M. Melloni on the Theory of Dew, 137 



sibly equal ; but we could never be justified in inferring lo- 

 gically from this experiment, that metals cool, during calm 

 and clear nights, as much as lamp-black and glass. 



To learn the true state of things, it is necessary, therefore, 

 to discard glass, and employ as substitutes, slender tubes of 

 tin which scarcely radiate at all, and keep the thermometer 

 sufficiently insulated from the heat of the ground ; it is re- 

 quisite, besides, to cover all points of the thermometer with 

 metallic plates. Thus, these plates being polished, the ther- 

 mometer gives as near as possible the temperature of the 

 air ; and when the covering is varnished, blackened, in con- 

 tact with leaves of vegetables or any other substances, we 

 obtain the degree of cold produced by the radiation of 

 this substance, by a simple comparison w4th the polished 

 thermometer. 



By means of instruments thus mounted, I have convinced 

 myself that the leaves of plants, glass, and varnish, and 

 lamp-blacky always cool, during calm and clear nights, from 

 1 to 2 degrees below the surrounding air. 



On observing this slight diminution of temperature, we 

 might be tempted to believe the decrease of 7 and 8 degrees, 

 reported by Wilson and Wells, to be greatly exaggerated. 

 But if we consider that the differences obtained by the two 

 English natural philosophers contain an element foreign to 

 the question : that the thermometers used for measuring the 

 surrounding temperature were raised from I"" 30, to 1™ 60, 

 while the instruments enveloped in the radiating substance 

 were placed near the ground, it is easy to understand to what 

 the great difference between their results and mine are 

 owing. 



In fact, Pictet's experiments have long since demonstrated, 

 that the temperature of the air decreases rapidly during calm 

 and clear nights as we approach the surface of the earth. The 

 single fact of this decrease ought therefore to render the 

 temperature of the radiating substance placed near the ground 

 inferior to that of the air, as indicated by a thermometer ele- 

 vated from 1™ 30, to 1"* 60 ; so that, in an arrangement of 

 this kind, the difference of the two instruments in no degree 



