witches’ broom disease (Crinipellis perniciosa) in 1989 motivated the conver-
sion of some of these traditional systems into pasture (Alger and Caldas 1992;
Tr evizan 1996). The official recognition of their potential for biodiversity con-
servation and ecotourism marked a fundamental change in political priorities
compared with campaigns of previous decades, in Brazil and in other cocoa-
growing regions, to thin these dense canopies of forest remnants in order to
increase cocoa yields (Johns 1999). It reflects particularly well the dual nature
of the cocoa agroforests as an agent of the conversion of natural forests into
agricultural ecosystems in this part of the Brazilian Atlantic forest and as one
of the most biodiversity-friendly land use options available to local farmers.
Of course, basically all upland agriculture in the humid tropics has to take
place on forestland and therefore ultimately at the expense of the forest. What
made the cocoa tree an important agent of the conversion of primary tropical
forests over the last four centuries, and especially in the twentieth century, is a
history of boom-and-bust cycles, combined with the tendency of the princi-
pal cocoa-growing regions to move from one place to another. Where these
cycles started, they led to the opening up of new forests, sometimes at a
tremendous speed. Where they ended, they left behind, in the best cases,
disease-infested groves of low productivity in a secondary forest environment
but often only poor fallows and pastures. These cycles were fueled by the
access to cheap forestland and often the labor force of immigrants.
In regions such as Bahia, southern Cameroon, southwest Nigeria, eastern
Ghana, and initially the Côte d’Ivoire, cocoa was grown in complex agro-
forests that are among the most diversified and forest-like of all agricultural
systems (see Chapter 10, this volume); in other cases, such as most of the Côte
d’Ivoire, western Ghana, Malaysia, and Sulawesi in Indonesia, cocoa was
grown in plantations with little or no shade, often almost monocultures. It is
obviously important for biodiversity, both on the plot and on the landscape
scale, whether forest is replaced by a tree crop monoculture or a complex agro-
forest with an understory of cocoa trees under the shade of old forest trees.
Even more important, however, for regional biodiversity is how these land use
types affect primary forest cover in the long term. Both the longevity of a tree
crop such as cocoa and the ease of replanting it on the same site are system
characteristics that are influenced by the degree of shading and may influence,
in turn, the long-term forest consumption by cocoa farms, as we shall see.
As attempts increase around the world to change the historical role of the
cocoa tree from a consumer of tropical forests into an instrument to improve
the livelihoods of tropical farmers and to conserve as much as possible of trop-
ical forests and their biodiversity, it may be instructive to review the factors
that have determined whether this crop was grown in complex agroforests or
monocultures, whether these systems were sustainable, and how they
responded to social, economic, and technological change. Although this chap-
108 II. The Ecological Economics of Agroforestry