Table 1.

Analyses of covariance testing for the effect of (A) plant height (log-transformed) and (B) identity (typica vs. elongata) on pitcher length for lower pitchers (R² = 0·81, residuals normally distributed: Shapiro statistic W = 0·99, P = 0·78) and upper pitchers (R² = 0·79, residuals normally distributed: W = 0·98, P = 0·40)

(A) Plant height
	Length of lower pitcher				Length of upper pitcher
Covariate	d.f.	SS	F	P	d.f.	SS	F	P
Log(plant height)	1	1234	273·4	<0·0001	1	34·91	4·94	0·0292
Type	1	10·4	2·3	0·1333	1	15	2·12	0·1491
Log(plant height) × type	1	29·8	6·6	0·0121	1	87·13	12·34	0·0008

(B) identity (typica vs. elongata)
	Length of lower pitcher				Length of upper pitcher
Parameter	Estimate	s.e.	T	P	Estimate	s.e.	T	P
Common intercept	−3·24	0·86	−3·8	0·0003	15·54	2·59	5·61	<0·0001
Slope for N. rafflesiana var. typica	4·74	0·33	14·31	<0·0001	0·01	0·51	0·02	0·9834
Slope for N. rafflesiana var. elongata	5·42	0·31	17·69	<0·0001	2·16	0·54	3·97	<0·0001