light availability and liana infestation between the last (4 years
post-treatment) and the first (pre-treatment) measurements. A
positive change in crown position denotes an improvement in light
availability over time, while a negative change in liana infestation
denotes a decrease in liana infestation over time. To determine the
effect of treatment on the change in light availability and liana
infestation, we used a Kruskal–Wallis test, followed by Mann–
Whitney tests of differences among treatments.
To determine the effect of treatment and functional group on
FCT growth rates, a two-way ANOVA was performed with
treatment and functional group as factors, and DBH at plot
installation as covariate. Additionally, to determine to what extent
variation in diameter growth rates can be attributed to individual
tree characteristics, a two-way ANOVA was performed using tree
characteristics (liana infestation and light availability, both 1 year
after logging as factors, and DBH as covariate. We used the tree
characteristics measured 1 year after logging in the ANCOVA
because this measurement was more highly correlated with
growth rate than the other measurements taken (data not shown).
The direct effects of silvicultural treatments on tree diameter
growth were analyzed using ANOVAs, with girdling (naturally free,
overtopped, liberated through logging, and liberated through
girdling) and liana cutting treatments (naturally free of lianas,
liberated from lianas, and infested with lianas) as factors. In
addition, to assess the combined effect of the two FCT liberation
treatments, all 12 possible combinations between the two
treatments were grouped in 5 categories: ‘‘no treatment’’,
including FCTs that needed one or both treatments but did not
receive either of them; ‘‘naturally free’’ were FCTs growing
naturally without lianas and without overtopping competitors;
‘‘both treatments’’ included FCT that were liberated actively both
from lianas and from overtopping trees either by girdling or
harvesting; ‘‘liberated through girdling/harvesting’’ included FCT
that were only liberated from overtopping competitors through
girdling or harvesting; and ‘‘liberated from lianas’’ included FCT
that were liberated from lianas only. This new grouping was
chosen to avoid having none or few individuals in some of the 12
possible combinations of the two treatments. To test if there were
differences among categories of the combined treatments in terms
of growth rates, an ANOVA was used.