. 2016 Jul 11;8:plw038. doi: 10.1093/aobpla/plw038

Table 2.

Network level metrics for the uninvaded, semi invaded, heavily invaded and combined interaction networks from analyses (i) treating bacterial genotypes as taxa (G) and (ii) identifying bacterial taxa on the basis of 98% 16S sequence similarity (98%). Five plant species were sampled at each site and eight across the combined dataset. Webs were significantly modular with high levels of interaction specialization.

	Bacterial taxa (B)		Connectance (I/P*B)		Interaction evenness (IE)		WNODF		Weighted modularity (Q)		Number of modules		Network specialization (H’₂)		Generality plant (Gp)		Generality bacteria (Gb)

	G	98%	G	98%	G	98%	G	98%	G	98%	G	98%	G	98%	G	98%	G	98%
Uninvaded	24	11	0.22	0.33	0.63	0.63	5.40^ns	22.31^ns	0.72^*	0.53^*	5	4	0.90	0.68	4.21	2.78	1.19	2.24
Semi invaded	20	7	0.24	0.34	0.61	0.59	8.00^ns	9.68^ns	0.74^*	0.60^*	5	3	0.88	0.89	3.63	1.68	1.20	1.98
Heavily invaded	17	6	0.27	0.40	0.62	0.60	5.88^ns	12.00^ns	0.71^*	0.59^*	5	3	0.85	0.86	3.36	1.55	1.30	2.04
Combined	44	16	0.15	0.23	0.60	0.57	7.92^ns	15.88^ns	0.78^*	0.59^*	7	5	0.89	0.73	5.61	2.31	1.26	2.49

P < 0.05 from z-scores. ns, indicates no significant nestedness from randomization tests.