Males of a sexually cannibalistic spider chemically assess relative female quality

Table 1 Summary of model selection analysis to reveal influential effects on the cannibalism rate

Model variables	K	AICc	ΔAICc	weight
♀ origin	2	74.1	0.00	0.366
♀origin + ♀ size class	4	75.1	0.94	0.228
♀ size classes	3	75.2	1.02	0.219
Null model	1	75.7	1.60	0.164
♀ origin + ♀ size class + ♀ origin:♀ size class	7	82.3	5.57	0.023

For each model, we used a binary logistic regression (N = 54) and listed them in descending order of their Akaike information criterion corrected for small samples (AICc). The model with the lowest AICc value was generally the most plausible model. We suggested other models as equally plausible if their AICc value differentiated less than 2 units from the most plausible model (grey background). From these models, we selected the model with the fewest parameters (K) (bold letters).ΔAICc = AICc difference to the most plausible model; weights = AICc weights

ISSN: 2730-7182