Skip to main content

Table 2 Results from regressing ln (relative forelimb length) on maximum-, median latitude, and temperature

From: An interspecific assessment of Bergmann’s rule in 22 mammalian families

Family

Predictor

n

Phylogentic half-life

Stationary variance

r 2

Optimal regression slope

AICc

AICc-θAICc

 

 

168.10

38.64

Emballonuridae

median latitude

34

160.10

0.004

9.693

41.09

2.45

max latitude

160.10

0.025

−18.349

40.36

1.71

 

temperature

 

160.10

0.015

−88.544

40.70

2.05

Median latitude - temperature regression: P-value = 0.0257; Adjusted R-squared = 0.1194; estimate = −1.9120

 

 

0.01

0.25

66.48

Molossidae

median latitude

40

0.01

0.25

0.020

0.008

68.13

1.65

max latitude

0.01

0.25

0.082

0.014

65.52

−0.96

 

temperature

 

0.01

0.25

0.048

−0.039

66.95

0.47

Median latitude - temperature regression: P-value = 1.31e–08; Adjusted R-squared = 0.5661; estimate = −2.5311

 

 

0.73

0.43

141.60

Phyllostomidae

median latitude

97

0.66

0.40

0.015

0.021

142.31

0.71

max latitude

0.86

0.45

0.031

0.033

140.81

−0.79

 

temperature

 

0.66

0.40

0.001

−0.017

143.72

2.12

Median latitude - temperature regression: P-value = 3.73e–08; Adjusted R-squared = 0.2665; estimate = −2.1620

 

 

0.13

0.50

93.80

Pteropodidae

median latitude

52

0.13

0.45

0.068

−0.012

95.95

2.15

max latitude

0.15

0.40

0.163

−0.025

90.80

−3.00

 

temperature

 

0.13

0.45

0.074

0.100

95.72

1.92

Median latitude - temperature regression: P-value = 3.04e–05; Adjusted R-squared = 0.2821; estimate = −2.4749

 

 

0.10

0.22

105.53

Vespertilionidae

median latitude

88

0.06

0.18

0.100

−0.004

105.80

0.27

max latitude

0.08

0.20

0.070

0.001

106.63

1.11

 

temperature

 

0.06

0.18

10.11

0.010

105.69

0.17

Median latitude - temperature regression: P-value = <2e–16; Adjusted R-squared = 0.8224; estimate = −1.9978

  1. Phylogenetic half-life estimates indicate the level of phylogenetic dependency in models where no predictors were included. In models with predictors, half-life gives the time (in lengths of the phylogeny) necessary to lose half the influence of the ancestral trait, i.e. rate of adaptation. All phylogenies are scaled to a total length of 1. Stationary variance gives the residual variance when the model has reaches a stochastic equilibrium, and r 2 gives the amount of variance explained by the optimal regression model. Optimal regression slope is the slope for which SLOUCH has removed the effect of phylogenetic inertia (slope expected under instant adaptation). AICc values are compared to the intercept-only model (AICc-θAICc) where larger negative values indicate the most improvement from the model without predictors. Also included are results from regressing median latitude on temperature for each family