276 ESSAY 



was 6, and between those of air and a fully 

 exposed part of the straw bed 9. 



EXPERIMENT 3rd. 



This was begun on the evening of the 16th 

 of October, and was likewise made agreeably 

 to the method related by Mr. Williams, 



Ice appeared in the pans, when the tempera- 

 ture of the air, at the height of 5J feet, was, 

 according to a naked thermometer, 37. 



On this night, I placed upon the straw bed a 

 dry earthen pan, among those which contained 

 water, and found the inside of its bottom to be 

 as much colder than the air, as the water was 

 in the other pans, before ice appeared in them. 

 After the water had begun to freeze, no proper 

 comparison could be made between its tem- 

 perature and that of the empty pan. This pan, 

 in the course of the night, attracted moisture, 

 which was afterwards converted into a film of 

 ice. 



But the chief fact established by the present 

 experiment was, that water may freeze at night, 

 in air of a temperature higher than 32, not 

 only without any loss of weight from evapora- 

 tion, but with a gain of weight from an opposite 

 process. 



I had observed that water, exposed early in 



