MADE AT BRUNSWICK, MAINE. 



37 



hot years culminating with 1840, where our record may possibly be defective. The 

 results for these years (1837, '38, '39, '40) should, therefore, be used with caution, 

 as the thermometer may have been affected with radiation. 



The numbers of Table IX are directly available for the study of the secular 

 change of the temperature, and if made out for a great number of localities cannot 

 fail to lead to valuable results. 



That the temperature has remained sensibly the same for the period during which 

 the observations were made, we can infer from the fact that between 1807 and 

 1832 inclusive, the mean annual temperature was 44°. 1, and between 1833 and 

 1859 inclusive it was 44°.7, the difference 0°.6 + 0°.5, with a probable error of the 

 same magnitude, is too insignificant to point to any change in the climate. 



Supposed Anomalies in the Annual Fluctuation of the Temperature. 

 It has been supposed^ that about the end of May there was a cold period and 

 about the end of October a warm period, forming apparently anomalies in the regu- 

 lar march of the annual fluctuation. To test this supposition the daily means from 

 51 years of observations were made out (vide Table I), and laid down graphically on 

 the two annexed diagrams, which cover, in time, the periods named above. These 

 diagrams show the daily irregularity which even from so long a series amounts to 

 ± 0°.8 in the daily mean. 



Mean daily iemjierature from fifty one years of observation. 











































































































































, 



_ 



^ 



■^ 





\ 



- 





























, 











/ 

































/ 



\ 



^ 



^ 





~- 



— 



^ 



















_ 







_ 



_ 



_- 



— 



/ 

































- 



































































^ 



































as i6 1718 19 20 2122 23 24 25 26 27 28 29 30 31 12 3 4 5 6 7 

 JVIay. June. 



'2Iean daily temperature from fifty < 



of obs. 





























































\ 



-^ 



/ 



\ 



























































■^ 



^ 



y 



\ 































































\ 





























































-^ 





\ 



_ 







^ 



/ 



■~~ 



























































\ 







/ 



\ 



\ 



















































































































^ 





•10 1112 13 141516 17 18 19 20 2122 23 24 25 25 27 28 29 30 311 2 3 4 5 

 October. Movemljer. 



The dots surrounded by a small circle indicate the temperature as computed by 

 the formula for the annual fluctuation and corrected for diurnal inequality. 



* Dr. Wilson, of Geneva, N. Y., was led to suppose, from twelve years of observation at that 

 place, that a cold period occurred about the end of May, and a warm one about the end of October. 



