amplitudes a and phases d can be obtained from the ^'s. The quantity e is the 

 random error or residual term. 



For fe = 1, equation (2) was fitted to each of the six sets of surface 

 temperature data. The results are shown in table 2. 



TABLE 2. HARMONIC ANALYSIS OF ANNUAL OSCILLATION 



Location 



Regression Coefficients 



Amplitude 

 a 



Phase 





^0 



i8. 



^. 



6 



PAPA 



8.54 



-3.51 



-1.81 



3.95 



154.9 



ECHO 



22.06 



-3.66 



-1.70 



4.03 



157.4 



Cape St. James 



9.11 



-2.03 



-1.84 



2.74 



139.9 



Triple Island 



9.23 



-2.23 



-2.26 



3.18 



136.6 



Langara Island 



8.64 



-2.13 



-1.97 



2.90 



139.2 



Scripps Pier 



16.91 



-1.93 



-2.60 



3.24 



128.4 



Certain measures related to the statistical fit are given in table 3. The 

 quantity R is the multiple correlation coefficient; oa and a^ are the standard 

 deviations of the observations about their mean and about the fitted regression 

 curve, respectively. The F-ratio indicates whether or not the regression curve 

 has significantly reduced the total sum of squares. 



TABLE 3. STATISTICAL FIT OF ANNUAL OSCILLATION 



Location 



R' 



<7o 



<^D 



ao* 



F-ratio 



F-ratio* 



Fo.o, 



PAPA 



0.88 



2.99 



1.05 



1.21 



811 



605 



4.6 



ECHO 



0.89 



2.98 



0.99 



1.17 



876 



626 



4.6 



Cape St. James 



0.79 



2.19 



1.00 



1.09 



553 



465 



4.6 



Triple Island 



0.82 



2.49 



1.06 





774 





4.6 



Langara Island 



0.79 



2.31 



1.05 



1.13 



612 



546 



4.6 



Scripps Pier 



0.76 



2.63 



1.28 





569 





4.6 



*Adjusted for nonrandom missing data 



Van Vliet' has shown that the variances of regression coefficients 

 should be increased if there cu-e nonrandom missing data. A similar situation 



19 



