Jones et al.: Population dynamics of Pogonias cromis 
453 
spondences between bomb radiocarbon chronologies 
from the atmosphere and those from otolith cores of 
black drum (Campana and Jones, 1998). Average 
birth date was arbitrarily taken to be 1 January 
(Jearld, 1983). To assess ageing precision, all hard 
parts (n =30) were read twice by each of two readers, 
and agreement between and within readers was 
evaluated by percent agreement methods (Beamish 
and Fournier, 1981; Chang, 1982). Disagreements 
were resolved by a third reading. 
To evaluate changes in otolith size in relation to 
fish total length and age, otoliths from 300 fish ( 1990 
collections; ages 0-57; 22.9-130.0 cm TL) were mea- 
sured for maximum length (otolith length [OL] +0.01 
mm), radius along the sulcal grove (otolith radius 
[OR] ±0.001 mm), maximum thickness (otolith width 
[OWID] ±0.01 mm), and weight (otolith weight [OWT] 
±0.001 g). Relation between otolith measurements 
and fish TL and age were evaluated by simple linear 
regression analysis. 
To evaluate growth, observed individual lengths- 
at-age were fitted to the von Bertalanffy growth func- 
tion, VBGF (Ricker, 1975), by using nonlinear regres- 
sion, SAS NLIN procedure DUD method (SAS, 1988). 
Likewise, individual weights-at-age were fitted to the 
VBGF. Model parameters were the following: L x , the 
mean asymptotic length; W x , the mean asymptotic 
weight; K and K', respectively; the Brody growth co- 
efficient on length and weight; and t 0 and t’ Q , the 
theoretical age at which the fish would have zero 
length on length and weight (Ricker, 1975). Growth 
curve parameters were compared between years and 
sexes with maximum likelihood ratio tests (Kimura, 
1980). 
Linear regression was used to determine length- 
weight relationships for fish ranging from 22.9 to 
130.0 cm TL and 0.6 to 49.4 kg TW. Differences be- 
tween sexes were tested with Rawlings’ ( 1988) tests 
of homogeneity of slopes and intercepts by using 
PROC REG in SAS (Littell et al., 1991). The hypoth- 
esis of isometric growth (Ricker, 1975) was tested 
with a Utest. 
Instantaneous total annual mortality rates, Z, were 
estimated from maximum age with Hoenig’s pooled 
regression equation (Hoenig, 1983), by calculating a 
theoretical total mortality for the entire life span fol- 
lowing the reasoning of Royce (1972), and with the 
regression method, i.e. with a catch curve combin- 
ing log e -transformed recreational and commercial 
abundance data. In the latter method, mortality es- 
timates were based on data from ages 21-43 and 21- 
59. Younger ages were truncated because the age 
group at the apex of the catch curve (age 20 ) may not 
have been fully recruited to the fishery ( Everhart and 
Youngs, 1981). Older ages were truncated at the first 
age class (age 44) with fewer than five fish following 
Chapman and Robson ( 1960). Data from 1990 to 1992 
were combined to minimize effects of variation in 
year-class strength (Robson and Chapman, 1961). 
The right limb of the catch curve was tested for devia- 
tion from linearity by analysis of variance (AN OVA). 
Estimates of Z were converted to total annual mor- 
tality rates (A=l-e 2 ; Ricker, 1975). 
All statistical analyses were performed with SAS 
(SAS, 1988). Rejection of the null hypothesis was 
based on a = 0.05, F-tests in ANCOVA were based on 
type-III sum of squares (Freund et al., 1986), and 
assumptions of linearity were checked with residual 
plots (Draper and Smith, 1981). Data were log 10 - 
transformed to correct for nonlinearity and hetero- 
geneity of variance when necessary. Log-transformed 
data are presented in graphs and tables in original 
units, unless otherwise stated. Variables that could 
not be normalized were compared with Wilcoxon’s 
two-sample test or a Kruskal-Wallis test for more 
than two samples, and large-sample approximate 2 - 
scores or were reported. 
Results 
Hard part comparisons 
All hardparts showed regular, concentric marks that 
could be interpreted as annuli. However, marks were 
not equally clear or consistent between all hard parts. 
Otoliths were the clearest and most precise of the 
hard parts to interpret. One hundred percent of 
otoliths, 36.7% of dorsal spines, and 63.7% of fin rays 
had marks clear enough to read. Between-reader 
precision was 100% for otoliths, 27.3% for dorsal 
spines, and 47.4% for fin rays. Compared with 
otoliths, dorsal spines and fin rays underestimated 
age; this underestimation worsened with increasing 
age (Kruskal-Wallis distribution-free multiple com- 
parison test, MSD=15.81,P<0.05). Underageing was 
especially marked with dorsal spines. On the basis 
of these results and otolith growth patterns (see next 
section), we deemed otoliths the clearest, most reli- 
able hard part, and used them for all ageing. 
Otolith size relationships to fish size and age 
Black drum otoliths continue to increase in size with 
fish length and age, apparently throughout life. All 
measures of otolith size — OL, OWT, OR, OWID — 
were significantly and positively related to fish length 
and age. Although black drum otoliths continue to 
increase in size, the relations of various otolith sizes 
to fish length and age were not consistent. Relations 
