procure seed. Widespread adoption
of improved germplasm followed.
Significantly, the key to success in
Kenya and Zimbabwe lay not only
in the development of appropriate
germplasm, but equally impor-
tantly, in the emergence of an
effective seed production and
distribution system capable of
delivering its product to small-
holders.

In certain cases the reason for low
use of improved germplasm is more
fundamental than poor seed pro-
duction and distribution: suitable
germplasm has not been fully
developed, let alone released to
farmers. The yield advantage
offered by improved materials re-
mains slight in some areas.
Dahniya et al. (1986) found that, in
on-farm trials conducted on a range
of sites in Sierra Leone, the local
check performed better than two
improved materials when no
fertilizer was applied. In such
cases, farmers are making a ra-
tional decision in electing not to
adopt improved materials, and
additional breeding work is
necessary.

Drought
Approximately 40% of the maize
area in sub-Saharan Africa experi-
ences occasional drought (defined
as causing average yield losses of
10-25/ ), whereas 25c experiences
frequent drought (causing average
yield losses of 25-50%). In eastern
Africa, southern Africa, and the
Sahelian zone of western Africa,
almost all ecologies in which maize
is produced are characterized by
unpredictable dry periods of one to
three weeks or more.

Maize farmers in Africa have
developed many strategies to cope
with drought, including selecting
drought resistant or drought toler-


ant germplasm, using water har-
vesting techniques to take maxi-
mum advantage of available mois-
ture, staggering maize planting
dates, and diversifying cropping
systems to reduce the risk of crop
failure. Strategies used by subsis-
tence farmers often are designed to
enhance yield stability (to ensure at
least enough production to meet
minimum household needs), even if
that means average yields will be
lower.

Current work on drought problems
in maize includes both germplasm
improvement and crop manage-
ment research. IITA, CIMMYT,
and several national breeding
programs are developing maize
populations with improved drought
resistance. IITA and a number of
national research systems in the
Sahelian countries collaborate on
drought work under an initiative
funded by the Semi-Arid Food
Grain Research and Development
(SAFGRAD) project. Scientists in
the SAFGRAD project have fol-
lowed two breeding strategies: they
have sought to improve drought
tolerance to mitigate the effects of
the dry spells that occur during the
growing season, and they have also
tried to reduce the time to maturity
so that maize escapes periods of un-
reliable rainfall at the beginning
and end of the rainy season.

At CIMMYT, four broadly adapted
elite populations have been chosen
for improvement as drought toler-
ant materials; the objective is to
provide national programs with late
white, late yellow, early white, and
early yellow maize populations to
use as sources of drought resis-
tance. In addition, cultivars that do
well despite drought conditions
have been combined to form the
Drought Tolerant Population
(DTP). Recurrent selection within
these populations is based on mor-
phological and physiological traits


that have been linked to drought
tolerance, including delayed leaf
senescence and reduced anthesis-
silking interval, in addition to high
grain yields under moisture stress
(Edmeades et al. 1988).

Crop management research for
drought environments has focused
on soil management techniques
such as mulching and ridging,
which are designed to increase and
preserve soil moisture levels, as
well as on techniques designed to
achieve optimal plant population
densities and spatial arrangements
for maize grown in dryland condi-
tions. Evidence from experiment
station trials as well as farmers'
fields indicates that tied ridging is
particularly effective for preserving
moisture and raising grain yields,
except in sandy soils (Figure 16).
However, the economic feasibility of
tied ridging must be evaluated case
by case, because the yield increase
does not always justify the signifi-
cant labor input required to con-
struct tied ridges.

Weeds
Most weed control in Africa is
performed manually or using
animal-drawn cultivators. The fre-
quency and timing of weeding vary,
depending on the severity of weeds
and the availability of labor. One or
two weedings are common in most
areas, although they are often
delayed because of seasonal labor
constraints. Weeding maize during
a critical period 10-30 days after
crop emergence greatly enhances
grain production; uncontrolled
weed growth during this period can
reduce maize yields by 40-60%
(Akobundu 1987). Weed competi-
tion is particularly problematic in
arid and semiarid zones, since
moisture lost to weeds translates
directly into yield losses in the
maize crop.