Friess and Sedberry: Age, growth, and spawning season for Beryx decadactylus 
23 
reached. Spawning seasonality was determined by 
plotting monthly frequencies of reproductive stages. 
Results 
Data from samples 
Of the 165 red bream sampled from the commer- 
cial fishery between April 2003 and January 2008, 
otoliths were collected from 163 specimens, gonads 
from 161, and fork length was measured for 164 fish 
(one specimen had a damaged tail). Otolith weights 
were obtained for 130 otoliths. No specimens could be 
sampled from the fishery in February and March in 
any year over the sample collection period because the 
wreckfish fishery is closed from January 15 through 
April 15. Sample sizes were highest for the months 
of July, August, September, and December (n=23-30) 
and lowest for April, October, and November (n=4-8). 
Because of low sample sizes, samples were pooled across 
years. 
Fork lengths ranged from 410 to 630 mm and were 
not normally distributed (Shapiro-Wilk test; P<0.001; 
Fig. 1). Thus, the median FL more accurately describes 
the central tendency of the data set. Male FLs ranged 
from 410 to 601 mm ( 7 ? = 6 1 ) , and female FLs ranged 
from 420 to 630 mm FL (ti = 98) (Fig. 1). The median 
FL for males (538 mm FL, standard error [SE] =7) was 
significantly different from that of females (550 mm 
FL, SE = 5; 2-tailed Wilcoxon rank sum test, P=0.017). 
Strong linear relationships for both sexes were de- 
tected between FL and TL (males: TL=1.05FL+63.79, 
r 2 = 0.983; females: TL = 1.03FL + 73.32, r 2 =0.973), 
FL and SL (males: SL = 0.91FL + 0. 25, r 2 = 0.983; fe- 
males: SL = 0.91FL— 5.44, r 2 =0.970), and FL and TBW 
(males: TBW=17.52FL-5588.2, r 2 = 0.952; females: 
7W=20.74FL-7323.1, r 2 =0.949). The otoliths provided 
from the Azores corresponded to fork lengths ranging 
from 190 to 400 mm. 
Age and growth 
On whole otoliths, growth increments were best visible 
on the anterior edge of the antisulcal surface close to a 
large, central opaque area but became virtually indistin- 
guishable closer to the otolith margin, especially in older 
specimens. Age determinations based on whole otoliths 
were therefore not attempted. In a few cases, one of the 
two otoliths collected from a specimen was misshapen, 
and crystals were clearly visible as lumps on the otolith 
surface. The second otolith would look normal and was 
used for aging. 
Thin transverse sections of red bream otoliths con- 
tained a large central opaque area, and the count of as- 
sumed daily growth rings indicated that the translucent 
zone outside the central opaque area marks the first 
annulus (first otolith: 389 daily rings, second otolith: 
371 daily rings). Thin sections of red bream otoliths 
contained a transition zone where increment width 
Fork length (mm) 
Figure II 
Length distribution of male and female red bream ( Beryx 
decadactylus) sampled from the commercial wreckfish fish- 
ery off the southeastern United States from 2003 to 2008. 
notably decreased and otoliths started increasing in 
thickness on the proximal surface rather than growing 
along the dorsoventral and anterioposterior axes (Fig. 2, 
A and B). 
Thin sections of red bream otoliths had poor clar- 
ity and were difficult to interpret. Bands were often 
unclear, split, or irregularly spaced, and increments 
became thinner and harder to see toward the edge of 
the otolith. The dorsal axis next to the sulcus acousti- 
cus was chosen as the preferred axis for aging because 
the banding pattern was most distinct there. In many 
otoliths readability decreased at various points along 
the chosen axis, and counting had to be shifted away 
from the sulcus and toward the dorsal tip by follow- 
ing an increment over to the new axis. Often, incre- 
ments, both close to the core and toward the edge of 
the otolith, were difficult to interpret in older as well 
as younger specimens because they were so poorly de- 
fined. In addition, semicrystalline fields were apparent 
in a number of thin sections, which added to difficulty 
in interpretation. 
Even though red bream otoliths were overall difficult 
to age, there were no otoliths in the sample that were 
identified as unreadable by both readers and, there- 
fore, all otoliths were aged and used in the analysis. 
Estimated ages ranged from 8 to 64 years for reading 
1, 10 to 71 years for reading 2, and 9 to 69 years for 
reading 3. The age variation for individual otoliths 
between readings was high and ranged from 0 to 23 
years, with a mean difference of 4 years. The age bias 
plots revealed that increment counts from reading 1 
tended to underestimate age with respect to reading 2, 
increasingly so at ages 30 and higher (Fig. 3A). Read- 
ing 3 tended to underestimate ages 36 and higher with 
respect to reading 2 (Fig 3B), and reading 1 underes- 
timated ages compared to reading 3, particularly ages 
up to 40 (Fig. 30. Pairwise comparisons of CVs were 
highly variable across ages but tended to decrease with 
increasing age (Fig. 3, D-F). The mean pairwise CV 
was lowest between readings 2 and 3 at 8.6%, and high- 
