assemblages observed in that patch, irrespective of the habitat type that
adjoins it.
Although abundance patterns over disturbance gradients are documented
for individual species by some authors, patterns at the guild level probably are
a more practical indicator of the main ecological processes affecting bird com-
munity characteristics. Several studies have shown that terrestrial forest birds,
particularly insectivores, are less species rich and abundant in fallows and crop
fields than in primary forest (Terborgh and Weske 1969; Johns 1991; Andrade
and Rubio-Torgler 1994). Bark-gleaning and foliage-gleaning insectivores
may also decrease in fallows, with a corresponding increase of sallying insecti-
vores such as flycatchers (Tyrannidae) and insectivore frugivores such as some
tanagers (Thraupidae; Blankenspoor 1991; Johns 1991; Andrade and Rubio-
Torgler 1994; Shankar Raman et al. 1998). Changes similar to these have been
observed in neotropical forest disturbed by logging (Thiollay 1992, 1997) or
fragmentation (Bierregaard and Stouffer 1997) and appear likely to be a gen-
eral response of bird communities to habitat modification and simplification.
The appearance in young second-growth and crop fields of granivores, as well
as doves and pigeons (Columbidae), which are generally absent from forest
habitats, is an important but not universal change in shifting cultivation land-
scapes (Johns 1991; Andrade and Rubio-Torgler 1994; Shankar Raman et al.
1998).
Because raptors are a single guild, compositional differences between land-
scapes of different structural diversity must be examined in terms of individ-
ual species (Anderson 2001). Only one raptor species was noticeably more
abundant in primary forest than in shifting cultivation and forest mosaic land-
scapes, although three species observed only once, or observed outside sample
plots, may be considered forest dependent; this group included the harpy eagle
(Harpia harpyja;Anderson 2001). Three species were more abundant in the
more heterogeneous landscapes, which included crop fields and fallows
(Anderson 2001). It was the appearance of these species, among other factors,
that accounted for the greater abundance and diversity of raptors in the more
structurally diverse landscapes.
What habitat factors contribute to the changes in species richness and
composition of vertebrate assemblages that are observed when shifting culti-
vation habitats are compared with primary forest? We will emphasize habitat
structure, food availability and foraging behavior, and microclimate. Our
focus is on factors linked to the lower abundances or absence of forest species
in agricultural habitats because of the conservation importance of these
species, not on the reasons why disturbed-habitat species colonize landscapes
because they are able to use crop fields and fallows.
It is well established that habitat structural diversity can influence animal
species diversity (a recent summary is provided by Begon et al. 1996). Ter-
borgh and Weske (1969) suggested variation in foliage height profiles as the
- The Biodiversity and Conservation Potential of Shifting Cultivation Landscapes 177