Statistical analysis
The cavitation resistance of petioles, leaves, petals, and peduncles (all organs were measured on each individual) was compared using two-way ANOVA to test the effect of organ type and individual on P50values in R. A log transformation was applied to fulfil assumptions of homoscedasticity and normality. Post hoc tests were computed using Tukey’s HSD.
The effect of disabling phloem water delivery by heat girdling on fruit expansion was tested using a two-way repeat measures ANOVA. The interaction term was determined to be non-significant and was removed from the model. Assumptions were tested for normality and no transformations were applied. A qq plot could not be fitted given the small data set, however, the consistency of results indicated that model assumptions could be met.
The cavitation resistance (quantified as the P50 and P88 ) and mean diameter of the largest 5 % of vessels in peduncles and stems (measured on separate individuals) was compared using a student’s T-test.

Results: