826 
Fishery Bulletin 96(4), 1998 
Table 1 
Scomber japonicus , S. australasicus, and S. scombrus. Sample name, size, location, haplotype diversity ih), and percent nucle- 
otide sequence diversity (n). 
Sample 
Haplotype 
Percent nucleotide 
Sample name 
size (n) 
Sample location 
diversity (h) 
sequence diversity (tt) 
Scomber scombrus 
S-MAS 
20 
Boston, Massachusetts 
0.85 
0.29 
S-ENG 
20 
Plymouth, England 
0.28 
0.07 
Total 
40 
0.58 
0.18 
Scomber australasicus 
A-RED 
15 
Sinai Peninsula south, Red Sea 
0.95 
0.41 
A-AUS 
18 
New South Wales, Australia 
0.86 
0.75 
A-NZL 
19 
Wellington, New Zealand 
0.75 
0.77 
A-JPN 
21 
Tokyo, Japan 
0.81 
0.30 
A-MEX 
20 
Revillagigedo Islands, Mexico 
0.59 
0.13 
Total 
93 
0.90 
2.18 
Scomber japonicus 
J-FLA 
20 
Panama City, Florida 
0.93 
0.40 
J-ARG 
18 
Mar del Plata, Argentina 
0.90 
0.50 
J-ISR 
20 
Mediterranean coast of Israel 
0.90 
0.38 
J-IVC 
20 
Abidjan, Ivory Coast 
0.91 
0.39 
J-SAF 
20 
Cape Town, South Africa 
0.81 
0.38 
J-TWN 
20 
Kaohsing, Taiwan 
0.86 
0.29 
J-JPN 
20 
Tokyo, Japan 
0.88 
0.35 
J-CAL 
20 
San Diego, California 
0.64 
0.14 
Total 
158 
0.95 
2.42 
Grand total 
291 
S.japonicus and S. australasicus group, and proceed- 
ing through the alphabet, and beginning with ‘Z’ and 
proceeding in reverse alphabetical order for S. 
scombrus. For the few S. scombrus restriction frag- 
ment patterns that also occurred in S. japonicus or 
S. australasicus, the latter species’ restriction morph 
letter designation was used. 
Estimates of nucleotide sequence divergence ( d ) 
among haplotypes were determined by using the 
approach of Nei and Li ( 1979) for fragment data, and 
Nei and Tajima (1981) and Nei and Miller (1990) for 
site data, with weighting based on the proportion of 
fragments or sites produced by each enzyme class 
(Nei and Tajima, 1983). Estimates of nucleon or hap- 
lotype diversity ih) and nucleotide sequence diver- 
sity in) were then calculated with the program DA of 
the statistical package REAP (McElroy et al., 1992). 
Corrected nucleotide sequence divergences (8) among 
populations were calculated by using the protocol of 
Nei and Li ( 1979). The homogeneity of haplotype dis- 
tributions among samples was evaluated by chi- 
square analyses using the Monte-Carlo method of 
Roff and Bentzen (1989) with 1000 randomizations 
of the data with REAP. 
Restriction sites were inferred from completely 
additive fragment patterns, and the information for 
each individual was coded using a presence-or-ab- 
sence matrix. Homology of restriction sites could not 
be assured between S. scombrus and the S. austra- 
lasicus-S. japonicus group owing to the relatively 
large genetic divergence between the species. Restric- 
tion site data were therefore evaluated only for phy- 
logenetic relationships within the S. australasicus- 
S. japonicus group, and an estimate of divergence 
between S. scombrus and the pooled data of S. 
australasicus and S. japonicus was determined by 
using restriction fragment data. Relationships among 
haplotypes were inferred by cluster analysis (un- 
weighted pair-group method, UPGMA, Norusis, 
1988). Neighbor-joining, parsimony analyses and 
consensus trees were generated with the PAUP* soft- 
ware program (test version). 
DNA amplification and sequence analysis 
Individuals representative of the major Scomber 
mtDNA matrilines, identified by restriction site 
analysis, and of Rastrelliger kanagurta, were selected 
for DNA sequence analysis. A 418-bp region of the 
cytochrome b gene was amplified from mtDNA-en- 
riched genomic DNA isolations by the polymerase 
chain reaction (PCR) with primers L15079 
