Table 3.

Test of differences among species, between annuals and biennials (life form), and between prostrate and upright annual species (growth form) in the attributes of adventitious sprouting

Response variable	Species	Life form	Growth form
No. of adventitious buds and shoots	390·6 (<0·001)	0·5 (n.s.)	12 202 (<0·001)
			12 187 (<0·001)
			smaller for upright
No. of adventitious shoots/no. of adventitious buds and shoots	36·5 (<0·001)	1·9 (n.s.)	1905 (<0·001)
			1900 (<0·001)
			smaller for upright
Presence of adventitious buds or shoots	403·2 (<0·001)	0·00 (n.s.)	3290 (<0·001)
			3290 (<0·001)
			probability smaller for upright
Cumulative length of adventitious shoots	34·6 (<0·001)	7·1 (0·038)	827 (<0·001)
		1·8 (n.s.)	825 (<0·001)
		larger for biennials	longer for upright
Length of the longest adventitious shoot	45·6 (<0·001)	9·9 (0·010)	762 (<0·001)
		1·2 (n.s.)	760 (<0·001)
		larger for biennials	longer for upright

Effect of individual predictors (columns) was examined in a separate model for each response variable (rows). The first row for each variable provides the χ² statistic and corresponding Type I error estimate for a likelihood-ratio test of the particular model term; the next row provides results from the corresponding model with phylogenetic correction (fitted only for previously significant effects, except for ‘Species’, where the correction was not appropriate). If any of the two models found a significant effect, then below is described the direction of the effect (for ‘Life form’ and ‘Growth form’ predictors).