Hawkins et al.: Genetic variation of Sebostes aleutianus and 5. borealis 



527 



170 00 -ISO 00' -renin' -160°00' -150°00 -140°00 -130 00' 



120 00' 



- 1(H) 00 



6000- 



50°00 ! 



F7\ 



ssia ) \ vv^. 



l"i 



Vsf "> 



i: 



14 



12 11 10 



Aleutian Is lands 



t 



N 



Oilliii 



50 00 



-170°00 



-I ' 



-150°00' 



-14o 00 



-13O'0O 



Figure 1 



Location of rougheye tSebastes aleutianus) and shortraker rockfish (Sebastes borealis) collection sites, which 

 correspond to locations in Table 1. 



Weinberg equilibrium within populations and in link- 

 age equilibrium with one another within populations. 

 These assumptions were tested with the PC program 

 GENEPOP (vers. 3.4, Univ. Montpellier, Montpellier, 

 France). 



Regional groups were separated into two groups of 

 rougheye rockfish types according to the multidimen- 

 sional scaling analysis and Bayesian clustering model 

 and were retested for Hardy-Weinberg equilibrium and 

 homogeneity of allelic frequencies (G-test) among regions 

 for each type. Chord distance (Cavalli-Sforza and Ed- 

 wards, 1967) for 25 loci was used to assess the overall 

 similarities of allelic frequencies for the two rougheye 

 rockfish types with multidimensional scaling analysis 

 (Rohlf, 2000). Only 25 loci were used because the Rus- 

 sian collection was missing data at 4 loci. Regions 9b. 

 14a, and 16a were therefore not included in these analy- 

 ses because of the limited number of loci available. 



Because the two rougheye rockfish types exhib- 

 ited a distinct yet puzzling pattern of distribution — 

 nearly all S. sp. cf. aleutianus in the Aleutian Is- 



lands, nearly all S. aleutianus in the central Gulf 

 of Alaska, and both types in sympatry in Southeast 

 Alaska — we collected rougheye rockfish at different 

 depths in 2001 (regions 5a-9a, 9b, 16a). We ran a 

 Mann-Whitney rank sum test (SigmaStat, vers. 2.0, 

 SPSS. Chicago, IL) to test for significant differences 

 of the mean, standard deviation, and range of depths 

 between the two rougheye rockfish types. A single depth 

 of 350 m was used to approximate depth of catch for the 

 2001 Southeast Alaska rougheye rockfish collections (re- 

 gions 5a, 6a, and 7a) because depths were reported only 

 as a range from 300 to 600 m. Had we chosen a deeper 

 average depth in the range, the difference in depth be- 

 tween the two rougheye rockfish types would have been 

 (and in actuality may be) even greater. Because the two 

 rougheye types were found in sympatry, we analyzed 

 the length data to determine if size differences existed 

 between the two types. Linear regressions were used to 

 examine the relationships between length (tip of snout 

 to fork of tail) and depth of capture of both shortraker 

 rockfish and the two rougheye rockfish types. 



