THE RATE OF DECREASE OF TEMPERATURE^ ETC. 97 



If, now, in order to smooth down still further the irre- 

 gularities arising from accidental causes at both the upper 

 and the lower stations, we take the means of groups of 

 three years, we have the following remarkable numbers : — 



o 



Mean of 1848-50=22-58 

 1849-51=22*80 



1 850-52=22*74 

 1851-53 = 23*23 

 1852-54=23*44 



i853-55=23'62 

 1854-56=23*16 

 i855-57 = "-95 

 1856-58=22*57 



In looking over these numbers, it seems impossible to 

 resist the conclusion that some influence has been in 

 operation by which the temperature of the higher station 

 was gradually reduced, as compared with the lower stations, 

 up to the beginning of the year 1854, and afterwards as 

 gradually increased to the close of the series. 



In the next Table I have given the mean annual tem- 

 peratures and the differences for the years 185 6- 1860 at 

 the two stations in England, Bywell and Allenheads : — 



Bywell. AUenheads. 



Mean ann. temp. Mean ann. temp. Diif. 



o 00 



1856 46*92 4^*78 4*14 



1857 49*50 44-82 4*68 



1858 48-47 43*57 4*90 



1859 - 48*81 43*90 4*91 



1 860 45*73 40*66 5*07 



Here we have a remarkably gradual and regular in- 

 crease of the difference from the commencement to the 

 end of the series, which contrasts strongly with the irre- 

 gularities in the changes of the actual temperatures of the 

 two stations ; and it will be noticed that this gradual in- 

 crease took place in the years during which a decrease 

 occurred in the difference between the Milan and Geneva 

 and Great St. Bernard Stations. 



SER. III. VOL. II. H 



