Table 4. Best subset models to predict hatch delay for great tits Parus major (GT) and blue tits Cyanistes caeruleus (BT) at Cambridge University Botanical Garden (CUBG), Cow Lane Nature Reserve (CL) and Brampton Wood Nature Reserve (BW) from mean weekly maximum daily temperature (T_max; ° C), minimum daily temperature (T_min; ° C), and rainfall (rain; mm).

Sp	Site	R2	M. Cp	S	E			C			I
					Tmax	Tmin	rain	Tmax	Tmin	rain	Tmax	Tmin	rain
GT	CUBG	15	0.5	1.24		X		X				X
	CL	32	1.9	1.33		X				X		X
	BW	72	4.4	1.54	X			X					X
BT	CUBG	48	1.6	1.04	X				X	X
	CL	25	7.5	1.96	X						X	X
	BW	84	1.5	1.44	X			X		X

The pertinent weeks were: i) the week before clutch completion, i.e. during egg laying (E); ii) the week of clutch completion when full incubation should start with the laying of the last egg (C); iii) the week after clutch completion which typically would be the second week of incubation (I). The best models were obtained with 3 of the 9 possible predictor variables; models with the lowest Mallows Cp (M. Cp) values and standard error of regression (S) were selected. For all models, hatch delay was negatively related to temperature and positively related to rainfall.