315 



In Table III (p. 314) tlic different correlation coefficients are 

 given and the distances between the stations expressed in degrees 

 of the great circle corresponding to about 111 k.m. 



For ascertaining meteorological conditions, the regression-equations 

 (preferably called meteorological condition equations) are of greater 

 importance than these general, interdependent correlation coefficients. 



X, = 0.238 a;, + 0.520A', + 0. Oil a-,-|-0.201 .«, + 0.292 a-, 

 .^, = 0.928 .B, + 0.416 .(,■, — O.O90 .b,— 0.485 .»,— 0.112 x, 

 X, = 680 .Bj + 0.109 .», + 0.242 .r,— 0.020 x, -0.336 x, 

 X, = 0.088 .Bj — 0.076 .B, + 594 .c, + 0.353 .b,— 0.150 x, 

 X, = 0.457 ,B^-0.259 ,b^- 0.054 .«, + 0.250 .r, + 0.396 x„ 

 X, = 0.929 .B, + 0.063 .B, -0.822 .», — 0.150 .«,+ 0.573 .b, 



R 



0.943 

 0.830 

 0.903 

 0.672 

 0.843 

 0.873 



(■1) 



The partial c.c. calculated from the coefficients of these equations 

 are given in Table IV, arranged according to their magnitude. 



TABLE IV, Partial correlation-coefficients. 



i 



Helder- Clermont . . . . | 0.594 



Helder— Christiansund . . 0.521 



Neufahrvv. -Christiansund . 0.476 



Helder— Valencia .... 0.470 



Milan— Clermont . . . . j 0.379 



Helder— Neufahrwasser . . i 0.303 



Milan Neufahrwasser . . 0.297 



Valencia-Clermont . . . 0.213 



Valencia— Christiansund. . ! 0.084 



Helder— Milan | 0.020 



Clermont -Neufahrwasser . -0.037 



Valencia— Milan .... I —0.085 



Milan— Christansund . . — 0.150 



Valencia -Neufahrwasser . i - 0.355 



I 



Clermont— Christiansund . I — 0.526 



From these results it appears that the ciioice of the stations was 

 good, except Milan which, although at about the same distance from 

 Helder as Clermont, still exercises a much smaller influence. 



Clermont and Milan being at a mutual distance of onlj- 4". 2. it 

 is possible that this result is due to purely arithmetical reasons : 

 the method followed involves that two stations near to each other 

 must be considered as one, because it depends on incalculable factors 

 how the common effect is distributed over either point, this being 

 of no importance for the result. 



If this were the case, however, the |)artial c.c. between Clermont 

 and Milan ought to be nearly equal to unity, which is contradicted 

 by the c.c. : 0.379. 



