Agriculture Reference
In-Depth Information
For species abundance, three categories are distinguished: (1) low abundance
where species are observed only once or twice and did not cover much of the gar-
den area, (2) medium abundance where several plants are recognized but covered
less than one fourth of the garden area, and (3) high abundance where many plants
occurred or covered more than one fourth than the garden area. The Shannon-
Wiever index was calculated to analyze the diversity of homegardens for each study
village with H' = −
(p
i
ln p
i
), where p
i
is the proportion of the occurrence of
species in a study village (expressed as a proportion of the total species occurrence
N; Kent and Cocker 1992). From that, the Equitability or Evenness was calculated
by E = H'/H'max, with H'max = ln s and where s is the number of species, to esti-
mate the homogenous distribution of plants in homegardens. Finally, the Whittaker's
β
Σ
diversity index was calculated to identify the differences in species composition
between plots within upland and lowland villages, using the following formula
(Coffey 2002): by
β
= S/
α
- 1, where S is the total number of species in the home-
garden sample and
is the average number of species in each subsample.
Semi-structured interviews were conducted to obtain information about plant
use, management and planting practices, economic benefits, reason for keeping a
homegarden and the cultivators' perception on major constraints in homegardening.
A list with 30 major reasons for having a homegarden was composed on the basis
of reasons spontaneously referred to by farmers in the field. Interviewees were
asked to select the ten most important reasons and to rank these from 10 (most
important) to 1 (least important). After completing the interviews, the average rank
scores for the ten most important reasons were identified for each village.
Differences among village means for various site and household variables, plant
density, evenness, and the Shannon and Whittaker indices were determined by
using the non-parametric multiple comparison analysis described by Neter et al.
1996. In order to identify different types of homegardens based on species combi-
nations, a hierarchical cluster analysis was performed with the complete linkage
(furthest neighbor) method and the chi-square as distance or similarity measure.
Species density (i.e., the number of individuals per species per unit area) was used
as main variable in the analysis. For the trees and non-tree species (the latter group
including
Musa
sp.), the analysis was made separately given the large differences
in species densities between the two groups.
α
2.3
Study Site and Household Characteristics
The villages are situated in the moist agro-climatic zone with an annual rainfall of
1,500 to 2,500 mm spread over a growing season of seven to eight months. The dry
season extends from January up to May or June. The upland village Dy Abra is
located in the least densely populated zone (<50 persons per km
2
; NSO 2001), along
the forest frontier and partly surrounded by hilly grassland and cultivated fields. It
is the most remote village among the five, with a travel time of at least one hour to
the nearest major (Tumauini) market (Table 2.2). During the wet season, the village
Search WWH ::
Custom Search