Table 3. Effect of food availability, season and river catchment on grey-sided vole selectivity.

Predictors	Responses
	Polygonaceae (n = 51)		Cornacea (n = 61)		Ericaceae (n = 75)		Poaceae (n = 76)		Salicaceae (n = 58)		forbs (n = 77)
	Est.	95% CI	Est.	95% CI	Est.	95% CI	Est.	95% CI	Est.	95% CI	Est.	95% CI
fixed effects
intercept	8.34	5.78, 10.90	3.91	0.22, 7.71	7.41	6.66, 8.15	−2.23	−5.54, 1.37	4.14	−1.61, 12.70	4.86	0.66, 9.32
Polygonaceae	−4.93	−14.28, 4.42
Cornaceae			0.43	−0.39, 1.38
Poaceae	0.23	0.02, 0.45	−0.29	−0.73, 0.16	−0.02	−0.09, 0.04	0.03	−0.27, 0.36	−0.08	−0.43, 0.20	0.03	−0.15, 0.23
Salicaceae			0.33	−0.07, 0.79	−0.03	−0.10, 0.04	0.16	−0.17, 0.48	−0.23	−0.54, 0.04	0.05	−0.16, 0.24
deidious Vaccinium	−0.01	−0.06, 0.04	0.03	−0.07, 0.13	0.004	−0.009, 0.02	0.01	−0.05, 0.08	−0.01	−0.06, 0.06	0.003	−0.03, 0.04
forbs	−0.58	−1.01, −0.16	−1.02	−2.57, 0.34	−0.04	−0.17, 0.09	−0.03	−0.66, 0.57	0.03	−0.44, 0.74	0.11	−0.28, 0.51
season (summer)	2.39	−0.82, 5.60	1.25	−3.10, 5.42	−1.12	−2.0, −0.22	3.22	−0.22, 7.32	1.93	−0.88, 6.68	0.63	−1.90, 2.89
river catchment (VJ)	0.16	−3.78, 4.09	−3.00	−10.33, 3.57	0.45	−0.60, 1.51	5.94	0.83, 11.00	−2.18	−6.79, 3.05	−0.88	−3.95, 2.19
random effects
grid	NA	(n = 17)	3.97	(n = 15, 24%)	NA	(n = 19)	3.30	(n = 19, 18%)	4.39	(n = 19, 40%)	1.50	(n = 19, 9%)
residual	4.62		7.12		1.68		6.98		5.45		4.58

Parameter estimates of linear mixed effect models for the effect of food availability (biomass g/m), season and river catchment on grey-sided vole selectivity, i.e. difference between stomach content and biomass proportions. Intercept is calculated for autumn, Komagelva (KO) and mean biomass of all continuous predictor variables. “Est.” refers to regression coefficients, measured at logit-scale. Random effects are presented as standard deviations, sample size (n) referring to the number of grids included in the analysis, % values to the percentage of residual variance assigned to grid. Estimates with bold indicate that 95% CI does not include 0, with italics that 95% CI includes zero at most 0.05 and effect size is >0.15. Models where data were insufficient to evaluate the random effect (NA), have been calculated as linear regressions with fixed effects only. “Forbs” as predictor variable represents availability of the functional group of forbs, except for models which have a forb family (Polygonaceae, Cornaceae) as response variable. For these, biomass of the respective family is excluded from that of forbs and used as a separate predictor. Empty cells indicate that predictor variable in question has not been included in the model. See Material and Methods for details.