12 



Fishery Bulletin 89(1). 1991 



Batch fecundity 



Batch fecundity in queenfish was posi- 

 tively related to female body size in each 

 year (Table 3; Fig. 1). Fecundity was 

 generally better related (based on higher 

 R 2 values) to somatic weight than body 

 length. Batch fecundity was dispropor- 

 tionately large in heavier females, as in- 

 dicated by the value of the slope in the 

 linear double-log plot (Fig. 1). Fecundity 

 also differed among years, even after ad- 

 justment for annual differences in female 

 size, with mean fecundity in 1984 signif- 

 icantly lower (by 20%) than mean fecun- 

 dity in the other four years (Tables 2, 3, 

 5; Fig.l). 



tn 



oi 



UJ too 



d 



z 



"O 9.5 



3 

 O 



CO 



1979 (.) 



LNF-1.125LN W*5.26 



R-BS4 



N-44 



P<001 



1980 (o) 

 LnF-1.190LnW*503 

 R-B44 

 N-126 

 P<.001 



1984 (•) 



LnF-1212 LnW*4.68 



R-.666 



N-71 



P<O01 



1985 (o) 



LnF-1249LnW*4.78 



R-.816 



N-77 



P<001 



1986 (4) 

 LNF-1.410LN W*4-14 

 R-.736 

 N-75 

 P<O01 



/3Q 



3.5 4,0 4.5 



Ln Somatic Weight (g) 



Figure 1 



Relationship between the log of batch fecundity (ln F) and log 

 female somatic weight (InW) during each of the five study 

 years. For illustration, mean fecundity data (+1 SE) are 

 plotted for each 10-g weight class. The allometric equation, 

 F = aW b (in log-linear form as InF = lna + bin W), and its 

 summary statistics are provided for each fitted regression line. 



Weight-specific fecundity 



Patterns of weight-specific fecundity (WSF, no. eggs 

 per g somatic weight) resembled those of batch fecun- 

 dity. WSF appeared to increase with female body 

 weight (ANCOVA of effects of somatic weight and year 

 on WSF: weight effect-F 1387 = 13.1, P<0.001), and 

 also seemed to vary among years (F 4 387 = 2.84, 

 P = 0.025). However, main effects were confounded by 

 a weakly significant weight-by-year interaction (F 4 ,3 8 7 

 = 2.57, P = 0.038). This heterogeneity of slopes pre- 

 vented adjustments for annual variations in female 

 weight and invalidated comparisons of intercepts 

 among all five years. Although WSF generally in- 

 creased with body weight, the disproportionate effect 

 of larger females varied among years. Most notable 

 was the particularly strong, positive influence of 

 somatic weight on WSF in 1986. If the 1986 data are 

 deleted and the ANCOVA analysis rerun, the slope 

 heterogeneity disappears (F 3 310 = 0.82, P = 0.49). 

 When main effects are reanalyzed, a strong year effect 

 (F 3313 = 6.02, P<0.001) becomes apparent, in addition 

 to that of somatic weight. This year effect disappears 

 (F 3 243 = 0.56, P = 0.57) if the 1984 data are removed. 

 Size-adjusted WSF in 1984 (mean ± SE = 264 + 15 

 eggs per g) clearly was less than that in 1979, 1980, 

 and 1985 (336 ± 8 eggs per g). 



Diameter vs. dry weight of eggs 



The median diameter of Gilson's-fixed, hydrated-state 

 eggs was significantly related to the mean dry weight 

 of these eggs (dry weight [in g, x 10~ 6 ] = 7.2 + 0.4931 

 egg diameter; r = 0.47, n = 46 females, P = 0.001). The 

 mass of hydrated-state eggs therefore was approx- 

 imately predicted (R 2 = 0.22) by egg diameter. While 

 appropriate, I acknowledge that a more direct and 



