CLARKE: FECUNDITY AND SPAWNING FREQUENCY OF NEHU 



that the significant differences for winter may have 

 resulted from chance alone. 



The regression statistics (Table 1) for data from 

 each season indicate a great deal of variability about 

 the functional relationships between fecundity and 

 length or weight or between the logarithms of these. 

 The correlation coefficients (r) for all regressions 

 were significantly (P < 0.05) different from zero, but 

 the coefficients of determination (r^) indicated that 

 only about half the variance of fecundity or In fecun- 

 dity was accounted for by the regression. The ex- 

 ponents from the logarithmic regressions of fecun- 

 dity on length are considerably higher than those 

 of the weight-length relationships, and the expo- 

 nents from the logarithmic regressions of fecundity 

 on weight are significantly greater than one. Both 

 indicate that fecundity is not linear vdth weight and 

 that the appropriate expressions for the functional 

 relationship with size are the power curves for 

 fecundity vs. weight. The exponents of the curves 

 for the two seasons were nearly the same, while the 

 summer- winter ratio of the preexponential factors 

 (antilogs of the regression intercepts), 1.56, was 

 almost identical with the ratio of mean relative 

 fecundities, 1.54. 



A small, but significant part of the variability in 

 fecundity within seasons was related to variation in 

 length- weight relationships of the fish. Using pre- 

 dictions of weight and fecundity from Model I (least 

 squares) logarithmic regressions on standard length, 

 I tested for correlations between relative deviations 

 (observed-predicted/predicted) of fecundity and 

 weight. The relative deviations were positively and 

 significantly correlated for both seasons (summer: 

 r = 0.44, P < 0.01; winter : r = 0.24, P < 0.05). 

 The coefficients of determination, however, indicate 

 that the variation in relative deviation from pre- 

 dicted weight accounted for small percentages of the 

 variation in deviation from predicted fecundity. 

 Maximum relative deviations in weight were ca. 

 + 20% about the predicted value, while deviations 

 in fecundity were much broader: ± 75% in summer 

 and ± 60% in winter. Thus there was a tendency for 

 relatively "fat" individuals to have higher fecundity, 

 but this did not account for much of the scatter in 

 the fecundity data. 



Nematodes were the only parasites noted fre- 

 quently, and their presence had a minor and insig- 

 nificant effect on fecundity. About half of the sum- 

 mer fish and about a third of the winter fish had 

 nematodes. For both seasons, the exponent from the 

 logarithmic regression of fecundity on weight was 

 higher for fish without nematodes than for those 



with them, but the 95% confidence limits over- 

 lapped. 



G/S values ranged from under 2% to about 12% 

 in summer fish and to about 7% in winter fish. For 

 females with maturing oocytes, G/S is a function of 

 both the number and size of oocytes. LeCluse (1979) 

 showed for Sardinops ocellata that ovum dry weight 

 does not increase once hydration begins, and my 

 own preliminary data indicated that this was also 

 true for nehu. Thus effects of variation in oocyte size 

 could be eliminated by considering only fish with 

 LMD >0.75 mm— the size at which hydration begins. 

 The mean G/S for such fish from winter was 4.8% 

 (n = 67: range: 2.4-7.1%) and from summer, 6.3% 

 {n = 44; range: 2.1-12.0%). Among fish with LMD 

 >0.75 mm, the exponents from logarithmic regres- 

 sions of gonad weight on fish weight were signifi- 

 cantly greater than one for both seasonal groups 

 (Table 1). 



Postovulatory Follicle Deterioration 



Although the number of specimens examined for 

 postovulatory follicles (POF) was limited (107 from 

 13 different samples), the results indicated that POF 

 were a reliable indicator of recent spawning up to 

 about 16 h after spawning. Among the 80 specimens 

 from 9 samples taken 1-5 h after estimated spav/n- 

 ing time, follicles were either present and obvious 

 or completely absent. Only seven mature females 

 were available from between midnight and dawn. 

 There were no traces of POF in one specimen; in 

 the others, POF were obvious but showed some 

 signs of degradation similar to that described for 

 northern anchovy, Engraulis mordax, by Hunter 

 and Goldberg (1980). Among the 20 specimens from 

 two samples taken 14-16 h after spawning, POF 

 were further degraded but still distinguishable from 

 other structures in half the fish, while the others 

 showed no traces. Judged from descriptions of POF 

 in E. mordax by Hunter and Goldberg, 14-16 h in 

 nehu appears roughly equivalent to 24 h in E. mor- 

 dax. Although controlled experiments such as those 

 of Hunter and Goldberg were not conducted, it 

 seems likely that, later in the day, POF cannot be 

 distinguished reliably enough to indicate spawning 

 the previous night. Since POF were either present 

 and very obvious or totally absent in fish collected 

 during the night, all traces of previous spawning are 

 apparently gone after about 24 h. 



Spawning 



Examination of fish from purse seine samples 



131 



