Table 1

Selection gradients for all traits and all years on Kidd cay.

W (2003) 33.8% survival

β/γ

SE

χ2

P-Value


Hindlimb

0.09

0.14

0.57

0.45

Forelimb

0.03

0.14

0.06

0.79

Hindlimb2

-0.34

0.22

3.19

0.07

Forelimb2

-0.30

0.26

1.96

0.16

Hindlimb × Forelimb

0.293

0.16

4.14

0.04


W (2003) Kidd cay


SVL

0.03

0.17

0.39

0.53

SVL2

0.16

0.20

0.57

0.45


W (2004 Animals replaced) 32% survival


Hindlimb

-0.13

0.15

0.77

0.37

Forelimb

-0.13

0.15

0.66

0.42

Hindlimb2

0.14

0.20

0.49

0.48

Forelimb2

0.24

0.26

1.07

0.30

Hindlimb × Forelimb

0.01

0.17

0.00

0.96


W (2004)


SVL

0.05

0.15

0.11

0.74

SVL2

0.60

0.22

6.88

0.009


W (2005) 45% survival


Hindlimb

0.04

0.12

0.04

0.84

Forelimb

-0.03

0.12

0.18

0.67

Hindlimb2

-0.006

0.16

0.00

0.96

Forelimb2

0.38

0.18

4.39

0.04

Hindlimb × Forelimb

-0.13

0.18

0.52

0.47

W (2005)

β/γ

SE

χ2

P-Value

SVL

-0.23

0.11

4.29

0.04

SVL2

-0.12

0.16

0.63

0.43


W (2006) 42% survival


Hindlimb

-0.03

0.10

0.00

0.98

Forelimb

0.13

0.10

0.44

0.50

Hindlimb2

0.12

0.12

0.80

0.37

Forelimb2

-0.10

0.14

0.75

0.38

Hindlimb × Forelimb

-0.21

0.09

5.64

0.02


W (2006)


SVL

0.09

0.13

0.46

0.50

SVL2

0.02

0.18

0.02

0.88


Data distributions were standardized (by year) to mean zero and unit variance. Hind and forelimb lengths refer to residual values (regression of log-transformed limb and snout-vent-lengths. W represents survival (0/1). Linear (β) and non-linear (γ) selection gradients come from general linear models, but significance tests were computed from logistic regressions. Quadratic gradients and associated standard errors were doubled [34].

Calsbeek et al. BMC Evolutionary Biology 2009 9:3   doi:10.1186/1471-2148-9-3

Open Data