Statistical Analysis
All statistical analyses and graphing were performed in R software
(version 3.2.2) (R Core Team, 2015) . The number of replicates varied
depending on the measured parameters. Leaf anatomy study was performed
in 3-5 replicates. The remaining measurements were performed in 5
replicates. If the dependent factor was not normal or homogeneous, the
data were transformed before using a two-way ANOVA. Two-way ANOVA
followed by contrasts was selected to estimate mean differences (LL
plants subtracted from HL plants) for each genotype separately for
individual parameters. Genotypes were combined for the ANCOVA followed
by contrast analysis to determine if changes in leaf water oxygen
isotope ratios co-varied with changes in other physiological and
anatomical traits (in terms of mean differences of HL and LL growth
treatment). Contrast analysis was considered the appropriate statistical
approach to interpret mean differences according to personal
consultation provided by the statistical facility at WSU.
According to Péclet effect theory, the extent to that the Péclet effect
is occurring varies with leaf water pool, so the relative importance of
each Péclet effect is dependent on its strength and water pool size
(Farquhar & Gan 2003;
Holloway-Phillips et al. 2016).
The two Péclet models fitted to the f sw versusE data were the xylem and the mesophyll and xylem Péclet models
where the unknowns were L v andϕ x in the xylem Péclet model andL m, L v, andϕ x for the mesophyll and xylem Péclet model. The
knowns were f sw (1 -
Δ18OLW/
Δ18Oe) and E . Model fitting was
done with the Levenberg-Marquardt Nonlinear Least-Squares Algorithm in R
using the minpack.lm package.