Table 1.

Summary of model selection results for factors affecting the onset of breeding for song sparrows on Mandarte Island, British Columbia. (Ln(L) is the value of the maximized log-likelihood function, AICc is Akaike's information criterion for small samples, ΔAICc is the scaled AICc relative to the top model, k is the number of model parameters and w_i are the Akaike weights that provide a measure of the relative support for each model. Age, age class (equal to or more than 1 year old); SOI, Southern Oscillation Index. All climate variables are values from January to April (average monthly SOI, average daily temperature and total precipitation). For the examination of fixed effects, nine candidate models were considered but only the top five are shown here. Random effects models were fit with restricted maximum likelihood (REML). ‘top fixed’ refers to model age+female density+SOI.)

model	ln(L)	AICc	ΔAICc	k	wi
(a) fixed effects only
age+female density+SOI	−3625.70	7259.45	0.00	4	1.00
female density+SOI	−3655.89	7317.81	58.35	3	0.00
age+SOI	−3665.92	7337.86	78.41	3	0.00
age+female density+temperature	−3678.92	7365.98	106.53	4	0.00
age+female density+precipitation	−3741.42	7490.88	231.43	4	0.00
(b) top fixed-effects model with random effects
top fixed+female in cohort×SOI	−3573.56	7163.27	0.00	8	1.00
top fixed+cohort×SOI	−3590.81	7195.74	32.47	7	0.00
top fixed+female in cohort	−3594.56	7201.21	37.94	6	0.00
top fixed+cohort	−3606.19	7222.44	59.17	5	0.00
top fixed	−3625.70	7259.45	96.18	4	0.00