332 
Fishery Bulletin 112(4) 
Table 2 
Results, at 3 spatial scales, of hypothesis testing on univariate shape indices of otoliths 
from female southern flounder ( Paralichthys lethostigma) that were collected in 1996 
and in 2009-12 in state waters in 2 basins, the South Atlantic (North Carolina, South 
Carolina, Georgia, and Florida) and the Gulf of Mexico (Florida, Alabama, Mississippi, 
Louisiana, and Texas), as part of state or federal agency sampling programs and from 
seafood dealers. 
Scale 
Variable 
Test statistic 
P - value 
Test statistic 
P-value 
Basin 
Ellipticity 
f=3.82 
<0.001 
Rectangularity 
t=-1.67 
0.100 
Form function 
t=-5.06 
<0.001 
Aspect ratio 
t= 3.74 
<0.001 
Within Gulf of Mexico 
Within South Atlantic 
State 
Ellipticity 
f=1.48 
0.145 
F=1.71 
0.184 
Rectangularity 
t=1.48 
0.149 
F=0.20 
0.815 
Form function 
t=- 2.06 
0.046 
F= 0.23 
0.797 
Aspect ratio 
t=1.37 
0.178 
F=2.09 
0.127 
Region i 
Ellipticity 
F=0.55 
0.581 
Rectangularity 
F=1.18 
0.313 
Form function 
F=8.08 
<0.001 
Aspect ratio 
F=0.51 
0.600 
1 Region within North Carolina. See text for further details. 
the distribution of reclassification success rates over- 
lapped with the null distribution of reclassification in 
each case (Fig. 4). 
Discussion 
Otolith shape and geographic distribution 
Using a combination of univariate shape indices and 
EFCs, we found strong evidence for the existence of 
different populations of southern flounder between the 
2 basins that we examined and less evidence for popu- 
lation structure as our spatial scope decreased (among 
states within each basin and among regions within 
North Carolina). Interestingly (and despite low samples 
sizes of fish from Florida waters in the Gulf of Mexico), 
neither of the within-basin analyses provided strong 
evidence of otolith shape differentiation, although the 
otoliths represented fish collected from the spatial ex- 
tremes within each basin distribution. Both of these 
observations are in agreement with recent genetic find- 
ings for this species (Anderson et ah, 2012). The use of 
EFA is considered one of the most reliable methods in 
otolith morphometric studies (Merigot et al., 2007), and 
it took only 10 coefficients to produce a rate of reclas- 
sification success of nearly 80% at the basin level (Gulf 
of Mexico versus South Atlantic). Additionally, because 
of the resampling aspect of the DFA, this result should 
be considered particularly robust. 
Of greater interest was a lack of evidence for stock 
differentiation at either the within-basin or within- 
state spatial scales. Although a number of factors 
could effectively mix individuals within each basin 
(or at finer spatial scales), tagging evidence does not 
support extensive mixing among or within states for 
younger, smaller fish, which typically remain within in- 
shore estuarine systems (Monaghan 7 ; Craig and Rice 8 ). 
Although phenotypic stocks remain possible on the 
basis of previous observations of spatial variation in 
somatic growth rates (e.g., Stunz et al., 2000; Fischer 
and Thompson, 2004) and differences in otolith size es- 
timates, our analysis of otolith shape did not clearly 
identify at the sub-basin level stocks of coastal south- 
ern flounder within either the South Atlantic or the 
Gulf of Mexico. We cannot rule out the possibility of 
undetected fine-scale structure; however, the strength 
with which otolith shape analyses are able to detect 
both large-scale genetic stock differences as well as 
fine-scale environmental differences indicates that 
within-basin structuring of southern flounder is likely 
weak. 
The interpretation of variation in otolith shapes re- 
quires consideration of several factors that can contrib- 
ute to otolith shape in the context of the species and 
study design. Otolith growth and shape are controlled 
dually by genetic and environmental influences (Vi- 
