Table 1.

Multiple regressions for copulation latency for each of the four species

Copulation latency
	Intercept			Relative frequency of heterospecifics			Total number of flies			Interaction: Relative frequency of heterospecifics × Total number of flies
	Estimate	t value	P-value	Coefficient	t value	P-value	Coefficient	t value	P-value	Coefficient	t value	P-value
Drosophila yakuba	20.99	16.639	<1 × 10−10	4.69 × 10−2	2.302	0.022	6.10 × 10−3	3.001	2.74 × 10−3	2.44 × 10−5	0.744	0.457
Drosophila santomea	18.75	17.404	<1 × 10−10	9.38 × 10−4	0.054	0.957	1.39 × 10−3	0.802	0.423	−1.03 × 10−5	−0.367	0.714
Drosophila simulans	27.56	24.868	<1 × 10−10	4.80 × 10−2	2.680	7.45 × 10−3	−1.00 × 10−2	−5.608	2.43 × 10−8	6.79 × 10−5	2.353	0.019
Drosophila sechellia	43.58	33.979	<1 × 10−10	2.35 × 10−2	1.135	0.259	2.25 × 10−3	1.086	0.280	−4.35 × 10−5	−1.301	0.196

A multiple linear regression was fitted to study the effect of number of heterospecifics (F_i), the population size (S_j), and the interaction between these two variables (F_i × S_j) in the copulation latency of each of the four species in experimental cages (). The two factors and the interaction were significant for D. yakuba and D. simulans but not for D. santomea or D. sechellia. Significant values (P < 0.05, df = 1496) are marked in bold.