McDonough et al.: Sexual differentiation and gonad development in Mugil cephalus 



605 



Aging techniques 



Age was determined by using the left 

 sagittal otolith, which was embedded in 

 epoxy resin. A 0.5-mm transverse section 

 encompassing the otolith core was cut with 

 an Isomet low speed saw with diamond 

 wafering blades. The thin section of otolith 

 embedded in the epoxy was observed with 

 a dissection microscope at 20x magnifica- 

 tion, and age was recorded as the number 

 of annular rings present. The otoliths were 

 initially aged by one reader. A second reader 

 then evaluated a subsample of specimens 

 from 1998 and 2000 and all the otoliths 

 from 1999. The two groups of ages were 

 compared by the percentage of agreement 

 between the different age determinations 

 and by a paired Mest that allowed a com- 

 parison of the means and variances of the 

 two groups iCampana et al., 1995). Ages 

 were then validated by marginal increment 

 analysis in order to establish the timing 

 and periodicity of increment deposition 

 (Campana. 2001). In addition, the precision 

 of the ages was compared by using average 

 percent error (APE) between the two sets of 

 ages. "Precision" was defined as the repro- 

 ducibility of age determinations (Beamish 

 and Fournier, 1981; Chang, 1982). Using 

 the Levenburg-Marquardt procedure (Zar, 

 1984). we determined the growth curve 

 with a nonlinear least squares regression 

 of total length on age. 



Results 



Age structure 



CL 20 



0123456789 10 



Age 



Figure 2 



Age-frequency distribution (expressed as a percentage) for striped 

 mullet 'Mugil cephalus L. i from South Carolina estuaries October 

 1997 to December 2000. n = 3760. 



We recorded the age of 3760 specimens 

 and examined these specimens histologi- 

 cally to determine sex and maturity stage. 

 An additional 2524 young-of-the-year (age 

 0) specimens were used for the nonlinear 

 regression of total length on age. as well 

 as the sex ratios by both size and age. The 

 age range for striped mullet in this study 

 was to 10 years, and 1- and 2-year-olds 

 dominated the age distribution (Fig. 2). 

 There was 81.7% agreement for age data 

 between the two readers, and 99.5% agree- 

 ment within one year for both readers. A 

 Mest indicated no significant difference 

 between the two sets of age estimations 

 (r=2.898. df=1.233. P<0.05). The average 

 percent error (APE) (Beamish and Fournier, 1981) 

 between the two sets of age estimations was 0.41%. 



Marginal increment analysis indicated that growth 

 increments were deposited during July 'Fig. 3). The 



0.35 



0.30- 



E 0.25 



E 



| 0.20 



i 0.15 



0.10 



0.05- 



0.00 



— i — i — i — n— i— 



1998 



1999 



2000 



Figure 3 



Mean marginal increment distance by month for striped mullet i Mugil 

 cephalus L. I from South Carolina estuaries, October 1997 to December 

 2000. n = 3760. Marginal increment equals the otolith section radius 

 minus the distance from the core to the last annular increment. 



total length at age regression demonstrated a strong 

 relationship <r 2 = 0.864, df=3759, F stat =21,742, P<0.05). 

 Despite this strong relationship, there was a wide 

 range of sizes among the 1-, 2-. and 3-year-olds (Fig. 4). 



