Skip to main content

Table 2 Model selection statistics for the pronotum width of male offspring from females of Onthophagus taurus

From: Maternal effects on male weaponry: female dung beetles produce major sons with longer horns when they perceive higher population density

Model

Pairwise sequential model comparisons

Fixed effects

df

log Likelihood

Likelihood ratio

P-value

Null

5

- 114.19

  

Brood mass weight

6

- 57.72

112.93

< 0.0001

Brood mass weight + M group

7

- 56.10

3.24

0.07

Brood mass weight + M group + PM group

8

- 56.18

0.16

0.69

Brood mass weight + M group + PM group + Brood mass weight : PM group

9

- 54.36

3.65

0.06

Brood mass weight + M group + PM group +

10

- 53.14

2.43

0.12

Brood mass weight : PM group +

    

Brood mass weight : M group

    

Brood mass weight + M group + PM group + Brood mass weight : PM group +

11

- 53.13

0.03

0.87

Brood mass weight : M group +

    

PM group : M group

    

Brood mass weight + M group + PM group +

12

- 53.08

0.09

0.76

Brood mass weight : PM group +

    

Brood mass weight : M group +

    

PM group : M group +

    

Brood mass weight : PM group : M group

    
  1. Females were assigned to experimental groups that differed in population density during the pre-mating period (PM), and in the number of possible mates during the mating period (M). The effect of brood mass weight was added as a covariate, the replicate mating chamber was added as a random effect, and female identity (nested within the replicate mating chamber) was also added as a random effect in all models. The random effects in the full model were 0.069 standard deviation for the replicate mating chamber and 0.115 standard deviation for female identity. The variance was modeled as a power function of the fitted values with an estimated parameter value of - 4.480 in the full model. Likelihood ratios were calculated as the absolute difference between the - 2 x log Likelihood of the two models being compared, and each model is being compared to the model that is one row above. Comparing these models in terms of their AIC values returns qualitatively similar results.