64 Rodolfo Dirzo and Karina Boege
Studies on the biology of tropical forests have a
strongbiasonTRFs,andherbivoryisnoexception.
For example, a review of the literature on patterns
of tropical herbivory shows that the number of
studies (published since 1970) dealin gwith TRF
plants is about three times greater than those on
TDF plants (Table 5.1). This imbalance compro-
mises our understandin gof both forest types. In
this study we attempt to broaden our current per-
spective on tropical plant–herbivore interactions
by developin ga comparative analysis of herbivory
in TRF versus TDF.
Herbivory is a central process in tropical for-
est biology. Phytophagous insects and their food
plants constitute a large proportion (at least 50%)
of the number of known species on earth (Price
1997) and both groups are disproportionately
represented in tropical forests (Dirzo and Raven
2003). On the other hand, although levels of her-
bivoryformosttropicalplantsarelowtomoderate,
on occasion, damage can approach complete defo-
liation both in TRF species (Dirzo and Mota 1997)
and in TDF species (Janzen 1981). In addition,
intense defoliation has been shown to affect sev-
eral components of plant fitness both in rain forest
species (e.g.,Piper arieianum, Marquis 1984) and
in dry forest species (e.g.,Erythroxylum havanense,
Dirzo and Domínguez 1995). Although the num-
ber of studies directly documentin gthe role of
herbivores as important selective pressures for
plants is very limited (see Marquis 1992), the
importance of herbivores as selective agents of
great preponderance in tropical forests has been
inferred from the over-representation of puta-
tive defensive compounds such as alkaloids (Levin
1976) when compared with extra-tropical plants.
Moreover, in a review specifically directed to assess
patterns of herbivory and defense in tropical ver-
sus temperate forest plants, Coley and Aide (1991)
conclude that herbivory and defense are signifi-
cantly greater in tropical forests than in temperate
forests.
While ecogeographic analyses of tropical her-
bivory have emphasized the tropical–temperate
comparison (Coley and Aide 1991), comparative
studies within the tropical realm are restricted
to interspecific comparisons within a given site
(Coley 1982, Dirzo 1984) or to limited citations of
observed levels of herbivory in TDFs as compared
with TRFs (Coley and Aide 1991, Dirzo and
Domínguez 1995, Coley and Barone 1996, Dyer
and Coley 2002). To our knowledge, no detailed
comparison of herbivory and defense between
TRF and TDF has been undertaken, and no spe-
cific theoretical frameworks for expected patterns
have been developed.THEORETICAL FRAMEWORKS
AND PREDICTIONS
Two specific theoretical constructs have been
developed to explain interspecific variation in her-
bivory (leadin gto differential defensive responses)
within a site: the resource availability hypothe-
sis (Coleyet al.1985, see also Janzen 1974) and
the plant apparency hypothesis (Feeny 1976). To
what extent can we use either or both of these
constructs to attempt to explain patterns of vari-
ation in herbivory and defense between TRF and
TDF plants? The first hypothesis posits that slow-
growing species, adapted to live in habitats of low
resource availability (e.g., shaded forest under-
story, poor-quality soils), in which the cost of
loss of tissue to herbivores is very high, should
have higher investments in defense than fast-
growing species adapted to persist in habitats of
high resource availability (i.e., forest gaps, more
fertile soils).This hypothesis is not directly applica-
ble for a comparison between tropical dry and wet
forests, given that, in addition to the differences in
the relevant limitin gresources between both types
of forest (light or soil nutrients in TRF and water
in TDF), it does not consider a central difference
between them: rainfall seasonality (see below).
The second hypothesis argues that the probabil-
ity of a plant bein gencountered by a herbivore
(apparency) determines the risk of herbivory and
the investment in defense: apparent plants (large,
lon glived, growin gin dense populations, etc.)
have a greater probability of being found and,
therefore, selection should favor greater invest-
ment in defense than in less apparent plants
(small,shortlived,growinginsparsepopulations).
This hypothesis could be challenged because of
the subjectivity of the concept of apparency and
the lack of a consistent fit with empirical data
(Stamp 2003). Moreover, the argued apparency