Fishery Bulletin 92(1), 1994 



would probably show the first mark and the core 

 virtually fused together. Since Atlantic croaker also 

 spawn over a long period in the Gulf of Mexico 

 (White and Chittenden, 1977), this might explain 

 why the first annulus was apparent in only a por- 

 tion of Barger's (1985) fish. 



Our interpretation of the first annulus is also con- 

 sistent with evidence from another ageing method. 

 Ross (1988) reported that some Atlantic croaker 

 from North Carolina showed an early, age-0 scale 

 mark, apparently formed during their first winter. 

 However, they were not counted as annuli. 



The high precision of repeated age readings and 

 validation of annuli almost to the maximum ob- 

 served age indicate that otolith sections represent 

 a very reliable method for ageing Atlantic croaker. 

 Identifying the first annulus may require some prac- 

 tice, but all other annuli are extremely clear and 

 easy to identify. Otolith sections do not have the 

 problems scales reportedly do, such as the occur- 

 rence of double marks (White and Chittenden, 1977; 

 Music and Pafford, 1984; Ross, 1988; Barbieri, 

 1993), or marks that are poorly defined and difficult 

 to distinguish (Joseph, 1972; Barger and Johnson, 

 1980; Barbieri, 1993). 



The pattern of otolith growth relative to fish 

 growth also indicated the high reliability of trans- 

 verse otolith sections for ageing Atlantic croaker. 

 Although otolith radius, the axis we used to read 

 annuli, showed a poor correlation with fish length, 

 the strong linear relationship between otolith radius 



and age indicates that otolith growth along this axis 

 seems to be continuous with age, independent offish 

 growth. This supports previous suggestions 

 (Mosegaard et al., 1988; Wright, 1991) that a pro- 

 cess other than somatic growth governs the rate of 

 otolith accretion. Because otoliths grow at a faster 

 rate than the body during slow somatic growth, they 

 are excellent structures for recording the seasonal 

 cycle and age in slow-growing and old fish, especially 

 those approaching asymptotic length (Casselman, 

 1990). The high correlation we found between otolith 

 radius and age for Atlantic croaker seems to confirm 

 this pattern. 



Growth and mortality 



High variability of observed lengths at age indicates 

 that length is a very poor predictor of age for Atlan- 

 tic croaker, especially beyond age 2. The wide range 

 in lengths at age can be attributed to a combination 

 of two factors: 1) most of Atlantic croaker's growth 

 occurs during the first two years, becoming asymp- 

 totic after age 2; and 2) fish are born at different 

 times during the extended spawning season and 

 display different growth rates. Warlen (1982) re- 

 ported that Atlantic croaker larvae caught later in 

 the spawning season had slower growth rates than 

 those taken during peak spawning. Since growth 

 decreases sharply after their first year, such differ- 

 ences in growth rates among young of the year is 

 likely to cause a large variation in lengths at age. 



Growth parameter estimates reported here do not 

 agree with previous reports for Atlantic croaker. 

 However, comparisons with previous studies are 

 difficult because other estimates were based on dif- 

 ferent ageing methods (White and Chittenden, 1977; 

 Ross, 1988), different otolith-ageing criteria (Barger, 

 1985; Hales and Reitz, 1992), or a period before any 

 significant fishery for Atlantic croaker occurred 

 (Hales and Reitz, 1992). Methods used to estimate 

 length-at-age data or to fit the von Bertalanffy 

 model also varied. Previous studies on Atlantic 

 croaker growth generally used back-calculated 

 rather than observed lengths at age. Although back- 

 calculation has been widely used and represents 

 standard methodology in age and growth studies 

 (Bagenal and Tesch, 1978; Jearld, 1983), recent evi- 

 dence indicates that it may generate biased results 

 (Campana, 1990; Ricker, 1992). By basing our 

 growth parameter estimates on observed lengths at 

 age of fish collected in September — the average 

 spawning period of Atlantic croaker in the Chesa- 

 peake Bay area — we avoided problems related to 

 back-calculation procedures or seasonal growth effects. 



Our total mortality estimates are the lowest ever 



