Sardine sampling was extended to include San Francisco by the 

 California Department of Fish and Game during the 1930-31 season. Owing 

 to a lack of personnel, semiweekly sampling was not carried on at first, 

 but whenever feasible 1,000 fish were measured in the middle of each 

 lunar month. Regular sampling was started in 1935-36 and carried on by 

 the California Department of Fish and Game until the 1938-39 season when 

 catches were sampled by both California and the U. S. Fish and Wildlife 

 Service . From that season to date all of the sampling in the San Fran- 

 cisco area has been done by the U. So Fish and Wildlife Service , During 

 the years of poor fishing in San Francisco and Monterey, samples were 

 taken when fish were available . 



The data are grouped in somewhat different time periods in the 

 early records , Since it was difficult to secure monthly catch records 

 for early years, the length frequencies of the samples were weighted 

 to quarterly catch data from 1919 through 1925° Although during this 

 period fishing continued throughout the year, only the fall, winter, and 

 spring landings were used in calculating the season total (April, May, 

 and June were omitted) „ This was done in order that the length composi- 

 tion of the catch might be comparable with later years when no regular 

 fishing was carried on at this season- 



From the 1926-2? through the 1937-38 seasons, catch records were 

 Available only by calendar montho All samples were weighted by calendar- 

 month catches „ From the 1938-39 season to date, all samples have been 

 weighted to the lunar-month catches . 



Because sampling was somewhat irregular in San Francisco from 1930- 

 31 through 1937-38, the weighted frequencies of San Francisco and Monterey 

 have been combined,. During several months when fish were landed, no 

 samples were taken in San Francisco. In these instances, the San Fran- 

 cisco catch was weighted to the Monterey samples. 



Beginning in 19U1-U2, the average weight of the 50-fish sample .from 

 each boat was obtained. The number of fish caught each week was estimated 

 from the total weight landed at each port divided by the average weight 

 of fish sampled during that period. For seasons prior to 19U1 estimates 

 of the number of fish caught during each quarter, month, or lunar month 

 were based on average length and weight data given by Clark (1928, pp. 37- 

 38). Average weights of each size group were multiplied by the frequen- 

 cies of the lengths in the monthly (or other) summaries to obtain the 

 total weight of the samples. Total weight divided by the total number 

 sampled yielded the average weight of fish taken during the period. The 

 total catch for the period divided by the average weight provided the 

 total number of fish landed. Totals are given in 100,000' s of fish. In 

 the few instances when the weighted total does not agree exactly with 

 the catch in numbers of fish as recorded in other publications, the 

 discrepancy is due to rounding of numbers. 



