Table 1.

Numerical experiments. We randomly select the two relevant colony efficiency values from a bivariate normal distribution. For Procedure 1, the two efficiency values are uncorrelated. For Procedure 2, they are correlated (with correlation 0.8). The results of the numerical experiment for Figures 3A and 8 are shown. For Figure 3A, which describes a recessive allele inducing non-reproductive workers, we randomly generate values for r_1/4 and r_1/2. For Figure 8, which describes a dominant allele inducing non-reproductive workers, we randomly generate values for r_1/2 and r₁. The table shows the likelihood of the four possible outcomes: non-reproductive workers (i) do not invade, (ii) invade for single mating but not for double mating, (iii) invade for double mating but not for single mating, and (iv) invade for both single and double mating. For this particular randomization experiment, double mating is more favorable than single mating for the invasion of non-reproductive workers. All p values are exactly as in the corresponding Figures.

DOI: http://dx.doi.org/10.7554/eLife.08918.013

	Does not	Invades for n = 1	Invades for n = 2	Invades for both
	invade	but not n = 2	but not n = 1	n = 1 and n = 2
Figure 3A, Proc. 1, recessive	0.7769	0.0644	0.1465	0.0122
Figure 3A, Proc. 2, recessive	0.8237	0.0177	0.0997	0.0589
Figure 8, Proc. 1, dominant	0.7944	0.0129	0.0830	0.1097
Figure 8, Proc. 2, dominant	0.7927	0.0146	0.0260	0.1667