on the Surfaces of Solid Bodies. 245 



from the circumstance that these substances condense aqueous 

 vapours in far larger quantity than the surface of the pile, and 

 that they therefore become more heated or more cooled than that ; 

 and hence their action on the pile is as great — or, according to 

 their thickness, greater than the direct action of the vapours 

 when they are immediately absorbed by the pile. 



The thickness of the plates must not exceed a certain measure 

 if the alteration of temperature is to be appreciable; but expe- 

 riments establish the conclusion, that all substances, however dif- 

 ferent they may be, are raised in temperature when air comes in 

 contact with them which is moister than that surrounding them, 

 and that they are depressed in temperature when they are exposed 

 to air which is drier than that by which they are surrounded. 



The very considerable heating or cooling which occurs on 

 altering the state of moisture of the atmosphere, led me to sus- 

 pect that this might possibly be appreciable by an air-thermo- 

 meter. Such an instrument was therefore used, consisting of 

 two glass bulbs connected by a narrow tube AB (Plate III. fig. 2). 

 Each of the bulbs was protected from accidental cooling by being- 

 surrounded by a small glass shade, C and D ; and into the tubu- 

 lure of each shade a glass tube was inserted, through which the 

 air might be blown in. The shades were closed below with cork 

 plates k k, which consisted of two pieces, and left a space about 

 the narrow tube of the air-thermometer for the escape of air 

 which had been blown in. On blowing the air of the room into 

 one of these small shades, no alteration in the height of the ther- 

 mometer took place. If the air had been previously dried, then 

 the bulb operated upon fell in temperature ; and if the air had 

 been saturated with aqueous vapour, the bulb rose in tempera- 

 ture, so that the difference of level of the liquid in the narrow 

 tube separating the bulbs amounted to from 4 to 6 millims. If 

 the blowing in of either one or the other kind of air were con- 

 tinued, the fluid assumed gradually its initial level. When, one 

 of the bulbs being coated with pine-soot, alternately a stream of 

 dry and then moist air was blown in upon it, a difference of level 

 of from 8 to 10 millims. was occasioned. 



A similar effect took place when alternately dry and moist air 

 was blown against the bulb of a mercurial thermometer divided 

 into half degrees. The thermometer was guarded against cur- 

 rents of air which might interfere, by being placed in a T-shaped 

 tube, being retained in its place by means of a cork (fig. 3). The 

 air was blown in at B. The difference in the height of the ther- 

 mometer xy, produced by the employment of dry and moist air, 

 was from 0°-2 to 0°*3 C. On blackening the bulb of the ther- 

 mometer, the difference rose to o, 6 C. 



These experiments show how considerable the heating and 



