McDonough et aL: Fecundity and spawning season of Mugil cephalus 



827 



1 2 



5 6 



Age (yr) 



10 



Figure 4 



Age-frequency distribution of fecund striped mullet in South Carolina estuaries from 1998 to 

 2000. n (number offish in sample) = 50 fish for 1998, n = 37 for 1999, and /; = 42 for 2000. 



of another study where size and age structure of striped 

 mullet in South Carohna was examined (Wenner and Mc- 

 Donough, unpubl. data^). A comparison of multiple read- 

 ings of the same group of otoliths assessed aging precision. 

 One year (1999) was chosen at random and all of the oto- 

 liths («=1234) from that year were aged by a second reader. 

 The ages of the two independent determinations were then 

 compared by using a one-way ANOVA and a ^test. The 

 variance statistic was 2.78 for the original ages and 2.81 

 for the second age reads, which were not significantly dif- 

 ferent (P=0.001) and both had almost identical normalized 

 residuals. Overall, there was an 83.4% agreement on ages 

 between readers. The results from the ANOVA (F= 1555.0, 

 df=10, P=0.000) and the t-tesi «=2.898, df=1233, signifi- 

 cance |2-tailed] =0.004) both confirmed that there was no 

 significant difference between the separate age determina- 

 tions. Therefore, the age recorded by the first reader for all 

 specimens was used in the analysis. 



The length-weight relationship for fecund female striped 

 mullet was compared by using a linear regression of (natu- 

 ral) log-transformed body weight against total length to 

 see if there were any differences between years. The 

 regression coefficients from each year were compared by 



Wenner, C. A., and C. J. McDonough. 2001. Cooperative 

 research on the biology and assessment of nearshore and estua- 

 rine fishes along the southeast coast of the U.S. Part IV: Striped 

 mullet, Mugil cephalus. Final rep, Grant no. NA77FF0550, 

 82 p. Marine Resources Research Institute, South Carolina 

 Dept. of Natural Resources, P.O. Box 12559 Charieston, S.C. 

 29422-2559. 



using a test of significance between more than two slopes 

 (Zar, 1984). The weight measurement used was total body 

 weight minus gonad weight (TW-OW=BW) because ovary 

 weight had a considerable influence on total body weight 

 in the fecund specimens (GSI values of 7.7 to 27.7). There 

 was no significant difference in the total length to Ln 

 body weight regressions between different years (F=9.22, 

 P=0.001, df=129). Because there was little difference in the 

 regressions between years and in order to increase sample 

 size, data from all three years were combined to obtain the 

 overall total-length to Ln-body-weight relationship of Ln 

 BW = -11.1 + 2.92 (TL) (Fig. 5). In contrast, the length and 

 body-weight-at-age relationships were highly variable; a 

 wide range of sizes occurred in the 2-, 3-, and 4-year age 

 classes (Fig. 6). The high degree of variability was also ex- 

 acerbated by the smaller number offish age 5 or older. 



The gonadosomatic index (GSI) for fecund mullet ranged 

 from 7.7 to 27.5 in 1998, 9.3 to 27.7 in 1999, and 9.5 to 26.6 

 in 2000. In contrast, the GSI for mature females that were 

 not undergoing any reproductive development ranged from 

 almost zero to 4 for all three years of the study. The rela- 

 tionship of GSI to size (TL or BW) was not very strong in 

 any year. However, GSI was positively correlated (P=0.01) 

 with oocyte diameter and negatively correlated with oo- 

 cyte density (Table 2) because of the inverse relationship 

 of oocyte density and oocyte diameter. The correlation coef- 

 ficient for GSI and age were very close to zero and slightly 

 negative (Table 2). 



Mean GSI by month for males and females (Fig. 7) in- 

 creased from October through April, peaking in November- 



