ON DEW, &c. 193 



become colder than the air, and condense the 

 watery vapour of the contiguous parts "of it, if 

 they should contain a sufficient quantity to 

 admit of this effect. But events similar to these 

 occur, when dew appears in an open and level 

 grass field, during a still and serene night. The 

 upper parts of the grass radiate their heat into 

 regions of empty space, which consequently 

 send back no heat in return ; its lower parts, 

 from the smallness of their conducting power, 

 transmit little of the earth's heat to the upper 

 parts, which at the same time receiving only a 

 small quantity from the atmosphere, and none 

 from any other lateral body, must remain colder 

 than the air, and condense into dew its watery 

 vapour, if this be sufficiently abundant, in re- 

 spect to the decreased temperature of the grass*. 



This subject may be further illustrated by a 

 reference to what happens in the experiment, 

 which has been used to prove the reflection of 

 cold. 



In the simplest form of this experiment, a 



* I have adopted in this explanation the hypothesis of Mr. 

 Prevost of Geneva, on the constant radiation of heat by bo- 

 dies in contact with the atmosphere, even at the time that 

 they are exposed to the influence of bodies warmer than them- 

 selves; as it appears to agree perfectly with all the pheno- 

 mena of the communication of heat, which do not depend 

 upon conduction. I shall hereafter make frequent use of this 

 hypothesis. 



O 



