On Night-Frosts. 21 



Time, 

 h. m. 



9 10 

 20 

 30 

 40 

 50 



10 

 10 

 20 

 30 

 40 

 50 



11 

 10 

 20 

 30 

 40 

 50 



12 

 10 

 20 

 30 



40 



The speed of refrigeratiou v varies only from 0",fifi at '.),io p. in. to 0'\22 

 at 10,30 p. m., after which time it rises again and at 11, 20 p. m. reaclies O^o 

 to decrease afterwards very slowly. 



It is evident that tlie falling dew had tirst diminished v between 9, 10 

 p. m. and 10,30 p. m., but later this cause lias ceased to work, ail tlie dew 

 liaving fallen, allowing v to increase again. The cause of this fresli decrease 

 must be ascribed partly to the formation of ice and partly to the condensation 

 of vapour in the upper layers of the atmosphère. This change in v may 

 also be traced in Curve IIo, but as there is a gap in the observations between 

 8,40 p. m. and 10, 40 p, m., we must refrain from a doser examination of it. 



