Smith et al.: A comparison of three genetic methods for stock discrimination of Hoplostethus atlanticus 
803 
Figure 2 
(A) Random amplified polymorphic DNA (RAPD) profiles in orange roughy generated with the primers A16 and E19. Lane 1 
contains a DNA size-ladder (2,072-100 bp), lanes 2-7 represent orange roughy amplifed with primer A16, lane 8 contains no DNA 
template, and lanes 9-13 represent orange roughy amplifed with primer E19. Each amplified sample of orange roughy contained 
approximately 50 ng of DNA. (B) RAPD profiles in orange roughy generated with the primers A16 and E19 at different concentra- 
tions of DNA template. Lane 1 contains a DNA size-ladder (2,072-100 bp), lane 2 no DNA, lanes 3 and 4 contain 12.5 ng DNA 
amplified with E19, lanes 5 and 6 contain 50 ng DNA amplified with E19, and lanes 7 and 8 contain 200 ng DNA amplified with 
E 19, lane 9 no DNA, lanes 10 and 11 contain 50 ng DNA amplified with A16, and lanes 12-14 contain 200 ng DNA amplified with 
A16. (C) RAPD profiles in orange roughy generated with the primer A14, lanes 1-3 contain 50 ng DNA, and lanes 4-6 represent 
the same samples at a concentration of 200 ng DNA. 
Statistical analyses 
Allozyme genotypes Genotypic frequencies were 
tested for Hardy- Weinberg equilibrium; weakly poly- 
morphic loci (frequency of most common allele >0.95) 
were excluded. Rare heterozygotes were pooled with 
their nearest electrophoretic neighbor to reduce the 
number of cells with less than five observations. Al- 
lele frequencies were tested for heterogeneity among 
populations with contingency % 2 tests with the 
BIOSYS software program (Swofford and Selander, 
1981). To test for geographic structure, contingency 
X 2 tests were undertaken on all pairwise combina- 
tions of populations. Probability levels were modi- 
fied by the Bonferroni procedure for multiple tests 
according to Rice (1989). 
The proportion of allozyme variation due to differ- 
entiation among populations was estimated with 
Nei’s gene-diversity statistic C ST ( Nei, 1973), which 
is a multiallele estimator of Wright’s F gT statistics 
(Wright, 1951). Gene diversity is equal to 
(H t -H s )/H t 
where H T = the total genetic diversity of all popu- 
lations; and 
H s = the mean genetic diversity per popula- 
tion, calculated from the average ex- 
pected heterozygosities. 
Sampling error will produce differences in allele fre- 
quencies, even when samples are drawn from the 
same population, therefore a randomization test was 
used to test for differences due to sampling error 
(Elliott and Ward, 1992). One thousand random- 
izations were used, and the probability was estimated 
from the number of randomizations that were equal 
to or greater than the observed G sr 
Gene diversity, G ST allows an estimation to be 
made of the number of migrants exchanged between 
populations per generation from the relation 
N e m = (1 /G st —1)/4, 
