Loss of genetic diversity due to genetic drift will occur in small populations. There are few examples of marine fish populations being reduced to the stage where genetic drift might be operating, but some populations of bivalves, such as giant clams and Trochus, have been severely depleted in the Indo Pacific Ocean as a result of overexploitation, and may be unable to survive without restocking. Protein electrophoresis of three species of giant clam has shown high levels of genetic diversity but the sampling was patchy and did not include samples from areas where the species are rare (Benzie 1993). In contrast rare and endangered freshwater teleosts in North America show reduced genetic diversity and several desert populations are characterised by zero variability (Echelle 1991). Low levels of heterozygosity in upper river populations of chinnok salmon Oncorhynchus tshawytscha from the Columbia River basin, Oregon and California are thought to be the result of natural and human-related bottlenecks (Winans 1989, Bartley et al. 1992b). The northern elephant seal Mirounga angustirostrus was heavily exploited last century and reduced to less than 100 individuals. The species shows reduced genetic variation measured by allozyme polymorphisms (Bonnell and Sealnder 1974) and mtDNA sequences (Hoelzel et al. 1993) in comparison with populations of southern elephant seals M.leonia which, although exploited, numbers did not fall below 1 000 individuals (Gales et al. 1989, Hoelzel et al. 1993).
Experimental evidence demonstrates that inbreeding reduces productivity and size of fish. Populations of the eastern mosquitofish Gambusia holbrooki were established in artificial ponds based on one virgin female and one virgin male fish. In three pools the fish were full sibs and in four pools the fish were unrelated. Populations established with full sibs have a predicted 25% less genetic variation than unrelated individuals. After three generations the populations founded on unrelated individuals produced 35 times more juveniles than populations founded on related individuals and size of males at sexual maturity was greater in unrelated than related populations (Ledberg 1990).
The expansion of the aquaculture industry and the move to hatchery production of seed for both farming and enhancement programmes has lead to a large number of hatcheries being established for marine and freshwater organisms. The high fecundity of these species allows the production of numerous seed from a few parents and the risk of loss of genetic diversity. Genetic studies of hatchery populations suggest that loss of genetic diversity is common and has been reported in hatchery salmonids in North America (Allendorf and Phelps 1980, Cross and King 1983, Leary et al. 1985, Verspoor 1988) and Europe (Ryman and Stahl 1980 Stahl 1983, Vuorinen 1984, Koljonen 1989), in sea bream in Japan (Taniguichi et al. 1983, Sugama et al. 1988) and in oysters (Gosling 1981, Gaffney et al. 1992, Hedgecock and Sly 1990), clams (Dillon and Manzi 1987, Benzie 1993), abalone (Smith and Conroy 1992) and shrimp (Sbordoni et al. 1986). In addition the life history characters of cultured Atlantic salmon Salmo salar often differ from those of wild stocks (Heggberget et al. 1993).
While loss of genetic diversity is of concern to aquaculturists the release, intentional or otherwise, of hatchery stocks has ecological and genetic implications for wild fisheries (Marnell 1986, Thomas and Mathieson 1993). The maintenance of some populations of Pacific salmon is dependent upon hatchery seed (Winans 1989) and more than 4 billion smolt are released annually in the North Pacific Ocean (Hindar et al. 1991). The genetic impact of large scale releases are uncertain but may have a negative impact on wild populations if the genetic differences between the hatchery stock and the wild population are great (Skaal et al. 1990, Hindar et al. 1991). This is especially relevant to salmonid populations that often are composed of genetically differentiated sub populations. A large number of genetic effects of cultured salmonids on native populations have been documented by Hindar et al. (1991). There is some evidence that introduced stocks of chum salmon Oncorhynchus keta do less well than native stocks and do not contribute to subsequent spawning runs (Altukhov and Salmenkova 1990), but in the Atlantic salmon Salmo salar electrophoretic evidence shows that escaped farm salmon interbreed with wild stocks (Crozier 1993). In Norwegian rivers significant numbers of escaped farmed salmon have been identified in the spawning rivers (Skaala et al. 1990) and in extreme situations up to 80% of river spawners may be of farm origin (Hindar et al. 1991). In this situation with small natural populations there is a risk that released fish will swamp the native gene pool.
The concerns raised over the genetic impact of hatchery produced salmonids on wild stocks serve as a warning of the pitfalls for marine enhancement programmes. Genetic approaches to avoiding inbreeding and loss of diversity due to drift in hatcheries are well known (FAO 1981, Soule 1986. 1987, Gall 1987, Munro 1993) but the holding and spawning of large numbers of broodstock can present technical difficulties for marine species (e.g. Gervis 1993). The use of local broodstock, or broodstock genetically similar to local populations, is preferred to reduce the risk of introducing foreign alleles in the enhanced population. In addition proper hatchery management including maximising Ne (effective population size) and minimising inbreeding will be necessary.
The concept of genetic marking of hatchery released fish has appeal to fisheries managers wishing to measure the short-term impact of the released fish on the natural population, but the practice could have a long term genetic impact on the population. Most proposals for genetic marking identify rare alleles as the genetic tag to be selectively bred into the hatchery seed, usually through crossing a few parents carrying the rare allele. Even the use of 50 wild parents carrying the rare allele does not guarantee that some are not related, and without some knowledge of the genetic make-up of the carriers of the rare allele, there is a risk of producing seed that exhibits some loss of diversity.
Several of the world's largest pelagic fisheries have collapsed in the past 50 years, although most have recovered slowly following a cessation of fishing (Beverton 1990). Limited genetic data are available for some stocks which have collapsed but need to be interpreted with caution. Low levels of genetic diversity in some mammalian populations have been explained by population bottlenecks, either natural (O'Brien et al. 1983) or due to overexploitation (Bonnell and Selander 1974, Hoelzel et al. 1993, Pemberton and Smith 1985). However, some other mammals which are not known to have passed through a recent bottleneck also exhibit low levels of genetic diversity (e.g. Simonsen 1982). Genetic data for the collapsed stocks were collected after the collapse and estimates of genetic change are based on a comparison with other populations or closely related species; it has not been possible to test directly for loss of specific rare alleles or reduction in average heterozygosity within populations. Loss of allelic diversity will occur before decreases in heterozygosity are detected (Waples et al. 1990) and short-term population bottlenecks will have little effect on heterozygosity, but may reduce the number of alleles present (Allendorf 1986). In addition the stocks may have collapsed from a commercial perspective, and fishing ceased due to economic reasons or management controls, but from a genetic perspective the species may have maintained large numbers of individuals. Examples of collapsed stocks are found in small pelagic species which at their peak have maintained very large population sizes with a mature biomass in excess of 1 million tonnes (Beverton 1990). The Icelandic herring which may have collapsed to 1/3 000 of its peak biomass would consist of around one million fish at its lowest recorded size, while other collapsed stocks have maintained larger population sizes (Beverton 1990).
The California sardine (Sardinops sagax caerulea) fishery collapsed in the late 1950s, and remained below 1/250 of its peak biomass for 20 years before showing signs of recovery (Beverton 1990). This species of sardine shows low genetic diversity relative to other clupeiods with a mean heterozygosity of 1.0% (Hedgecock et al. 1989). In fifteen other species of clupeoid the average heterozygosity is 7.1%, and ranges from 4.0 to 10.1% (Hedgecock et al. 1989). However, Hedgecock et al. (1989) suggested that this low genetic diversity is unlikely to be due to the collapse of the fishery as other populations of Pacific sardine in Baja California and the Gulf of California, which were unaffected by the collapse, also show low heterozygosities. However, the low level of genetic variation in S. sagax caerula and the relatively small sample sizes have not permitted a test of the distribution of rare alleles which may be a more sensitive measure of genetic change in large populations. Other species of sardine which do not appear to have suffered a recent collapse also exhibit low heterozygosities (Kinsey et al. 1994, Menzes 1994).
The fishery for the Japanese sardine Sardinops melanosticta collapsed in the 1940s, falling from a peak catch of 2 700 000 tonnes in 1937 to less than 10 000 tonnes in 1965. The species disappeared in the northern part of its range, but subsequently recovered with catches peaking at over 1 000 000 tonnes per annum by the late 1970s (Kondo 1980). This species has an average heterozygosity of 6.4%, similar to two other clupeiods tested in the same laboratory (Fujio and Kato 1979). Even at the low point following the collapse of the fishery catches remained at thousands of tonnes per annum. Likewise with the South African pilchard Sardinops ocellata fishery, which collapsed in the 1970s from peak catches of over one million tonnes per annum to 11, 000 tonnes per annum in 1980, the average hetreozygosity (5.2%) is similar to other clupeoids (Grant 1985).
The Georges Bank herring Clupea harengus fishery collapsed in the 1970s and the stock may have decreased to 1/1 000 of its peak biomass (Beverton 1990). Spawning herring were recorded on Georges Bank for the first time in 1986 and samples were collected for genetic analyses. The Georges Bank herring were found to be different to neighbouring populations in the Gulf of Maine at one enzyme locus and, coupled with a different year class structure to neighbouring populations, suggests that the stock recovered through resurgence and not recolonization (Stephenson and Kornfield 1990). The heterozygosity levels at two enzyme loci and the level of mitochondrial DNA diversity were similar to other spawning groups in the Northwest Atlantic Ocean, indicating no loss of genetic diversity following the collapse of the fishery (Stephenson and Kornfield 1990), although this comparison is made against other heavily exploited stocks of herring.
The bluefin tuna Thunnus maccoyii is heavily exploited in the Southern Oceans; the fishery has not collapsed but the parental biomass is only 10–20% of the 1965 level (Report 1993). Tissue samples collected in the New Zealand fishery during 1982–83 were tested for 14 variable enzyme loci and repeat tests made in 1992. There was no significant genetic change, measured as average heterozygosity and average number of alleles, over this ten- year period of declining biomass (Smith unpub). In contrast stocks of the orange roughy which have been reduced by 60–70% of the virgin biomass have shown a significant loss of genetic heterozygosity, but this has been attributed to selection and not drift, as the roughy stocks maintain large spawning populations and support reduced but significant (several thousand tonnes per annum) fisheries (Smith et al. 1991).