy by superincumbent 



