Hammond 2000; Harrison 2000). Shifting cultivation has persisted for a cen-
tury or more on extremely poor oxisols in the Bragantina region of the Brazil-
ian Amazon (Alencar et al. 1994; Smith et al. 1999a), and in vast regions of
Central Africa and New Guinea it has been one of the major types of agricul-
ture for centuries or millennia (Dounias 2001; Kocher-Schmid 2001). The
ecological, social, economic, and political contexts under which shifting culti-
vation persists over the long term are an interesting area for research. De
Wachter (1997) studied the traditional swidden cultivation system practiced
by the Badjwe people at the periphery of the Dja Reserve in Cameroon. Mod-
eling suggests that under this land use system, the current population density
of 4.2 inhabitants per square kilometer of arable land could grow to up to 38
inhabitants per square kilometer and still maintain, after more than 50 years
and within 5 km of roads, a landscape mosaic with 50 percent of the land
under shifting cultivation, 43 percent under primary forests (including 26 per-
cent of swamp forests), and 7 percent of pure Raphiapalm stands. Under a
scenario of intensification of cash crop–oriented activities (cocoa, plantain),
the numbers fall to 18 inhabitants per square kilometer and about 35 years to
maintain a similar mosaic within 5 km of the road.
The Shortening of Fallow Periods: Causes and Consequences
The length of fallow periods is a key factor in the degree to which biodiversity
accumulates on individual patches in swidden landscapes and the degree to
which landscapes as a whole maintain certain components of biodiversity; in
many senses, the longer the fallow, the better. This simple rule leads us in an
equally simple way to one of the major conflicts regarding biodiversity in shift-
ing cultivation landscapes. In the classic Boserup model (Boserup 1965), agri-
cultural gains were accomplished through land expansion, increased labor,
and, critically from our point of view, shortening of fallow periods (but see
Stone and Downum 1999 for counterexamples). Many recent studies report a
tropics-wide tendency toward shorter fallows in shifting cultivation, and this
is one of the best-documented aspects of the subject from both ecological and
agronomic standpoints (Jones and O’Neill 1993; de Jong et al. 2001).
In recent discussions, Smith et al. (1997, 1999a, 1999b, 2001) evaluate the
factors that underlie the shortening of fallow periods by farmers. They observe
that farmers set fallow periods by weighing the costs and benefits associated
with different fallow lengths. Longer fallows have potential benefits that
include higher yields from the cropping period, lower labor costs for weed
control during that period (weed control is one of the most important func-
tions of the fallow), and greater opportunities for harvesting the products of
unplanned biodiversity during the fallow. The incentive to increase fallow
length may also arise because as the time during which land has been under
the shifting cultivation cycle increases, the productivity of fallows and the
- The Biodiversity and Conservation Potential of Shifting Cultivation Landscapes 159