Table 2.

Relationships between chick age, hatching rank, heterozygosity and growth in body weight, wing and tarsus length

Explained variable	Hz index used	A*H*R	AIC	A*H in A-chicks	A*H in B-chicks
Body weight	H	<0.0001	3267.5	0.68	<0.0001
	SH	<0.0001	3271.7	0.44	<0.0001
	IR	<0.0001	3273.3	0.86	<0.0001
	H'	0.020	3284.8	0.12	0.0020
	SH'	0.046	3285.9	0.094	0.0038
	IR'	0.064	3289.2	0.12	0.0055
Wing length	H	<0.0001	2612.3	0.58	<0.0001
	SH	0.0001	2615.0	0.91	<0.0001
	IR	<0.0001	2615.4	0.39	<0.0001
	H'	0.013	2621.7	0.31	0.0007
	SH'	0.037	2621.6	0.15	0.0013
	IR'	0.038	2626.8	0.41	0.0033
Tarsus length	H	0.0017	2758.8	0.71	0.0025
	IR	0.0013	2760.4	0.95	0.011
	SH	0.012	2762.2	0.28	0.0045
	H'	0.027	2765.2	0.75	0.04
	IR'	0.014	2767.2	0.20	0.054
	SH'	0.089	2766.3	0.97	0.089

We used AIC for model selection. Parameters: A (Age), H (Heterozygosity, estimated by the index in the column "Hz index used") and R (hatching Rank, a binary effect). For heterozygosity indices: H, SH and IR are calculated over all loci, while H', SH' and IR' are calculated without OHW loci and K31. With either heterozygosity index, the structure of the selected model was: A+H+R+AH+HR+AHR+A² plus the random effect of chick identity. Since the interaction A*H*R was significant, we also tested the significance of the A*H effect in each rank (A or B-chicks). P-values below 0.05 are in bold. (Note: a more complete version, with the mean effect size for heterozygosity for each analysis, is given as additional file 1).