Table 1. The f ₄ -statistic (±s.e.) results for comparisons of different topologies.

	Four-taxon tree	f4-statistic	Z-score	P-value
(a)	(M,O; R,T)	−0.0070±0.0015	−4.392	<0.001
	(M,R; O,T)	−0.00611±0.0016	−3.752	<0.001
	(M,T; O,R)	0.0091±0.0011	0.840	0.401
	(O,A; R,T)	0.0064±0.0018	3.590	<0.001
	(O,T; R,A)	0.0049±0.0018	2.751	0.006
	(O,R; T,A)	−0.0015±0.0014	−1.057	0.291
(b)	(R,A; O,M)	0.0089±0.0022	3.992	<0.001
	(R,M; O,A)	0.0095±0.0022	4.255	<0.001
	(R,O; M,A)	0.0006±0.0015	0.377	0.706
	(M,T; R,A)	−0.0040±0.0014	−2.785	0.005
	(M,R; T,A)	−0.0034±0.0017	−2.028	0.043
	(M,A; R,T)	−0.0006±0.0015	−0.414	0.679
(c)	(M,T; O,A)	−0.0031±0.0014	−2.227	0.026
	(M,O; T,A)	−0.0019±0.0016	−1.216	0.224
	(M,A; O,T)	−0.0012±0.0014	−0.851	0.395

A total of 461 jackknife blocks of 5 SNPs were used to estimate s.e. Two-tailed P-values are reported for each Z-score. (a) The f₄-statistic test results for comparisons that include three Pacific ecotypes with one outgroup. Note that in each combination of four populations, one topology (bold) is clearly supported over the two rejected topologies. In each case it is the topology that groups the offshore and resident ecotypes. (b) The f₄-statistic test results for comparisons that include two Pacific ecotypes with the two outgroups. The topologies that group the North Pacific ecotypes (bold) are supported over the two rejected topologies. (c) The f₄-statistic test results for two Pacific transient and offshore ecotypes, with the two outgroups. Note that there is no one clearly supported topology. Further interpretation and visualisation of the drift paths are given in the Supplementary Material. Results over different jackknife block sizes are given in Supplementary Table S4.