VariaHens of Temperature, ^e. in the Atmosphere. 171 



but probably from barometric observations. 



The mean error or difference from observations is l°.l. 

 This is a degree of coincidence which, in observations subject 

 to so many causes of error as these are, could scarcely be 

 expected. Were we. to allow only i° for the superior radiation 

 and influence of bodies at the surface of the earth, even this 

 trifling difference would disappear; or were the observations 

 at the same places to be repeated at night instead of in the 

 day, the time at which it is very likely they were made, it is 

 highly probable the difference would be reversed, and be posi- 

 tive instead of negative. For my part, I have no doubt that 

 the apparently more rapid diminution of temperature near the 

 surface than higher up, is owing to the observations having 

 been made in the day time ; and that the contrary would 

 happen were they made in the night, especially if the weather 

 was calm and clear. Dr. Wells's experiments, which in the 

 night manifested an increase of temperature of sometimes 12 

 or 16 degrees at the elevation of only a few feet, are a strong 

 confirmation of this. 



