Gharrett et al.: Phylogeographic analysis of mitochondrial DNA variation in Oncorhynchus kisutch 
533 
Figure 3 
Haplotype tree for coho salmon mtDNA restriction site variation. Haplotypes 
A-H, including A', C', and E', were observed in the preliminary survey (70 
fish); haplotypes 1-0 were observed in the subsequent survey ( 188 additional 
fish); and haplotype P was detected in digests conducted to resolve positions 
on the tree of the latter haplotypes. Arrows indicate the direction of site gain. 
Ambiguities in the topology were resolved (Castelloe and Templeton, 1994) 
and the nested structure was constructed (Templeton and Sing, 1993). Hap- 
lotypes A', C', and E' were pooled with A, C, and E, respectively, for the nested 
clade analysis. The results of the analysis are provided in Table 5. 
chatka River (n=17) and Delta Clearwater River (n= 21) 
populations. To make analysis of larger numbers of sam- 
ples practical, we focused on restriction sites in five 
mtDNA regions that defined the most abundant eight 
(A-H) of the 11 haplotypes. Haplotypes A', C', and E', 
each of which was represented by a single individual, were 
eliminated. In this survey, we analyzed site variation for 
the following PCR region by restriction endonuclease com- 
binations: 
12S/16S 
rRNA Cyt b/D-loop ND5/ND6 ND1/ND2 ND3/ND4 
Cfo I BstN I Ava I Dde I Cfo I 
Dde I Dde I Sty I 
This survey detected 62 restriction sites (56.47 on aver- 
age in each haplotype), which correspond to 262.67 nucleo- 
tides (238.22 on average). The proportion of the genome 
screened was a maximum of 1.58% (1.43% on average). To 
completely resolve the placement of haplotypes I, J, K, and 
L in the “gene tree,” we digested their ND5/ND6 regions 
with Mbo I and their COI/COII regions with Dde I. An ad- 
ditional haplotype (P) was resolved. In total, the expanded 
survey resolved three additional haplotypes (I, J, and P) 
within the A-D clade and five additional haplotypes (K, L, 
M, N, and O) within the E-H clade (Table 4, Appendix 2, 
and Fig. 3). Haplotypes of both clusters appeared in most 
drainages. It is notable that four drainages included hap- 
lotypes of only a single clade: Delta Clearwater had 20 of 
haplotype H and one of the related haplotype O; Indian 
River had 19 of haplotype A and one of haplotype C; Bern- 
ers River had five of hapotype A, 11 of haplotype C, one 
of haplotype I, two of haplotype J, and one of haplotype 
P; and the Little Susitna collection had four A haplotypes, 
15 C haplotypes, and one I haplotype. Within drainages, 
haplotype diversities ranged from 0.10 to 0.73 and nucleo- 
tide diversities ranged from 0.22 to 9.07 substitutions per 
1000 bp. The 13 collections exhibited strong heterogeneity 
(P<10^, Table 4). 
In a Fitch-Margoliash phenogram that estimates rela- 
tionships among drainages based on haplotype frequen- 
cies (Fig. 4), Delta Clearwater River differed strongly from 
the other collections, and the collections from systems 
that drain into the Bering Sea and from Karluk Lake on 
Kodiak Island, clustered separately. The remaining col- 
