Table 2. Effects of population and herbivore removal on damage from herbivores and fitness components in three Lythrum salicaria populations in the year following the experimental treatment (2014).

Flowering status (flowering vs. not flowering) was analysed with a generalized linear model (χ2 given); the other response variables were analysed with ANOVA (F-ratios given). The effect of herbivore removal was analysed separately by population with contrasts when a significant population × treatment interaction was detected. Significant effects are in bold.

					Forsmark		Skagsudden
Response variable	Source of variation	df	F / χ2	P	t	P	t	P
Proportion of leaf area removed	Population	1,331	451.64	<0.0001
	Herbivore removal	1,331	0.0003	0.99
	Pop × Herb rem	1,331	0.14	0.71
Early-season plant height	Population	1,356	46.64	<0.0001
	Herbivore removal	1,356	1.97	0.16
	Pop × Herb rem	1,356	1.05	0.31
Late-season plant height	Population	1,333	50.95	<0.0001
	Herbivore removal	1,333	2.01	0.16	-1.42	0.16	1.56	0.12
	Pop × Herb rem	1,333	4.43	0.04
Flowering status	Population	1,335	106.17	<0.0001
	Herbivore removal	1,334	0.11	0.74
	Pop × Herb rem	1,333	1.25	0.26
# fruits per plant	Population	1,333	104.20	<0.0001
	Herbivore removal	1,333	1.67	0.20	-1.29	0.20	2.35	0.02
	Pop × Herb rem	1,333	6.65	0.01