Table 3.

ANOVA results (Type III Sum of Squares) for effects of topsoil removal (r), hay transfer (h), and their interaction effect (r × h) on N concentration and δ¹⁵N in plant leaves.

	Factor	d.f.	%N		δ15N
			F	p	F	p
(A) ALL SPECIES
	r	1	65.025	<0.001	194.035	<0.001
	h	1	61.378	<0.001	17.477	<0.001
	r × h	1	15.664	<0.001	0.018	0.892
(B)FUNCTIONAL IDENTITY
non-legumes	r	1	34.737	<0.001	783.733	<0.001
	h	1	57.936	<0.001	77.090	<0.001
	r × h	1	25.901	<0.001	0.110	0.740
legumes	r	1	57.228	<0.001	12.082	0.001
	h	1	29.693	<0.001	4.028	0.047
	r × h	1	0.631	0.428	0.113	0.737
(C) SPECIES IDENTITY
ant	r	1	16.379	<0.001	0.001	0.976
	h	1	8.646	0.005	3.320	0.076
	r × h	1	0.434	0.514	2.746	0.105
dor	r	0
	h	1	0.391	0.541	0.383	0.544
	r × h	0
hel	r	1	4.501	0.037	74.492	<0.001
	h	1	15.994	<0.001	32.447	<0.001
	r × h	0
lot	r	1	2.192	0.146	1.944	0.170
	h	1	0.800	0.376	2.974	0.092
	r × h	0
tri	only in −r−h	na
gal	r	1	40.278	<0.001	203.595	<0.001
	h	1	49.685	<0.001	11.124	0.001
	r × h	0
roots	r	1	0.679	0.418	42.869	<0.001
	h	1	1.860	0.185	1.366	0.254
	r × h	1	1.457	0.239	0.476	0.497

Results show (a) all data analyzed together, (b) data split into plant functional identity (legumes versus non-legumes) and then tested for restoration treatment effects and (c) data analyzed per species and tested for effects of restoration treatments. Significant effects are shown in bold. Please note that this is not a split plot design, but that we use the table to present separate analyses at different hierarchical levels.

Target species: ant = Anthyllis vulneraria, dor = Dorycnium germanicum, hel = Helianthemum nummularium and mesic species: lot = Lotus corniculatus, tri = Trifolium pratense (not tested because it only occurred in one treatment), gal = Galium mollugo, roots = unspecific combined root samples.