Macewicz et al.: Fecundity, egg deposition, and mortality of Loltgo opalscens 



321 



(n.=1275) sampled from catch during the period 

 1998-99 (Fig. 12); this estimate is equivalent 

 to approximately 58% of the potential fecundity 

 calculated from mean length (129 mm). 



Our second approach was to estimate total 

 fecundity (E YD , standing stock of oocytes and 

 ova) indirectly using the combined formalin wet 

 weight of the ovary and the oviduct, in addition to 

 mantle condition. Combining ovary and oviduct in 

 one weight is more efficient than weighing them 

 separately because much less time is required for 

 dissection. Dorsal mantle length was also con- 

 sidered as a variable but it was not significant. 

 The final equation for the total standing stock of 

 oocytes and ova in a mature female squid is 



E YD = 378.28e 12 33C + ° 2447G - °- 24CG > 



(18) 



where C = mantle condition index; and 



G = gonad (ovary and oviduct) weight. 



5000 r 



4000 - 



o 3000 - 



2000 



1000 



E Y = 220.453e< 1 99C + 000791.) 

 for L= 129 mm 



Mature Females 



Collected Dec. 1998-Dec.1999 



n= 1275 



Mean C (0.625) ±2SE 



1 — 1 — 1 — r 



0.2 



0.3 



0.4 



1 — 1 — r~ 

 0.5 0.6 



~~ 1 — 1 — r 



0.7 0.E 



~\ T~ 



0.9 



1 1 



1.0 



Mantle condition index (C) mg/mm 2 



The predicted fecundity related well to the observed 

 with a pseudo r 2 of 0.60 (df=143). We also used 

 generalized additive models to estimate fecundity 

 (GAM, pseudo r 2 =0.64), as well as regression on 

 the first principal component which explained 867r of the 

 total variation (pseudo r 2 =0.55). Although the GAM gave 

 a slightly higher pseudo r 2 than the parametric nonlinear 

 regression, we chose the later for easier interpretation and 

 implementation. A pattern existed in the residuals from our 

 model (Fig. 13); the model overestimated some fecundities 

 at high mantle condition and underestimated fecundity 

 at low mantle condition. This pattern in the residuals is 

 probably related to the differences in density and size of 

 oocytes in the ovary. Regardless of the minor problem with 

 the residuals, this proxy (Eq. 18) for the standing stock of 

 oocyte and ova is preferred because it gives a much more 

 precise estimate at the minor additional cost of preserv- 

 ing and subsequently determining the combined weight 

 of ovary and oviduct. Although formalin weight of ovary 

 and oviduct are not presently monitored in the fishery, it 

 is a variable that could be added to fishery protocols at a 

 minor increase in cost. Another benefit of this more pre- 

 cise approach using E YD is that oviduct is included in the 

 estimate. If an estimate of the removal of fecundity by the 

 fishery is needed, ova must be included. Because ova are not 

 included in Equation 17, to add them requires using the ovi- 

 duct classification system (Table 1) to estimate the average 

 number of ova — a system that is imprecise but cheap. One 

 could, of course, use Equation 17 for E Y and either count the 

 ova in the oviduct or weigh the oviduct, but that would take 

 more work than applying Equation 18 for E YD . 



Discussion 



Potential fecundity 



Our estimate of Loligo opalescens potential fecundity is 

 based on a regression of the standing stock of oocytes on 



Figure 12 



Standing stock of oocytes in the ovary (£>•) as a function of mantle 

 condition index (C) for a 129-mm mature female L. opalescens as 

 predicted by Equation 17 (equation also given on top of panel; L is 

 dorsal mantle length). Dashed lines are ±2SE. The mean E Y for the 

 females taken in the fishery was 2221 oocytes. 



dorsal mantle length for mature preovulatory females 

 having yolked oocytes in their ovaries. The accuracy of 

 this approach depends upon the assumption that these 

 females are at the point in life when the standing stock 

 of oocytes in their ovaries is equivalent to their potential 

 lifetime fecundity. This key assumption would not hold if 

 some of the mature squid classed histologically as pre- 

 ovulatory had in fact spawned. We do not know how long 

 postovulatory follicles are distinguishable from atretic 

 structures in the ovary of L. opalescens and, as far as we 

 know, the rate of degeneration has not been determined for 

 any loliginid. We know from our work on anchovy (Hunter 

 and Goldberg, 1980; Hunter and Macewicz, 1985a), 

 although it is not a cephalopod, that postovulatory follicles 

 are distinguishable from atretic structures in the ovary of 

 anchovy for about two to three days after spawning when 

 the water temperature is about 16°C. This means that for 

 undetected spawning to occur in L. opalescens, the inter- 

 val between ovulation periods would likely need to exceed 

 three days. This may be a minimum estimate because L. 

 opalescens spawn at lower temperatures (9-13 C, Butler 2 ) 

 than do anchovy. Definitely a laboratory study on the 

 rate of degeneration is necessary because postovulatory 

 follicles in fish degenerate slower at lower temperatures 

 (Fitzhugh and Hettler, 1995). In addition to the absence 

 of postovulatory follicles, the oviduct must be empty for a 

 spawning act to be undetected. Undetected ovulation and 

 spawning seems unlikely because females with multiple 

 stages of postovulatory follicles were common (87% of 247 

 mature females), females with only old postovulatory fol- 

 licles were not detected, and the average life span on the 

 spawning grounds may only be a few days. 



Atretic losses of oocytes are another possible bias in 

 estimating potential fecundity. Atresia (degeneration and 

 resorption of an oocyte and its follicle) appears to be a 



