568 
Fishery Bulletin 95(3), 1997 
cance level = aJ(l+i), where i is the number of tests 
already performed. When a test is significant, it and 
all subsequent tests are deemed significant. 
Cluster analysis of the allozyme allele frequency 
data and the mtBNA haplotype frequency data used 
the UPGMA (unweighted pair-group method using 
averages) algorithm with Nei’s (1978) unbiased ge- 
netic distance measure, as implemented in BIOSYS-1 
(Swafford and Selander, 1981). 
Estimates of mtDNA nucleotide sequence diversity 
and divergence (Nei and Tajima, 1981; Nei, 1987) 
were made with REAP vers. 4.0 (see McElroy et al., 
1992), and population divergences were clustered by 
using UPGMA. 
Results 
The Seychelles muscle samples were partially de- 
graded on arrival, and could not be confidently 
screened for allozyme determinations, although 
mtBNA analysis presented no problems. Because it 
is sometimes difficult to distinguish tuna species, we 
usually find a small percentage (3-5%) of non-yel- 
lowfin tunas among nominal yellowfin tuna collec- 
tions. These misidentified fish can be recognized by 
aberrant allozyme (Graves et al., 1988; Elliott and 
Ward, 1995) and mtBNA patterns (Grewe 3 ). For ex- 
ample, five (3.7%) of the 135 Philippines samples 
collected in 1994 proved to be bigeye tuna, Thunnus 
obesus. However, at times, the proportion of mis- 
identified fish can be much higher: 18 (46.2%) of the 
39 “yellowfin tuna” from Sri Lanka proved to be big- 
eye tuna. The misidentified fish were excluded from 
the following analyses. 
Allozyme allele frequencies at four polymorphic loci 
(ADA*, FH*, GPI-A*, GPI-B *) for eight collections 
(Table 1) and mtBNA haplotype frequencies for nine 
collections (Table 2) were determined. 
No significant deviations from Hardy-Weinberg 
expectations were recorded for any allozyme locus. 
Heterogeneity chi-square analyses (Table 3) of allele 
frequencies revealed no significant differentiation for 
three loci (ADA*, FH*, and GPI-B*), but highly sig- 
nificant heterogeneity at the fourth locus, GPI-A* 
(P<0.001, a=0.0125). Genetic diversity ( G ST ) analyses 
(Table 3) indicated that for ADA*, FH*, and GPI-B*, 
3 Grewe, P. M. 1993. CSIRO Division of Marine Research, 
Hobart, Tasmania, Australia. Unpubl. data. 
Table 1 
Allozyme allele frequencies and sample sizes ( n ). GOM = Gulf of Mexico, S.Lan. = Sri Lanka, Philipp. = Philippines, Cl. 
Sea = Coral Sea, Calif. = California. 
Locus 
Allele 
Atlantic 
Indian 
Pacific 
GOM 
S. Lan. 
Philipp. 
Cl. Sea 
Kiribati 
Hawaii 
Calif. 
Mexico 
ADA* 
125 
0.005 

0.003 





115 
0.414 
0.310 
0.330 
0.306 
0.399 
0.361 
0.317 
0.359 
100 
0.548 
0.643 
0.622 
0.638 
0.567 
0.609 
0.671 
0.628 
85 
0.033 
0.048 
0.045 
0.056 
0.034 
0.030 
0.012 
0.013 
n 
105 
21 
176 
98 
89 
115 
41 
39 
FH* 
130 
0.118 
0.091 
0.081 
0.117 
0.086 
0.076 
0.051 
0.075 
100 
0.875 
0.909 
0.910 
0.878 
0.900 
0.920 
0.949 
0.925 
75 
0.007 
— 
0.009 
0.005 
0.014 
0.004 
— 
— 
n 
68 
11 
111 
98 
70 
112 
39 
29 
GPI-B* 
-20 
0.015 




0.004 


-60 
0.180 
0.167 
0.176 
0.163 
0.233 
0.113 
0.187 
0.231 
-100 
0.806 
0.833 
0.824 
0.837 
0.767 
0.878 
0.813 
0.769 
-125 
— 
— 
— 
— 
— 
0.004 
— 
— 
n 
103 
21 
176 
98 
88 
115 
40 
39 
GPI-A * 
145 


0.003 





135 
0.045 
— 
0.015 
0.036 
0.011 
0.035 
0.122 
0.077 
100 
0.624 
0.286 
0.651 
0.683 
0.673 
0.609 
0.305 
0.231 
75 
0.332 
0.714 
0.328 
0.281 
0.316 
0.357 
0.573 
0.692 
40 
— 
— 
0.003 
— 
— 
— 
— 
— 
n 
101 
21 
175 
98 
87 
115 
41 
39 
