22 
Fishery Bulletin 1 14(1) 
Figure 2 
(A) Initial image taken under light microscopy of otolith from larval Atlantic croaker ( Micro - 
pogonias undulatus) collected from Bayou Tartellan, Louisiana, from October 2006 through 
March 2007 and from September 2007 through March 2008. The black line from the center 
to the lower left represents the radius used for reading. (B) Otolith image after digital im- 
age enhancement. The black line from the center to the lower left is the same radius as the 
one on the previous image. (C) Initial grayscale measurement along the radius from the 
center of the otolith to the edge. (D) Grayscale measurements along the same radius after 
low pass filtering. 
the method described by Barbieri et al. (1994a, 1994b). 
All dissections were conducted with an Olympus SZX12 
stereoscope (Olympus Corp., Tokyo) with a lx objective 
lens. Both left and right sagittal otoliths were removed 
and placed on a slide with Permount mounting medi- 
um (Thermo Fisher Scientific, Waltham, MA), the left 
otolith concave side up and the right otolith concave 
side down. Otoliths were polished with 0.3-pm alumina 
paste and a microcloth to reveal the core. Otoliths were 
etched with 0.1-N hydrochloric acid for between 10 and 
20 s to facilitate readability under a compound stereo- 
scope. Digital images were taken at a magnification 
from 500x to 1250x with an Olympus BX41 compound 
stereoscope equipped with a phase contrast filter to 
highlight light- and dark-ring discontinuity zones of 
otoliths with oil immersion (Fig. 2A). 
Adobe Photoshop CS4, vers. 11.0 (Adobe Systems 
Inc., San Jose, CA) was used to convert the image from 
color to grayscale and to enhance differences between 
light and dark rings by increasing contrast and bright- 
ness (Fig. 2B). Measurements of otolith radii lengths, 
the distance from the otolith nucleus to the distal edge, 
and of grayscale values for each radius were conduct- 
ed with Image J image analysis software, vers. 1.44p, 
(National Institutes of Health, Bethesda, MD). Mea- 
surement of each radius produced a calibrated length 
and corresponding grayscale values that ranged from 
0 (black) to 255 (white) along that radius. A central 
radius and 2 radii to the left and right of the central 
radii, offset by a single pixel each, were measured. All 
5 radii were averaged to produce the grayscale values 
later used for filter analysis, to avoid the bias that 
could be introduced by choice of the radius for reading 
(Morales-Nin et al., 1998). Radius length and grayscale 
data were collected for each otolith that was imaged 
(Fig. 20. 
Image data were imported into MATLAB, vers. 
7.6.0.324 R2008a (The MathWorks Inc., Natick, MA) for 
filtering and nominal age determination. Low pass fil- 
ter structure was determined with a fast Fourier trans- 
form (FFT) to transform the initial radii measurements 
into frequencies to identify and exclude high-frequency 
subdaily discontinuities from the otolith radius. The 
low-pass filter was fitted iteratively to the individual 
otoliths, on the basis of the understanding that the Ny- 
quist frequency is the daily otolith increment accreted 
by the larvae. As noted by Morales-Nin et al. (1998), 
this iterative fitting is done for each otolith because of 
increments of varying radius length between otoliths, 
differences in magnification, and variable growth rates 
for individuals. An inverse FFT was then performed to 
transform the signal from the frequency domain back 
