Gel electrophoresis coupled with histochemical staining of specific proteins was developed in the 1960s and became the most widely used method for measuring genetic variation in natural populations. The technique is relatively inexpensive and provides a measure of many variable and non variable genes in individuals and populations. The gel phenotypes are easy to interpret and there are computer programmes available for data analyses. Most surveys of genetic diversity in marine species have used proteins, and diversity is measured as the average heterozygosity over many protein loci (heterozygosity is the proportion of individuals that are heterozygous at a single gene locus).
Protein genes represent only about 10% of the genome, so that variation measured by protein electrophoresis may not be representative of the whole genome. The relationship between protein heterozygosity and fitness is uncertain although there is evidence for a positive relationship between heterozygosity and life history characters such as growth rate, disease resistance, and developmental stability (Mitton and Grant 1984, Allendorf and Leary 1986, Danzmann et al. 1989).
The rapid advances in molecular biology have provided a range of techniques for direct examination of variation in DNA. To date most populations studies have used restriction fragment length polymorphisms of the mitochondrial genome. The mitochondrial (mt) DNA is small and relatively easy to purify, and the fragments generated with restriction enzyme digests are easy to interpret. Variations in fragment numbers are generated by additions and deletions of restriction sites, and in fragment lengths by insertions or deletions of blocks of bases. Similar techniques can be applied to nuclear (n) DNA, but the considerably larger size of nDNA means that small pieces of the genome have to be analyzed with specific probes. Several regions of the nuclear genome contain multiple repeats of short minisatellite sequences which are resolved as DNA fingerprints. The hypervariable nature of these variable number tandem repeats has lead to widespread use of DNA fingerprinting in forensic studies, but the technique has had limited application in marine population studies (Baker et al. 1992).
The development of the polymerase chain reaction, PCR, method has provided the means to amplify small fragments of the genome. With appropriate size primers the method can be used to screen for genetic variation in individuals and populations, alternatively amplified fragments can be sequenced. Application of these new genetic methods may produce new insights into the genetic structure of natural populations, as did protein electrophoresis in the 1970s and 1980s, although to date the methods have not been used widely with marine species.
Karyological methods can be used to measure genetic variation, either as chromosome number or banding polymorphisms. The techniques are laborious in comparison with electrophoretic techniques and require the use of live fish for chromosome preparations, thereby reducing their potential application with many marine species.
Morphological characters, the tools of traditional taxonomy, have been used to describe variation among individuals and populations. The characters used are meristic (countable) such as number of fin rays or vertebrae and morphometric (measurable) expressed as ratios of standard length or fork length. Morphological characters have limitations for describing intraspecific genetic diversity as they are polygenic and expression can be modified by the environment. Their use in population-stock identification studies has been superseded for the most part by the development of direct genetic methods.
The life history strategies and characteristics of a species have been determined by evolutionary and ecological processes. These strategies and characters will determine both how a population responds to exploitation and be modified by exploitation (Garrod and Knights 1979, Garrod and Horwood 1984). The life history traits determining fitness in fish populations are growth rate, age and size at first maturity, life span and fecundity. These traits all exhibit phenotypic plasticity and respond adaptively to changes in the environment, in addition the traits are heritable and population means can be changed through a differential mortality. The relative importance of phenotypic plasticity and genetic change is unknown for most species (Wootton 1990) and because of the difficulties of resolving the genetic and environmental components, life history characters have not been as well studied as protein markers in marine species. It is likely that life history characters are influenced by a large number of segregating loci (Bentsen 1994). Heritability estimates are available for some life history characters for species used in aquaculture (see references in Wilkins and Gosling 1983, Gall and Busack 1986, Crandell and Gall 1993a 1993b), and coupled with the results from selection programmes (e.g. Gjedrem 1983) demonstrate the genetic base of these characters. However, transplant experiments with cutthroat trout Salmo clarkii and charr Salvelinus malma have shown that fish can adjust age at maturity non genetically to changes in growth rate (Jonsson et al. 1984).
Invertebrates generally have higher levels of genetic diversity than vertebrates as measured by protein electrophoresis (Nevo 1978). Within the vertebrates amphibia have the highest and teleosts the lowest levels of genetic diversity (Ward et al. 1991). The ecological significance of these findings have been debated for marine species (Nevo 1978, 1983, Nelson and Hedgecock 1980, Smith and Fujio 1982, Mitton and Lewis 1989, Waples 1991). Marine invertebrates show wide variation in levels of genetic diversity. In 26 species of mollusc heterozygosities range from 2 to 32% (Johannesson et al. 1989). Crustacea have lower levels of genetic diversity ranging from 0.4 to 10.9% in 44 species of decapod (Nelson and Hedgecock 1980), from 0.8 to 6.4% in six species of tropical decapod and two species of tropical Stomatopod from the Gulf of Carpentaria (Redfield et al. 1980), and from 0.6 to 3.33% in 13 species of Australasian prawns (Mulley and Latter 1980).
In marine teleosts heterozygosities range from 0.0 in the anglerfishes Lophius litulor (Fujio and Kato 1979) and L. piscatorius (Leslie and Grant 1991), Liparis tanakai (Fujio and Kato 1979) and three species of Cottidae (Johnson and Utter 1976) to more than 17% in the pelagic Cololabris saira (Fujio and Kato 1979) and coastal Fundulus heteroclitus (Mitton and Koehn 1975). The mean heterozygosity for 106 marine species was 5.5% with high levels in Clupeiformes, Atheriniformes, and Pleuronectiformes and low levels in Gadiformes and Scorpaeniformes (Smith and Fujio 1982). Elasmobranchs have low heterozygosities (MacDonald 1988, Smith 1986). There have been relatively few studies of mtDNA diversity in marine species and most species tested have low intraspecific sequence diversities (Ovenden 1990), although the Japanese scallop Pactinopectin yessoensis has high diversity (Boulding et al. 1993).
Many electrophoretic studies of allozymes have been undertaken for the purpose of stock identification or delineation while more recent studies have included the use of direct DNA markers. As might be expected in the marine environment there is less genetic differentiation among populations of teleosts than there is with anadromous and freshwater species. There are fewer isolating barriers to gene flow, which occurs through either larval drift or adult movement, in the continuous realm of the oceans. The proportion of genetic diversity due to population subdivision rises from 1.6% in marine species to 3.7% in anadromous and to 29.4% in freshwater species (Gyllensten 1985). Likewise the level of genetic differentiation measured with mitochondrial DNA is lower in marine than freshwater fishes (Avise et al. 1987). Nevertheless discrete genetic stocks of marine fishes are recognised with proteins (see review in Smith et al. 1990). Genetic differentiation is negatively correlated with dispersal ability in some species of inshore fishes (Waples 1987), starfishes (Williams and Benzie 1993) and crustacea (Mulley and Latter 1981a 1981b). The use of more sensitive DNA techniques is revealing finer population structure, for example a major genetic break was detected in populations of the horseshoe crab Limulus polyphemus off Florida tested with mitochondrial DNA (Saunders et al. 1986) but not with allozymes (Selander et al. 1970), similar results were found over the same areas for the American oyster Crassostrea virginica (Buroker 1983, Reeb and Avise 1990). In the deep water teleost orange roughy Hoplostethus atlanticus allozyme studies of Australian and New Zealand populations have revealed little genetic sub division (Smith 1986, Elliot and Ward 1992) whereas mitochondrial DNA studies have revealed genetic sub division (Smolenski et al. 1993, Smith and McVeagh unpub.). A large mtDNA sequence divergence was found between Arctic and coastal cod Gadus morhua (1.8–5.6%) but a low divergence between coastal localities (< 1%) in the NE Atlantic Ocean (Dahle 1991).
Some species of marine invertebrates with sedentary adult stages, but pelagic larval stages, have shown significant genetic differentiation over short (< 5 km) geographical distances, while others show large scale genetic uniformity of allele frequencies over hundreds to thousands of kilometres (Benzie and Williams 1992, Burton 1983, Hedgecock 1986, Kordos and Burton 1993, Williams and Benzie 1993). Localised differentiation, termed genetic patchiness (Johnson and Black 1982), appears to originate from pre- recruitment processes in the sea urchin Echinometra mathaei (Watts et al. 1990), limpet Siphonaria jeanae (Johnson and Black 1982, 1984), gastropod Drupella cornus (Johnson et al. 1993) and queen conch Strombus gigas (Campton et al. 1992; Mitton et al. 1989). However, micro-geographical genetic changes in other invertebrates are thought to be due to post-settlement selection (e.g. abalone Haliotis rubra, Brown 1991), although some observations on spatial genetic change in the genus Mytilus are due to species mixing (Koehn 1991). The most convincing example of post-settlement selection is from the studies of Koehn and co workers on the leucine aminopeptidase (LAP) polymorphism in the blue mussel Mytilus edulis. Steep clines in LAP allele frequencies in M. edulis along salinity gradients in Long Island Sound (Koehn et al. 980) and at Cape Cod (Koehn et al. 1976) and in Mytilus trossulus on the west coast of North America (McDonald and Siebenaller 1989) are due to post recruitment selection on ocean derived seed (Hilbish 1985, Hilbish and Koehn 1985). In spite of gene flow between populations genetic differentiation occurs in response to strong selection.
Temporal changes in allele frequencies have been reported in several species of marine teleost through repeat sampling of the same locality over time or the sampling of discrete year classes. Given that most genetic studies of marine populations have utilised protein electrophoresis, estimates of temporal genetic changes are restricted to periods within the past 25 years. One of the longest studied markers is the haemoglobin polymorphism in the Atlantic cod Gadus morhua for which allele frequencies in Norwegian populations have remained stable over a 25-year period (Gjosaeter et al. 1992; Jorstad and Naevdal 1989). Populations of the red drum Scianenops ocellatus in the Gulf of Mexico show stability in allozyme and mtDNA genotypes among year classes (Gold et al. 1993).
In contrast genetic changes have been noted over short time periods between year classes of seabream Chrysophrys auratus, (Smith 1979), tarakihi Cheilodactylus macropterus (Gauldie and Johnston 1980), killifish Fundulus heteroclitus (Mitton and Koehn 1975), and the crested blenny Anoplarchus purpurescens (Johnson 1977). Major shifts in allele frequencies at one enzyme locus were reported in a reef fish, the damselfish Stegastes partitus, from Florida over two generations (Lacson and Morizot 1991). A population of damselfish with atypical allele frequencies was found to have typical allele frequencies when resampled three years later. It was suggested that typical allele frequencies were re-established by high levels of gene flow into a perturbed population (Lacson and Morizot 1991). Genetic changes observed in populations of chinook salmon Oncorynchus tshawytscha from the Pacific coast of Oregon appear to result from genetic differences between batches of hatchery released fish (Bartley et al. 1992b, Waples and Teel 1990).
In the Atlantic eel Anguilla rostrata, which has a single spawning ground, there are significant genetic differences between adults and elvers, and among localities along the east coast of the United States (Williams et al. 1973). These genetic differences must develop during the elver stages when they drift from the common spawning ground in the Sargasso Sea (Williams et al. 1973).
Homozygous excess has been reported in allozyme studies of marine organisms, especially in marine molluscs where the excess is most notable in larval and juvenile stages (Singh and Green 1984, Zouros and Foltz 1984). Such excess occurs in large populations and is unlikely to be due to genetic drift or technical scoring errors, and the cause of this widespread phenomenon remains obscure (Zouros and Romero-Dorey 1988).
Several allozyme studies have revealed cryptic species in coastal fisheries and have shown that resources considered to be single taxa consist of two or more species. Examples of cryptic species have been found in squid (Brierley et al. 1993, Carvalho et al. 1992, Smith et al. 1981), octopus (Levy et al. 1988), bivalves (Grant et al. 1984, Richardson et al. 1982, Sarver et al. 1992), swellfishes (Masuda et al. 1987), silversides (Prodohl and Levy 1989), lizard fishes (Shaklee et al. 1982, Waples 1981, Yamaoka et al. 1989), bone fishes (Shaklee and Tamaru 1981, Shaklee et al. 1982), and small pelagics (Daly and Richardson 1980, Smith and Robertson 1981).
Conversely, the lack of genetic differences between colour morphs of the small serranid fishes of the genus Hypoplectrus suggest that they compose a single species (Graves and Rosenblatt 1980). Likewise the lack of genetic differences between specimens of the pelagic armourheads Pseudopentaceros wheeleri and P. pectoralis from the North Pacific Ocean lead to the conclusion that the armourhead consists of a single metamorphic species with different morphologies between life-history stages (Humphreys et al. 1989). Lack of genetic differences at 33 loci between two species of rock lobster Jasus edwardsii from New Zealand and J. novaehollandiae from Tasmania indicates that these are conspecific populations (Smith et al. 1980).
The effects of pollution on coastal resources are often dramatic with mass mortalities in local stocks, reduction in species diversity, and changes in species composition. Local areas may be closed to harvesting. Sources of pollution include heavy metals, pesticides, oils and detergents, and thermal and radioactive discharges. There are limited reports of genetic changes due to marine pollution, in part because of the difficulty of measuring genetic changes in fish populations in which recruitment can be from outside the polluted area. Most examples of pollution induced genetic changes are for species with limited dispersal abilities; molluscs may be recruited from outside the area of pollution but the juvenile and adult stages are sessile.
The extensive studies of Nevo and coworkers on genetics of pollution in the Mediterranean have shown that genetic changes occur in natural populations of marine organisms exposed to local pollution events (Nevo et al. 1984, 1987). As a result genetic markers have been proposed as a monitoring tool for marine pollution (Ben-Shlomo and Nevo 1988; Nevo et al. 1984). Laboratory studies on molluscs and crustacea have demonstrated differential survival of allozyme genotypes exposed to heavy metals (Hvilsom 1983; Lavie and Nevo 1982, 1986; Nevo et al. 1981). Similar changes in gene frequencies have been detected in marine organisms exposed to crude-oil (Battaglia et al. 1980; Nevo and Lavie 1989; Nevo et al. 1978) although Fevolden and Garner (1986) found no evidence for genotypic selection in Mytilus edulis exposed to low concentrations of oil in Norwegian fjords. In laboratory tests on pairs of species exposed to marine pollutants those species with the higher level of genetic diversity showed greater survival (Nevo et al. 1986).
There are fewer reports of genetic changes in teleosts due to pollution. Temporal shifts in allozyme frequencies were greater at polluted than none polluted sites in Baltic populations of the fourhorn sculpin Myoxocephalus quadricornis (Gyllensten and Ryman 1988). However, it is not clear if these genetic changes are due to a direct selective mortality or due to invasion of polluted sites by new stock (Gyllensten and Ryman 1988).
Most marine fishes and invertebrates are iteroparous, reproduce over several years, whereas the Pacific salmon, genus Oncorhynchus, are semelparous, reproducing once in their lifetime. Many species are highly fecund and the life history includes a dispersive phase (Fogarty et al. 1991). However, there are considerable interspecific variations in lifespan and fecundity, and many of the widely distributed species exhibit long spawning seasons with the time of spawning varying latitudinally. In the pink salmon Onorhynchus keta the time of return for spawning has a genetic component (Gharret and Smoker 1993). In the European scallop Pecten maximus transplant experiments have demonstrated a genetic component to spawning period and there are intra-stock differences for this character (Cochard and Devauchelle 1993, Mackie and Ansell 1993). Size at sexual maturity varies between intraspecific stocks in skate Raja radiata (Templeman 1987) and American plaice Hippoglossoides platessoides (Roff 1982). Herring Clupea harengus and European plaice Pleuronectes platessa show geographic variations in fecundity (Mann and Mills 1979) and cod Gadus morhua intraspecific differences in age at maturity (Garrod and Horwood 1984).
There is a significant correlation between the age at first reproduction, mortality and growth rate in teleosts and it has been suggested that these correlations are the result of an evolutionary trade-off between growth, reproduction and survival (Roff 1984). Fishing theory predicts that an increase in fishing mortality will produce an increase in growth and recruitment. Whether the onset of sexual maturity is determined by age or size it will be changed by increased fishing mortality: faster growing fish will reach the minimum size more quickly and mature at an earlier age, alternatively if the onset of maturity is determined by age, then fish will mature at a larger size. The complex and poorly understood relationships between the genetic components of growth rate and the size and age of first maturity, and the non-genetic responses of these traits to changes in population density and other environmental parameters, especially water temperature, make it difficult to separate the genetic and non genetic impact of fishing on natural populations.