FISHERY BULLETIN: VOL. 74, NO. 2 



Lf = 2.500e ^^2' ^'"^ ^ 1-0 



AGE (months) 



Figure 2. — Growth of anchovy reared in the laboratory. The Gompertz growth 

 model of the formL, = Lq exp {C[l - exp (-an] } is used to describe the data. Solid 

 line is based on Lq fixed at 2.500 mm, and broken line is based on Lq estimated, 

 2.062 mm. Data from Schumann-Ill (unpubl. data. Southwest Fisheries Center, 

 La Jolla, Calif). The mean (circle), one standard deviation on each side of the 

 mean, and sample size are shown. 



Table 2. — Estimated age and average standard length of 

 northern anchovies reared in the laboratory by R. Leong (pers. 

 commun., Southwest Fisheries Center, La Jolla, Calif.) 



The Gompertz growth model was applied, but 

 did not adequately fit the data. This is charac- 

 teristic of asymptotic models like the Gompertz 

 model when all the data points are for a segment 

 of the growth curve where growth is relatively 

 slow and the plot of the data exhibits little 

 curvature. 



GROWTH FROM HATCHING 

 TO ADULT STAGE 



Growth Curve 



As indicated earlier, anchovies reared by 

 Shumann-III grew slightly faster than those of 

 Schumann-II, probably due to slight differences 

 in the rearing environment and procedures. Be- 

 cause our goal was to construct a general growth 

 curve and the differences in the data were rela- 

 tively slight, we elected to disregard the differ- 

 ence and pooled the data from the three experi- 

 ments (Schumann-II, Schumann-Ill, and Leong). 

 The Gompertz growth model was applied to the 

 pooled data. The results (Figure 3) indicate that 

 the model does not adequately describe the data. 

 For example, the model overestimates the sizes of 

 fish at about 4 to 12 mo old and underestimates 

 the sizes of fish older than about 13 mo. These 

 biases are caused by the steplike growth pattern 

 which produces plateaus at about 6 mo and 19 mo 

 of age. 



To account for this steplike growth pattern, a 



274 



