Table 1.

Results from the five different hypotheses on modularity (Fig. 1) for the three pollination syndromes.

Modularity hypothesis	Buzz-bee (n = 16)			Mixed-vertebrate (n = 8)			Passerine (n = 6)			Merianieae
	CR	p	Z	CR	p	Z	CR	p	Z	CR	p
Hypothesis 1	0.815	0.001	2.353	1.100	0.653	0.286	1.077	0.318	0.287	1.527	1.000
Hypothesis 2	0.858	0.026	2.069	1.011	0.319	0.506	1.025	0.192	0.499	0.993	0.25
Hypothesis 3	0.935	0.051	1.889	0.994	0.203	0.767	1.012	0.101	1.117	0.963	0.113
Hypothesis 4	0.787	0.001	5.727	0.947	0.036	2.196	0.831	0.001	13.172	1.020	0.345
Hypothesis 5	0.812	0.001	3.423	0.977	0.070	1.579	0.917	0.005	4.270	0.977	0.124

Highest degrees of modularity are present in the buzz-bee syndrome and lowest in the mixed-vertebrate syndrome, analyses of evolutionary modularity accounting for phylogenetic relatedness (column Merianieae) show significant modularity in Hypotheses 3, 4 and 5

p – p-value < 0.05 (in italics and bold) indicates significantly smaller CR than expected when no modularity is present

CR covariance ratio, Z effect sizes of CR