Table 6. 



FISHERY BULLETIN: VOL, 77. NO 3 



-Fecundity (eggs per gram female) eEtimates for northern anchovy females collected in February 1978 off 



southern California. 



'Only the nonhydrated eggs m most advanced mode are included 



'Females having both hydrated eggs and postovulatory follicles were included 



advanced eggs in a weighed sample of the gonad. 

 The number of hydrated eggs per gram of ovary 

 did not vary with fish weight and was 2,880 ±373 

 ( ±2 SE) eggs/g of ovary. If any weight-related bias 

 existed, it probably was related to the gonad 

 weight-female weight relation. To ascertain how 

 ovary weight varied with female weight, we sepa- 

 rated the data into three classes on the basis of 

 mean diameter of eggs in the most advanced mode 

 and plotted ovary weight as a function of female 

 weight for each class (Figure 5). The relation was 

 slightly curvilinear; the departure from linearity 

 was most ovbious in the 0.51-0.60 mm egg diame- 

 ter class. Considering the variability in the 

 number of mature eggs per gram of ovary, and the 

 slight departure from linearity, no practical pur- 

 pose is achieved in expressing fecundity as a func- 

 tion of female weight rather than as a direct pro- 

 portion of weight, although direct proportionality 

 is somewhat less accurate for extreme weight 

 classes. 



The relation between gonad weight and female 

 weight differed somewhat among the diameter 

 classes as can be seen in Figure 5. This might be 

 expected because the weight of the ovary should 

 increase somewhat with the average diameter of 

 the eggs in the most advanced mode. We analyzed 

 the data using multiple regression to determine if 

 the In doge) gonad weight (G) could be estimated 

 from the diameter of eggs in the most advanced 

 mode(D), In ovary-free female weight (W), and the 

 interaction term (D In G ). Both female weight and 

 the interaction term had a significant effect on 

 gonad weight, whereas diameter alone did not. 

 The final multiple regression equation was: 



D = 



In G + 4.213 - 1.069 In W 



0.555 In W 



InG 



-4.213 + 1.069 In U' + 0.555 D In W 



where r^ = 0.92. Solving for diameter we obtain 

 650 



This equation may be useful for estimation of 

 maturity stages for anchovy from weight relation- 

 ships; 60% of the estimates of mean diameter were 

 within ±0.1 mm of the observed values and thf 

 residuals were distributed evenly. The equation is 

 more useful than gonad index (ovary weight 

 female weight), which is commonly employed to 

 assess maturity, because it produces a number 

 that can be directly related to reproduction and 

 avoids a weight bias for extreme weight classes. 

 The weight bias in gonad index is apparent by 

 examination of data in Figure 5 (lowest); a 30 g 

 female with eggs of 0.65 mm in the most advanced 

 mode has a gonad index of 0.064 whereas that of a 

 10 g female at the same stage of maturity has an 

 index of 0.040. The equation also identifies 

 females with hydrated eggs; the average diameter 

 of eggs estimated by the equation for females with 

 hydrated eggs was 1.20±0.12 mm in = 22) and is 

 close to the mean of spawned eggs ( 1.34 mm). Ob- 

 viously, such equations would be specific to popu- 

 lations having similar weight relations, but it does 

 seem a useful approach for assessing maturity. 



DISCUSSION 



This paper provides a method for direct estima- 

 tion of the frequency of spawning of a multiple 

 spawning pelagic fish population. From such es- 

 timates it may be possible to directly estimate 

 spawning biomass from the abundance of eggs and 

 larvae over a short segment of the breeding sea- 

 son. One of the major assumptions underlying the 

 estimate is that a representative sample of 

 females is obtained. Spawning frequency would be 

 overestimated if nonspawning females were in re- 



