Candy et al.: Dividing population genetic distance data by partitioning optimization 
49 
Table 1 
Chinook salmon (Oncorhynchus tshawytsclia ) sampling data by region, stock code, population, years sampled, and annual and 
total sample size used in the bi-partitioning optimization using restricted growth strings (bi-PORGS) analysis. (H) indicates 
major hatchery facilities. 
Region 
Stock code 
Population 
Years sampled 
Annual n 
Total n 
Quatsino Sound 
1 
Marble 
1994, 1996, 1999, 2000 
58, 98, 149 202 
507 
2 
Colonial 
1999, 2004 
40, 18 
58 
Nootka Sound 
3 
Zeballos 
2002, 2004 
4,30 
34 
4 
Tahsis 
1996, 1999, 2002, 2003 
72, 87, 104 47 
310 
5 
Conuma (H) 
1988, 1996, 1997, 1998 
47, 214, 143 52 
456 
6 
Tlupana 
2002, 2003 
34, 32 
66 
7 
Gold 
1983, 1985, 1986 
9, 13, 71 
93 
8 
Burman 
1976, 1985, 1986, 1989, 
8, 20, 2, 35 19, 56, 35 34, 51, 13 
273 
Clayoquot Sound 
9 
Tranquil 
1990, 1991, 1992, 2000, 
2002, 2003 
1996, 1999 
209, 133 
342 
10 
Kennedy 
1992, 2005 
49, 190 
239 
Barkley Sound 
11 
Nahmint 
1996, 2001, 2002, 2003, 
27, 56, 51 124, 135 
346 
12A 
Somas 
2004 
1973 
155 
155 
12B 
Robertson (H) 
1996, 2003 
155, 183 
338 
13 
Thornton 
1992, 1999, 2000, 2001 
37, 147, 150 184 
518 
14 
Toquart 
1999, 2000 
70, 17 
87 
15 
Sarita 
1996, 1997, 2001 
112, 157, 146 
415 
Southwest 
16 
Nitinat (H) 
1989, 1996, 2003 
53, 153, 140 
346 
Vancouver Island 
17 
San Juan 
2001, 2002 
80, 116 
196 
18 
Sooke 
2004 
58 
58 
the observed values of the original proximity matrix by 
using the gap statistic (Eq. 5; Fig. 2, D and E). 
Genetic data for Chinook salmon 
from the west coast of Vancouver Island 
Next we applied PORGS to genetic data derived from Chi- 
nook salmon populations. Tissue or scale samples were 
taken from Chinook salmon during times of broodstock 
collection and spawner enumeration along the west coast 
of Vancouver Island (Fig. 1). Genomic DNA was isolated 
from samples taken at 18 sites since the early 1970s 
(Table 1) and polymerase chain reaction analysis was 
conducted to amplify 12 microsatellite markers ( Ogo2 , 
Ogo4, Oke4, OkilOO, OtslOO, OtslOl, Otsl04, Otsl07, 
Ots2, Ots9, Omy325, and Ssal97) by standard methods 
(Beacham et ah, 2006a). All samples were combined over 
years except for one sample from the Robertson Creek 
Hatchery which was used to test the temporal stability of 
the microsatellites for population differentiation. Above 
the confluence with Sproat River, the Somas becomes 
Stamp River. Robertson Creek Hatchery lies on Stamp 
River downstream from Great Central Lake. The sample 
collected from Somas River in 1973 was considered to be 
the same stock as that represented by the sample from 
Robertson Creek Hatchery taken in 1996 and 2003. 
The Robertson Creek Hatchery stock was founded from 
Somas River fish collected from 1972 to 1976. and from 
additional broodstock collected from the river since then 
(Table 1, stock codes 12a and 12b). 
For this analysis, we used Weir and Cockerham’s 
(1984) co-ancestory coefficient 6, a widely used measure 
of genetic differentiation (Waples and Gaggiotti, 2006). 
A pairwise estimate of 6 was calculated from data on 
multilocus genotypic distance between populations by 
using FSTAT software (Goudet, 1995). 
History of Chinook salmon brookstock 
from the west coast of Vancouver Island 
The west coast of Vancouver Island has three major 
hatcheries, Robertson Creek, Nitinat River, and Conuma 
River, which began enhancement of Chinook salmon in 
1972, 1990, and 1979, respectively (Cross et ah, 1991). 
Juveniles from these hatcheries have been transplanted 
into several river systems (Table 2). Robertson Creek 
Hatchery provided the founder stock for Thornton Creek 
Hatchery with transfers from brood years 1982-84. 
Since then, the hatchery-supported run has been perpet- 
uated by returns to Thornton Creek. The Nitinat River 
Hatchery has transferred Nitinat River Chinook salmon 
to Toquart River (broods 1990-97 and 1999-2001) and 
