Table 3.

Percentage of missing data at which the P value of the correlation between the ranking of the communities calculated without missing data and with imputed data became not significant for five of the 100 sets of communities, using the MICE, the dissimilarity or the Relationships imputation methods, or the deletion of species with one missing trait value

	Methods of imputation, resp. deletion
	MICE	Dissimilarity	Relationships	Deletion
Functional identity
H	\	45	\	11
LDMC	\	43	\	6
SM	14	31	15	10
SLA	\	42	\	7
Functional range
H	39	40	45	14
LDMC	\	\	\	33
SM	17	32	12	7
SLA	\	\	\	23
Functional divergence
H	\	25	33	10
LDMC	\	31	37	8
SM	\	40	5	7
SLA	\	32	37	10

H, vegetative height; LDMC, leaf dry matter content; MICE, Multivariate Imputation by Chained Equations; SM, seed mass; SLA, specific leaf area.