36 



Fishery Bulletin 94(1), 1996 



tioned between samples (G ST , 

 of G ST were in the 5.8-8.3% 



Gene diversity analysis of all seven samples (Table 

 6) indicated that over 92.4% of the mtDNA variation 

 detected was within samples, hence 7.6% was parti- 



. Jackknifed estimates 

 range, and the mean 

 estimate was significantly different from zero 

 (* s =22.51, df=6, P<0.001). G ST values for pairwise 

 comparisons of samples were in the 2.6—19.4% range, 

 the highest values from comparisons involving the 

 Aegean Sea group (Table 7). G ST was subdivided into 

 variation between seas (G SEA ) and between sampling 

 ports within seas (for G ps , in the Adriatic only). G SEA 

 accounted for 77% of inter-sample variation, leaving 

 23% partitioned between Adriatic ports (Table 6). The 

 indication was that the degree of genetic differentia- 

 tion was greater on a larger geographic scale. 



Discussion 



Restriction analysis of PCR-amplified mtDNA is a 

 relatively simple technique which enables the rapid 

 screening of large numbers offish for DNA-level poly- 



morphism. The wide range of "universal" and tele- 

 ost-specific primer sequences in the literature (e.g. 

 Avise, 1994) means that one or more fragments that 

 amplify successfully should be obtainable for most spe- 

 cies. Additionally, the production of /ig quantities of 

 DNA by PCR and the silver-staining technique remove 

 the necessity for isotopic labelling of restriction frag- 

 ments, an improvement both in terms of cost and safety. 

 Our results show that variability in anchovy ND5/ 

 6 genes is notably high: 53 haplotypes were detected 

 in only 140 fish, with a mean nucleotide diversity of 

 0.0164. Hauser et al. (in press) have reported simi- 

 lar levels of variation for the ND5/6 region in 

 kapenta, Limnothrissa miodon, a freshwater clu- 

 peoid. Using the same primers as those in this study 

 and a set of six enzymes, they reported 85 haplotypes 

 in 362 fish, with nucleotide diversities of 0.0137 and 

 0.0098 for Lake Tanganyika and Lake Kivu fish, re- 

 spectively, and commented that the order Clu- 

 peiformes may show generally high levels of mtDNA 

 variation. These figures contrast with a restriction 

 study that showed the ND5/6 region to be monomor- 

 phic in Atlantic salmon, Salmo salar (O'Connell, 

 1993), while work on brown trout, Salmo trutta (Hall, 

 1993 ), has demonstrated intermediate levels of varia- 

 tion (19 haplotypes in 219 fish with the use of seven 

 enzymes). Recent work on the ND5/6 genes of Trini- 

 dadian guppies, Poecilia reticulata, has also shown 

 low levels of intrapopulation diversity, attributed to 

 the effects of local extinctions and founder events 

 (Shaw 4 ). We believe that the relatively high level of 

 variation described here may be representative of the 

 entire anchovy mtDNA genome because our pilot 

 study of the NDl/16s region showed that variability 

 was of a similar order to ND5/6. Chapman et al. 

 ( 1994), in a recent study of the genus Alosa, noted 

 that the NDl/16s region evolved more quickly than cy- 



4 Shaw, P. W. 1994. School of Biological Sciences, Univ. Wales 

 Swansea, Singleton Park, Swansea, U.K. Personal commun. 



Table 3 



Nucleotide divergence (Nei, 1987 [Eq. 10.21]) between samples of European anchovy, Engraulis enerasieolus. 



Nucleotide divergence 



Sample 



Trieste 



\ncona 



Vieste 



Ionian 



Sicily 



Tyrrhenian 



0.00423 



