762 



Fishery Bulletin 100(4) 



Second, the Prince William Sound and Middleton Island 

 populations may warrant consideration as a distinct stock, 

 which should be managed independently of the Shelikof 

 Strait population. 



The significant genetic differences between population 

 pairs in the Gulf of Alaska are alternately supported by 

 allozymes and mtDNA. That is, the Shelikof Strait and 

 Prince William Sound populations differ significantly 

 according to the allozyme data but not the mtDNA data, 

 and the Shelikof Strait and Middleton Island populations 

 differ significantly according to the mtDNA data but not 

 the allozyme data (Table 5). We believe that the significant 

 variation obsei-ved for each population pair is, in fact, real 

 but is not supported by both marker types because these 

 populations differ at only three loci iSOD-2*, MPl' , mtD- 

 NA). Thus, it is reasonable to expect differences between 

 the populations at one locus (in this case, the marker 

 class) and not the others based solely on the random ef- 

 fects of genetic drift. 



Temporal change in genetic variation: 

 Bogoslof Island, Prince William Sound 



In addition to evidence of spatial structure within the Gulf 

 of Alaska, we found evidence of interannual variation in 

 genetic diversity among replicate samples from Bogoslof 

 Island (mtDNA) and Prince William Sound (allozyme 

 SOD-2*). In each case the statistically significant varia- 

 tion was evident at only one marker, although in each 

 case a second marker showed noticeable, but not statis- 

 tically significant, variation between years (e.g. Bogoslof 

 Island, MP/*— 0=0.011, P=0.066; Prince William Sound, 

 mtDNA— 0j,.7^O. 012, P=0.055). We believe there are four 

 likely sources for this apparent temporal shift. 



First, walleye pollock may experience high variability in 

 reproductive success among spawning adults as has been 

 shown, for example, in oysters (Hedgecock, 1994). Walleye 

 pollock have many life history traits such as semiplank- 

 tonic larvae, high fecundity, and external fertilization in a 

 variable environment (Hedgecock, 1994), that may result 

 in highly variable reproductive success and may contrib- 

 ute to temporal instability, particularly in small census 

 populations such as that in Prince William Sound. 



Second, walleye pollock may have varying degrees and 

 types of philopatry. For example, juvenile walleye pollock 

 may recruit to any one of a number of spawning aggrega- 

 tions and then exhibit repeat spawning behavior as an 

 adult (adult philopatry). Adult philopatry has been used 

 to explain the temporal instability observed in Atlantic 

 herring (Clupea harengus, IMcQuinn, 1997] ) and could 

 explain the apparent heterozygote deficit at microsatel- 

 lites in the Bogoslof Island 1998 and Unimak Pass 1998. 

 Alternatively, the degree of philopatry in walleye pollock 

 may be extremely low despite the spatial and temporal 

 predictability of spawning aggregations. 



Third, a source-sink type metapopulation relationship 

 may exist between some walleye pollock populations (e.g., 

 Shelikof Strait and Prince William Sound; Unimak Pass 

 and Bogoslof Island). Bailey et al., (1999) suggested that 

 a metapopulation theory may explain existing data on 



population structure and dynamics of walleye pollock. 

 The temporal genetic variation we observed in Prince 

 William Sound and Bogoslof Island is consistent with 

 this theory. Other circumstantial evidence supporting 

 a source-sink relationship among some walleye pollock 

 populations includes the discovery of walleye pollock in 

 Prince William Sound (sink) concurrent with the increase 

 in biomass of walleye pollock in the neighboring Shelikof 

 Strait (source). 



Finally, our supposed temporal replicates may actually 

 represent samples from different spawning aggregates that 

 spawn in the same location but at different times of the 

 year. This finding suggests that fine-scale, intra-annual, ge- 

 netic variation exists among walleye Pollock that spawn in 

 the same area but perhaps days or weeks apart. 



Summary 



Our results show that the genetic variation, as measured 

 by F^j, among spatially distinct spawning aggregations 

 of walleye pollock is generally low and is similar to the 

 spatial genetic variation observed in other gadids such as 

 Atlantic cod and European hake, Merhiccius merluccius 

 (e.g. Ruzzante et al, 1998; Lundy et al., 2000). In contrast, 

 evidence of temporal genetic variation in gadid fishes is 

 equivocal. Although some populations of European hake 

 exhibit interannual variation in gene frequencies (Lundy 

 et al., 2000), Ruzzante et al. (1997) found no evidence of 

 temporal instability in Atlantic cod. For walleye pollock 

 more detailed studies are needed to determine the extent 

 to which all or some of the factors described above explain 

 the temporal variability observed within some spawning 

 aggregations. A priority of future genetic studies of wall- 

 eye pollock should be to examine both fine-scale temporal 

 variation (days and weeks within a year) and the genetic 

 relationship among age cohorts from spatially distinct 

 spawning aggregations. Because these studies will require 

 extensive sampling, we recommend detailed examination 

 of one or two spawning aggregations, such as those at 

 Prince William Sound and Shelikof Strait. To further 

 evaluate the metapopulation theory, candidate samples 

 should include potential source populations (e.g. Shelikof 

 Strait and Unimak Pass) and sink populations (e.g. Bogo- 

 slof Island and Prince William Sound). Future studies 

 should also include a statistical examination of single 

 locus 6 values to assess if SOD-2* exceeds expectation 

 and is perhaps influenced by a locus-specific force such as 

 diversifying selection (e.g. Beaumont and Nichols, 1996). 



Acknowledgments 



Funding for this study was provided by the Alaska 

 Department of Fish and Game general funds and test fish 

 funds, and by the Exxon Valdez Oil Spill Trustee Council 

 (Restoration Study 99252). Genetic data were collected 

 by Judy Berger, Barbara Debevec, and Eric Kretschmer 

 of the Alaska Department of Fish and Game. Substantial 

 assistance in organizing sample collections was provided 



