Midway et al. : Stock structure of Paralichthys lethostigma inferred from otolith shape analysis 
335 
cr 
(U 
250 
200 
150 
100 
50 
0 
0.2 
0.4 
0.6 
0.8 
1.0 
Reclassification success 
Figure 4 
Distributions of randomized reclassification (i.e., null) success rates (shown as the gray bars on the left in each 
panel) and actual linear discriminant randomization (shown as the hatched bars on the right) for each of 4 geo- 
graphic analyses of otolith shapes from female southern flounder ( Paralichthys lethostigma) collected in 1996 and 
in 2009-12 in 2 basins, the South Atlantic (North Carolina, South Carolina, Georgia, and Florida) and the Gulf of 
Mexico (Florida, Alabama, Mississippi, Louisiana, and Texas). Distributions are shown for 4 geographic groups: (A) 
both ocean basins, (B) Gulf of Mexico states, (C) South Atlantic states, and (D) North Carolina regions. Only the 
basin-level analysis indicated clear separation of distributions; each of the within-basin analyses and the within- 
North Carolina analyses revealed some degree of distributional overlap (illustrated by the gray, hatched bars). 
lihood of differing environmental conditions between 
the U.S. South Atlantic and the Gulf of Mexico led us 
to hypothesize the existence of distinct otolith shapes 
for southern flounder at the spatial scale of the basin, 
and we did detect distinct shapes. On the basis of past 
observations of high site fidelity from tag-return data 
in estuarine systems (Monaghan 7 ; Craig and Rice 8 ), we 
expected to find differences in otolith shape among fish 
from separate areas within each basin and even po- 
tentially among fish from different estuarine systems 
within North Carolina. We considered North Carolina 
a good model with which to test for structure at a finer 
spatial scale because the inshore waters of the state 
are made up of a range of system types, from large sys- 
tems in central and northern regions (e.g., Pamlico and 
Albemarle Sounds) that contain extensive oligohaline 
reaches to small, river-based estuaries in the southern 
region. 
Given that even relatively fine-scale shifts in habi- 
tat have been observed to generate heterogeneity in 
otolith shape (e.g., Vignon, 2012), we hypothesized 
that deviation in geography and hydrography among 
systems within North Carolina might produce enough 
environmental variability to influence growth rates 
and, therefore, the shape of otoliths from southern 
flounder. However, we detected only a weak signal of 
spatial structuring in otolith shape of southern floun- 
der among locales within each basin and among re- 
gions within North Carolina. We did achieve a rate of 
reclassification success that was higher than expected 
in each case, an outcome suggestive of localized envi- 
ronmental effects, but our findings also indicate that 
