SAKAGAWA and KIMURA: GROWTH OF LABORATORY-REARED ANCHOVY 



scribed the data better. In the Kramer-Zweifel 

 analysis the length at zero age,Lo, was fixed at 2.5 

 mm, the average size at hatching. We also applied 

 the Gompertz grovd:h model to data of Schumann- 

 II. Kramer and Zw^eifel (1970) used data only for 17 

 days of growth. We used all of the data of Schu- 

 mann-II, which included sampling through 22 

 days of growth, and fitted the model first with Lq 

 fixed at 2.5 mm and again without this constraint, 

 i.e.,Lo was estimated. The results (Figure 1) indi- 

 cate that there is not much difference in the curves 

 withLo fixed or estimated within the range of the 

 data. Outside the range of the data, the curves 

 diverge considerably and there is a substantial 

 difference; the curve withLo estimated has a lower 

 asymptotic length (61 mm) than the curve withLo 

 fixed at 2.5 mm (asymptotic length of about 696 

 mm). 



Zweifel and Lasker^ showed that a two-phase 

 Gompertz curve described the data from Schu- 

 mann-II better than a single-phase Gompertz 

 curve. The separation of the phases occurred at 

 about 6 days of age, the onset of feeding in an- 

 chovy larvae. 



Schumann-Ill reared anchovies for a longer 

 period than Schumann-II. Fish reared by Schu- 



^Zweifel, J. R. and R. Lasker. 1974. Prenatal and postnatal 

 growth of fishes — a general model. Unpubl. manuscr. Southwest 

 Fisheries Center, La JoUa, CA 92038. 



mann-II, however, were larger than those reared 

 by Schumann-Ill at similar ages. For example, at 

 0.5 mo of age fish reared by Schumann-II aver- 

 aged 12.1 mm long and fish reared by Schu- 

 mann-Ill, 8.2 mm long. Although the sample size 

 is small, this difference is statistically significant 

 at the 1% probability level. Differences in rearing 

 procedures, i.e., diet and temperature of water, 

 probably produced the difference in growth 

 (Kramer and Zweifel 1970; Lasker et al. 1970). 



The Gompertz growth model was applied to 

 data from Schumann-Ill first withLo fixed at 2.5 

 mm and then with Lo estimated (Figure 2). As in 

 the case with data from Schumann-II, this model 

 describes the grovd;h data reasonably well, and 

 the curve with Lo estimated has a lower asymp- 

 totic length (81 mm) than the curve with Lq 

 fixed (asymptotic length of about 93 mm). 



GROWTH DURING JUVENILE 

 TO ADULT STAGE 



Anchovies reared by Leong (pers. commun.) 

 were juveniles at the start of the experiment and 

 grew to an average size of 117.7 mm in 474 days 

 (Table 2). Growth was in steplike stages charac- 

 terized by rapid growth followed by a leveling off. 

 The first stage was between 4 and 12 mo of age 

 and the second was between 12 and about 20 mo 

 of age. 



25 



ccoQM ^-0.570t, 

 L» = 2.500e ^-^29(1-6 ) 



L,-2.140e 3-546 (l-e-'-279t) 



0.0 



Q2 03 0.4 0.5 0.6 

 AGE (months) 



0.7 



0.8 



0.9 



FIGURE 1.— Growth of anchovy larvae reared in the laboratory. The Gompertz 

 growth model of the form, L , = Loexp {C [ 1 - exp (-at)] } is used to describe the data. 

 Solid line is based on Lofixed at 2.500 mm and broken line is based onLoestimated, 

 2.140 mm. Data from Schumann-II (Kramer and Zweifel 1970). The mean (circle), 

 one standard deviation on each side of the mean, and sample size are shown. 



273 



