823 
Global phylogeography of mackerels 
of the genus Scomber 
Daniel R. Scales* 
Bruce B. Collette*' ** 
John E. Graves* 
* School of Marine Science 
Virginia institute of Marine Science 
College of William and Mary 
Gloucester Point, Virginia 23062 
** National Systematics Laboratory 
National Marine Fisheries Service, NOAA 
National Museum of Natural History 
Washington, DC, 20560 
Present address (for D, R. Scoles): Neurogenetics Laboratory, Division of Neurology 
Cedars-Sinai Medical Center 
UCLA School of Medicine 
8700 Beverly Boulevard 
Los Angeles, California 90048 
E-mail address (for D. R. Scoles): scolesd@csmc.edu 
Abstract .—Inter- and intraspecific 
genetic relationships among and within 
three species of mackerels of the genus 
Scomber were investigated by restric- 
tion site analysis of the whole mito- 
chondrial (mt) DNA genome and direct 
sequence analysis of the mitochondrial 
cytochrome b gene. A total of 15 samples, 
averaging 19 individuals each, were col- 
lected from geographically isolated 
populations throughout the ranges of 
S. scombrus (two samples), S. austra- 
lasicus (five samples), and S. japonicus 
(eight samples). Restriction site analy- 
sis with 12 restriction enzymes re- 
vealed substantial genetic variation 
within each species. Sample haplotype 
diversities ranged from 0.28 to 0.95, 
and nucleotide sequence diversities 
from 0.13% to 0.76%. Spatial partition- 
ing of genetic variation was observed 
in each of the species. Eastern and 
western North Atlantic samples of S. 
scombrus exhibited significant hetero- 
geneity in the distribution of mtDNA 
haplotypes, but no fixed restriction site 
differences were observed between 
samples. Similarly, no fixed restriction 
site differences occurred among samples 
of S. japonicus in the Atlantic Ocean, 
although there were significant differ- 
ences in the distribution of haplotypes 
among samples. In contrast, samples of 
S. japonicus from within the Pacific 
Ocean were characterized by fixed re- 
striction site differences. North and 
South Pacific samples of S. austra- 
lasicus were highly divergent, and one 
of two divergent mtDNA matrilines was 
restricted to samples from the South 
Pacific. A 420-bp segment of the cyto- 
chrome b gene was sequenced for rep- 
resentatives of each of the major 
mtDNA lineages identified by restric- 
tion site analysis. Scomber scombrus 
differed from S. australasicus and S. 
japonicus by more than 11% net nucle- 
otide sequence divergence, considerably 
greater than the 3.5% sequence diver- 
gence between S. australasicus and S. 
japonicus. Levels of interspecific ge- 
netic divergences based on restriction 
site data were similar in pattern, but 
were approximately 20% lower in mag- 
nitude when based on the cytochrome 
b sequences. Parsimony analysis and 
neighbor-joining of restriction site data, 
and parsimony analysis of cytochrome b 
sequences showed similar paraphyletic 
patterns in both S. japonicus and S. 
australasicus. Levels of divergence 
among samples of S. japonicus were 
similar to those between samples of .S'. 
australasicus and S. japonicus. Com- 
plete partitioning of halpotypes among 
some samples of S. japonicus that are 
morphologically distinct suggests that 
Atlantic and Indo-Pacific populations of 
S. japonicus may need to be recognized 
as separate species. 
Manuscript accepted 26 January 1998. 
Fish. Bull. 96: 823-842 (1998). 
Population structure is largely in- 
fluenced by the biological character- 
istics of a species and the attributes 
of its environment that promote or 
impede gene flow. In marine fishes, 
dispersal of planktonic early life 
history stages and adult vagility can 
facilitate genetic exchange over 
great distances (Rosenblatt, 1963; 
Shulman and Bermingham, 1995). 
Features of the marine environ- 
ment that limit gene flow are often 
equally large in scale, such as 
current systems, major changes in 
temperature or salinity, or the 
presence of large land masses 
(Sinclair, 1988). Accordingly, many 
genetic analyses of broadly distrib- 
uted marine fishes have shown little 
divergence among conspecific 
samples from geographically dis- 
tant locations. 
Limited population structure has 
been demonstrated for a variety of 
marine fishes. Electrophoretic 
analyses of allozymes have revealed 
little population divergence among 
collections of milkfish {Chanos 
chanos ) separated by up to 10,000 
km (Winans, 1980), damselfish, 
Stegastes fasciolatus, sampled from 
throughout the 2500 km Hawaiian 
archipelago (Shaklee, 1984), in 
twelve species of tropical marine 
shore fishes sampled from both 
sides of the Pacific Barrier, a 5000- 
km expanse of deep ocean separat- 
ing central and eastern Pacific shal- 
low water habitats (Rosenblatt and 
Waples, 1986), or in five species of 
damselfishes collected from isolated 
Caribbean reefs (Lacson, 1992 ). 
Similarly, analyses of mitochondrial 
DNA have revealed little genetic dif- 
ferentiation among populations of 
five species of tropical reef fishes 
sampled from throughout the Car- 
ibbean ( Shulman and Bermingham, 
1995), or within three cosmopolitan 
species of tunas (Graves et al., 1984; 
Graves and Dizon, 1989; Scoles and 
Graves, 1993). 
Much less information is avail- 
able on the population structure of 
broadly distributed marine species 
that occur in fragmented or disjunct 
distributions. Recent studies of 
striped mullet (Mugil cephalus, 
