ON AQUEOUS AND IGNEOUS METEORS. 553 



so that at the straights of Gibraltar, a current runs inwards at the surface 

 and outwards near the bottom, for the same reason as the air, when it is 

 denser in a passage than in the adjoining room, blows a candle towards the 

 room at the lower part of the door, and draws it towards the passage at the 

 upper. Had there been a continual current inwards through the Straights, 

 at all parts, the Mediterranean must in the course of ages have become a 

 rock of salt. It is indeed remarkable that all lakes, into which rivers run 

 without any further discharge, are more or less salt, as well as lakes in 

 general near the sea : but where a river runs through a lake into the sea, 

 it must necessarily, in the course of time, have carried the salt of the lake 

 with it, if it had ever existed. 



Experiments on the deposition of moisture, like those of Mr. Dalton, are 

 liable to a slight inaccuracy, on account of the effects of an apparent 

 elective attraction, by means of which, some substances seem to attract 

 humidity at a temperature a little higher than others. Thus, a surface of 

 metal often remains dry, in the neighbourhood of a piece of glass which is 

 covered with moisture. It is certain that some substances attract moisture 

 from the air, even when the quantity which it contains is incomparably 

 less than that which would saturate it, since it is on this circumstance 

 that the construction of hygrometers depends ; and it is probably by a 

 property somewhat similar, that even surfaces of different kinds possess 

 different attractive powers for moisture nearly ready to be deposited. It is, 

 however, only necessary to employ, for Mr. Dalton's experiment, a sub- 

 stance which has a very weak attraction for moisture ; and any kind of 

 metal will perhaps be found sufficiently correct in its indications. 



It has been observed, that a piece of metal, placed on glass, usually pro- 

 tects also the opposite side of the glass from the deposition of dew ; and 

 Mr. Benedict Prevost has shown, that in general, whenever the metal is 

 placed on the warmer side of the glass, the humidity is deposited more 

 copiously either on itself, or on the glass near it ; that when it is on the 

 colder side, it neither receives the humidity, nor permits its deposition on 

 the glass ; but that the addition of a second piece of glass, over the metal, 

 destroys the effect, and a second piece of metal restores it. It appears that, 

 from its properties with respect to radiant heat, the metallic surface pro- 

 duces these effects, by preventing the ready communication either of heat 

 or of cold to the glass.* 



The quantity of invisible moisture, contained in air, may be, in some 

 degree, estimated from the indications of hygrometers, although these in- 

 struments have hitherto remained in a state of great imperfection. A sponge, 

 a quantity of caustic potash, or of sulfuric acid, or a stone of a peculiar 

 nature, has sometimes been employed for determining the degree of moisture 

 of the air, from which it acquires a certain augmentation of its weight. 

 A cord dipped in brine,t or the beard of an oat, is also often used for the 

 same purpose : the degree in which it untwists, from the effect of moisture, 

 being shown by an index. But the extension of a hair, or of a slip of 

 -' 



* B. Prevost on Dew, Ann. de Chimie, xliv. 75. 

 t Smeaton, Ph. Tr. 1771, p. 198. 



