206 OBSERVATIONS ON THE MEAN TEMPERATURE 



This singular agreement between the observed and calcula- 

 ted results, and the equilibrium of the positive and negative 

 errors, shews that the formula embraces the leading causes 

 which affect the mean temperature of the west of Europe. 

 The sum of all the positive errors is 13°. 12, and the sum of 

 the negative errors 9.11; and their total amount is 22°. 23, 

 which gives only an average error of icths of one degree of 

 Fahrenheit upon each observation. 



The results of Mayer's formula give all the errors positive 

 except one, and their sum is no less than 70°. 7, being 2°.3 

 for each observation. The error of the formula becomes so 

 great as 9° between the parallels of 60° and 70°, which 

 proves, in the most convincing manner, that the temperature 

 of 32°, which Mayer assigns to the North Pole, is very far 

 above the truth. The formula which I have given above 

 makes the polar temperature so low as zero, or 0° of Fahren- 

 heit's scale, differing 32° from the preceding measure ; but the 

 circumstance of its representing with accuracy the mean tem- 

 peratures at very high latitudes, inspires us with some confi- 

 dence even in this extreme result. 



In this state of uncertainty respecting the probable tempe- 

 rature of the North Pole, and of the accessible parallels be- 

 tween 70° and 80°, I proposed to examine the most northern 

 meteorological journals, with the view of finding some general 

 rule by which the mean temperature of the year might be de- 

 duced from one or two months observations. I had previous- 

 ly communicated my formula to Mr Scoresby, and requested 

 from him some information respecting the temperature of the 

 Greenland Seas ; and I had the satisfaction of finding, that this 

 subject had engaged his most particular attention, and that he 

 had actually deduced the mean temperature of the parallels of 



76* 45, 



