research, but that have received benefits from on-farm research and participated in
tests; (c) farmers not involved in any research/testing activity (control group). For
each village, we compiled a list of households falling in each category, and selected
respondents based on the estimated proportion of each group in each village.
Similarly, we conducted 21 focus group discussions (FGDs) involving 15–20
participants in each village. In particular, the household survey respondents formed
part of the FGD, and were joined by other local community members that could not
participate in the interviews but were designated by the villages informants. These
FGDs were used to develop participatory agro-ecological map of each village, and
for the general discussion on the demand and supply for each type of innovation.
To establish agro-diversity levels (Sects. 10.3.2–10.3.4), we selected the study
systems based on spatio-temporal processes that take place within the croppingfallow-forest conversion cycle. The main criteria include the:
– Density of crop species
– Number of economic plant species
– Economic importance and density of agricultural and non-agricultural plant
species
– Spatial deployment of the cropping fallow forest conversion cycle
– Length of the fallow period corresponding to another rotation
Based on these criteria, we selected five land use systems that were informed by
the nine systems identified in the study area through observation, FGDs and surveys.
The five land use types followed the cropping fallow forest conversion cycle namely:
(a) Cucumeropsis agroforestry systems (i.e. melon seed); (b) mixed food crop
agroforestry systems (or mixed groundnut-based crop farms); (c) cocoa agroforestry
systems, (d) young pre-forest fallow systems (5 years), and (e) young secondary
forest systems (15 years). We collected land use and biophysical data from different
plots in these land uses, adapting the multidisciplinary landscape assessment
approach of Sheil et al. (2003). In total we sampled 184 plots, distributed as follows:
– Cocoa agroforestry systems: 41 plots with dimensions of 20 m  20 m (0.04 ha)
(total area ¼ 1.64 ha)
– Cucumeropsis agroforestry systems: 36 plots with dimensions of 20 m  20 m
(total area ¼ 1.44 ha);
– Young pre-forest fallows and young secondary forest systems: 30 plots with
dimensions of 20 m  20 m (total area ¼ 2.40 ha both two land uses);
– Mixed food-crop agroforestry systems: 47 plots with dimensions of 40 m  5 m
(total area ¼ 0.94 ha).
In each farm, one to four plots were sampled in cocoa and Cucumeropsis
agroforestry systems, depending on total farm size. In each farm, one plot was
systematically sampled for the regrowth fallow and regrowth forests. This was to
avoid problems related to the boundaries and size of the former farm. Each of the
20 m  20 m plot was subdivided into four sub-units of 20 m  5 m (0.01 ha), and
324
W. A. Mala et al.
tests; (c) farmers not involved in any research/testing activity (control group). For
each village, we compiled a list of households falling in each category, and selected
respondents based on the estimated proportion of each group in each village.
Similarly, we conducted 21 focus group discussions (FGDs) involving 15–20
participants in each village. In particular, the household survey respondents formed
part of the FGD, and were joined by other local community members that could not
participate in the interviews but were designated by the villages informants. These
FGDs were used to develop participatory agro-ecological map of each village, and
for the general discussion on the demand and supply for each type of innovation.
To establish agro-diversity levels (Sects. 10.3.2–10.3.4), we selected the study
systems based on spatio-temporal processes that take place within the croppingfallow-forest conversion cycle. The main criteria include the:
– Density of crop species
– Number of economic plant species
– Economic importance and density of agricultural and non-agricultural plant
species
– Spatial deployment of the cropping fallow forest conversion cycle
– Length of the fallow period corresponding to another rotation
Based on these criteria, we selected five land use systems that were informed by
the nine systems identified in the study area through observation, FGDs and surveys.
The five land use types followed the cropping fallow forest conversion cycle namely:
(a) Cucumeropsis agroforestry systems (i.e. melon seed); (b) mixed food crop
agroforestry systems (or mixed groundnut-based crop farms); (c) cocoa agroforestry
systems, (d) young pre-forest fallow systems (5 years), and (e) young secondary
forest systems (15 years). We collected land use and biophysical data from different
plots in these land uses, adapting the multidisciplinary landscape assessment
approach of Sheil et al. (2003). In total we sampled 184 plots, distributed as follows:
– Cocoa agroforestry systems: 41 plots with dimensions of 20 m  20 m (0.04 ha)
(total area ¼ 1.64 ha)
– Cucumeropsis agroforestry systems: 36 plots with dimensions of 20 m  20 m
(total area ¼ 1.44 ha);
– Young pre-forest fallows and young secondary forest systems: 30 plots with
dimensions of 20 m  20 m (total area ¼ 2.40 ha both two land uses);
– Mixed food-crop agroforestry systems: 47 plots with dimensions of 40 m  5 m
(total area ¼ 0.94 ha).
In each farm, one to four plots were sampled in cocoa and Cucumeropsis
agroforestry systems, depending on total farm size. In each farm, one plot was
systematically sampled for the regrowth fallow and regrowth forests. This was to
avoid problems related to the boundaries and size of the former farm. Each of the
20 m  20 m plot was subdivided into four sub-units of 20 m  5 m (0.01 ha), and
324
W. A. Mala et al.