12 Mr. J. Jeffrey s's Experiments on Exosmose mth reference 



which yielded the experiments. This ware underwent, from 

 a dull red up to the time it was at a full white heat, a regular 

 progressive contraction of its parts, by which time it had lost 

 one fourth or one fifth in length. Commencing with pores 

 so open that they would transmit water or gas with little re- 

 sistance, it became at the heat of fused cast iron as close and 

 dense in texture as glass ; and when cooled, was equally im- 

 permeable by fluids under any pressure, exerted during any 

 length of time. 



Such a substance in its passage through every intermediate 

 degree of density, from that at which all fluids were allowed 

 to pass through it, to that at which they were all confined, 

 might be expected to arrive at some one, at whifih it would 

 detect any difference in the size of the molecules of two fluids 

 pressing against it for passage, by detaining the one but 

 transmitting the other. 



The results were curious. Vessels formed of this composi- 

 tion thick enough to bear the pressure within, were filled with 

 water, containing a small quantity of mineral alkali, and into 

 which was condensed about twelve times its bulk of carbonic 

 acid gas under a pressure of nearly eight atmospheres ; their 

 mouths being effectually secured so that no leakage whatever 

 took place through them. The alkali varied from half a 

 drachm to a drachm and a half of carbonate in twenty 

 ounces. Such vessels as were in their most porous state, 

 having been only hardened at a full red heat, allowed both 

 the water and the gas to be ejected by the repulsive force 

 through their pores, the liquid being projected in every di- 

 rection in the form of a fine rain or cloud, until the pressure 

 within was relieved. 



Such vessels again as were in a properly effective condition, 

 having been exposed to their maturing heat, retained their 

 contents without any apparent loss for a year ; and when 

 placed near a fire would burst with a violent explosion, with- 

 out allowing any exudation to take place. 



A third kind of the same composition, apparently interme- 

 diate as to density between the two already noticed, allowed 

 the gas alone to make its escape through the pores, rather 

 quickly at first, but as the pressure lessened more gradually, 

 until an equilibrium was nearly established, the whole or 

 nearly the whole of the liquid being retained. 



In a fourth kind a density was attained much above the 

 last, but short of that of the entire impermeability of the se- 

 cond kind described, and this texture yielded results very 

 curious in themselves, and especially so when contrasted with 

 those last described. Here instead of the gas escaping, and 



