200 Stefan A. Schnitzeretal.Figure 12.2 Lianas in the understory on Barro
Colorado Island, Panama. Photograph by S. Schnitzer
(2005).
species competes optimally at a particular combi-
nation of resources, then within-ga presource gra-
dients may allow species to coexist in equilibrium
over the landscape, provided that species are not
seed limited and can disperse their seeds to newly
formed gaps (Dallinget al. 1998, Hubbellet al.
1999, Muller-Landau Chapter 11, this volume).
Species may also specialize along a gradient of gap
sizes, with some species specializing on small gaps
and others on large gaps (Denslow 1980, Orians
1982, Brokaw 1987). Under this latter scenario,
gaps promote species’ coexistence by providing
a heterogeneous environment at the landscape
scale, with gaps of different sizes providing vari-
ous levels of resources on which different species
specialize. In both cases, the resource niche view
(reviewed by Chase and Leibold 2003, Kitajima
andPoorterChapter10, thisvolume)requiresthat
gaps create sufficiently large resource gradients,
either within the ga por among ga ps, to allow
species to stably coexist at equilibrium (Ricklefs
1977).
These three proposed mechanisms by which
gaps are proposed to maintain diversity provide
the following testable predictions: (1) resources
will be measurably more heterogeneous within
gaps (or among gaps of different sizes and char-
acteristics) than in a comparable area within the
intact forest; (2) some species will require the
increase in resource quantity or heterogeneity
from canopy gaps to establish and survive(e.g., Dallinget al. 2001); (3) some species will
require the enhanced resources available in a gap
to initiate reproduction; for example, understory
trees, shrubs, and herbs that may be able to estab-
lish in the absence of a ga pbut fail to re produce in
the shaded understory; and (4) individuals within
gaps (or near the gap edge) will have substantially
higher fecundity or a larger proportion of them
will reach reproductive maturity on a per area or
per stem basis than those individuals in the intact
forest. Thus, even if thinning reduces diversity as
gaps close, gaps may maintain diversity if they are
sites of high fecundity for many species (Schnitzer
2001). These second two predictions of the gap
hypothesis remain little studied.EMPIRICAL TESTS: IS THERE
EVIDENCE TO SUPPORT THE GAP
HYPOTHESIS?
There is evidence that gaps maintain a significant
level of tropical plant species diversity, particularly
for some plant groups. For example, pioneer tree
species require gaps for colonization and regener-
ation and are almost always absent in the intact
shaded understory (e.g., Brokaw 1985a, 1987,
Clarket al.1993, Whitmore and Brown 1996,
Schnitzer and Carson 2000, 2001). Thus, the
cyclical and predictable disturbance from treefall
gaps is necessary for pioneer trees to remain
in the community. Under the idealized succes-
sional pathway in tropical forests, pioneer trees
recruit into gaps soon after gap formation and
are later replaced by shade-tolerant species. Pio-
neer tree diversity may be maintained by resource
partitioning if species are uniquely adapted to
resources in different zones of a single ga por
among gaps of different sizes. For example, in a
rainforest at Los Tuxtlas in Mexico, Popmaet al.
(1988) reported that many pioneer species had
clear preferences for regeneration in either the
ga pcenter or edge, but not both. In a Panama-
nian moist forest, Brokaw (1987) reported that
Trema micrantha,Miconia argentea, andCecropia
insignisall specialized in gaps of different sizes
(see also Barton 1989, Van der Meeret al. 1998,
Pearsonet al. 2003b,c). While gap-size partition-
ing may occur when gaps differ greatly in size