54 
Fishery Bulletin 99(1 ) 
Table 3 (continued) 
Species 
ND3/ND4 haplotypes 
ftsvN I 
Bstl J I 
Cfo I 
Dde I 
Hind II 
Hint I 
Mho I 
Msp I 
Rsa I 
Sty I 
7d 
F 
A 
D 
E 
A 
A 
D 
B 
C 
B 
8 
F 
C 
D 
E 
A 
A 
D 
B 
B 
C 
9 
F 
C 
D 
E 
A 
A 
K 
E 
B 
C 
10a 
F 
C 
B 
N 
A 
E 
K 
D 
D 
C 
10b 
F 
C 
B 
N 
A 
F 
K 
D 
D 
C 
11a 
C 
C 
A 
B 
A 
D 
1 
B 
D 
C 
lib 
C 
C 
D 
B 
A 
D 
1 
B 
D 
C 
12 
F 
C 
D 
D 
A 
A 
F 
B 
B 
C 
13 
A 
D 
A 
M 
A 
A 
F 
B 
F 
C 
14 
G 
C 
D 
C 
A 
A 
K 
D 
B 
D 
15a 
F 
B 
C 
L 
A 
G 
C 
C 
B 
A 
15b 
F 
B 
C 
L 
A 
G 
J 
C 
B 
A 
16 
J 
G 
F 
Q 
A 
A 
N 
G 
H 
A 
17a 
H 
E 
E 
0 
A 
H 
M 
F 
G 
F 
17b 
1 
E 
E 
0 
A 
H 
M 
F 
G 
F 
17c 
H 
F 
E 
0 
A 
H 
M 
F 
G 
E 
17d 
H 
E 
E 
P 
A 
H 
M 
F 
G 
F 
can be resolved by using Msp I, and S. zacentrus and S. 
variegatus (see above). We do not recommend using Dde I 
because it has many sites, often produces small fragments 
requiring both agarose and polyacrylamide gels for reso- 
lution, and is, therefore, time consuming to analyze. How- 
ever, the restriction patterns of Dde I are nearly species 
specific. 
Discussion 
Sufficient interspecies restriction site variation occurred 
in the ND3/ND4 and 12S/16S mtDNA regions in Sebastol- 
obus alascanus , Helicolenus hilgendorfi , and 15 Sebastes 
species to distinguish among them. Intraspecific varia- 
tion was observed in nine of the seventeen species, but it 
did not interfere with our ability to distinguish between 
species. We used the interspecific variation to devise a 
strategy to identify the species we studied. Intraspecific 
variability can serve as a basis for stock identification. 
A broader survey, particularly for S. zacentrus and S. var- 
iegatus, might reveal overlaps in haplotype compositions 
that compromise the ability to distinguish between some 
species pairs. This would be most likely if there were gene 
flow between the species or if the species had recently di- 
verged. Otherwise, extending the analysis to other mtDNA 
regions and additional restriction endonucleases should in- 
crease resolution. Of course, additional intraspecific varia- 
tion has the potential to obscure the topology of trees. To 
test this possibility, we examined trees that included the 
additional haplotypes observed in samples of 40 to 126 indi- 
viduals each from S. caurinus (n=79),S. cileutianus (u = 126), 
and S. borealis 0?=40) (data not shown). The additional 
haplotypes (5, 13, and 5, respectively) increased the num- 
ber of branches at the tip of the species limbs but did not in- 
fluence or obscure relationships with other species. We are 
currently investigating the population structure of S. aleu- 
tianus, S. borealis, S. alutus, S. caurinus, and Sebastolobus 
alascanus by using mtDNA restriction site variation. 
Because of the similarity of many Sebastes species, there 
is a chance that very similar species can be misidentified. 
In fact, a young dusky rockfish (S. ciliatus) and a young 
yellowtail rockfish {S. flavidus ) were misidentified in the 
field as black rockfish (S. melanops) prior to our mtDNA 
analysis. Also, it is possible that closely related species 
may hybridize (e.g. Seeb, 1998). Because hybrids carry on- 
ly the maternal lineage and because only the maternal 
contributor can be identified, mtDNA analysis is a poor 
tool for identifying hybrids. 
In addition to providing a tool that can distinguish 
among a variety of rockfish species, the data appear to 
provide criteria that may prove useful in unraveling some 
questions about rockfish systematics. Both outgroups are 
distinct from Sebastes ; H. hilgendorfi is more closely re- 
lated than Sebastolobus alascanus. The 15 Sebastes spe- 
cies studied include eight subgenera, five of which were 
represented by two or more species. Despite the uncertain- 
ty in some of the subgenus assignments, 2 our analyses of 
mtDNA restriction sites show some concordance with sub- 
generic assignments. Unfortunately, the only recently re- 
viewed subgenus is Sebast077ius (Chen, 1971), for which 
we have only a single representative ( S . lielvo7naculatus). 
A phylogeny of subgenera is unavailable. 
Several species pairs were persistent in the analyses. 
Within Sebastes, S. tnaliger and S. caurinus (subgenus 
Pteropodus) were distinct from the other Sebastes species. 
