202 



DISCUSSION OP THE ANNUAL FLUCTUATION 



must be confined within much narrower limits than the variation in the epoch of 

 maximum cold. 



On the western coast the records of three stations were examined (San Diego, 

 San Francisco, and Sitka), but, owing to the shortness of the record, only a glimpse 

 of the existence of an inequality could be obtained with an indication of the 

 occurrence of the extreme shift in winter later than in 1844. 



Takinff means of the values for the dilferent stations, for winter and summer, we 

 obtain the foUowins: results: — 





Cold 



Season. 



Warm 



Season. 





Cold Season. 



Warm 



Season. 



Epochs. 



No. of 



Mean of 



No. of 



Mean of 



Epochs. 



No. of 



Mean of 



No. of 



Mean of 





Stations. 



Jan.— Feb. 



Stations. 



July — Aug. 





Stations. 



Jan. — Feb. 



Stations. 



July— Aug. 



1786-90 



2 



+ I?o 



2 



+ 1-5 



1S31-35 



10 



— 1°6 



10 



-t-2°o 



1791-95 



3 



— i.o 



3 



+ 1 



I 



1836-40 



9 



—1.8 



8 



+ 2.9 



1796-1S00 



4 



—1.8 



4 



+ 1 



5 



1841-45 



10 



[+0.6] 



10 



4-1.1 



1801-05 



4 



—2.7 



4 



+ 1 







1846-50 



10 



0.0 



10 



+ 1-4 



1806-10 



6 



[-3-9] 



6 



1+0 



b| 



1851-55 



9 



—3-4 



9 



+ 2.4 



1811-15 



6 



—3.6 



6 



+ 1 



S 



1856-60 



8 



—3-9 



8 



+3-0 



1S16-20 



4 



—2-4 



6 



+ 1 



9 



1861-65 











1S21-25 



6 



— 3-0 



6 



+ 1 



9 



1866-69 











1S26-30 



6 



— 30 



S 



+ 1 



8 













Extreme values are indicated by being contained within brackets, and they 

 point approximately to the epochs 1809 and 1844, when the greatest cold fell on 

 the average early in January and about the middle of February, respectively. 

 Respecting the epoch of greatest heat, the figures leave us in no doubt, though the 

 probability would seem to be in favor of a correspo7iding lateness about 1808 and 

 an earlier occurrence in the position of the maximum at some rather undefined 

 later epoch. 



If the preceding result could be considered as well established, the cycle of the 

 shifting of these dates of maximum cold (and heat) would be about twice 35 years. 



Tables of observed extremes of temperature, for every month, for a series of years. 



To complete our information respecting the annual fluctuation of the tempera- 

 ture, it is necessary to examine the extreme variations from the normal values ; 

 with this view the following table of monthly extremes has been prepared for a 

 number of selected stations. They comprise nearly all the longer series, for which 

 maxima and minima have been tabulated ; the extreme values given are those 

 found in the record, entered at the regular hours of observation, as adopted by the 

 respective observers, the cases of maxima and minima thermometers being very 

 restricted. They do not, therefore, exhibit the absolute extremes, but only 

 approximations to them; besides, the intervals of time over which the series 

 extend are far too restricted to entitle the extremes to be regarded as anything 

 more than approximations. For the geographical position, and the actual duration 

 of each series, after the deduction of breaks, the reader will have to consult the 



