Scoles et a I.: Global phylogeography of Scomber 
833 
more distant outgroup, Rastrelliger kanagurta, and 
a preliminary parsimony tree drawing. Analysis of 
58 informative characters yielded 1031 equally par- 
simonious trees. Each of these 1031 trees shared the 
same topology among the five major matrilines and 
differed only within groups, with most rearrange- 
ments occurring among Atlantic S. japonicus 
haplotypes. Branches defining the five matrilines 
were supported at 100% in a majority-rule consen- 
sus that illustrated the relationship of all haplotypes 
of S. japonicus and S. australasicus and revealed a 
topology that was not different from that obtained 
by neighbor-joining (Fig. 4). 
The Red Sea sample had haplotypes intermediate 
to S. japonicus and S. australasicus, highlighted by 
character state reversals. Two reversals occurred 
among the Red Sea S. australasicus and the Japan- 
Taiwan S. japonicus clades, and the unique clade of 
S. australasicus. An Hpal site that is absent in eight 
of the 10 Red Sea haplotypes (nos. 60, 62-68), nine 
of the 10 unique S. australasicus haplotypes (nos. 
69-73, 75-78), and all of the S. japonicus haplotypes 
from Japan and Taiwan, is present in all other 
haplotypes. A Puull site that is absent in six of the 
10 Red Sea haplotypes (nos. 63-68) and all haplo- 
types of the unique S. australasicus clade is present 
in all other haplotypes. Consequently, corrected 
nucleotide sequence divergences among the Red Sea 
and unique S. australasicus haplotypes are lower 
than those among Red Sea and Atlantic S. japonicus 
haplotypes (Table 4). Although UPGMA cluster 
analysis grouped the Red Sea haplotypes with the 
unique <S. australasicus haplotypes (data not shown), 
neighbor-joining resulted in the placement of the Red 
Sea sample with Atlantic S. japonicus , concordant 
with the majority-rule consensus. Neighbor-joining 
is favored over UPGMA because it removes the as- 
sumption that the data are ultrametric (the condi- 
tion that all the taxa are equidistant from the root) 
(Swofford and Olsen, 1990). 
Parsimony analysis of the cytochrome b sequences 
(Table 5) yielded a tree topology that was similar to 
the patterns observed by parsimony and neighbor- 
joining of mtDNA restriction site data of S. japonicus 
and S. australasicus. However, the analysis showed 
greater similarity between the unique and ubiqui- 
tous S. australasicus clades that were not separated 
by Pacific S. japonicus. A majority-rule consensus of 
two equally parsimonious trees of 86 informative 
characters (of 93 variable positions) of cytochrome b 
sequences showed thatS. scombrus is the sister group 
to S. japonicus and S. australasicus (Fig. 5). Sequence 
divergences among the cytochrome b sequences were 
1.7-3. 1% in S. australasicus , 0.8-4. 3% in S. 
japonicus, and 2.8% in S. scombrus. The cytochrome 
b divergences were 2. 8-4.0% between haplotypes of 
S. japonicus and S. australasicus, and ranged from 
14.6 to 17.6% between the S. scombrus and the S. 
japonicus-S. australasicus group. The divergence of 
cytochrome b haplotypes between Rastrelliger 
kanagurta and S. scombrus was 21.5% and between 
Rastrelliger kanagurta and the S. japonicus-S. 
australasicus group was 17.1-18.8%. Neighbor-join- 
ing analysis of divergences among the cytochrome b 
sequences revealed a tree that had equal topology to 
that in Fig. 5 (data not shown). Parsimony and neigh- 
bor-joining of mtDNA restriction site data showed a 
paraphyletic pattern in both S. japonicus and S. 
Table 4 
Percent net nucleotide sequence divergences (8) among samples of Scomber 
mtDNA restriction site data. 
australasicus and Scomber japon 
cus derived from 
J-FLA 
J-ARG 
J-ISR 
J-IVC 
J-SAF 
J-TWN 
J-JPN 
J-CAL 
A-RED 
A-AUS 
A-NZL 
A-JPN A-MEX 
J-FLA 
— 
J-ARG 
0.042 
— 
J-ISR 
0.137 
0.071 
— 
J-IVC 
0.078 
0.025 
-0.003 
— 
J-SAF 
0.165 
0.099 
0.007 
0.015 
— 
J-TWN 
1.554 
1.472 
1.493 
1.522 
1.483 
— 
J-JPN 
1.576 
1.497 
1.515 
1.544 
1.503 
-0.010 
— 
J-CAL 
1.209 
1.128 
1.151 
1.179 
1.176 
0.284 
0.319 
— 
A-RED 
0.752 
0.706 
0.724 
0.728 
0.723 
1.136 
1.159 
1.077 
— 
A-AUS 
0.934 
0.907 
0.874 
0.892 
0.861 
0.974 
0.990 
0.886 
0.489 
— 
A-NZL 
0.926 
0.909 
0.878 
0.894 
0.860 
0.991 
1.004 
0.872 
0.528 
-0.019 
— 
A-JPN 
1.640 
1.599 
1.560 
1.579 
1.505 
1.523 
1.529 
1.533 
1.176 
0.554 
0.477 
— 
A-MEX 
1.688 
1.654 
1.604 
1.622 
1.543 
1.602 
1.606 
1.607 
1.267 
0.619 
0.533 
0.021 
