694 



Fishery Bulletin 92|4). 1994 



mersal eggs tend to have more protein than lipid 

 (Flachter and Pandian, 1968), which results in nega- 

 tive buoyancy. This could account for the high pro- 

 portion of protein in winter flounder eggs in contrast 

 to the high proportion of lipid in cobia, striped bass, 

 and red drum eggs. Another possible explanation for 

 the two very different patterns of biochemical com- 

 position is that cobia, striped bass, and red drum are 

 warm-temperate species whereas winter flounder is 

 a cold-water species. Cobia (Ditty and Shaw, 1992), 

 striped bass (Harrell et al., 1990), and red drum 

 (Vetter et al., 1983) have short incubation times: 24 

 hours at 29°C, 48 hours at 18°C, and 22 hours at 23°C, 

 respectively. Winter flounder has a much longer in- 

 cubation time, 11-20 days at 4-6°C (Cetta and 

 Capuzzo, 1982). The difference in incubation times 

 for different species is due in part to the effect of 

 temperature on metabolic rate of the developing 

 embryos. Catabolism of specific endogenous energy 

 stores in fish eggs is known to be related to the tem- 

 perature of incubation. Lipid tends to be consumed 

 in higher quantities at higher temperatures but pro- 

 tein consumption dominates at lower temperatures 

 (Heming and Buddington, 1988). Therefore, it is not 

 surprising to see different patterns of biochemical 

 composition in light of the temperature history dur- 

 ing early development of these different species. 



In addition to reporting the changes in biochemi- 

 cal composition during cobia ovarian development, 

 we also examined the cyclical variation in ovary size. 

 This was done by means of the gonosomatic index 

 ( GSI ), a commonly used ratio that normalizes gonad 

 size among animals of different size classes in order 

 to assess their reproductive state. The GSI was de- 

 termined for each female cobia sampled in this study 

 and compared both to stage of ovarian development 



and to month of capture. The majority of cobia landed 

 in April and May had ovaries in stage-3 condition 

 (-60%). This was reflected in the high mean GSI for 

 those months. By July and August fewer cobia, -30% 

 and 0%, respectively, had stage-3 ovaries; this was 

 reflected by the declining mean GSI. In September 

 the increase in cobia with ovaries in prespawning 

 condition was indicated by the slight increase in GSI. 



It is not clear why there was a greater proportion 

 of stage- 1 and stage-2 ovaries in August and Sep- 

 tember. Possible explanations include 1 ) difficulty in 

 distinguishing stage-2 ovaries from stage-2' ovaries; 

 2) presence of resident young, small fish that were 

 immature at the beginning of the summer but which 

 grew to maturity late in the season; or 3 ) an influx of 

 older, late-arriving cobia from unknown areas. We 

 believe that a combination of the first two explana- 

 tions is most likely. Some of the late summer/early 

 fall fish with ovaries classified as stage-2 fish may 

 well have spawned a batch of eggs earlier in the sea- 

 son and therefore were actually stage 2'. But after 

 the POF and any unspent stage-3 oocytes are re- 

 sorbed, it is not possible to distinguish between a 

 stage-2 and stage-2' ovary. On the other hand, the 

 late summer and early fall stage- 1 fish were small; 

 fork length was 94.8 ± 5.3 cm and 102.8 ± 7.9 cm in 

 August and September, respectively. Based on cobia 

 growth equations, 2 it is highly unlikely that these 

 fish could have spawned the previous year, and they 

 were probably too immature to have spawned ear- 

 lier in the same year. It is not known whether these 

 fish would have spawned in the fall or whether they 

 would have overwintered without further ovarian 

 development. 



Lotz et al. 2 suggested that cobia spawn over some 

 unspecified period of time during the May to Sep- 



