TABLE 1

Power of tests (cases examined by Hudson et al. 1992)

Power
n1n24Nu4Nc4NmKS*Z*Hsχ2Snn
3555.00.02.00.580.620.620.780.77
30100.790.830.830.910.94
25150.870.900.900.960.98
20200.880.920.910.970.98
10105.00.320.340.320.360.42
20.00.460.440.210.210.46
15150.00.470.520.520.630.66
20.00.700.660.540.600.74
25250.01.00.991.000.991.001.00
20.01.001.000.991.001.00
0.02.00.940.960.950.991.00
20.01.000.990.981.001.00
0.05.00.690.750.790.910.92
20.00.900.880.870.950.96
0.010.00.410.460.530.680.69
20.00.680.630.700.810.81
50500.05.00.910.950.971.001.00
20.01.000.991.001.001.00
0.010.00.710.780.850.960.97
20.00.950.930.971.001.00
25250.1560.05.00.170.170.160.190.22
0.6240.180.180.170.190.23
0.3130.00.300.290.280.330.37
1.250.310.300.300.350.38
0.6250.00.410.410.400.500.53
2.50.460.460.440.560.57
1.250.00.530.540.530.670.69
5.00.620.640.640.780.79
2.50.00.610.660.680.830.84
10.00.780.770.790.910.91
5.00.00.690.750.790.910.92
20.00.900.880.870.950.96
10.00.00.760.830.870.950.96
40.00.980.960.850.900.98
15.00.00.780.860.880.960.97
60.00.990.980.740.760.99
  • For each row of this table, 4000 independent samples were generated under a symmetric two-island model. For each of these samples, 4000 random permutations were carried out to estimate the P value of each of the statistics for the sample. n1 and n2 are the sample sizes from locality one and locality two, respectively. N is the population size of each subpopulation. u is the neutral mutation rate per generation. c is the per generation recombination rate between the ends of the segment sequenced. m is the migration fraction per generation. KS*, Z*, and HS are the sequence-based statistics considered by Hudson et al. (1992). χ2 is the chi-square statistic and Snn is the nearest-neighbor statistic. The power estimates are the proportion of samples with estimated P value <0.05.