FISHERY BULLETIN: VOL. 73, NO. 4 



In this paper we have used a different statistical 

 technique than Peterson's and have reached a 

 different result regarding lag differences. We 

 have computed the correlation coefficient for 

 varying lag times. This technique uses higher 

 resolution in the data than the techniques used by 

 Peterson and results in a graphical presentation of 

 the magnitude of correlation at each lag. The 

 results show that significant correlation extends 

 for longer lag times than V2 or l'/2 yr and reverse 

 the order of the difference in lag times along the 

 coast. Also, we have used the same technique to 

 describe the relationship between catches in 

 different years. These results show that the 

 primary cause of the cyclic nature of the variations 

 may not be upwelling. 



MATERIALS AND METHODS 



The data used were the same as those used by 

 Peterson (1973). The catch data are based 

 primarily on the annual reports of the Pacific 

 Marine Fisheries Commission and the California 

 Department of Fish and Game (see Peterson 1973 

 for detailed discussion of sources). The total catch 

 for each year (Figure 1) is believed to represent 

 almost all that year's legal-size male crabs (Pacific 

 Marine Fisheries Commission 1965:38). The 

 completeness of this yearly sample makes the 



Dungeness crab an attractive species for analysis 

 of population dynamics. 



The upwelling index developed by Bakun (1973) 

 is based upon measurements of atmospheric pres- 

 sure. Upwelling and downwelling are the result of 

 Ekman transport which is, in turn, due to surface 

 winds. Bakun developed a method of estimating 

 the intensity and direction of Ekman transport by 

 relating the wind speed and direction to at- 

 mospheric pressure measurements. Using this 

 method, he computed monthly averages of the 

 component of Ekman transport perpendicular to 

 and away from the shoreline. This upwelling index 

 is assumed to indicate the average monthly up- 

 welling. Peterson (1973) summed the monthly 

 Bakun upwelling index over the summer up- 

 welling season to compute a seasonal upwelling 

 index (Figure 2). The upwelling indices were 

 summed over the time periods from May through 

 September for California. 



These data were analyzed using a technique 

 commonly employed in analysis of random 

 processes to determine the degree to which two 

 processes (or separated points of one process) 

 covary. The formula used was: 



RJi) = 



1 



"xy' 



n-i 



("-0 j=i 



x(j)i/U+i)/S,S^ 



10 



\0 



o 



s 



H 



^ 



California 



Oregon 



Washington 



-5 



Figure l.-Crab catch data for Washington, Oregon, and California (mean removed). 



902 



