ON DEW, &c. 165 



in contact with the metal. But even when the 

 ball of a thermometer was applied directly to a 

 metal, on a clear and calm night, a temperature 

 was marked by it, commonly 2 and 3, and some- 

 times more degrees less than that marked by 

 a similar , thermometer, inclosed in gilt paper, 

 and similarly placed. I found it likewise ne- 

 cessary, in this inquiry, to correct the tempera- 

 ture of the air, as given by a naked thermo- 

 meter. For, on still and serene nights, a ther- 

 mometer inclosed in a case of gilt or silvered 

 paper, and suspended in the air 4 feet above 

 the grassplat, was usually observed to be 1 J or 

 2 higher than a bare thermometer, of the same 

 construction, suspended near to it. The dif- 

 ference of two such thermometers, thus placed, 

 was once observed by me to be 2|-, and once 

 3^. It may be thought, perhaps, that these 

 differences were caused by the metalled case 

 obstructing the transmission of the temperature 

 of the air to the inclosed instrument. But that 

 this was not the reason is shewn by my observ- 

 ing, that on cloudy nights there existed no dif- 

 ference between the two thermometers ; that, 

 even on clear nights, a thermometer contained 

 in a case of white paper, somewhat thicker than 

 the metalled, was always nearly of the same 

 temperature with a naked one which was sus- 

 pended close to it 5 and that, when a difference 



