examined restriction fragment length polymorphisms (RFLPs) of three PCR-amplified mtDNA fragments from eight species/subspecies of Thunnus : (1) cyt b (355 bp, 15 enzymes, n = 132); (2) 12S ribosomal RNA (450 bp, 20 enzymes, n = 16); and (3) part of the coding regions of the ATPase and COIII genes (ATCO, 940 bp, 20 enzymes, n = 131), as well as regions adjacent to the coding regions. Cyt b RFLP fragment differences separated these tunas into four groups. The groupings identified from the cyt b data by Chow and Inoue (1993) were discordant with groupings inferred from direct sequencing of cyt b (Block et al., 1993) and from analyses of morphological variation (Collette et al., 1984). Chow and Inoue (1993) also reported that ATCO fragments indicated differences among species. One of the 18 northern Pacific bluefin tuna had cyt b and ATCO fragment patterns identical to those of northern Atlantic bluefin tuna. The authors suggested that this reflects incomplete genetic differentiation between northern Atlantic and Pacific bluefin tuna. A more reasonable hypothesis is that the exceptional mtDNA haplotype reflects movement of Atlantic bluefin tuna into the Pacific Ocean basin. The possibility of movement of bluefin tuna between Atlantic and Pacific ocean basins merits further investigation.

Information on biochemical and molecular genetics of Atlantic (Thunnus thynnus) and southern (T. maccoyi) bluefin tuna populations is limited. One early study of frequencies of alleles coding for the protein transferrin among four samples of southern bluefin tuna in Australian waters showed significant allele-frequency differences (G = 21.02, degrees of freedom = 3, P < 0.001; Fujino and Kang 1968). In Atlantic bluefin tuna, Edmunds and Sammons (1971, 1973) found allele-frequency homogeneity at the superoxide dismutase (SOD) locus among samples from the western Atlantic Ocean (New Jersey, n = 269; Rhode Island, n = 87; and Nova Scotia, n = 25) and between these samples and one from the Bay of Biscay (n = 675) in the eastern Atlantic Ocean (G-test, P > 0.05). Edmunds and Sammons (1973) pooled these samples into a single sample (n = 1,056) and tested the pooled sample for fit to Hardy-Weinberg proportions. No deviation from expected equilibrium proportions was found. This was taken as evidence for the lack of genetic differentiation between eastern and western Atlantic bluefin tuna. The oceanwide fit of SOD genotypes to proportions expected from random mating is consistent with the hypothesis that a single population of bluefin tuna occurs in the North Atlantic Ocean.

Using allele frequencies at three nuclear gene loci, Phipps (1980) tested the hypothesis that early-and late-arriving bluefin tuna in St. Margaret's Bay, Canada, did not differ genetically. Most bluefin tuna enter the bay in two waves, one in July and a smaller one in mid-September/October (429 were sampled in the first wave, 16 in the second wave). No departures from expected Hardy-Weinberg proportions within or among samples were found at SOD-1, SOD-2,