habitats (Gascon 1993; Kotze and Samways 2001). Others, including many
insect species, respond rapidly to edge-related factors. In the Amazon, insect
abundance and diversity usually increase in the understory near edges, proba-
bly because of increased understory plant density and productivity (Fowler et
al. 1993; Malcolm 1997b). However, insect abundance is lower in the upper
forest strata, presumably because the density of overstory vegetation is reduced
by recurring canopy disturbances (Malcolm 1997b). Similarly, species adapted
for humid, dark forest interiors, including certain beetles (Didham et al.
1998), ants (Carvalho and Vasconcelos 1999), and butterflies (Brown and
Hutchings 1997), decline in abundance near edges.
Birds that forage in treefall gaps, such as some arboreal insectivores, hum-
mingbirds, and habitat generalists, often become abundant near edges (Stouf-
fer and Bierregaard 1995a, 1995b; Dale et al. 2000). However, a number of
insectivorous understory birds avoid edges (Quintela 1985; Dale et al. 2000),
especially solitary species, obligatory ant followers, and those that forage in
mixed-species flocks (S. G. Laurance 2000). Animals that nest or forage on
fallen dead trees, including wood-decomposing insects (Souza and Brown
1994) and certain marsupials (Malcolm 1991) and rodents (Ready et al.
1983), are also favored and increase in abundance near edges. The abundance
and species richness of small mammals increase in Amazonian fragments, pre-
sumably in response to greater availability of insect prey along the edge (Mal-
colm 1997a). However, an opposite response was detected in Brazilian
Atlantic forests (Fonseca 1988), tropical Australia (Laurance 1994), and Thai-
land (Lynam 1997). Similarly, although ants generally increase in abundance
near forest edges in central Amazonian rainforests (Didham 1997a; Carvalho
and Vasconcelos 1999), they decline in dry tropical forests of Madagascar
(Olson and Andriamiadana 1996), suggesting that local climatic factors can
affect species responses to habitat fragmentation. In at least some cases, the
nature of the edge response depends on edge age. In Colombian montane
forests, for example, new and old forest edges had different fruit abundance
and different communities of fruit-eating birds (Restrepo et al. 1999).
Given the great diversity of edge effects, it is not surprising that different
edge phenomena penetrate to varying distances inside fragments. In central
Amazonia, different kinds of edge effects have been shown to penetrate any-
where from 10 m to at least 400 m into fragment interiors (Figure 2.2). The
penetration distance (d) of an edge effect is a key parameter because if deter-
mined empirically it can be used with a mathematical core-area model (Lau-
rance and Yensen 1991) to predict the vulnerability of any fragment to that par-
ticular edge effect. In the central Amazon, the furthest-penetrating edge effect
documented to date is wind damage to forests (Figure 2.2), detectable up to
400 m from edges (Laurance et al. 1998a; Lewis 1998). However, recent evi-
dence reveals that certain other edge effects, such as destructive fires and inva-
sions of feral animals, can penetrate at least several thousand meters into trop-
- Ecological Effects of Habitat Fragmentation in the Tropics 39