Table 2. Results of models evaluating the effects of sex and the interaction of sex and number of false head (FH) components, in addition to number of FH components and collection of origin (MZFC or CNIN), on the probability of symmetrical damage at the species (a) and genus (b) level.

In both models, sex and the interaction between sex and number of FH components were non-significant and therefore were removed from the initial models (a, c), leaving number of FH components and collection of origin as the only variables in the minimum adequate models (b, d). Significant P-values are in bold type.

Fixed Effects	β± SEa	Za	PZa	Pχ2b
(a) Species level initial (complete) model
Number of FH components	0.7459 ± 0.2299	3.244	1.18 × 10−3	7.652 × 10−3
Collection (MZFC-CNIN)	0.3989 ± 0.2129	1.874	0.06095	0.055
Sex	1.1220 ± 0.6293	1.783	0.07459	0.8001
Number of FH components * Sex	−0.3851 ± 0.2145	−1.795	0.07259	0.0724
Random effects: ${\sigma }_{\mathrm{species}}^2$= 0.5674 n = 306 total observations from 105 species
(b) Species level minimum adequate model
Number of FH components	0.4999 ± 0.1827	2.736	6.21 × 10−3	–
Collection (MZFC-CNIN)	0.3955 ± 0.2126	1.860	0.06282	–
Random effects: ${\sigma }_{\mathrm{species}}^2$= 0.5478 n = 306 total observations from 105 species
c) Genus level initial (complete) model
Number of FH components	0.6274 ± 0.2840	2.209	0.0272	0.0488
Collection (MZFC-CNIN)	0.3999 ± 0.2035	1.966	0.0493	0.0424
Sex	0.6065 ± 0.6769	0.896	0.3703	0.8300
Number of FH components * Sex	−0.2165 ± 0.2479	−0.873	0.3825	0.3842
Random effects: ${\sigma }_{\mathrm{genus}}^2$= 0.4399 n = 150 total observations from 47 genera
(d) Genus level minimum adequate model
Number of FH components	0.4936 0.2415	2.044	0.041	–
Collection (MZFC-CNIN)	0.3989 0.2035	1.960	0.050	–
Random effects: ${\sigma }_{\mathrm{genus}}^2$= 0.4404 n = 150 total observations from 47 genera

Notes.

^aParameters from model.

^bP-value resulting from χ² nested model comparisons following removal of each variable during backwards stepwise simplification.