68 Rodolfo Dirzo and Karina Boegebecomes irrelevant from the dry forest perspec-
tive, given that during the dry season, when the
plants are not apparent, herbivores (the target of
the apparency signal) are not active either, imply-
in gthat abiotic control, rather than apparency,
may be a more important determinant of the risk
of herbivory. Although some elements of these
hypotheses can be applicable to our TRF/TDF
comparison, their application is only partial, as
we discuss below.
We propose, instead, a scenario of contrasting
responses of herbivory and defense based on rain-
fall seasonality as the startin gpoint leadin gto
a contrast in resource availability for herbivores
between TDFs and TRFs (Figure 5.1). Precipita-
tion, an important component in the physical
environment of bothTRFs andTDFs, operates as a
fundamentaldriverof contrastingplantphenolog-
ical responses, in particular foliage availability, the
resource base for leaf-feedin gherbivores. Folia ge
availability, as affected by rainfall seasonality, will
in turn lead to differences in herbivore risk of
attack/damage, which will determine differences
in the potential impact of herbivores on plants.
Under these conditions, differential selective pres-
sures for the evolution of defense and, ultimately,
patterns of herbivory, should be expected.The spe-
cific predictions regarding these variables for TRF
and TDF are as follows (Figure 5.1):
1 Foliage availability. Under the conditions of
stron gseasonality in TDFs folia ge availability to
herbivores should be episodic, restricted to 50%
of the time or less, while availability of foliage
should be continuous throughout the year in the
vast majority of plants in TRFs.
2 Ris kof attac k. Given the above contrast, and
assumin gno initially (i.e., back in evolutionary
time) marked differences in the levels of defense
between TDF and TRF plants, the risk of attack
should be restricted to a few months of the year
for TDF plants, and should be continuous (i.e., all
months of the year) for TRF plants.
3 Potential impact of herbivores. In addition, con-
trasts in rainfall seasonality will promote marked
differences in the lifespan of leaves in plants
from both types of forest, with consequences
for herbivory and plant fitness. The higher leaf
turnover rate and shorter leaf longevity of
dry forest plants (durin gthe rainy season) as
compared with those of the slow-growingrain for-
est species (Coley and Aide 1991) should allow
for a greater capacity to replace tissue loss, ren-
derin gthe fitness costs of herbivory lower in the
former.
4 Selective pressure for defense. The expected con-
trast in impact of herbivory should lead to differ-
ences in the selective pressure regime exerted by
herbivores. The lower impact of herbivory in TDF
plants as compared with TRF plants is expected to
operate as a selective pressure of relatively lower
intensityfortheformerandoneof higherintensity
for the latter.
5 Investment in defense. Accordingly, investment
in defense is predicted to be comparatively lower
in plants of seasonally dry tropical forests, as com-
pared with plants of aseasonal/less seasonal rain
forests.
6 Herbivory. Consequently, our final prediction
is for the level of herbivory to be comparatively
higher in plants of seasonally dry forests, whereas
plants of tropical rain forests are predicted to
sustain lower levels of herbivory.
The apparent contradiction of the fact that high
levels of herbivory do not promote the evolution
of high levels of defense in plants of TDF could be
explained, as we argued above, by the deployment
of attributes that confer some degree of toler-
ance to herbivory, in particular their higher leaf
turnover rates and shorter leaf lifespans. In con-
trast, the low levels of herbivory in TRF plants
do not favor the evolution of reduced defense
given the continuous risk of attack/damage and
the higher fitness costs of herbivory in this envi-
ronment. These contrasts are compatible with
the theoretical expectations of plants to evolve a
varietyof responsestoherbivory,rangingfromtol-
erance of varyin gde grees, to defense, or a given
combination thereof (Strauss and Agrawal 1999,
Stoweet al.2000, Fornoniet al.2003). Moreover,
our arguments do not imply that plants of theTDF
shouldbedevoidof defenses;weargue,rather,that
in comparison with TRF species, levels of defense
are likely to be lower and levels of tolerance are
predicted to be higher in TDF species. The abi-
otic control of herbivore populations (population
reductions durin gthe dry season) concomitant
with foliage availability should enhance the effect
of rainfall seasonality.