A STUDY OF CORRELATIONS AilONG TERRESTRIAL TEMPERATURES. 



361 



Since n = 4, and r, the number of monthly terms, is 24, 



»i^S/ = 35.8 

 Thus (9) gives the equation 



288v = A = + 6.9 



Carrying this computation through all the time-terms we have the following results 



Period 



1872-7.3 



1874-82 

 1883-89 



isno-98 



1899-00 



Sum 



Equation for r ' 



288r„==+ 7 

 129(3 +104 



1G80 4-127 



3240 -I- 8!) 



480 + 18 



6984r„==-f-345 



A positive correlation is well shown, leading to the mean result 



V = -0493 



T„ = d= 0.22° C. = ± 0.4° Fahr. 



^Vhen we add in the equation from Dove's work the final equation is 



9396v = 401 



whence 



T„=±0°.21 



The existence of the positive correlation is beyond serious question, but before we 

 accept it as cosmical, we must learn whether it holds between the more distant 

 stations, as well as between those in neighboring great geographic zones. 



As no correlation but a cosmical one can exist between the North American and 

 the other regions, we first compare that with the others. The table shows that simul- 

 taneous temperatures in North and South America are available from 1872 to 1898, a 

 period of 324 months. Forming the suni of the 324 products I'l^a we find the result 



2yy' = A= -I- 15.3 



Proceeding in the same way with the other stations the collected i-esults are : 



North America — South America; r = 324 Sec = + 15.3 



" " —India; " 348 " - 2.8 



" " — Batavia; " 348 " +18.4 



— Australia ; 



— Apia ; 



Sum 



" 216 

 " 132 



-I- 



0.0 

 1.0 



+ 31.9 



I 



