Smith et al.: A comparison of three genetic methods for stock discrimination of Hoplostethus atlanticus 
805 
Table 2 
Heterogeneity x 2 pairwise comparisons for allozyme loci, mtDNA haplotypes, and random amplified polymorphic DNA (RAPD) 
fragments, among four populations of orange roughy. For the allozyme and RAPD data only those loci and fragments that were 
significant applying a Bonferroni-modified probability level are given. 
mtDNA haplotype RAPD primer fragments 
Pair Allozyme loci and probabilities probablities and probabilities 
Ritchie and Box 
NS 
NS 
NS 
Ritchie and Waitaki 
Est-1* 
PcO.001 
0.001 
E19-3 
P<0.001 
Ritchie and Puysegur 
Gpi-2* 
PcO.001, Idh-2* P<0.001 
NS 
A16-1 
P<0.001 
Box and Waitaki 
Est-1* 
P< 0.001, Idh-l*P= 0.001 
NS 
E19-3 
P<0.001 
Idh-2* 
PcO.001 
Box and Puysegur 
Idh-2* 
P<0.001 
NS 
A16-1 
P<0.001 
Waitaki and Puysegur 
Est-1* 
PcO.001 
0.003 
E19-3 
P<0.001 
fled P for 11 loci (Table 1). To test for geographic struc- 
ture, additional x 2 tests were carried out on all 
pairwise combinations of populations. There was a 
significant heterogeneity for at least one locus be- 
tween all population pairs, except Ritchie Bank and 
Box (Table 2). 
The heterogeneity in the total data was confirmed 
by the gene diversity analysis (Table 1). When a 
Bonferroni-modified P is applied, the 5 loci show a 
G st significantly greater than that due to sampling 
error. Over all eleven loci G ST was 0.020 (Table 1), 
indicating that around 2% of the observed genetic 
variation was due to differences among populations. 
From this estimate of G ST , and by subtracting the 
G sTnull ’ ^e minimum number of effective migrants 
per generation ( N e m ) was 13.2 (Table 3). Individual 
pairs of Nrn varied from 15.7 (Box and Waitaki) to 
124 (Ritchie and Box). 
mtDNA 
The estimated size of the PCR amplified B-loop was 
1,500 base pairs and that of the cytochrome b was 
500 bp. Four restriction enzymes, BstU I, Cfo I, Msp 
I, and Nla III, produced two or more fragment pat- 
terns with the D-loop primers (e.g. Fig. 3) and were 
tested in all fish. For each area, a few fish samples 
failed to produce an amplification product; the same 
fish samples also failed to produce an amplification 
product with the RAPD primers. Four restriction 
enzymes, Alu I, Bfa I, Rsa I, and Taq I, showed varia- 
tion in the first 24 fish tested with the cytochrome b 
primers, but the variation was limited to a single 
individual with each restriction enzyme. No further 
amplifications were undertaken with this set of prim- 
ers. The numbers of haplotypes observed at each site 
are shown in Appendix Table 2. There is a signifi- 
cant heterogeneity in the total data (P=0.001), with 
only 1 out of 1,000 randomizations exceeding the 
Table 3 
The estimated number of migrants exchanged per genera- 
tion ( Njn ) for allozyme, mtDNA, and random amplified 
polymorphic DNA (RAPD) data sets of orange roughy. 
Population 
Allozyme 
mtDNA 
RAPD 
Ritchie and Box 
124.0 
277.8 
75.0 
Ritchie and Waitaki 
14.7 
7.2 
7.7 
Ritchie and Puysegur 
19.4 
35.7 
18.0 
Box and Waitaki 
15.7 
18.3 
7.8 
Box and Puysegur 
25.3 
36.5 
16.0 
Waitaki and Puysegur 
75.5 
9.6 
6.5 
Total 
13.2 
9.8 
7.0 
original x 2 value (Table 1). In pairwise comparisons 
of the four spawning populations (Table 2), signifi- 
cant differences were found between Ritchie Bank 
and Waitaki (P<0.001) and between Waitaki and 
Puysegur (P=Q.Q01), but not in the other pairwise 
comparisons. 
Gene diversity was estimated to be 0.057 (Table 
1), which is significantly greater than that due to 
sampling error, and indicates that around 6% of the 
observed genetic variation is due to differences 
among populations. From this estimate of G gT , and 
by subtracting the G STnul[ , the minimum number of 
female migrants per generation (Af e m^) among the 
four populations was estimated to be 9.8 (Table 3). 
The pairwise values varied from 7.2 (Ritchie and 
Waitaki) to 277.8 (Ritchie and Box). 
RAPD 
Seven primers tested in 24 orange roughy produced 
clear DNA fragments and the same profiles in re- 
peat tests. The primers (and their sequences 5' to 3') 
