Roumillat and Brouwer: Reproductive dynamics of Cynoscion nebulosus 



All 



thus potentially adding bias to our SF estimates. Even 

 though estimates of SF based on FOM were performed 

 a posteriori, we chose to report them strictly for com- 

 parison to other studies with this method. Because 

 sampling for females exhibiting FOM was accomplished 

 in a directed fashion, statistical comparisons were not 

 attempted. 



Batch fecundity and relative fecundity 



Observations taken over a decade of sampling the 

 Charleston Harbor estuarine system showed that in 

 females captured in shallow water (<1.5 m) during the 

 spawning season, FOM began at about 1200 h (Wenner 6 ). 

 Similarly, Crabtree and Adams' reported FOM begin- 

 ning in Florida spotted seatrout at about mid-day. Low- 

 erre-Barbieri et al. 8 found hydrated females in shallow 

 water in the vicinity of aggregations of drumming males 

 in deeper water. We also speculated that from mid- 

 to late afternoon hydrated females moved along the 

 marsh edge toward deeper water spawning aggrega- 

 tions (8-25 m). Hydrophone surveys conducted in the 

 Charleston Harbor area over several years (Riekerk et 

 al. 3 ; Wenner 6 ) indicated that noise production typically 

 began around 1800 h and ceased around 2200 h. Because 

 this behavior has been associated with spawning in this 

 and other sciaenids (Mok and Gilmore, 1983; Holt et al., 

 1985; Saucier et al., 1992; Saucier and Baltz, 1993), we 

 assumed that spawning began at 1800 h and stopped at 

 2200 h. Thus we were able to target spotted seatrout in 

 the mid- to late afternoon specifically to capture females 

 with oocytes in the late stages of FOM for batch fecun- 

 dity (BF) estimation. Because we have consistently iden- 

 tified recently spawned females in shallow areas, they 

 apparently return to the marsh edge where they once 

 again become available for capture with our sampling 

 gear. Our stratified random sampling of estuarine areas 

 along the coast (described previously) was designed to 

 representatively sample these recently spawned females 

 for SF estimation. 



We conducted BF sampling during two consecutive 

 afternoons fortnightly from the middle of April through 

 the first week of September 1998, 1999, and 2000. We 

 deployed a trammel net from a shallow water boat as 

 described above at preselected sites in Charleston Har- 

 bor in depths ranging from 1.0 to 1.5 meters during the 

 afternoon (1400-1800 h EDT) high tide. 



7 Crabtree R. E., and D. H. Adams. 1998. Spawning and 

 fecundity of spotted seatrout, Cynoscion nebulosus, in the 

 Indian River Lagoon. Florida. In Investigations into near- 

 shore and estuarine gamefish abundance, ecology, and life 

 history in Florida, p. 526-566. Tech. Rep. for Fed. Aid in 

 Sport Fish Rest. Act Project F-59. Florida Marine Research 

 Institute, Department of Environmental Protection, 100 

 Eighth Ave. SE, St. Petersburg, FL 33701. 



8 Lowerre-Barbieri, S. K., L. R. Barbieri, and J. J. Albers. 

 1999. Reproductive parameters needed to evaluate recruit- 

 ment overfishing of spotted seatrout in the southeastern 

 U.S. Final report to the Saltonstall-Kennedy (S-K) Grant 

 Program (grant no. NA77FD0074), 23 p. 



Restricting our sampling to the hours immediately 

 preceding the evening spawning event ensured that 

 those females preparing to spawn were available for 

 capture. Male spotted seatrout, identified by their 

 drumming sounds, caught during this targeted effort 

 were measured and released at the site of capture. We 

 supplemented samples for BF estimation with specimens 

 from local sportfishing tournaments held during sum- 

 mer months in the Charleston Harbor area. 



We processed samples in the laboratory as previously 

 described. If ovaries appeared by macroscopic examina- 

 tion to contain hydrated oocytes, they were fixed in 10% 

 buffered seawater formalin for potential counts (Hunter 

 et al., 1985). The appropriateness of these ovaries for 

 BF counts was subsequently determined by examining 

 the corresponding histological preparation. 



To ensure that only those oocytes destined to be ovu- 

 lated during the upcoming spawning event were counted, 

 we chose to use only those oocytes undergoing FOM that 

 could be easily separated by size from late vitellogenic 

 oocytes (Nieland et al.. 2002; Lowerre-Barbieri et al. 8 ). 

 If we observed numerous recent POFs in the histologi- 

 cal sample, the corresponding whole ovary was not used 

 for oocyte counts (because their presence indicated that 

 ovulation had occurred). We reweighed ovaries (approxi- 

 mately 2 weeks after fixation) to the nearest 0.01 g and 

 randomly extracted three 130-150 mg aliquots from 

 eight potential locations in the ovary (each lobe was par- 

 titioned into quarters lengthwise). We stored subsamples 

 in 50% isopropyl and counted oocytes under a Nikon 

 SMZ-U dissecting microscope at 12 x magnification. We 

 counted each subsample twice by using a Bogorov tray 

 and a hand-held counter and conducted a third count if 

 the two initial counts were dissimilar by more than 10%. 

 We used the mean number of oocytes in each subsample 

 to calculate mean oocyte density (number of oocytes per 

 gram preserved ovary weight) and total numbers of oo- 

 cytes in the ovary. We compared mean oocyte densities 

 among the four regions of each ovarian lobe and between 

 the two lobes by using a two-way analysis of variance 

 (ANOVA). Because our variances were heteroscedas- 

 tic, we used nonparametric ANOVA (Kruskal-Wallis or 

 ANOVA on ranks) for comparisons of mean BF among 

 ages, months, and years. To investigate the relationships 

 between BF and length, somatic weight (ovary-free body 

 weight), and age, we used linear regression. 



Relative fecundity (RF) was calculated as the num- 

 ber of oocytes per gram somatic weight (ovary-free). 

 To select samples for inclusion in RF calculations, we 

 looked for the presence of nuclear migration in histo- 

 logical preparations. We used this criterion to ensure 

 that oocytes of similar morphological dynamics would 

 be used, minimizing the potential for error. We used 

 the Kruskal-Wallis test to investigate the effect of age 

 on RF. Because sample sizes were quite uneven among 

 months, we chose to compare RF between the beginning 

 and end of the spawning season (May and August). This 

 comparison was done by using a Mann-Whitney test. 

 To corroborate any trends in RF, we also conducted 

 diameter measurements on the preserved (10% buffered 



