676 
Fishery Bulletin 99(4) 
ulation structuring in southern flounder. The estimated 
^st value of 0.088 for southern flounder was greater than 
that found for Sciaenops oeellatus (F ST =0.022, Gold et ah, 
1994), Cynoscion nebulosus (F ST = 0.009, King and Pate, 
1992), or Pogonius chromis (F ST =0.013, Karel, unpubl. 
data) in this region. Physical or biotic factors may have 
caused greater isolation for southern flounder than for 
other nearshore species. Cluster analysis suggested that 
Table 3 
Hierarchical analyses of molecular variation (AMOVA) 
among mtDNA composite haplotypes of southern flounder 
( Paralichthys lethostigma ) from the U.S. Atlantic coast and 
Gulf of Mexico. 
Source of variation 
Variance 
% variation 
P' 
All sites 
Among coasts 
0.00026 
0.06 
0.04 
Among sites 
within coasts 
-0.00020 
-0.04 
0.92 
Within sites 
0.44985 
99.99 
0.74 
Eastern versus 
western cluster 
Among clusters 
0.00042 
0.09 
0.07 
Among sites 
within regions 
-0.00034 
-0.08 
0.10 
Within sites 
0.44985 
99.98 
0.76 
' Probability of finding a more extreme variance component by 
chance alone (1000 permutations). 
the region of greatest differentiation occurs along the mid- 
dle Texas coast. Similar genetic structure in Crassostrea 
virginica may be explained by seasonal current patterns 
in Corpus Christi Bay, Texas (King et al., 1994). A com- 
parable mechanism may operate in southern flounder; off- 
shore currents may have resulted in reduced dispersion 
of eggs and larvae between the upper and middle Texas 
coasts. Currents off the Texas coast are seasonally vari- 
able and complex (Cochrane and Kelly, 1986) and may aid 
in reducing egg and larval dispersion between regions of 
the western Gulf of Mexico. 
Many marine species spawn in the open ocean, have 
eggs and larvae with an extensive planktonic stage, or are 
highly mobile as adults. It is not surprising that such or- 
ganisms are often panmictic, or exhibit subdivision on on- 
ly broad levels (e.g. Sciaenops oeellatus-, Gold et al., 1994). 
When genetic differentiation is found in marine organ- 
isms (e.g. Cynoscion nebulosus', King and Pate, 1992), ex- 
tensive regions of clinal change in allele frequency may be 
seen and may be an adaptive feature (King and Zimmer- 
man, 1993). A critical question for fishery management 
is how much differentiation is necessary to indicate bi- 
ologically significant population structuring. Gold et al. 
(1994), for instance, found that red drum were subdivided 
(albeit weakly) between Atlantic and Gulf of Mexico sub- 
populations despite relatively high levels of gene flow be- 
tween the populations and failure of cluster analyses to 
consistently segregate localities into proper geographic re- 
gions. However, significant differences in allele frequen- 
cies for two loci were found and the researchers were able 
to demonstrate, through hierarchical gene-diversity anal- 
ysis, that 20% of the variation in gene diversity was re- 
lated to within-region differences. 
Table 4 
Tests <x 2 ) of pairwise population differentiation for southern flounder (Paralichthys lethostigma ) samples obtained from sites on 
the U.S. Gulf and Atlantic coasts based on variability at eight allozyme loci. The x 2 value tests the hypothesis of no difference in 
allelic distribution across collection localities. The associated probability ( P ) is given in parentheses. Inf. = infinite. 
North Carolina 
Florida 
Alabama 
Mississippi 
Sabine 
Galveston 
Matagorda 
Florida 
12.875 
(0.38) 
Alabama 
22.791 
(<0.01) 
4.478 
(0.81) 
Mississippi 
28.820 
(<0.01) 
13.350 
(0.21) 
7.723 
(0.66) 
Sabine 
26.087 
(0.01) 
17.208 
(0.14) 
18.942 
(0.09) 
21.959 
(0.04) 
Galveston 
59.146 
(<0.01) 
40.903 
(<0.01) 
42.314 
(<0.01 ) 
37.970 
(<0.01) 
44.407 
(<0.01) 
Matagorda 
Inf. 
(<0.01) 
Inf. 
(<0.01> 
Inf. 
(<0.01) 
Inf. 
(<0.01) 
Inf. Inf. 
(<0.01) 
(<0.01) 
Laguna 
Inf. 
Inf. 
Inf. 
42.178 
Inf. 
Inf. 
26.956 
Madre 
(<0.01) 
(<0.01 ) 
(<0.01) 
(<0.01) 
(<0.01) 
(<0.01) 
(0.07) 
