670 



Fishery Bulletin 91(4), 1993 



Development and maturity of ovaries 



Ovarian development of P. granulosa was described in 

 detail by Lovrich (1991) and is similar to that of L. 

 santolla (Vinuesa, 1984). The only difference with the 

 latter was that oocytes in the ovaries of female P. granu- 

 losa that had recently molted and spawned were al- 

 ready in an early vitellogenic phase. These oocytes 

 were surrounded by fibrous connective tissue and ra- 

 diated outwards from the germ strand. 



Two groups of females, separated by differences in 

 oocyte diameter (OD) and gonadosomatic index (GSI), 

 were present in virtually all of the sampled months 

 (Figs. 5 and 6). Discriminant analysis on a matrix con- 

 sisting of CL, CH, OD and GSI (n=220 females sampled 

 during 1990) gave 2 groups. The first included POM, 

 EIM, or MIM females with eggs in ES I and ES II, the 

 second comprised AIM or PRM females with eggs in 

 ES III, ES rV, or ES V. The resulting discriminant 

 function was 



y = 0.083 CL + 0.08 CH - 0.727 OD - 0.323 GSI, 



where coefficients were standardized by the standard 

 deviations within groups. The canonical correlation was 

 0.798. The reproductive variables OD and GSI domi- 

 nated the function, whereas morphological measure- 

 ments contributed negligibly (canonical loadings: CL: 

 0.17; CH: 0.155; OD: -0.955; GSI: -0.866). 



An a posteriori classification was generated from the 

 discriminant function. After applying the discriminant 

 function to data for individual females we found that 



mean scores (i.e., y in function above) were signifi- 

 cantly different for each group (Games and Howell 

 test, t'=2A9; P<0.01). We thus conclude that there are 

 two distinct groups differing in shell condition and 

 development of ovaries and eggs. These two groups 

 are a key feature of a biennial reproductive cycle. 



Females with asynchronous embryonic development 

 were grouped by the discriminant analysis with fe- 

 males in their first year of the reproductive cycle. In 

 fact, for these females, mean GSI was 2.78% in August 

 and 3.38% in October, representing 45 to 55% of maxi- 

 mum GSI (fully developed ovaries in October, Fig. 5), 

 respectively. Their average OD was 1.24 mm in August 

 and 1.36 mm in October, whereas the maximum ex- 

 pected at the end of the reproductive cycle was 1.8 mm 

 (Fig. 6). 



Fecundity and reproductive effort 



Fecundity was studied in relation to carapace length 

 in the Ushuaia (Fig. 7) and Becasses Islands. Clutches 

 in late stages (ES IV or ES V) were analyzed sepa- 

 rately to avoid underestimations due to possible loss 

 of eggs (Kuris, 1991). Regressions of fecundity on cara- 

 pace length were significant (Table 2; Fig. 7) and the 

 slopes did not differ between females carrying clutches 

 with eggs in ES I-II, in ES V, or with asynchronously 

 developing eggs (from Becasses Islands) (F=1.12; 

 P=0.33). ANCOVA indicated that females with ES V 

 eggs or with asynchronously developing eggs had fewer 

 eggs per clutch than those with ES-I and ES-II eggs 

 (F=7.29; P<0.001) (Table 2). Adjusted fecundity was 



JASONMAMJJASOD 



(20) (21) (14) (21) (18) (21) (29) (29) (30) (30) (38) (31) (53) (9) 



1989 



1990 



Figure 5 



Relative frequency of oocyte diameter for female Paralomis granulosa caught in the Beagle Channel. 

 Sample size is indicated in parentheses below each month. 



