Table 2.

Information on the genetic structure of the population at generation 500 for 15 evolutionary runs under random versus m-trait mating (run numbers sorted on m separately for cases A. and B.; genetic structure of runs 1 and 15 for both cases illustrated in Fig. 2). The heterozygote deviance value (see Method section) is with respect to the parameter α, while the % variance explained by our principal components analysis (PCA) is with respect to the first two components

	A. Random matinga			B. m-Trait matinga
Run #	m	Heteroz. devianceb	% PCA Var	m	Heteroz. devianceb	% PCA Var
1	0.19	0.07	54	0.02	−0.99	80
2	0.31	0.08	53	0.02	−0.97	79
3	0.38	−0.06	54	0.03	−0.98	78
4	0.40	−0.32	58	0.04	−0.99	86
5	0.41	0.03	58	0.04	−0.97	78
6	0.44	0.10	55	0.05	−0.96	78
7	0.45	0.56	55	0.05	−0.96	87
8	0.48	0.04	53	0.06	−0.98	64
9	0.54	−0.14	57	0.06	−0.96	75
10	0.61	NAc	55	0.09	−0.95	70
11	0.66	0.10	53	0.11	−0.96	62
12	0.67	0.15	54	0.11	−0.94	60
13	0.76	0.13	57	0.25	−0.84	57
14	0.78	−0.17	57	0.28	−0.83	58
15	0.93	0.19	52	0.28	−0.86	62

^aStatistical comparisons of A. and B. are highly significant for all three measures using a using a two-tailed non-parameteric Mann-Witney U test

^bA negative value implies a deficit of heterozygotes compared with Hardy-Weinberg equilibrium

^cNo value is given because insufficient genetic structure evolved in the parameter α to identify fewer than 5 distinct alleles