Table 1.

Multiple effects, their interactions and fit statistics for each model examined for best explaining the variation in the proportion of sea lamprey that moved upstream across treatments.

effect(s)	interaction(s)	logLik	AIC	Δi AIC	$\exp (-0.5\times {\mathrm{\Delta }}_i)$	wi
odourant; temp	odourant×temp	−25.3	−21.3	0.00	1.00	0.22a
temp		−12.6	−6.6	14.70	0.48	0.10
	odourant×period	−4.8	1.2	22.50	0.32	0.07
	temp×period	−6.8	−0.8	20.50	0.36	0.08
odourant; temp; period	odourant×temp; odourant×period; odourant×temp×period	−5.8	−1.8	19.50	0.38	0.08
temp; period	temp×period	−5.8	0.2	21.50	0.34	0.07
	odourant×temp	−7.9	−3.9	17.40	0.42	0.09
odourant; period	odourant×period	−4.8	1.2	22.50	0.32	0.07
period		−1.6	4.4	25.70	0.28	0.06
odourant		−12.6	−6.6	14.70	0.48	0.10
date		3.7	7.7	29.00	0.23	0.05

^aFinal model selected with corresponding AIC and logLik values. Δ_i AIC and Akaike weights (w_i) were calculated using equations (3.1) and (3.2). Relative likelihoods were calculated with the following equation: $\exp (-0.5\times {\mathrm{\Delta }}_i)$. Odourant refers to 3kPZS, methanol (control) or larval extract treatments. Period refers to early (20.00–23.00 h) or late (00.00–03.00 h) trials. Temp refers to stream temperature (°C).