Table 3.

Niche overlap metric (Schoener's D) and results of niche equivalency tests, niche similarity tests, and indices of niche change in E-space (PCAenv) for five cytotypes of A. oleraceum.

Allium cytotypes		Niche overlap (D)	Niche equivalency	Niche similarity		Indices of niche change (95th percentile)
1	2			1 → 2	2 → 1	Expansion (E)	Stability (Se)	Unfilling (U)	Stability (Sn)
3x	4x	0.161	ns	More similar	More similar	0.101	0.899	0.000	1.000
	5x	0.227	ns	ns	ns	0.068	0.932	0.005	0.995
	6x	0.264	ns	ns	ns	0.077	0.923	0.107	0.893
	7x	0.056	Less equivalent	ns	ns	0.000	1.000	0.509	0.491
4x	5x	0.806	More equivalent	More similar	More similar	0.001	0.999	0.039	0.961
	6x	0.439	ns	ns	ns	0.000	1.000	0.232	0.768
	7x	0.224	ns	ns	ns	0.001	0.999	0.254	0.746
5x	6x	0.540	ns	More similar	More similar	0.000	1.000	0.145	0.855
	7x	0.251	ns	ns	ns	0.012	0.988	0.258	0.742
6x	7x	0.371	ns	ns	ns	0.234	0.766	0.243	0.757

The niche similarity test was repeated in both directions, i.e., by resampling first the occurrences of higher ploidy level and then those of lower ploidy level. Expansion (E) and unfilling (U) indices represent proportions of the niche of higher ploidy level which is not occupied by the lower ploidy level and vice versa. Niche stability (S_n, S_e) represents the proportion of the niche of either lower (S_n) or higher ploidy level (S_e) shared with the other cytotype. Indices of niche change were calculated at the intersection of the 95th percentile of both compared environmental densities. We determined significance of niche equivalency and niche similarity tests if the observed D of the cytotypes being compared was within the lower (“less equivalent/similar”) or upper 2.5% (“more equivalent/similar”) quantile of a null distribution of simulated values of D, respectively (ns, non-significant).