new cocoa plots at another site, outside the plantation, where the shade inten-
sity is easier to regulate. Ideally, this would be a previously planted area under
secondary forest, as in the old rotational system, but in practice farmers often
use primary forest if available.
Production shifts at village and regional scales provide historical evidence
of this process. Between 1960 and 1963, the East province and the depart-
ment of Dja et Lobo, where cocoa is grown mainly in agroforests, produced a
yearly average of 8,000 and 9,000 metric tons of cocoa, respectively. Until
1984, annual production had fallen to 5,300 metric tons per year, while cocoa
production in the southwest province jumped from 7,300 to 27,000 metric
tons per year and that in the M’Bam department from 6,200 to 10,900 met-
ric tons per year (Losch et al. 1991). In the latter regions, cocoa is produced
mainly by migrants in lightly shaded systems. Production shifts also occurred
within the departments: in the Nyong et M’foumou department, where cocoa
is produced in ancient chocolate forests, the main cocoa production centers in
the 1970s were the road from Akonolinga to Yaoundé and Endom in the
south; by the late 1980s, cocoa production had moved to other districts such
as Nwane Soo and Ayos Fang Biloun in the north of the department (F. Ruf,
unpublished survey data, 1990).
Available data suggest that noncocoa revenues from these agroforests usu-
ally are insufficient to compensate the farmers for reduced cocoa revenues at
times of low cocoa prices. Therefore, when the cocoa price collapsed in the
1990s, farmers resorted to new forest clearing oriented toward food crop pro-
duction as a survival strategy (J. Gockowski, pers. comm., 1998). However,
during this time of economic crisis cocoa farmers close to the urban market of
Yaoundé were successful at diversifying their farms by planting mandarin
orange trees, often in places where cocoa trees had died and were difficult to
replant (Aulong et al. 1999; Gockowski and Dury 1999), and this could indi-
cate a way to further commercially oriented diversification of the cocoa farms
in other regions. In the 1980s, farmers in the Nyong et M’foumou depart-
ment mentioned the tree Voacanga africana(Obahtoan) as a source of an
exportable medicinal product, but this export trade has ceased, probably
because of a lack of certification and nonconformity to the legislation of the
European market (Arditi et al. 1989). An interesting species is the African
plum tree (Dacryodes edulis), which according to a survey of 300 farmers in
southern Cameroon was planted by 83 percent of the respondents in their
cocoa farms (Sonwa et al. 2000).
In conclusion, although cocoa agroforests have successfully conserved part
of the forest environment of southern Cameroon, they have not been able to
sustain farm revenues at times of crisis, although examples of successful eco-
nomic diversification of cocoa farms are emerging. Replanting problems in
agroforests are different from those in no-shade systems, but they do exist and
may have contributed, together with immigrations into the southwest and
128 II. The Ecological Economics of Agroforestry