Table 2 Variance components for local dispersal distance and exploration rate in great tits.

(b) Bivariate model
	N	Variance components					h ²
		Additive genetic	Natal brood	Birth plot	Birth year	Residual
Dispersal distance	1,230	0.022 (0.0091)	0.012 (0.0080)	0.0021 (0.0020)	0.0016 (0.0014)	0.12 (0.010)	0.14 (0.057)
Exploration rate	806	20.12 (8.33)	11.94 (7.25)	0.26 (0.68)	0.56 (0.72)	39.88 (8.18)	0.28 (0.11)
		Components of covariance					r _G	P -value
Dispersal and exploration rate	Distance	0.66** (0.24)	−0.14 (0.23)	0	0	−0.29 (0.29)	0.99 (0.40)	0.0062

Univariate (a) and bivariate (b) animal models were used to estimate the variance components of dispersal distance and exploration rate, from which narrow-sense heritability (h²) estimates were calculated. Sex was included as a fixed effect on dispersal distance to account for sex differences in dispersal distance (see Fig. 3). The covariances and the genetic correlation (r_G) between dispersal distance and exploration rate were estimated with the bivariate animal model (b). Birth plot and birth year effects explained negligibly small fractions of the variance in both dispersal distance and exploration rate (also see Fig. 4), and therefore we fixed the covariances for the birth plot and birth year effects at 0. A model actually estimating these covariances gave similar results (Supplementary Table S8). S.e. are in parentheses. All analyses and reported estimates are based on ¹⁰log(x+10) transformed dispersal distances (in metres) and residual exploration rates from a linear regression on test date (from 1 July; see Methods). Using untransformed dispersal distances gave similar results (Supplementary Table S6). The significance of model terms (except for the residual variance terms) was determined using log-likelihood-ratio tests. For the bivariate model, this test allowed for assessing the significance of the covariances only. **P<0.01 and ^#P<0.1.

Search