Table 1. Description of models tested.

Model	Population where selection occurred	$t_{AB}$	$t_{ABC}$	$t_{\mathrm{M}}$	s	$N_{\mathrm{e}}$
A	Ancestral population	500	2,000	1,800	0.1	10,000
B	Ancestral population	1,000	4,000	2,500	0.1	10,000
C	Ancestral population	2,000	4,000	3,500	0.1	10,000
D	Ancestral population	3,000	8,000	5,000	0.1	10,000
E	Ancestral population	2,000	16,000	8,000	0.1	10,000
F	Ancestral population	4,000	16,000	8,000	0.1	10,000
I	Daughter population a	2,000	4,000	1,000	0.1	10,000
J	Daughter population a	3,000	8,000	2,000	0.1	10,000

All times are in generations. Selection in the “ancestral population” refers to a selective sweep where the beneficial mutation and fixation occurred before the split time of the two most closely related populations. Selection in “daughter population a” refers to a selective sweep that occurred in one of the two most closely related populations (a), after their split from each other. $t_{AB},$ split time (in generations ago) of populations a and b; $t_{ABC},$ split time of population c and the ancestral population of a and b; $t_{\mathrm{M}},$ time at which the selected mutation is introduced; s, selection coefficient; $N_{\mathrm{e}},$ effective population size.