74 Rodolfo Dirzo and Karina Boege
data considerin gdifferences in phenolo gy and
growth rate in TDF and TRF plants, respectively
(Figure 5.3, lower panel, center), detected a signif-
icant contrast only between slow-growing versus
fast-growing species – an expected result, but
of secondary interest to our hypothesis. Thus,
while total phenolics supported the original pre-
dictions (see Figure 5.1), condensed tannins did
not. However, in addition to the limitations of
the dataset we used for these comparisons, it
is necessary to bear in mind that these two
groups of compounds, assayed on such a variety
of species, may not necessarily reflect defensive
responses, given the variety of secondary metabo-
lites known to be important as anti-herbivore
mechanisms(RosenthalandJanzen1979).Again,
this is an aspect that warrants further investiga-
tion involvin gadditional sites and other defensive
compounds.
Leaf toughness, in contrast, exhibited a very
consistent difference between TDF and TRF
species. The overall difference was highly signif-
icant (Figure 5.3, upper panel, right): rain for-
est species were approximately six times tougher
than dry forest species, and this difference arises
despite the fact that plants from both forest types
in this comparison included different phenologies
and growth rates. Therefore, when data are dis-
aggregated, among-group differences are highly
significant (Kruskal–Wallisχ^2 =69.9, d.f.=3,
P=0.0001) and the contrast between decidu-
ous plants from TDF and slow-growing species
from TRF is even more marked – a 10-fold dif-
ference (Mann–Whitney’sZ=−6.41, d.f.=57,
P<0.0001; Figure 5.3, lower panel, right). Fur-
thermore, evergreen species from dry forest were
about two-fold tougher than their deciduous
counterparts. Regarding the comparison between
slow-growing and fast-growing species from TRF
(Figure 5.3, lower panel, right) we detected that
slow-growth species were 1.75 times as tough as
rapid-growth species.
In all three cases of comparison of defensive
characteristics between the two groups of TRF
plants (i.e., slow- and fast-growth species), dif-
ferences were highly significant, indicating that
defensive attributes were significantly greater in
slow-growing species, as would be expected from
previous studies and the resource availability
hypothesis (Janzen and Waterman 1984, Coley
et al. 1985).
Despite our consistent findings, it is important
to take into consideration that the three attributes
we used to infer defense may not necessarily
reflect ultimate evolutionary responses resulting
from previous selective pressure exerted by herbi-
vores on plants. The importance of the distinction
between proximal and ultimate factors responsible
for currently observed characteristics in plant–
animal interactions has been discussed in other
studies (Dirzo 1984, Farrell and Mitter 1993) and
warrants further work in the context of our water
availability/phenology hypothesis.PHYLOGENETIC INERTIA
Our results have shown that plants from TRF sus-
tainsignificantlylowerlevelsof herbivoryandpos-
sess attributes that confer, at least at the proximal
level, greater defense. Moreover, results from the
intra-site comparisons in the case of TDFs, look-
in gat deciduous versus ever green species, support
such an argument.The latter comparison, in addi-
tion to highlighting the internal heterogeneity
of TDF plants, provided a useful comparison by
exertin gsome control of environmental,extrinsic
factors. However, otherintrinsicfactors that might
be responsible for the observed patterns may need
to be taken into account. In particular, recent
studies argue for the importance of controlling
for phylogenetic inertia (see Armbruster 1992):
observed ecological patterns may be, at least
partly, the result of differences due to phylogeny.
Unfortunately, our dataset is very limited in allow-
in gassessment of the importance of phylo geny in
our comparisons. Nevertheless, our dataset allows
for five intra-generic comparisons of herbivory
between related species from TDF (deciduous
species)andTRF(slow-growthspecies)(Lonchocar-
pus,Cordia,Erythroxylum,Trichilia, andPiper) and
three intra-family comparisons (two species of
each of these families: Sapindaceae, Apocynaceae,
Euphorbiaceae). If all eight pairs of taxa are used
(Figure 5.4), we observe that in six of the com-
parisons herbivory is greater in plants from TDF,
as compared with their TRF relatives, while in one