McDonough et al.: Fecundity and spawning season o\ Mugil cephalus 



831 



recently spawned ovaries (characterized by the presence 

 of postovulatory foHicles). Atretic ovaries were observed 

 from December through May. There were no postovu- 

 latory follicles observed, indicating that any atretic 

 ovaries were not from recently spawned fish. The fish 

 with atretic ovaries were characteristically emaciated 

 for their size (TL and BW) and were most common from 

 January through March. The presence of females with 

 atretic ovaries starting in December is strong evidence 

 that spawning occurred in November, if not earlier, 

 and females with atretic ovaries caught as late as May 

 demonstrate that spawning may still occur as late as 

 April. Additional evidence for the October through April 

 spawning period has also been shown in backcalculated 

 birth dates for juvenile striped mullet by daily growth 

 increments (McDonough and Wenner, 2003). This 

 evidence supports the concept of offshore spawning 

 in striped mullet and a yet undetermined time period 

 required for moving from the estuaries to the spawning 

 areas and for returning again to the estuaries. Other 

 authors have come to the same conclusion from similar 

 evidence in estuaries throughout the southeast (Jacot, 

 1920; Broadhead, 1956; Anderson, 1958; Stenger, 1959; 

 Shireman, 1975; Dindo and MacGregor, 1981; Greeley 

 et al., 1987; Render et al., 1995; Hettler et al., 1997). 

 All of the fecundity specimens were caught from Oc- 

 tober through February when the mean monthly GSI 

 was highest. Pien and Liao (1975) found that mullet 

 oocytes reached a hydrated size of 900 to 1000 ^m. The 

 size of oocytes used for fecundity counts in the present 

 study ranged from 463 to 682 ^m. The maximum size of 

 oocytes in the tertiary stage of vitellogenesis from our 

 study was 600 ^m or greater. This result agrees with 

 those of previous studies where the maximum size of 

 oocytes prior to either hydration or atresia (if spawning 

 did not occur) ranged from 600 to 700 ^m (Shehadeh et 

 al., 1973b; Kuo et al., 1974). There was no evidence of 

 prespawning atresia in any of the specimens used for 

 fecundity estimates. 



The appropriateness of using a GSI alone to determine 

 the level of reproductive development has been questioned, 

 particularly for serial or asynchronous spawning fishes (De 

 Vlaming et al., 1980; Hunter and Macewicz, 1985). Striped 

 mullet can have a wide range of GSI values that range from 

 practically zero to over thirty (Render et al., 1995). The GSI 

 range for females in our study ranged from almost zero to 

 27.7. Because of the high variability in GSI with size, it 

 does not appear appropriate to use GSI alone in order to 

 assess reproductive development in striped mullet. When 

 used in conjunction with histological analysis and mean 

 oocyte diameter of tertiary-stage oocytes, GSI does provide 

 excellent supporting evidence of reproductive schedules 

 and spawning season duration. GSI is probably more ap- 

 propriately used for isochronal spawning fishes than for 

 serial spawning fishes because of the uniform development 

 of oocytes in the former. However, it is still difficult to meet 

 all the basic assumptions of the GSI index as given by De 

 Vlaming et al. ( 1980) because of the high variability of GSI 

 with size. Another technique that has been used in aquacul- 

 ture situations to assess maturity and sex involves the use 



g. 



o 



6 

 c^ 



>^ 



g 200 



0) 

 3 

 CT 



Developing males 

 n = 340 



n 



JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 



B 



Developing females 

 n = 277 



I 



n 



JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC 



Figure 8 



Frequency distribution of reproductively developing male (A) 

 and female (B) striped mullet by month in South Carolina 

 estuaries from 1998 to 2000. 



of a cannula to remove oocytes from the ovaries of live fish 

 which are then evaluated (Shehadeh, et al, 1973a; Kuo et 

 al., 1974). This technique, although useful for determining 

 sex and the extent or stage of reproductive development, 

 would be inappropriate for estimating potential fecundity. 

 Historically, reproductively developing mullet have been 

 found in southeast United States waters from November 

 through December (Jacot, 1920; Anderson, 1958; Stenger, 

 1959). During our study, reproductively developing mullet 

 were caught in the Charleston Harbor Estuary from Oc- 

 tober through February. Gonad development of these fish 

 was discernible through gross morphological observation, 

 and histological sections showed that vitellogenesis was 

 well underway. Other studies have shown that previtel- 

 logenic oocytes were usually less than 160 jum and that 

 the onset of vitellogenesis began when the oocytes reached 

 a size of 180 pm (Dindo and MacGregor, 1981; Greeley et 

 al., 1987; Tamaru et al., 1994; Render et al., 1995). Devel- 

 oping individuals caught during our study that were not 

 used for fecundity counts had less-developed ovaries, GSI 

 values less than 7, and mean oocyte diameters less than 



