162 TABLES OF DIFFERENCES OF MEAN TEMPERATURES. 



the mean temperature derived from the ordinary hours of observation (7 A. M. and 

 2 and 9 P. M.) in order to refer them to the true daily mean. 



A table of the daily fluctuation for this place vv'ould answer for most stations 

 situated within the elevated and arid region generally known as the great interior 

 basin, as well as for the regions of the upper Rio Grande and of western Texas. 



In some instances the recorded mean monthly dift'erence between the morning 

 and afternoon temperatures rises to 40°, and if the observations are to be trusted 

 to 45°; the corresponding daily incidental range is equally great and for the regions 

 described above it is not uncommon to meet, in the morning, with a temperature 

 below the freezing point and to experience in the afternoon of the same day a heat 

 rising to 70 or 80° Fah. 



Variahility of the temperature at any hour of the day from the normal value of 



that liour. 



To complete the investigation of the general laws of the daily fluctuation we have 

 yet to inquire into the amount of digression of the monthly mean of any observed 

 hourly temperature when compared with its normal value. 



These irregular variations are most readily ascertained by a comparison of the 

 monthly means for each hour of the day, given separately for a series of years, with 

 the mean of the combined years for each hour. By this method we completely free 

 our results from the effects of the annual fluctuation, and have the advantage of 

 presenting the probable error to the hourly temperatures, as given in the first set 

 of tables for each month, provided the particular table was derived from a single 

 year of observations ; if the tables are constructed from n years, the probable errors 

 require a division by y' n in order to represent the probable uncertainties of their 

 tabular numbers. 



With a special view to this investigation the Mohawk table of hourly tempera- 

 tures is given in full, from 1860 to 1868, only six years of hourly observations, 

 hovk^ever, could be utilized for the present purpose. At Philadelphia, the Girard 

 College series furnished hourly means for nearly 6 years from 1840 to 1845. At 

 Sitka a series of hourly observations (with omissions of 5 readings in each day) 

 was taken from the records of the observatory, for 5 years, selecting 1847-8-9 

 and 1862-3-4. For Toronto, Can., the results are copied from Table Vll^ of the 



1 Tl 



e following is, in p 



art, a 



Eopy of the Toronto table 















Hour. 







-c 





















> 



-^ 'S 



> 





ti 



^ 

 fe 



1 



< 



B 



k 



^ 



< 



0. 



6 



> 



1 



R 



3S.S 



l^-s 









o 



^ 



























J, 















^ 



2 P. M. 



3-49 



2.47 



2.43 



1.94 



2.3« 



2.20 



2.36 



1.66 



'■95 



1.69 



1.45 



3.12 



2.44 



2.08 



4 " 



3-31 



2-54 



2-59 



1.76 



2.36 



2-13 



2.09 



1-37 



1. 71 



1.46 



1.30 



3.10 



2.38 



1.90 



lo " 



3-53 



3-S2 



2.69 



i.=;fa 



1.82 



1.7b 



1.36 



1. 10 



1. 21 



I-S4 



1.26 



3.02 



2-59 



1.47 



Mdn't 



3.b7 



3.«.S 



2.76 



1.52 



1.76 



1.88 



1-33 



1.14 



1.07 



1.S6 



1.30 



3-04 



2.70 



1-45 



6 A.M. 



3-90 



3.t'S 



2.98 



1.32 



1.72 



1.8S 



1-59 



1.09 



I.2S 



1.48 



1.24 



3.20 



2.74 



1.47 



8 " 



3-«9 



3-57 



2.85 



i.3« 



I-9S 



1.99 



1.67 



1. 01 



1.26 



1-59 



1.23 



3.12 



2.71 



1-54 



All hours 



3-63 



3-27 



2.72 



1.58 



2.00 



1-97 



•■73 



1.23 



1. 41 



'■55 



1.30 



3.10 



2.59 



1.65 



