Abstract:

Large within and between farm variability of banana {Musa spp., group AAA) yields in Uganda
are caused interalia by a complex of soil fertility differences, pests, and light availability. These
biophysical factors are again influenced by crop management practices, which show large spatial
variability. Previous studies did not sufficiently take into account the large spatial variability in
biophysical factors that prevail under on-farm conditions. The objective of the study was to
quantify the relative contributions of soil fertility factors, pests and diseases, and crop
management practices (e.g. planting density) to banana yield losses within and between farms
that cover a range of wealth and slope classes. From a total of 50 farms covered under a
preliminary diagnostic survey in Ntungamo district (southwest Uganda), 10 farms representing a
gradation of biophysical (slope position) and socioeconomic (wealth status) variability were
selected for on-farm monitoring of plant performance in Butare sub county. On each farm, 15
mats of the most dominant genotype (cv. Enyerti) were selected for the study along a transect
parallel to the slope such that they fell within a range of relative distances (RD), close, mid and
remote from the homestead. The crop management, growth from emergence to harvest,
pest/disease status and yields of two ratoon suckers were quantified. Soil samples were collected
per mat at the start of the study and analyzed for physical and chemical properties following
standard laboratory methods. Boundary-line analysis revealed that soil fertility factors
collectively contributed to 60% yield limitations, pests accounted for 30.0% and crop
management 10.0% of total yield limitations. Among the limiting factors, banana weevil
{Cosmopolites sordidus (Germar)) damage accounted for the highest (26.7%) yield limitation,
soil K/Mg ratio (20.0%), soil pH (16.7%), excessive leaf pruning (10.0%), clay fraction (10.0%),
exch. Ca (10.0%), Ca/K+Mg+Ca and nematodes each 3.3%. Banana streak virus, black Sigatoka
and leaf spots depressed bunch weights by 21.5%, 17.9% and 16.9%, respectively in severelyinfested
plants compared to disease-free ones. Low pH and inadequate Ca limited yield
improvements only among poor fanners as opposed to high weevil pressure among medium and
rich fanners. The former had nematodes as an additional constraint and high soil K/Mg ratio,
which also occurred among resource rich fanners. Mat spacing was significantly (p < 0.05, R2 =
0.904) negatively correlated with bunch weight, and the heaviest bunch occurred at 2.6 m x 2.6
m (1512 mats ha'1) compared to the recommended 3.0 m x 3.0 m. However, the poor farmers
who had significantly (p < 0.001) higher plant densities than wealthier ones indicated, proper
I
Wealth classes, Diagnostic survey
xiii
spacing should be a greater concern among poorer farms. Pests and foliar diseases could not be
attributed, but leaf pruning, to reduced number of functional leaves. Farmers should therefore
aim at maintaining nine leaves at bud emergence to maximize yields, amidst the heavy
defoliation. Poor farmers should invest in soil fertility management practices that improve soil
pH and exchangeable bases availability in addition to proper spacing and weeding since their
soils (were of similar inherent fertility in all farm categories) were relatively less fertile than
those for wealthier farmers accordingly. Wealthier fanners should allocate their resources for
pest control more especially banana weevil and for soil inputs for a good K/Mg balance. They
should circumvent heavy defoliation through moderate leaf pruning (gradually remove diseased
leaves) and avoid cutting of green leaves for preparing steamed food.