22 
Fishery Bulletin 109(1 ) 
quent analyses and will be referred to as “reading 3” 
throughout the rest of this article. 
Thin sections from two otoliths were polished to the 
core and presumed daily growth rings were counted 
under 40 x magnification to help with the interpretation 
of what was thought to be the first annulus. Presumed 
daily growth rings were not validated. In addition to 
thin section interpretation, some of the uncut second 
otoliths were randomly selected and viewed whole under 
the dissecting microscope to qualitatively evaluate the 
ease of determining growth bands on the whole otolith 
surface. 
To validate ages, eighteen of the uncut, second oto- 
liths were selected for bomb radiocarbon analysis. 
These otoliths were chosen for readability and to span 
the range of observed ages that were based on thin 
section growth band counts, including otoliths from 
specimens with estimated birth years spanning the 
prebomb to postbomb period. The selected otoliths were 
embedded in epoxy resin and sectioned transversely 
through the core to 1 mm thickness. The sections were 
then taped to plates, and the cores were isolated by 
using a Dremel model 221 variable speed rotary tool 
(Robert Bosch Tool Corporation, Mt. Prospect, IL). Core 
isolation was visually aided because the central opaque 
area of the first year of growth in red bream otoliths is 
clearly visible on whole and sectioned otoliths. 
The extracted cores were rinsed in 10% HN0 3 for 
15-30 s, ultrasonically cleaned with a Branson ultra- 
sonic cleaner B-22-4 (Branson Ultrasonics Corpora- 
tion, Danbury, CT) with distilled water, air-dried, and 
weighed to the nearest 0.1 mg (Baker and Wilson, 
2001). Each core was placed in a glass vial that had 
been cleaned with 10% HNO g , and the samples were 
sent to the National Ocean Sciences Accelerator Mass 
Spectrometry (NOSAMS) laboratory at Woods Hole 
Oceanographic Institution. NOSAMS provided values 
of delta Carbon-14 (4 14 C) for each sample that were 
then plotted against otolith-derived birth years and 
compared to a reference chronology of validated ages 
for haddock (Melanogrammus aeglefinus ) collected from 
Newfoundland (Campana, 1997). The timing of initial 
radiocarbon increase and mean year of increase were 
calculated with the deterministic model developed by 
Hamel et al. (2008) which models the pulse of radio- 
carbon from nuclear testing as a Gaussian curve over 
time and couples it with a continuous exponential decay 
process to describe radiocarbon dispersion and dilution. 
Aging precision and bias between paired readings 
were examined graphically with age bias plots, and 
the coefficient of variation (CV) was calculated as a 
measure of the relative ease of aging red bream oto- 
liths (Campana et al., 1995). Age frequencies were 
computed, and a standard linear regression relating 
increment count to otolith weight was also conducted. 
The von Bertalanffy growth function (VBGF; von Ber- 
talanffy, 1938) was fitted to unweighted length-at-age 
data by using a random effects (RE) model with gamma 
population distribution likelihood as implemented in 
the program IGOR+, a Microsoft Excel®-based applica- 
tion developed by Cope and Punt (2007). The random 
effects model is based on a likelihood function that 
takes into account multiple reads for each otolith and 
thus incorporates both process and interpretation error 
into growth parameter estimations (Cope and Punt, 
2007). Age and length information from all three read- 
ings could therefore be included in the estimation of 
the VBGF. The standard von Bertalanffy growth pa- 
rameters k (the Brody coefficient), L ^ (the theoretical 
mean maximum length), and t 0 (the theoretical age 
at length zero) were calculated with IGOR+ program. 
The model was run separately for males and females 
and for both sexes combined. The resulting sex-specific 
growth functions were compared by using likelihood 
tests (Kimura, 1980). An additional run was made in- 
cluding the age and length data for the 22 specimens 
from the Azores to assess how much the model param- 
eters would change by including the smaller fish. 
Natural mortality was estimated by using the equa- 
tions developed by Hoenig (1983) and Pauly (1980) and 
by using the IGOR+ program. Hoenig’s longevity-based 
estimator, In Z=1.46-1.01 ln*G majc ), uses maximum age 
(t max ), and Pauly’s equation, log 10 M = -0.0066 - 0.279 
log 10 L x + 0.6543 log 10 k + 0.463 log 10 T, uses von Ber- 
talanffy growth parameters and water temperature, T, 
which in this case was the average annual water tem- 
perature for Beryx habitat from the literature. IGOR+ 
calculates total mortality by using the catch curve of 
the gamma-distributed “true ages” estimated in the 
RE model. 
Reproductive biology 
Gonads were weighed and processed according to the 
standard procedure used by the Marine Resources 
Monitoring Assessment and Prediction (MARMAP) 
Program at the Marine Resources Research Institute 
of the South Carolina Department of Natural Resources 
(White et al., 1998). A portion of the posterior gonad 
was removed, fixed in 10% formalin for 7-14 days, and 
transferred to 50% isopropanol for an additional 1-2 
weeks. Tissues were then dehydrated, cleared, and 
blocked in paraffin under vacuum infiltration by using 
a Leica ASP300 tissue processor (Leica Microsystems 
Inc., Bannockburn, IL). Blocks were allowed to cool in 
a freezer, and three 7-pm cross sections were cut with 
a Leica RM2255 rotary microtome (Leica Microsys- 
tems Inc., Bannockburn, IL). These sections were then 
transferred to a microscope slide and allowed to dry 
overnight before they were stained with hematoxylin 
and eosin. The sections were viewed and interpreted 
under a Nikon Eclipse 55i compound microscope (Nikon 
Instruments Inc., Melville, NY), and reproductive stages 
were assigned independently by two readers without 
knowledge of specimen age, length, or collection date, 
according to criteria described by Harris et al. (2004). 
Females in spawning condition were identified by the 
presence of hydrated oocytes and postovulatory follicles. 
Specimens, for which interpretation between readers 
differed, were re-examined jointly, and a consensus was 
