570 
Fishery Bulletin 95(3), 1997 
Figure 1 
Map of sample sites showing GPI-A* gene frequencies in yellowfin tuna, Thunnus albacares. Larger circles represent our data (Table 1), 
the three smaller circles (Bismark Sea, Roca Partido, and Ecuador) data are from Sharp (1978). The location of the Seychelles sample, 
examined for mtDNA variation but not for GPI-A* variation, is identified. The shaded area represents the approximate global distribu- 
tion of yellowfin tuna. 
from the Atlantic Ocean collection suggests that these 
fish constitute separate stocks. Thus the GPI-A* 
data, taken together with the spatial orientation of 
these collections, indicate the existence of at least four 
yellowfin tuna stocks: Atlantic Ocean, Indian Ocean, 
west-central Pacific Ocean, and east Pacific Ocean. 
Six of the 24 mtDNA haplotypes (CO, QA, WA, ZB, 
A2B, Q2B, see Table 2) were not recorded in the ear- 
lier survey of Ward et al. (1994) but were rare (fre- 
quencies less than 3.5%). Fragment sizes for most 
haplotypes are given in Ward et al. (1994), but a full 
list is available on request. Haplotype (nucleon) di- 
versities per collection ranged from 0.633 to 0.727 
(mean estimate of 0.683) (Table 2). Percent nucle- 
otide diversities per collection ranged from 0.998 to 
1.263 (mean estimate of 1.061) (Table 2). 
Table 3 
Analyses of genetic differentiation among the samples. 
Locus 
Number of 
fish 
Number of 
alleles/haplotypes 
Heterogeneity x 2 
analysis 
Genetic diversity analysis 
t 
P 
G st 
^ ST.nulA 
P 
ADA* 
684 
4 
17.326 
0.666 
0.006 
0.008 ±0.004 
0.569 
FH* 
538 
3 
9.241 
0.821 
0.006 
0.011 ±0.008 
0.736 
GPI-B* 
680 
4 
29.256 
0.128 
0.008 
0.008 ±0.005 
0.370 
GPI-A* 
677 
5 
131.416 
<0.001 
0.118 
0.008 ±0.004 
<0.001 
mtDNA 
767 
24 
227.743 
0.048 
0.023 
0.015 ±0.005 
0.071 
