228 



TABLES or MONTHLY EXTREMES 



The monthly absolute range is least m summer and greatest m whiter, a result 

 which has already been reached in a difierent way in reference to variations in the 

 monthly means, and the ratios indicate a regular progression in the yearly period ; 

 the January variability in the temperature is one and a half times as great as the 

 July variability. 



The 11 stations given in the table for Florida yield the following results: — 



Averages. 



Jan. 



Feb. 



85° 4 

 32-4 

 S3-0 



Mar. 



88°. 

 39-3 

 48.7 



Apr. 







90.2 

 45-9 

 44-3 



May. 

 94.1 



55-1 

 39-° 



June. 



97.0 

 64.x 

 32-9 



July. 



95° 7 

 68.8 

 26.9 



Aug. 



97° I 

 66.7 



3°-4 



Sept. 



95° 5 

 63.0 



325 



Oct. 



90° 8 

 49.1 

 41.7 



Nov. 



87% 

 40.2 

 47.0 



Dec. 



85° 4 

 33-0 

 52.4 



Highest temperature . . . 

 Lowest temperature . . . 

 Absolute monthly range . . 



Ratio, the average being 41.9 



84.2 

 3°-4 



53-8 



1-3 



1-3 



1.2 



I.I 



0.9 



0.8 



0.6 



0.7 



0.8 



I.O 



I.I 



1.2 



We have the same regularity in the law of the annual progression, but the ratio 

 of the variability in January to that of July is as 2 to 1. The average variability 

 during the year in the latitude of New York is to the variability in the latitude of 

 Florida as 69 to 42. 



Tabulation of the Mean Annual Temperature in the United States, and at some 

 places in British North America, for a succession of years, from the earliest records 

 to the close of the year 1870. 



The object of this tabulation Avas to furnish, in a convenient form, a basis for 

 discussions relating to the study of the variations of our climate — as far as the same 

 depends on temperature — during long intervals, involving questions of permanency, 

 of periodic variations, of irregular fluctuations, and other relations. The tables 

 will, therefore, be of permanent value, since they furnish the earliest material 

 available, and they have consequently been made as complete as possible, at least 

 within the area of the United States. The arrangement is that by States and 

 Territories and by stations in each, the whole in alphabetical order. 



In conformity with previous investigation the annual means have been corrected, 

 as far as that could be done now, for daily variation, excepting those few cases 

 where the hours of observation were unknown, as indicated by foot notes. To give 

 to the tables the fullest extent compatible with accuracy, broken records (extending 

 over less than one year) have been completed by interpolation, but only when obser- 

 vations were found recorded during at least 9 months of the calendar year. This 

 interpolation for 1, 2, or 3 months (as the case may be) was eff'ected as follows: 

 comparison by differences was made with records complete during the period at an 

 adjacent station or at 7iear places for some months preceding and following the 

 lacuna, and the average diff'erence was applied to the record to furnish the inter- 

 polated value for the incomplete station. If no suitable adjacent station for com- 

 parison could be found, the general mean from the whole series for the particular 

 months or month was substituted in the place of the blank record. The first 



