Proiect D 
Promoting food security and income 
generation through sustainable production and 
commercialization of starchy and grain staples 
in Eastern and Southern Africa 
M 
o 
o 
N 
......, 
~ 
o 
c.. 
Q) 
c:: 
CO 
::J 
C 
c 
<C 
\,TA 
Research tu Nourish Africa 
Intemationallnstitute of Tropical Agriculture I Institut international d'agriculture tropicale Ef1 
Contents 
Preface 
Project goal .............................................................................................................. 1 
Project purpose ..................................................................................................... 1 
Outputs 
1 Market opportunities identified for setting R&D agenda and formulating policy 
options ......................................................................................................... 1 
2 Commercially viable small· and medium·scale agroenterprises that expand 
trade of the starchy and grain staples enhanced ............................................... 13 
3 Broad·based and special trait genotypes and populations targeting the major 
production systems and market opportunities developed and disseminated ........... 18 
4 Environmentally safe integrated plant protection technologies that reduce 
pre· and postharvest losses due to pests and diseases developed and applied ....... 34 
5 Crop and natural resource management practices for sustainable, 
competitive, and commercially based production systems developed and 
disseminated ................................................................................................ 73 
6 Public and private sector partnerships that promote information and technology 
exchange with emphasis on scaling up strategies for increased trade developed 
and/or strengthened .......................................................... , ................................... 77 
7 Capacities of NARS and other stakeholders to generate, evaluate 
and disseminate appropriate knowledge and intervention technologies 
for target sub sectors enhanced ...................................................................... 80 
Publications ..................................................................................................... 84 
Refereed journal articles ..................................................................................... 90 
Preface 
IITA's research-for-development agenda is divided into six project themes, around which 
these project annual reports are prepared. These projects themes address different aspects 
of attaining sustainable increases in productivity of dominant farming systems and utili-
zation practices in the various agroecologies of sub-Saharan Africa (SSA). Research and 
training activities carried out in the six projects are being implemented together with 
national program partners in order to increase the well-being of resource-poor people in 
SSA through higher levels offood production, better income, nutritional status, and reduced 
drudgery particularly for women. 
Project D 
Promoting food security and income generation 
through sustainable production and commercial-
ization of starchy and grain staples in Eastern and 
Southern Africa 
by 
Adebayo Abass, Steffen Abele, Maria Andrade, Piet van Asten, Danny Coyne, 
Carine Dochez, Thomas Dubois, Morag Ferguson, Anneke Fermont, Shaun Ferris, 
Cliff Gold, Rachid Hanna, Bill Khizzah, Edward Kanju, Sicco Kolijn, James 
Legg, Nzola Mahungu, Victor Manyong, Chiedza Muchopa, Caroline Nankinga, 
Bjoern Niere, Pheneas Ntawuruhunga, Suleman Okech, Michael Pillay, Phillip 
Ragama, Abdou Tenkouano, Muaka Toko, John Wendt (Project Coordinator), and 
Jim Whyte 
Project goal 
Enhanced food security, improved livelihoods, and socioeconomic development in Eastern 
and Southern Africa. 
Project purpose 
Develop and promote, in collaboration with public and private sector stakeholders, the 
adoption of improved technologies and market support services for sustainable increases 
in production, utilization, and commercialization of starchy and grain staples. 
Outputs 
1 Market opportunities identified for setting R&D agenda and 
formulating policy options 
1.1 Opportunities for soybean processing and marketing in Rwanda 
by S. Abele and S. Ferris, in collaboration with J. Magnay (Uganda Grain Traders) 
and A. Turner (Chemonics) 
Given the critical situation of income and food security in Rwanda,ADAR seeks to improve 
rural income and food security by fostering the agribusiness sector. The project goal was to 
determine the market potential for soybeans and its derivatives and to design a process· 
ing industry that could profitably supply the quantity of soybean products needed on the 
domestic and possible external markets. Consequently, the objectives of the study were: 
• To determine the market potential for soybean products in Rwanda and the subregion. 
Products envisaged at the beginning of the study were mainly soybean oil and soybean 
flour. Further areas of interest that came up during the study were mixed food products 
based on soybeans and cereals. Market potentials were evaluated by assessing prices 
of the products studied and their substitutes, i.e., their competitors. 
-1-
• To design a competitive processing plant for soybeans in the potential project area of 
Kibuye. This design was done on the basis of an initial business plan provided by the 
client. Criteria for the design of the plant were the products to be identified as having 
the best market potential in the first stage ofthe study. The second task in designing 
the plant was to choose an optimal location between Kibuye and Kigali. 
• To determine the feasibility of improving soybean primary production in Kibuye. The 
initial outline of the project included a development component for an outgrowers' 
scheme of soybean producers in Kibuye. The major issue in this decision was whether 
it was more profitable for the project to make the raw product within a vertical integra· 
tion or to buy it from the market. 
Activities, methodologies, and database. The field study was conducted during a visit of the 
team from 24 to 29 March 2003. During this study, data on prices for soybean products and 
substitutes were collected in the study region as well as in Gitarama and Kigali. A group 
interview with farmers in the envisaged production region near Kibuye was conducted to 
gather information on their production schemes and on soybean production in particular. 
Further information was gained through statistical publications from the Ministry of Plan, 
the Ministry of Agriculture, the Customs office, and the National Bank of Rwanda. Addi· 
tionally, data were taken from the Food and Agriculture Organization, and from secondary 
literature. 
Based on the data, the market size of soybeans and soybean products was estimated by 
setting up commodity balances. Primary production schemes were assessed by running 
scenarios on gross margins. The business plan was set up with respect to investment and 
operating costs, and figures for return on investment and cash flow were estimated under 
different market and finance scenarios. 
Findings on soybean demand. An analysis of Rwandan household expenditure data with 
respect to soybean consumption revealed the relationship between income and private 
households' soybean consumption in Rwanda. As consumption is scattered and unevenly 
distributed, precise econometric estimations of the income elasticity could not be made. By 
defining income classes and calculating the average consumption within these classes, at 
least a rough idea on income elasticity and future demand could be gained. 
The data show that soybean consumption is increasing with increasing income. Yet, there 
seem to be two different sections of income and consumption relationships. In the upper 
income class section (from 500 000 Frw onwards), consumption seems to have different 
development paths and elasticity. This could indicate that in the sections below and above 
500 000 Frw of income, different types of soybean products are consumed. However, in 
both sections, initial income elasticity is relatively high, but then decreasing very rapidly. 
Elasticity in the lower income section is on the average 1.7, in the upper section it can be 
estimated as one. This means that at given income growth rates, in the lower section, it 
would take 12 years to double the expenditure for soybeans-from less than one kg per 
household per year to about two, in the upper section it would take 20 years to double soy· 
bean product consumption from now about two kg per year to four. The statistical data also 
show that soybean oil, currently consumed at a level of 7000 L a year, does not constitute 
a significant market for a commercial enterprise. 
-2-
Apart from the private market for soybean products, there is also an institutional market 
created by international organizations like the World Food Program (WFP), UNICEF, and 
the Government. WFP distribute about 2500 t of Corn-Soya-Blend (CSB) and Unimix, a 
blend of extruded whole maize and whole soybeans on the Rwandan market. This equals 
about 700 t of soybeans per annum. UNICEF supplements milk which could be replaced by 
about 200 t ofUNIMIX which is an equivalent of 60 t of soybean raw material. Contrary to 
the previous market discussion, forecasts offuture developments cannot be easily made as 
this market is not dependent on income and preference development, but on political situa-
tions and the will of donors. Further, market transactions are rather based on negotiations 
than on the ''invisible hand" which is supposed to arrange prices on a free market. This 
makes it difficult to estimate price and quantity developments. It can be said that the WFP 
will continue their work in East Africa and thus the market will be sustained. Concerning 
price developments, the costs will surely be affected by inflation. However, it depends on 
the negotiating power of the supplier to what extent the inflated costs can be covered by 
higher prices on the output side. It should also be emphasized that revenues especially are 
heavily dependent on regional competition, and price margins are rather due to logistical 
differences (e.g., transport costs of the finished products) than to overall inflation, so that 
Rwandan inflation (except for raw material prices) is a minor issue. 
From the above conclusions, the following observations can be made: 
• Present soybean demand traded on private commercial markets is rather weak. 
• Prospects for the future are also not very promising, as income elasticity of demand 
is not enough to assure high (absolute) growth rates of soybean products on the 
private commercial markets. 
• The only stronger, stable demand for processed and soybean-based food products 
comes from the institutional markets. 
Rwandan soybean production. Rwanda produces about 17 000 tonnes a year of soybeans, 
but production is unevenly distributed within the country. The highest production is found 
in Gitarama, about 50 km from Kigali, with slightly more than 9000 t per year, most of 
which is home consumed and only a small part traded. Data show that production is highest 
in Gitarama, both in terms of quantities produced and in terms of households producing 
soybean. This implies that the region of Gitarama has a comparative advantage in soybeans, 
while Gitarama's advantage is probably due to the consumption preferences in Gitarama 
itself (which has the second largest population density in Rwanda) and its proximity to 
Kigali, so that we could call the production in Gitarama "demand-driven". The study site, 
Kibuye, does not have a significant comparative advantage, neither in terms of consump-
tion nor in terms of production. 
Primary production at the study site. Primary production of soybean in Kibuye is done 
on smallholder farms, mainly own less than 1 ha, if not less than 0.5 ha of land. Besides 
soybean, other crops planted are coffee, which is the only cash crop, banana, sweetpotato, 
manioc, and others. The production, including soybean, is mainly subsistence oriented 
with only the surplus sold off at times of good harvest. Chemical input use is restricted to 
coffee production, for which farmers buy the inputs on credit and repay when selling their 
harvest to the same traders. Inputs used on other crops are own produced like animal 
manure. Seeds are also own produced so that one could assume that they are ofJow quality. 
The data obtained from the field study give an overview of the current soybean production 
-3-
system and outlines the possible improvements when using improved seeds and fertilizing 
techniques. Data show that there is a potential to increase both soybean production and 
profitability through intensification. It also shows that the calculation is quite sensitive to 
output price changes and to different assumptions on yields. A decline in output prices of 
less than 20 percent could lead to a situation where gross margins are cut back to the state 
of traditional production. This price decline could occur when institutionally driven inten· 
sification leads to a sharp increase of production without a sufficient absorption capacity 
of the local market. The size of the possible decline is quite probable, as price data show 
volatilities up to 32%. The same holds for yield deviations. The Ministry of Agriculture 
gives a range of 1200 to 1500 kg yield per hectare. Going to the lower edge of the yield 
scale, the returns would again be only 80 percent of the initial prospects as in the low 
price scenario. This throws up the question why the farmers should invest in a technology 
that would yield only a 20 percent increase of gross margins under favorable conditions. 
However, other studies show that Rwandan farmers tend to renounce up to 40 percent of 
their yield potential for hedging against risks. In other words, risk adverse farmers might 
tend to not apply highly intensive technologies. 
It thus seems that there a high effort is required to obtain the desired results of a function· 
ing outgrowers' scheme: 
• Farmers have to be trained to intensify production. 
• Extension has to be conducted with special care on farm and household economics and 
objectives of farmers. 
• Inputs have to be provided at reasonable prices, mostly to remote areas. 
• Contracts have to be made to link farmers to soybean production. 
• The scope of production has to be planned to assure inputs for the processor at profit-
able prices and to maintain profitable output prices for the farmers to sustain their 
production. 
It is doubtful if this can be done efficiently by the processor. It was thus recommended to 
use the market signals given by the demand of the processor and other buyers to set the 
incentives for the farmers. 
Plant design. The selection criteria for the processing site as chosen in the study were: 
• Proximity to markets capable of absorbing the product at profitable prices. 
• Proximity to resources like energy, maintenance facilities etc. 
• Proximity to institutional buyers like public institutions, and to development and 
aid programs like world food program. This was a result of the market observations 
made. 
• Good institutional background and support from the community (availability of land 
titles and production permits). 
Given these criteria, it was concluded that the best site would be Kigali for the following 
reasons: 
• In Kigali, the prices for soybean products are generally higher than elsewhere, except 
for Gitarama, but the latter does not have Kigali's advantages. 
• Proximity to institutional buyers is closest. This implies not only proximity to the 
outlets, but also low transaction costs in negotiating with buyers and assuring proper 
interaction in quality management. 
-4-
• Provision of resources is best. especially in terms of stability of water and electricity 
supply, as well as availability of maintenance and repair services. 
• Institutional background and support from the community is best in Kigali. It is prob· 
ably lowest in Kibuye, where the community seems to encourage tourist activities rather 
than industrial investments. 
Profits and cash flow assessment of the plant. Following the market and site assessment. 
the study evaluated the profitability of the proposed firm. Criteria for assessment were 
(a) profitability, (b) liquidity or cash flow, (c) returns on owned capital, d) the net present 
value over periods of five and ten years, and e) the internal rate of return. All of them 
were calculated under different finance scenarios that are given by the abovementioned 
banking conditions, as well as different market scenarios and the consequent coverage of 
the plant's potentiaL 
The analysis showed that, when designed for producing blended soybean products, the 
plant can operate profitably already with a low market share or potential coverage. For 
finance, the best option with respect to cash flow and profitability is to obtain a loan from 
the Rural Development Bank. However, this requires higher own capital than loans from 
private creditors, so that in this option, initial returns on investments are lower. 
1.2 Market opportunities identification study in southern Sudan 
by S. Abele and S. Ferris, in collaboration with CRS I USAID 
The aim of this market opportunity study is to identify and describe the current and 
expected market potential and demand for a broad range of commodities and items that 
will be addressed by the six training centers under the recently initiated Southern Suda· 
nese Agricultural Rehabilitation Program (S·SARP). 
The study is being undertaken in three parts. Phase I, is focused on gathering secondary 
information on six selected sectors, crops, agroprocessing, livestock, forestry, fish, and 
wildlife; and conducting a rapid market opportunities identification (Mal) analysis to 
evaluate and prioritize products with the most promising marketing options in each sub· 
sector. The results of this first analysis will be discussed extensively with the partners 
and stakeholders of the SSARP consortium, as these results will be used to direct the next 
phase of the study. 
Phase II of the study will conduct a more detailed market chain analysis of those products 
which were identified as having highest market potential, for small, medium, and poten-
tially large-scale intervention. The aim of this second level of analysis is to gain a more 
comprehensive understanding of the critical constraints to marketing and to find the most 
promising areas of investment in specific supply chains. The information from Phase II 
will be used by the S-SARP consortium to develop business/agroenterprise plans. Phase 
III of the process will be to work with the SSARP consortium to use the findings from the 
marketing studies to develop specific training plans for the six S-SARP agribusiness centers 
to exploit the prioritized market opportunities. 
Methodology of analysis. Due to the very specific nature of the situation in Sudan and 
the scale of the territory, the process of analyzing market opportunities was unlike most 
marketing studies. In most cases secondary data is easily sourced, markets and financial 
structures are functioning. infrastructure is in place, and key informants can he located and 
-5-
interviewed with ease in their normal place of work. Due to the ongoing insecurity, none of 
these conditions are found in Sudan and therefore the study team was required to develop 
a more complicated analytical framework. The analysis was also further complicated by 
the scale of operation being reviewed and the potential for investment at local, national, 
regional, and international levels. 
To accommodate these conditions, the analysis was conducted with information and analysis 
from several perspectives and included an analysis which compared current conditions and 
scenarios of "what might be" when peace and economic functionality returns. 
To undertake this type of analysis, several types of analysis were combined including: 
• Analysis of secondary data. 
• Development of economic filters and use of secondary data to model potential trade 
scenariOs. 
• Survey work of current market conditions in S·Sudan. 
• Analysis from commodity traders. 
• Analysis from expert practioners on markets and current import/export opportuni· 
ties. 
A "triangulation process" of the data was then used as a basis to prioritize commodities 
within sub sectors in regard to market opportunities and to formulate the basic ideas for 
the way forward, i.e., the next stage in the analysis. The starting point of the MOl analy· 
sis is to draw up a "Long List" of products and commodity options which are already or 
could be produced and marketed from the territory, i.e., S-Sudan. The MOl analysis then 
attempts to use market demand as the key filter in prioritizing commodities and products 
such that the initial "long list" is distilled down to a manageable "short list" of products, 
which have strong market demand and can be studied in more detail. The results from 
the MOl analysis of commodities, products, and sectors were drawn up as a basic long list 
and a short list after being analyzed through market filters. The prioritized grain crops 
are maize and rice; oil crops, palm oil, sesame, sunflower, and groundnuts; root crops, cas-
sava; stimulants, tea and tobacco. SSARP should focus on making gains in food security 
and once marketable surpluses are available, should work with local farmers' groups to 
find the best market channels. As indicated in the livestock section, SSARP could acceler-
ate the process of market support by developing a reliable market information service to 
facilitate the marketing process. As also indicated in the Trader's analysis, export options 
for hides and skins could be a major source of income, particularly if the products are of 
high quality. For all the markets, a critical constraint is lack of market information and 
this is an aspect that could be addressed by SSARP relatively quickly. 
The forest reserves in southern Sudan, particularly in the Equatorial region, are known to 
be one of the most lucrative assets that the region can immediately exploit for the domestic, 
regional, and international markets. In Eastern Mrica, the stocks of tropical hardwoods 
are rapidly being exhausted and this is effectively raising the value of the Sudanese forest 
reserves. In regard to the exploitation of the tropical hardwoods into the regional and 
international markets, all efforts should be made to increase the level of value addition 
that is carried out in Sudan. The domestic markets are based on the opportunities for sub-
stituting what is already being imported. The listings of goods for the international market 
are based on strengthening existing operations. The potential for all of these markets is, 
however, much dependent upon exporting goods only of the very highest quality and this 
-6-
means that considerable investment should be made in highly professional operations, such 
that quality is strictly controlled and that the operators focus on niche, high end market 
supply chains. 
Way forward: areas for Phase II marketing and training possibilities 
Summary of "Results from the market opportunities identification survey: options and 
commodities to consider for the phase II marketing study" are as follows: 
• Undertake subs ector analyses for crops with highest market demand/potential within 
the country and for regional markets. These crops are sorghum, maize, rice, sesame, 
and cassava, with a view to developing these sectors on a large scale. 
• Develop business plans for larger scale, plantation operations, possibly linked to out-
grower schemes including the production and processing of oil palm, sunflower, rice, 
sugar, cotton, coffee, tea, gum Arabic, and tobacco, as well as livestock and fish. These 
business plans should be developed in partnership with interested and/or existing 
investors. 
• Development of investment plans for niche products and markets including Arabica 
coffee, lulu, spices, dried fruit, to be done in collaboration with NGOs that have agro-
enterprise skills and plans/funds to organize community-based production systems. 
1.3 Survey on banana-coffee systems: economics of farming and marketing 
systems 
by S. Abele, S. Okech, C. Gold, and C. Nankinga, in collaboration with A. Nabuye 
(Coffee Research Institute (CORI), Uganda) A. Karugaba (Mbarara District Agri-
cultural Extension Department, Uganda), and W. Tushemereirwe (Kawanda Agri-
cultural Research Station, Uganda) 
The purpose of the study is to provide first insights into the yet underresearched farming 
systems in Western Uganda that are based on banana and coffee. These two crops are the 
economic bases for most of the people living in the rural areas of Western Uganda. They 
provide both cash (coffee and banana) and, for the case of banana, food and by-products. 
Yet, the system faces decline in various ways: Both banana and coffee are affected by pests 
and diseases that threaten the whole system and the livelihood of the rural communities. 
Moreover, price declines, especially on global coffee markets, have a negative impact on the 
systems. It is thus of vital interest to generate knowledge about the banana-coffee system, 
on the interaction of these crops in agronomic and economic terms, and on how theses sys-
tems can be stabilized and further developed. This pre study aims at gaining information 
to built up the hypotheses of a following detailed diagnostic survey and analysis. It also 
attempts, as far as this part of the study is concerned, to determine first assumptions on 
the economic behavior and decision-making patterns of small-scale farmers within this 
system. 
The study assesses the economics of three sections of the farming and marketing system. 
The fIrst section deals with the farming system, throwing specific emphasis on the banana 
and the coffee production systems and their interactions, both in terms of competition 
and complementarities. The farming section further includes other on-farm activities of 
cropping and livestock, off-farm activities, characteristics of farms especially in land and 
labor organization, constraints and management, farm sizes, decision-making, market ori-
entation, and on-farm constraints such as liquidity and risk. The marketing section deals 
-7-
with patterns and problems of marketing and farmers' organizations. The third section on 
postharvest processing addresses the status quo of postharvest technologies in the region 
as perceived during the survey. 
The methodology used to obtain the information was the conducting of group interviews on 
three sites in Mbarara during AprillMay 2003. Group interviews were held during meetings 
with 15-40 farmers. Further, single farmers were visited on their fields and interviewed 
in a nonstandardized way. 
Farm and household economics. Farm sizes are rather small in all the sites, usually aver-
aging about 0.8 ha (2 acres), but ranging from 0.1 to 47 ha and from 0.1 to 1.6 ha across 
sites. While in the rust area, land is mainly consolidated, in other sites, land holdings are 
dispersed due to the heritage system. Also in some sites, large size farms or landholdings 
exist that are assigned to their owners by the state. On these holdings, farmers can settle 
but have to pay rent. About ten percent of the population on the sites is landless. It was 
mentioned throughout the sites that land is scarce and that this scarcity is one ofthe main 
reasons for intercropping. 
As far as labor organization is concerned, hired labor is used on all sites, mainly for weed-
ng and pruning, but also for soil preparation. On one site, labor is scarce during the high 
cropping season, in Ndaija, mainly during the coffee-picking season, while in other sites, 
labor scarcity does not seem to be much of a problem, probably due to the remoteness that 
makes off-farm alternatives scarce. Labor is mostly hired for working on cash crops (banana 
and coffee), for food crops, and for soil preparation. Contracts may be per piece oflabor, per 
land area, daily, monthly, or, rather rarely, permanently. Payment is mainly made in cash, 
but sometimes also in kind, especially in times of food scarcity. The other major constraint 
apart from land is liquidity that imposes a bottleneck to input use (including labor, so that 
management of coffee and banana is affected) and investments. 
Cropping patterns are quite similar across sites. The most prevalent system is intercrop-
ping, with a few exceptions in banana and coffee that are discussed further below. The 
main reason for intercropping is land scarcity, and it appears that land-use management in 
terms of complementary intercrops (e.g., legumes to fix nitrogen) is hardly known amongst 
farmers; at least it was never explicitly mentioned during the survey. Sale cropping mainly 
occurs in banana and in coffee systems. Although they were predominantly intercropped 
in the fust two sites, coffee was mainly planted sale in the third site. 
Crops differ only slightly from site to site. Apart from banana and coffee, millet, maize, 
groundnut, beans, and horticultural products like avocado, orange, pineapple, mango, 
passion fruit, and guava are grown. In all the sites, one has to discriminate cash and food 
crops. Coffee and banana are major cash crops on all the sites. Food crops are banana, 
cassava, millet and the horticultural products, whereas cash crops are (according to the 
farmers'ranking) banana, beans, coffee, livestock products (especially milk), millet, and 
groundnut. As stated above, most of these crops are intercropped with at least one other 
crop. Although the main reason mentioned for this is land scarcity, intercropping systems 
can be quite complex, as the following example shows. By visiting one farmer, we found 
beans are grown in a banana plantation. But the beans are not grown by the owner of the 
plantation, but by the day laborers, mostly women. The day laborers take the grains quasi 
a salary, while the banana plantation benefits through the application oflabor and from the 
nitrogen fixing effect ofthe beans as the residues are used as mulch. The same complexity 
-8-
can be shown for banana--coffee interactions. To start with some (yet) separate aspects; we 
look at the economic functions of banana and coffee as mentioned by the farmers. 
Most bananas grown at all sites are ofthe matooke (cooking banana) type, followed by brew-
ing banana (mbiri), large desert bananas (bogoya) and, on a smaller scale, small desert or 
apple bananas (ndizi). Yet, the shares shift across sites. In the first site, about 75% of the 
bananas grown is matooke, about 11% are mbiri, the same share is held by bogoya, and the 
remaining 3%, is held by ndizi. In the second site, matooke holds about 80% of production, 
the rest is mainly dedicated to bogoya. The other types are not considered as profitable by 
the farmers at this site. In the third site, between 40 and 60% of the bananas grown are 
matooke, followed by mbiri (2~30%), which is seen as very profitable at this site, bogoya 
(1~20%), and ndizi (10%). The main agronomic constraint to banana production is banana 
weevil, followed by wilts, nematodes, droughts, and windbreaks. 
As for coffee, the dominant variety is robusta; the higher quality yielding arabica is very 
rarely found. An improved variety of robusta, colono, is also found on the sites, but to a 
lesser extent than robusta, as colono requires high management (which is not available 
due to liquidity constraints and the consequent labor shortage), and is susceptible to all 
kinds of agroclimatic risks, especially drought. The biggest problem, apart from manage-
ment, seems to be coffee-wilt. One of the agronomic advantages of coffee is that it can be 
grown on marginal soils. 
Banana and coffee are, beside beans, the two most important cash crops in the region. But 
this is not the only feature they have in common. These crops interact in many ways, be 
it in economic terms or in agronomic terms as intercrops. 
Concerning the agronomic interaction, we have to first look at the banana--coffee-inter-
cropping systems. Again, sites differ in the scope of banana--coffee intercropped from 70% 
of the farmers having coffee and banana intercropped, to about 30% of farmers who inter-
crop banana and coffee. The rest of the farmers grow them on pure stands. These figures 
depict well the controversy of opinions and perceptions about agronomic banana--coffee 
interactions_ There appear to be many agronomic factors affecting the relationship between 
banana and coffee, both positive and negative, both applying to intercropping and sole 
cropping relationships. 
First, there seems to be a certain competition between the crops on intercropped stands. 
They obviously compete for soil fertility, while coffee is often believed to deplete soils and 
eventually kill intercropped bananas. Banana seems to throw shade on coffee, which is 
advantageous for the younger plants or for colono--as colono is trained to have lower 
bushes-but disadvantageous for older and higher coffee plants. Bananas are also seen to 
retain moisture and thus provide a certain shelter for coffee in times of drought. Further, 
both banana and coffee provide mulch for each other, bananas in terms of leaves and cut 
stems, coffee in terms of coffee-husks (which cannot be spread on coffee due to the conse-
quent spreading of coffee wilt). 
The most important economic relationship between the two crops is that they provide 
seasonal liquidity (especially coffee) for their mutual benefit on management. During the 
whole year, banana sales provide liquidity for hiring labor for banana and coffee manage-
ment, coffee provides liquidity for banana management during the coffee-harvest season. 
The same controversy as for agronomic interactions is found when it comes to the discus-
sion of economic features of the two crops and their profitability. 
-9-
Comparing the mere cash revenues of banana and coffee, coffee generally seems to be 
the slightly more profitable crop. Here, it can be seen that coffee yields, according to site, 
between 600 000 and 1 400000 Ush/acre/year in cash, while banana yields between 400 
000 and 1 000 000 in cash on the same area. But if the additional amount of food is calcu-
lated and valued that banana gives, the picture changes rapidly. Assuming that 15 to 50% 
of the harvest is home consumed, we can add a value of about 150 000 to 320 000 Ush to 
the economic yields of banana, which makes banana almost as profitable as coffee. It can 
thus be assumed that farmers renounce possible coffee profits for the sake ofliquidity and 
food security, although this renunciation is at current prices rather small. 
This leads to a summary of economic functions that comprise but go beyond profitability for 
both crops. While coffee provides cash even from marginal sites, it is also seen as a saving 
device and a source offunds for investments that have to be made once a year. Banana gives 
cash, but also provides food security. Another major economic function of bananas is that 
they can be harvested throughout the year and thus provide a continuous flow of liquidity. 
Finally, banana provides by·products like leaves that can be used as mulch, animal feed, 
and fibers. This underlines the preference for banana versus coffee of most farmers. Even 
if coffee prices rose, so most of the farmers stated, banana would still be cultivated to cover 
permanent liquidity needs and to assure food security. 
Marketing, postharvest, and processing of banana and coffee. In this section, the farmers 
were asked about the products marketed, the organization of banana and coffee marketing, 
problems with this organization, price developments, and seasonality of the two products. 
Further, problems of farmers' organizations were discussed. Another topic was postharvest 
and processing. 
In general, the farming systems on the survey sites can be considered as both market and 
subsistence oriented. Apart from coffee, of which the marketed share is naturally 100%, 
be it fresh or dried, of other crops like beans, 50% of the harvest is marketed. In the case 
of banana, 50-85% of the harvest is sold, depending on the site and, of course, on the rela-
tionship between land endowment, land productivity, and the nutrition requirements of 
the owner and his or her family. 
The way the farm produce is harvested is quite homogeneous across the sites, with a slight 
difference in coffee marketing at the first site. Generally, products are sold at the farm gate, 
without having a market place. Brokers go around the farms and negotiate the price on 
behalf of the traders. As soon as there is agreement on the price, the crops are cut (in the 
case of the banana) and then sold on the spot. The system is the same for all other crops in 
two of the sites, while there is a slight difference on the third site, where coffee is taken to 
the trading post, gathered, and stored before selling to traders. This is the only exception 
to the rule of farm gate selling. 
Farmers encounter many problems in this system. Above all, there is a huge asymmetry 
in information, as banana prices in Kampala are only known by the brokers or traders, 
not by the farmers. The farmers have to take what they are offered by the brokers, the 
latter having at least a temporal monopoly at the farm gate. This depresses the prices 
the farmers earn for their product, as "the buyer determines the price". Other problems 
encountered by the farmers are renegotiation of prices when bunches are already cut (e.g., 
at collection centers), that they have to pay brokers in-kind (i.e., with bunches), as well 
as theft during the high price season, failure of payment, and losses of cut bunches while 
-10-
waiting for brokers to come. The problems encountered in coffee marketing are mainly 
similar concerning the weak marketing position of the farmers. 
Banana, especially matooke, prices show significant seasonality. From September to 
March, prices range from 2000 to 3000 Ush per bunch (depending on size and quality of 
the bunch). They go down to 5(}-250 Ush per bunc, during the season when large quanti-
ties are harvested, i.e., April to August, yet price fluctuations slightly differ across sites. 
Moreover, additional price deductions of up to 25% during the rainy seasons are reported 
due to poor accessibility through roads. 
Other banana types, e.g., bogoya, show lower seasonal price variability. Coffee has only one 
main harvesting season in the region, so there is no seasonal variation. However, through 
the harvesting season, price fluctuations are reported, astonishingly of two different kinds. 
While at one of the sites, prices at the beginning of the season are lower (at about 200 
Ushlkg) and rise towards the end of the season to about 300 Ushlkg, in other sites prices 
fall during the season from 300 to 150 Ushlkg. The explanations for this difference have to 
be sought in the liquidity pressure at the beginning of the season, which seems to be high 
in the first case: farmers have no cash reserves and are thus forced to sell off at almost 
any price. Later in the season, liquidity pressure is no longer so high; small quantities of 
coffee are available, while during the season, quantities rise and thus prices decline. It 
would be interesting to see why these different price developments in adjacent regions are 
not balanced by interregional trade. 
Besides these seasonal price volatilities, one has to consider long·term price developments 
of both banana and coffee. Also here, perceptions differ across sites. While at one site, both 
banana and coffee prices are considered to have declined through the last decade and thus 
negatively affected profitability of both crops, at the other two sites, banana, particularly 
matooke, is considered as getting higher prices as ten years ago. This is said to be due to 
better infrastructures and higher demand in the urban regions. During the same period 
coffee seems to have experienced a sharp decline in price. It was mentioned that matooke 
prices have risen from 500 to 2500 Ush per bunch. Coffee prices were reported to have 
declined from 800 to 300 U shlkg·. 
Organizations and marketing boards, e.g., a cooperative, that could counterbalance these 
negative developments, especially in coffee prices, and increase the market power in other 
products' trade do not exist in any of the three sites. Cooperatives seem to have disinte· 
grated during the early nineties, mostly due to corruption and mismanagement. While at 
one site farmers are about to revitalize the cooperative system, and hope for more market 
power and better prices, farmers in other sites refrain from doing so. They argue that 
during high price seasons, there is no need for collective action, and during the low price 
seasons, collective action would be pointless due to the overall surplus and price depression 
throughout the region. In economic terms, these farmers see themselves as price takers 
in a highly atomized market. 
As far as postharvest activities and processing are concerned, these are considerably low at 
all sites. Coffee is partly dried, partly sold fresh. There does not seem to be many problems 
in storing coffee, especially when dried. Banana is processed into beer in case of the mbiri 
variety, which seems to be a quite profitable business. Future requirements mentioned by 
the farmers were mainly for processing matooke to improve storage capability and reduce 
postharvest losses, so that low price seasons can be overcome by adding value or storing 
until the higher price season. 
-11-
In conclusion, it can be said that systems, both farming and marketing, are complex and 
diverse and a lot of research is required to provide optimal support for the development 
of these systems. It seems that both the economic and agronomic potential especially of 
intercropping systems is not yet fully exploited. Banana and coffee playa major economic 
role as cash crops (both) and food crop (banana) on all the sites. Whether they are grown 
intercropped with each other or cultivated on pure stands, banana and coffee have many 
economic and agronomic relationships, both positive and negative. These relationships have 
to be further developed and improved. As for the economic part, the following assumption 
can be made: 
Behavior patterns of the farmers are not only for profit orientation, but liquidity and food 
security also play an important role in economic decision· making. The major constraints to 
production are land scarcity and poor marketing options and therefore a lack of liquidity, 
which leads to poor crop management and poor exploitation ofthe economic potential 
Economic hypotheses to be tested during further research are the following: 
• Improved crop management could increase land productivity and thus remove the 
bottleneck of land scarcity. 
• Improved marketing and organization would increase cash income and thus remove 
the bottleneck of liquidity and add resources (esp. labor) to improve management and 
input utilization. 
1.4 Cassava market studies 
by N. Mahungu 
Market surveys to quantify demand and product type are being conducted in Malawi, 
Tanzania, and Zambia. This will complement the findings of the cassava and sweetpotato 
subsector analysis conducted in these countries in 2002 whose final reports have been 
distributed to stakeholders and important libraries in these countries. 
To improve on marketing of cassava products, SARRNET is also working with the Malawi 
Bureau of Standards to formulate grades and standards for cassava flour and cassava 
starch. The grades and standards for edible flour have been published which gives speci· 
fications on safety levels, packaging, transportation, storage, and labeling requirements. 
In Zambia work is also done in close collaboration with the Zambia Bureau of Standards 
(ZBS). 
1.5 Research on harmonized policies for increased commercialization and 
trade among the SADC countries 
by V. Manyong, G. Muchopa, T. Takavarasha (FANRPAN), and FANRPAN country 
nodes of five SADG countries 
Countries such as Angola, Malawi, Mozambique, and Zambia have suffered from a series 
of natural disasters that have also been compounded by poor short-term policy decision· 
making in the past decade. Often policies are country specific while the effects of climate, 
pests, or trade span borders. In periods of severe crisis, such as the most recent drought, 
specific trade policies appear to be inappropriate to address issues of a regional dimen-
sion. It is therefore important to better understand both country level agricultural policies 
and regional policies and their contribution to the mitigation of the food crisis that has 
emerged. 
-12-
This study is targeting five countries of the SADe region (Botswana, Malawi, Mozambique, 
Tanzania, and Zambia). The beneficiaries ofthis study are the smallholders who can diver-
sify their rural economy in order to become less vulnerable to natural disasters. The goal 
is to provide critical policy analysis leading to dialog and enhancement of policies within 
selected SADe countries. 
Methodology/implementation strategy. The study involves the GIS mapping ofthe countries 
in development domains and identification of commodities with a comparative advantage 
in the target countries and the region. Many policy consultations will be organized at the 
country level and for the region as a whole. The study will include policy analysis, policy 
technology formulation, communication to national governments, a regional Ministerial 
conference, and follow up (with SADe) of Ministerial decisions with national govern-
ments_ 
Expected results/deliverables in 2004. Three intermediate results are expected: (1) Policy 
constraints evaluation and prioritization, (ii) Effective policy dialog initiated and strength-
ened in and across the target countries, and (iii) Mechanism established for monitoring 
and evaluation of policy technologies. 
Milestones for coming year. Organization of a regional methodology workshop, country 
technical reports followed by country policy workshops, synthesis in a regional report and 
organization of a Ministerial conference, and development of a database for agricultural 
statistics in the four pilot countries. 
Key findings to date in 2003. The workplan for the implementation module was developed 
and discussed during the inaugural meeting of the project, which took place at Lilongwe 
on November 2003. Representatives of FARNPAN nodes from five countries endorsed the 
workplan: Botswana, Malawi, Mozambique, Tanzania, and Zambia. 
Georeferenced data were collected from various sources in the websites. Preliminary GIS 
analysis was conducted for five potential countries using those secondary data. Results 
indicated that Botswana falls outside of potential development domains for this project. 
Therefore, this country was not retained for the first year of this project while four coun-
tries were selected: Malawi, Mozambique, Tanzania, and Zambia. 
2 Commercially viable small- and medium-scale agroenterprises 
that expand trade of the starchy and grain staples enhanced 
2.1 Rural agroprocessing of cassava flour 
by S. Kolijn, v: Rweyendella, and national research partners (Root Crops team SRI· 
Kibaha and TFNC) 
Based on the successful establishment of a cassava chip·processing group in Bungu Village, 
Rufiji District, 140 km south of Dar es Salaam, three additional processing groups were 
established in Bungu ward. The three new groups started processing dried cassava chips 
in February-March 2003 and were successfully linked with the identified flour millers in 
Dar es Salaam for selling into urban supermarkets and food stores. 
Two SUA students studied and documented the performances and operations of the four 
groups and a report was developed in AprillMay 2003. 
Key findings and development of events. Making chips has raised the farmer's income 
significantly from around 280 USDlha (in case of selling the fresh roots to local traders) to 
-13-
above 800 USDlha when processing and marketing of dried cassava chips to urban flour 
millers. The vertical integration through value adding has created additional local rural 
labor (harvesting, peeling, and processing) and has resulted in expansion of the total 
production of cassava among the members (all farmers intended to double or triple their 
area under cassava production), less dependency of local traders, ability to process older 
stocks of cassava fields into chips (traders in fresh cassava don't like to buy cassava> 12 
months). The four groups organized themselves into a processing association to oversee and 
coordinate the transportation and marketing of chips to Dar es Salaam, and coordinate the 
communication with SARRNET, government authorities, and traders. The Rufiji District 
Council backed the initiative further by providing micro loans to the four groups as a kind 
of operations capital fund. 
Despite the fact that the four groups have only 50 members, many more additional non· 
member villagers have benefited from the presence of local chipping processing facilities 
as the group buy cassava roots from nonmembers andlor nonmembers can get their roots 
processed there for a fee or can get employed as casual workers (peeling, harvesting, etc.). 
By the end of 2003 the four groups had produced around 60 t of dried chips. 
In MayfJune 2003 SARRNET started exploring further market outlets in Dar es Salaam 
in an effort to expand urban demand (especially among the lower income groups) but, due 
the much lower maize prices, this effort remained largely without much impactfsuccess. 
(Maize prices retail remains under 0.30 USDlkg). Mid·2003, farmers hesitated to process 
more chips due to higher prices of fresh cassava roots due to lower stocks in their fields. 
Despite the fact that cassava flour is only available at a relatively high retail price (0.40-
0.55 USDfKg) and has only been bought by the middle·income consumers who are more 
quality concerned; the full market potential of high quality cassava flour still needs to be 
explored and developed for the much larger lower income consumer group who are looking 
for an affordable starch food (to be consumed in ugali form like maize). 
This requires more efficient production and marketing of cassava chipsfflour (right now 
the two flour millers take more than 200 USDft for their milling and marketing operations 
while the larger scale maize and wheat flour normally take up to 20%). 
Intermech Engineering in Morogoro further developed its expertise and capacity with 
regular technical backstopping by me, manufacturing the various cassava processing 
equipment. The fruitful collaboration was further strengthened through the joint visits to 
farmer groups, postharvest meetings, and agroequipmentfagricultural showsfexhibitions. 
SARRNET facilitated further in the importation of petrol engines at an affordable price. 
Many visitors paid a visit to Bungu during 2003 to get first hand experience about the 
project approach and implementation, and the experience of the four farmers' groups. The 
following NGOs and individuals visited Bungu: Norwegians People Aid Program-Dodoma 
team, Concern, Action Aid, 12 farmers from The Initiative for Development and Equity in 
Mrican Agriculture (IDEAA) Project Malawi, representatives of the Ministry of Agricul· 
ture, Rockefeller Foundation, IMPACTIUSAID team, commercial flour millers, and Peter 
Chisawillo (Intermech Engineering). The general outcome of these visits is that several 
NGOs are currently implementing this pilot· processing project in other areas of Tanzania 
and Malawi. 
In May and July 2003 the team visited Mtwara Region (southern Tanzania) to assess the 
cassava sector in Southern Tanzania, and sensitize Provincial and District Governors 
-14-
and agricultural staff on what could be done with the introduction of a market-oriented 
processing strategy to their regions_ The CFC funded project on small-scale cassava flour 
processing is building further on these initial meetings. 
In January 2004, David and Marlene Mowbray documented the work of two of the four 
groups in Bungu. (The lead story in the IITAAnnual report 2003 will cover this). 
2_2 Development of a large-scale cassava starch industry in Tanzania 
by S. Kolijn, E. Kanju, V. Rweyendella, Root Crops team SRI-Kibaha, and Dr 
Subbiah (METL) 
During 2003 the partnership with METL in developing a larger scale, commercial cassava 
starch industry was further strengthened by the ongoing agronomy, selection of most suit-
able varieties, and long-term, soil-fertility trials. 
This work has been developed in close collaboration with the Root and Tuber program at 
SRI-Kibaha. The fertility trials, using MOP, NPK, and a combination ofthe two, were har-
vested during November 2003 and despite the extreme low rainfall throughout 2003 (less 
than 260 mm!) the yields were above expectations (maximum 35 t of fresh roots). Further 
new, detailed soil fertility-production trials, with the help and input of Anneke Fermont, 
had to be postponed due to the lack of rain in March 2003. 
This unique and new partnership has attracted the interest of many collaborators and 
donors and hopefully will lead to additional donor funding (e.g., RF). IITA could provide 
technical backstopping in the development of an outgrowers scheme to supply the 20 000 
t native starch capacity. 
Technical inputs by Anneke Fermont, John Wendt, Edward Kanju (IITA), and local staff: 
Mr Vianey Rweyendella, and SRI-Kibaha). 
2.3 Introduction of food extrusion technology in Tanzania 
by S. Kolijn and S. Ferris 
One of the approved FoodNet projects was an extrusion project developed for and with 
Powerfoods Ltd; a medium-scale flour miller in Dar es Salaam. Food extrusion technology 
is aimed at improving the food quality of flours by a quick and intensive heat treatment 
(based on shear and compaction). The extruder and cleaner unit were imported from UK 
and the equipment was commissioned in March 2003. In May the DG visited the site and 
met Mrs Anna Temu, the General Manager of the company. By June, just after the official 
inauguration of the equipment, Powerfoods got its first contracts to supply UNHCF and 
WFP as it now could meet the UN food standards for UNIMIX and other fortified flour 
blends. All standard flours retailed by Powerfoods now undergo the extrusion process. This 
technology has boosted the business of Powerfoods, as the company is now able to compete 
with imported foodstuffs like weaning food, high quality flour porridges, and refugee flour 
blends. In addition to the FoodNet extrusion project in Dar es Salaam two local agroprocess-
ing manufactures developed a strong interest in food extrusion after the FoodNet sponsored 
study tour to Vietnam (November 2002) and, in 2003, several small units were constructed 
and tested by Pemba Engineering (Mwanza) and Intermech (Morogoro). 
-15-
2.4 Inaugural stakeholders' workshop of the Common Fund for Commodities 
(CFC)-funded regional project on "Small-scale cassava processing and 
vertical integration of the cassava subsector in Eastern and Southern 
Africa" 
by A. Abass, R. Ranaivoson, A. Zacarias, N. Mlingi, A. Agona, M. Zulu, and J. 
Whyte 
The Small·scale cassava processing and vertical integration of the cassava subsector in 
Eastern and Southern African countries (ESAC) of Madagascar, Mozambique, Tanzania, 
Uganda, and Zambia is a cassava commercialization project within the R4D agenda of 
UTA. The project seeks to develop the income generating potential of cassava as a cash 
crop in these countries. The strategy to achieve this is to be tested in pilot operations by 
making simple market·oriented processing technologies and appropriate marketing strate· 
gies available to smallholder farmers, processors, and farmer cooperatives. It is expected 
that participating smallholder farmers and processors will be able to transform highly 
perishable fresh cassava roots into shelf· stable, market· grade, intermediate products like 
chips, starch, or flour and be able to access both the national and the regional markets. 
Phase I of the project, is being executed in collaboration with five national institutions; 
Centre National de Recherches Appliquees au Developpement (FOFIFA) in Madagascar, 
Instituto Nacional de Investigacao Agropecuaria in Mozambique, Tanzania Food and 
Nutrition Centre (TFNC), Kawanda Agricultural Research Institute (KARl) in Uganda, 
and Department of Research and Specialist Services, Food Conservation Unit in Zambia. 
The inaugural stakeholders' workshop of the project was held in Dar es Salaam, Tanzania 
from 24 to 25 November 2003. The meeting brought together a coalition of stakeholders 
from the five participating countries, the International Institute of Tropical Agriculture 
(UTA· Malawi, Tanzania, Uganda), the Common Fund for Commodities (CFC·Netherlands), 
Food and Agriculture Organization (FAO·Ghana), the Natural Resources Institute (UK), 
regional cassava networks (i.e., SouthernMrica Root Crops Research Network-SARRNET 
and Eastern Mrica Root Crops Research Network-EARRNET), UNIDO, other national 
and international organizations. The meeting discussed the work plan for Project Year 
1 (PYI) and appointed the Steering Committee (SC) members. Some activities approved 
for PYI include mobilization and formation of farmers, processors, and end·users into 
groups, training sessions for farmers and processors, introduction of technologies and 
machinery for production of high quality cassava flour (Madagascar, Tanzania, Zambia), 
rali (Mozambique) and chips (Uganda), and market creation for the products. The 12 
member SC was elected in a manner that gave the se a wide representation of the entire 
stakeholder' categories. The CFC, FAO, NRI, UTA, EARRNET, SARRNET, private sector, 
public sector, research/universities, farmers/processors, and NGOs were all represented in 
the SC charged with the responsibility to oversee the project activities for the three·year 
period of Phase I. 
2.5 Training on cassava processing, market expansion for cassava products, 
and development of a combined cassava grater and dewatering device 
by A. Abass in collaboration with V. Rweyendela and P. Chisawilo 
Most farmers in Eastern and Southern Mrican countries (ESAC) have very little knowl· 
edge of cassava processing and are therefore unable to take advantage of the huge latent 
market opportunities for processed cassava in the industrial sector. Where processing is 
done, the quality and safety of the products are extremely poor. This is responsible for the 
-16-
rebuff of cassava in the industry and prevalence of diseases associated with consumption 
of improperly processed cassava foods. The situation is made worse by lack of appropriate 
processing machinery such as graters and dewatering devices. 
By third quarter of2003, IITAconducted training of trainers for staff ofTanzania-NARS and 
some farmers' groups on the grating technique for production of high quality cassava flour 
(HQCF), starch and gari. Prior to this, the chipping technique that was popularized during 
the SARRNET project phase was already in commercial use but the grating technique was 
not. Seven extension and research staff of Tanzania-NARS were trained on production of 
industrial cassava starch while three farmers' groups in Msongola area of the Coast region 
of Tanzania were trained and assisted to set up HQCF and gari processing enterprises. 
The Root and Tuber group, farmers, and processors have commenced test marketing of 
these cassava products to textile mills (starch); supermarkets, biscuit factories (HQCF) , 
and consumers (gan). The textile factories are at the moment demanding for more starch 
to conduct larger tests. 
Rapid dewatering is one of the key operations in the production of HQCF and certainly 
the most tedious. It is the second most important rate-determining factor after drying in 
the production of HQCF. With IITA backstopping, Intermech Engineering Ltd based in 
Morogoro, Tanzania embarked on a project to design and fabricate a "combined grating-
dewatering device" which would make it possible to carryout both grating and dewatering 
operations in a single, rapid, and continuous operation. The device comprises of a transport 
system, which also squeezes the cassava mash through a series of screens thereby enabling 
the water to be removed (Fig. 1). 
About 50% dewatering efficiency was achieved with the first design but further improve-
ments on the device are continuing. It is hoped that the technology, when perfected, will be 
transferred/disseminated to other parts ofSSA where cassava processing and consumption 
are being promoted. 
Figure 1. The new cassava-dewatering device by Intermech Engineering Ltd, Tanzania. 
-17 -
2.6 Cassava pilot production and processing facilities 
by N. Mahungu 
Four pilot production and processing sites for processing high quality cassava flour and 
chips for industrial use have been formed. These are Kapili, Mthiramanja, Mbawa, and 
Mpamba in Malawi. In Tanzania, in addition to the Ubungu farmers associations, pilot pro-
cessing centers have been established in Kigoma and Tanga regions, as well as Mansa town. 
In Zambia, the sites are in Western (Kaoma), Northwestern, and Luapula provinces. 
In Malawi, two levels of pilot processing sites have been established. The first level is the 
one described above where individual farmers are organized into associations and are 
using improved production and processing technologies and making high quality chips. 
The second level is referred to as a secondary pilot· processing center, which will involve 
promotion of mid-entrepreneurs. The secondary pilot processing center will be buying cas-
sava processed products from the primary pilot processing center to further process into 
flour, package, and display/sell in various markets including supermarkets. 
Improved processing machines such as graters, chippers, and leaf choppers are being pro-
moted in these pilot-processing centers. Training of local fabricators for these machines 
is still ongoing. 
SARRNET is promoting cassava·based products such as cassava silage for livestock feed 
with dairy farmers in Malawi in five sites: Doroba, Chimbiya, Chigumula, Chileka, and 
Mpemba. Twelve farmers are involved including one large· scale commercial dairy farm-
Katete Farm-where 5 ha have been planted with cassava for cassava silage. 
SARRNET is promoting both sweet and bitter cassava varieties including yellow cassava 
varieties, which have Beta·carotene. The sweet varieties such as Mbundumali are being 
promoted for the fresh market but also for the bakery and biscuit industries. The bitter 
varieties Sauti, Yizaso, and Mkondezi in Malawi are being promoted for industrial appli-
cation (chips and flour) such as in the plywood industry. One variety, Mkondezi, has been 
found to have a better binding effect in the plywood industry compared to other cassava 
varieties. The bitter varieties are also used in processing fermented flours. 
3 Broad-based and special trait genotypes and populations tar-
geting the major production systems and market opportunities 
developed and disseminated 
3.1 Identification of sources of resistance to Radopholus similis in Musa 
germplasm 
by C. Dochez, M. Pillay, J. Dusabe and J. Whyte, in collaboration with D. De Waele 
(KU Leuven, Belgium) 
The East Mrican highland banana is the most important staple crop in the East Mrican 
Great Lakes Region. However, nematodes are a serious constraint to sustainable banana 
production. Radopholus similis is commonly known as the most damaging nematode spe-
cies, causing severe yield losses. The use of host-plant resistance provides promising pros-
pects as a basis towards sustainable nematode management and improved banana produc-
tion. Using a rapid and reliable screening method based on the inoculation of individual 
roots, screening of available germplasm for resistance to nematodes has been markedly 
-18-
improved and efficiency increased. This screening method was used to evaluate available 
germplasm and newly developed hybrids for resistance to R. similis. All five clone sets of 
the East African highland bananas, including cooking and beer types, are susceptible to 
R. similis. Crosses between East African highland bananas and Calcutta 4, a diploid wild 
banana with resistance to R. similis, have resulted in resistant tetraploid hybrids. Resis· 
tance was also identified in several diploid hybrids, which were used to further improve 
the tetraploid hybrids. 1b date, five secondary triploids with resistance and seven with 
partial resistance to R. similis have been identified. In addition, new sources of resistance 
have been identified, mainly among germplasm from Papua New Guinea. An overview of 
the host responses of the evaluated germ plasm is summarized in Table 1. 
Table 1. Overview of host responses of landraces, wild bananas and UTA's breeding 
materials to RadopholuB similis based on the individual root inoculation method. 
Resistant 
Reference cultivars 
Yangambi km5 (AAA) 
Gras Michel (AAA) 
Calcutts 4 (AA) 
Partially resistant 
East African highland bananas (AAA) 
Tetraploid hybrids 
TMHx917K-2 
TMHx 660K-l 
TMHx 4349S-2 
TMHx 1977K-l 
Qi{>loids 
TMP2x 1549-7, TMB2x 3107S-4 
TMB2x 1968-2, TMB2x 2582S-1 
TMB2x 5265S-1, TMB2x 25378-1 
TMB2x 8075-7, TMB2x 4443S-1 
SH 3142, TMB2x 25698-1 
TMB2x 9128-3, TMB2x 2569S-2 
Secondary triploids 
72698-15, 76048-4 
83868-5t-,. 83868-22 
12010Pr79 
Wild bananas and landraces 
Marau, Pora Para, Kokopa, Pisang 
Mas, Saba, Gia Hiu, M.a.ssp. 
burmaniea, Galea, M.a.ssp. 
malaeeneis, Vudu papua 
TMHx222K-l 
TMHx 1438K-l 
TMB2x 1297-3 
TMP2x 1518-4 
12010PP30 
2156K-41,2409K-3 
8386S-15,9494S-1O 
8386S-19,83868-50 
Pitu, Yalim, M. balbi-
siana, Caehaca, Yanin 
Yefan 
Susceptible 
Valery (AAA) 
Endirira, Enzirabahima, Kazirakwe, 
Mbwazirume, Nakawere, Nfuuka, 
Nakitembe, Nakabululu, Musakala, 
Kabula, Tereza, Kabucuragye, 
Enyeru, Kibuzi, Nakyetengu, 
Kisansa, Namwezi, Nakayonga 
TMHx 401K-l, TMHx 246K-l 
TMHx 1201K-l, TMHx 376K-7 
TMHx 365K-l, TMHx 199K-4 
TMHx 2048K-2 
TMB2x 7197-2 
TMB2x 8848·1 
TMB2x 9839-1 
TMP2x 9722-1 
8386S-4 83868-27 
1201OPP81, 12010PP85 
7798S-2 
Gunih, Mshale, Pama, 
Durningi, Kikundi, Wambo, Merik, 
M.a.ssp. truncata, Pagatau, 
Garunga, Ambiri 
3.2 Identification of mechanisms of resistance to nematodes in Musa 
by C_ Dochez, M. Pillay, J. Dusabe, A. Tenkouano, and J. Whyte, in collaboration 
with D. De Waele (KULeuven, Belgium) and Makerere University Kampala, 
Uganda 
Studies on the mechanisms of resistance will help the breeding program to select for a 
desired feature. Lignin might form a physical barrier for nematode penetration, i.e., pre-
infectional resistance. Formation of phenolic compounds might refer to induced resistance 
as a response to nematode penetration. In a first set of experiments, the attraction and 
-19-
penetration ability of R. simi lis was compared among resistant and susceptible cultivars. 
No significant differences in attraction and penetration of R. similis were observed between 
resistant and susceptible cultivars. Similar invasion rates of R. similis on resistant and 
susceptible cultivars, suggests that the resistance is not due to physical or mechanical 
barriers. Histochemical experiments were carried out to detect differences in lignin and 
phenolic compounds between susceptible and resistant culti vars. At different times after 
inoculation with R. similis of resistant and susceptible cultivars, root sections were taken 
from near the corm, in the middle of the root, and at the root tip. Detection of lignin was 
done using safranine with counter stain fast green. Phenolic cells were detected using ferric 
chloride. An increased number of phenolic cells were observed after nematode inoculation 
in resistant cultivars. The highest number of phenolic compounds was observed in the 
resistant diploid hybrid TMB2x 9128·3 and the reference cultivar Yangambi Km5, both 
showing resistance to four R. similis populations from Uganda. No lignified cells were 
observed in the cortex of the root. Lignification was only observed in the endodermis, 
especially in susceptible cultivars. 
3.3 Genetic analysis of nematode resistance in a segregating population of 
Musa 
by C. Dochez, A. Tenkouano, M. Pillay, and J. Whyte, in collaboration with Dirk De 
Waele (KU Leuuen, Belgium) 
A diploid banana population (TMB2x 6142·1 x TMB2x 8075· 7) with 81 progeny was screened 
for resistance/susceptibility to R. similis to find out if the gene(s) for nematode resistance/ 
susceptibility may be segregating in this population. Screening of this population and 
determination of segregation ratios will provide information on the inheritance and number 
of alleles controlling this trait in banana. Eighty·one progeny have been tested with the 
individual root inoculation method, and the reproduction ratio of R. similis on each progeny 
was compared with the reproduction ratio of R. similis on the resistant reference cultivar 
Yangambi km5 and the susceptible reference cultivar Valery. The female parent TMB2x 
6142-1 is a cross between Nyamwihogora (EAHB) and Long Tavoy and is susceptible to R. 
similis. The male parent TMB2x 8075·7 is resistant to R. similis and derived from the cross 
between SH 3362 (PJB pedigree) and Calcutta 4, which are both resistant to R. similis. 
Of the 81 hybrids evaluated, 37 were resistant, 13 were partially resistant, and 31 were 
susceptible to R. similis. After chi-square analysis, these results indicate that resistance to 
R. similis is controlled by two dominant genes, A and B, both with additive and interactive 
effects, whereby recessive bb suppresses dominant A (either A- or B- required for partial 
resistance, both A· and B· confer full resistance, but bb suppresses A-). 
3.4 Identification of sources of resistance to Pratylenchus goodeyi in Musa 
germplasm in Rwanda 
by C. Dochez, J. Dusabe, M. Pillay, and J. Whyte, in collaboration with Dirk De 
Waele (KU Leuuen, Belgium) and Sueta Gaidashoua (ISAR, Rwanda) 
A yield loss trial in a farmer's field was established at Kibuye, Rwanda, in September 
2003. This area showed the highest densities of Pratylenchus goodeyi during a survey 
carried out in 2001. No other important nematode species (as Radopholus similis and 
Helicotylenchus multicinctus) were present in this area. Three varieties ofthe EastAfrican 
highland bananas (Mbwazirume, Ingaji, and Intuntu) were planted in a nematode infested 
plot and in a noninfested plot. Nemacur is used to control nematodes in the noninfested 
plot at a rate of 15 g active ingredient (phenamiphos) twice a year. Six plants per variety 
-20-
were planted in three replications. Within the infested and noninfested plots, subplots with 
mulch and no mulch were established. Data will be taken at nine months after planting, 
at flowering, and at harvest of the motherplant and fIrst ratoon. 
A screening method is being developed for evaluating Musa germplasm for resistance to 
P. goodeyi. 
3.5 Studies on expression of host plant resistance 
by C. Gold, in collaboration with G. Night and A. Powers 
Antibiosis has been advanced as the most important pathway in the resistance of banana-
to-banana weevil. Treatment of eggs with sap Oatex) from resistant cultivars has been 
reported to reduce hatchability. However, it is not clear whether the negative effects of sap 
are due to its physical properties, chemical elements, or both. Furthermore, an antibiotic 
mechanism in some cultivars was suggested based on high performance liquid chromatog-
raphy (HPLC) analysis of extracts from various cultivars. Increased developmental time, 
reduced vigor, and increased mortality of larvae were observed when larvae were reared 
on corm tissue of resistant cultivars. Antibiosis may be expressed either as a chemical 
toxin or antifeedant. 
In addition to chemical composition of plants, physical factors such as corm hardness may 
confer antibiotic resistance to banana weevil. Increased corm hardness may lead to failure 
of larvae to tunnel and feed. Tissue toughness may lead to wearing of larval mandibles 
and reduced feeding. 
In banana, patterns of corm damage by weevils suggest uneven distribution of plant 
defenses. Distribution of damage in the plant may also influence yield loss. Additionally, 
differences in oviposition and larval performance on different phenological stages have 
been reported. By evaluating resistance at different growth stages, useful patterns can be 
identified that have applications for breeding programs. Understanding of these phono-
logical patterns is also important for purposes of pest management and timing of control 
measures. 
Expression of resistance is a result of interplay between genetic traits and environmental 
factors. One of the environmental factors that may influence resistance of banana plants 
to the weevil is plant nutrition. Given a limited set of resources, plants might change their 
allocation pattern in response to such factors as herbivore attack. 
The objectives of this ongoing student research activity are: (l) to characterize the chemical 
and physical factors responsible for banana resistance to weevil; (2) to study the distribu-
tion of resistance factors within the plant and variation of resistance with plant phenology 
for selected cultivars; and (3) to determine the influence of plant nutrition on expression 
of resistance. 
Observations were made on characteristics of sap: color, quantity, viscosity and speed of 
coagulation, and dispersion. The study was carried out in the dry season between Septem-
ber and October 2002. The more weevil resistant cultivars tended to have a higher number 
of drops per unit area compared to the more susceptible cultivars. Also, sap of resistant 
cultivars thickened on wounding of the corm. Atwalira, a susceptible cultivar, did not pro-
duce sap. These observations will be repeated in the fIrst rainy season of 2003. The effect 
of sap on egg hatchability was also studied. It was hypothesized that if sap effects were 
chemical, embryo development would fail. If, however, the effects were physical, embryo 
- 21-
development would occur but eclosion might be reduced. As of yet, no conclusive results 
have been obtained, in part, because the controls did not perform well (hatchability was 
only 70%). This experiment will be repeated. Eclosion rates will also be determined for 
eggs treated with methanol and water·based extracts from the pseudostems and corms of 
different cultivars. 
Feeding deterrence is one way in which antibiosis may be manifested. An experiment was 
set up to observe settling and tunneling time of newly hatched larvae offered corms of dif· 
ferent cultivars (Atwalira, Kabula, Mbwazirume, Calcutta, Kauai, Keying, and Yangambi 
Km 5). Settling time was increased on resistant cultivars. While on susceptible cultivars 
(Atwalira, Kabula, Mbwazirume) all larvae penetrated the corm completely in 1-1.5 hours, 
up to 40% of the larvae had not penetrated the corms of resistant cultivars (Kauai, Keying, 
and Yangambi Km 5) in the same time interval. 
Antibiotic effects of resistance are being tested further in larval feeding bioassays. An 
experiment is underway to study survivorship and development rate of larvae raised on 
corms of different cultivars. Additionally, the influence on adult weight of feeding larvae 
on different cultivars will be determined. 
The objectives of an ongoing study are to study mechanisms of resistance to banana weevil, 
to determine the distribution of resistance in different parts of the plant and in different 
plant phenological stages, and to determine the influence of plant nutrition on resistance 
expreSSIOn . 
. The distribution of resistance factors within the plant and in different plant phenological 
stages will be determined using egg and larval bioassays described above. HPLC analysis 
are also being carried out on different parts of the plant and different plant phonological 
stages. 
Larval bioassays on corm tissue indicate that larvae feeding on resistant cultivars generally 
have longer developmental periods. For instance, larval developmental period on Kauai 
was 37 days while that on Atwalira was 25 days. Adult weights were similar. Survivorship 
data was not consistent with resistance patterns observed in the field. A pot experiment 
is being conducted to further study resistance in different cultivars. 
Larvae feeding on pseudostem had longer developmental periods (59 days) than those 
feeding on the corm (31 days). Also, survivorship of individuals feeding on pseudostem was 
low (59% for early instars; 29% at adult stage) compared to that oflarvae feeding on corm 
tissue (77% for early instars; 53% at adult stage). Adult weight of weevils raised on corm 
tissue was 0.054 g compared to 0.033 g for those on pseudostem tissue. Studies on variation 
of resistance with plant phenology are underway both in the laboratory and in the field. 
In another study, egg hatchability (eclosion rates) was influenced by cultivar and tissue 
type. For cortex tissue, the lowest hatchability was in Calcutta 4 (68%) and highest in 
Atwalira (83%). For the pseudostem, Calcutta 4 again gave the lowest hatchability (84%) 
and Atwalira and FHIA 03 the highest (88%). High mortality and delayed larval develop-
ment was observed on resistant cultivars (Keying, Kauai, and Calcutta 4) compared to a 
susceptible cultivar (Atwalira). While the larval stage lasted 27 days inAtwalira, it ranged 
from 30 to 37 days among the resistant cultivars. Adult weights were similar on the dif-
ferent cultivars. 
A study on the influence of plant nutrition on expression of resistance to banana weevil will 
be carried out in a field experiment planted in December 2002 at the KawandaAgricultural 
-22-
Research Institute, Uganda. The cultivars planted are Atwalira, Nsowe, FHIA 03, Caven· 
dish, Keying, and Yangambi Km 5. The nutrition levels are low and optimum nutrients. 
3.6 The identification of candidate genes for developing transgenic resistance 
to banana weevil in East Mrican Highland banana 
by M. Pillay and C.S. Gold. in collaboration with A. Kiggundu*. A. Viljoen. and K. 
Kunnert 
This project is being addressed through an ongoing PhD study under the lead ofthe Uni· 
versity of Pretoria and in collaboration with IPGRI. The primary objective is to explore the 
feasibility of developing transgenic plants with resistance to banana weevil. 
The use of proteinase inhibitors to engineer plants for resistance to a wide range of insect 
pests has recently gained much attention. Several studies have demonstrated effectiveness 
in the use of such inhibitor genes for transgenic control of various pests including insects 
and nematodes. Proteinase inhibitors which operate by inhibiting the gut enzymes that 
break down proteins in the insect's diet, have been isolated in several plant species and are 
believed to function as defense compounds against insect and pathogen attack. Exploring 
plants own defense mechanisms, by engineering and the introduction of insect defense pro· 
teins from other plants, is seen as a more natural strategy. Several genes encoding various 
proteinase inhibitors have been introduced into crop plants with successful improvement in 
resistance to pests and these plants are now at various stages offield·testing worldwide. 
The general objectives of this project are to (1) evaluate a strategy for the use of cysteine 
proteinase inhibitors as a first target gene for genetic engineering of banana for resistance 
to banana weevil; and (2) to optimize inhibition activity of plant cystatins to obtain novel 
inhibitors with increased activity and thus acquire intellectual property rights. Specifi· 
cally, we wish to: (a) identify the major class of proteolytic enzyme activity in the mid.gut 
of banana weevil and thus potential proteinase inhibitors of plant origin; (b) develop in·vivo 
and in· vitro bioassays for testing purified proteinase inhibitors against gut proteinase activ· 
ity. and on larval growth and development; (c) engineer through site directed mutagenesis, 
two plant cystatins with improved and specific inhibition of banana weevil gut proteinase 
enzymes, both in·vivo and in·vitro; and (d) design gene constructs for engineering bananas 
with novel proteinase inhibitor genes. 
A system for the dissection of adult and larval banana weevil mid guts and subsequent 
extraction of their active proteinases has been developed. This was followed by the devel· 
opment of a fluorimetric in·vivo assay to evaluate proteinase activity. optimum pH, and 
inhibition of the gut extracts. The assay uses a synthetic peptide which releases a highly 
fluorescent compound on breakdown by proteolytic enzymes present in the gut extracts. 
Activity is measured by a fluorescence spectrophotometer or can be scored visually under 
UVlight. 
The major proteinase enzyme class in both adult and larval banana weevils was established 
to be of the cysteine type as has been reported in many other Coleopteran insects. Cyste· 
ine proteolytic activity in weevil gut extracts was highly inhibited, in·vitro, by synthetic 
cysteine specific inhibitor E·64 and two plant cystatins; Oryzacystatin I (OC·I) from seeds 
of rice (Oryza sativa) and Papaya cystatin from leaves of paw paw (Carica papaya). The 
two phytocystatins were obtained in large amounts by cloning, and expression in E. coli 
bacteria before purification as fusion proteins. 
-23-
An in-vivo bioassay system based on vacuum infiltration of banana stems with cystatin 
solutions and subsequent rearing of weevil larvae on the stems was developed_ Preliminary 
experiments with this system have shown that larvae reared on cystatin infiltrated stems 
experienced up to 76% reduction in early larval growth and development_ This assay system 
is a significant achievement since there is no known artificial diet for banana weeviL 
3_7 Micropropagation of planting material in MUSQ 
by M_ Pillay, in collaboration with W. Tushemereirwe 
Routine micropropagation of introduced germplasm from the Musa Transit Center, Belgium 
and other planting material was carried out_ The genotypes used for micropropagation and 
the number ofpropagules derived from them are presented in Table 2_ 
Table 2_ List of banana genotypes propagated during 2003_ 
Genotypes 
Mpologoma, Kisansa, Nakitembe, Namaliga, 
Mbwazirume, Mudwale, Atwalira, William, 
KM-5, FHlA-17, 18,23,25 
P.Nangka, Cultivar Rose, PA03-22, PA03-
44, GCTCV-215, GCTCV-119, PV42, PV-53, 
Robusta, Malaccensis, Ruyondo, Truncata, 
Who-o-gu, P. awak, M balbisiana Cameroon, 
Toowoolee, P.Tongat, P. raja, Nzizi, Foulah 4, 
Too_ 
Source 
Local 
ITC 
TMBx 5295-1, FHlA-l, FHlA-23, FHlA-25, SH ITC 
3436-9 
Nine secondary triploids plus a landrace EET 
Twelve PNG accessions plus two Musa acumi- ITC 
nata banksii, Calcutta 4 and M. balbisiana 
# propa- Purpose 
gules 
38100 Field trials for tbe 
National Banana 
Research Program 
190 Breeding 
In progress For distribution in 
Malawi and Zambia 
In progress Preliminary yield 
trial 
In progress Screening for banana 
bacterial wilt 
3_8 Germplasm propagation, exchange, and conservation 
by M. Pillay, in collaboration with W. Tushemereirwe (NARO) 
The tissue culture laboratory in Kawanda is responsible for embryo rescue and micro-
propagation for NARD and IITA A total of 25 643 hybrid seeds from the various crosses 
made at Sendusu and Kawanda were dissected for embryo culture. The characteristics of 
the seeds and their embryo germination rate are shown in Table 3_ In addition, 2095 Gamma 
irradiated seeds of Calcutta 4 were cultured to observe their rate of germination. 
During this period, 2927 hybrids from various crosses were weaned and hardened for field 
trials_ 
3.9 Evaluation of seed fertility in MUSQ 
by M_ Pillay and A Tenkouano 
FHIA 23 and FHIA 25 were used as female parents and crossed with male fertile triploid 
East Mrican banana landraces Tereza, Nakasabira, and Namwezi. No seeds were obtained 
from these crosses_ However, FHIA 23 produced seeds when crossed with Calcutta 4 sug-
gesting that it is female fertile_ The FHIA hybrids are now being crossed with the diploid 
male parents including 8075-7, 9128-3, and 7197-2. The cross 1438K-1 x FHIA23 produced 
170 seeds, while 1201K-1 x FHIA 25 produced a single seed. The female fertility of the 
triploid 1968-2 is being ascertained by crossing it with Calcutta 4. 
-24-
Table 3. Characteristics of hybrid seeds for emhryo rescue at Kawanda (January-
November 2003). 
Female' Male' Seeds Embryos Germi· Embryo Embryo ger· 
nation recovery (%) mination (%) 
Matooke land- Improved 2x 293 199 24 67.9 12.1 
races 
Matooke land- Calcutta-4 10 4 0 40.0 0 
races 
TMHx I'::Eroved 2x 11595 7121 717 61.4 10.1 
TMHx C cutta-4 4566 3263 164 71.5 5.0 
TMHx Kokopo 25 6 2 24.0 33.3 
TMHx FHIA-23 140 112 24 80.0 21.4 
TMHx TMHx 523 334 30 63.9 9.0 
TMBx 22788-4 Balbisiana 367 64 20 17.4 31.3 
Imf.roved 2x Imnroved2x 250 73 17 29.2 23.3 
Ca cutta-4 P. ilin 606 346 80 57.1 23.1 
'Ibtal 18375 11522 1078 62.7 9.4 
'Matooke landraces: Entukura, Enzirabahima, IITA-Nakyetengu, Kabucuragye, Nakasabira, Nakayonga, 
Namwezi. Nfuuka and Tereza. 
'Improved diploids (2x): TMB2x 6142-1, 7197-2, 8075-7, 9128-3, 9719· 7, SH 3142, 3217,3362. 
'TMHx: 199K-4, 222K-l, 365K·l, 376K-7, 401K-l, 660K-l, 917K-2, 120IK-l, 1438K-1. 
3.10 Musa improvement at diploid level 
by M. Pillay and A. Tenkouano 
Eight diploids were identified by flow cytometry and selected for further improvement. 
These diploids, their parents, and some of their characteristics are listed in Table 4. 
Table 4. List of diploid hybrids with 'matooke' characteristics selected Cor further 
improvement. 
Genotype Parentage Remarks 
5610S·1 Kabucuragye x TMB2x 7197·2 Big bunch, being crossed with tetraploids, 
Jlroduced 41 seeds with TMHx 660K-l 
Nonparthenocarpic, little pollen, female fertile 
Nonparthenocarpic, being crossed with tetra-
861S-1 
1537K-l 
8817S·1 
7927S·1 
12187S-12 
12495S-3 
10969S-1 
Namwezi x Calcutta 4 
Kabucuragye x Calcutta 4 
Kazirakwe x 8075-7 
Nak,Yetengu x 7197-2 
376K-7 xo105-1 
917 K-2 x 9719-2 
376K-7 x 5105-1 
l'.loids 
llejng pollinated on tetraploids 
Bemg propagated 
Ploiay determined by chromosome counting 
Ploidy determined by chromosome counting 
Ploidy determined by chromosome counting 
3.11 Evaluation of Fusarium wilt resistance in Musa 
byM. Pillay 
Panama disease (Fusarium wilt) is the one of major constraints affecting Pisang Awak 
CABB) and Sukali N dizi CAAB) in eastern and southern Africa. Nine genotypes comprising 
Pisang lilin, Fougamou, Pisang Awak, SH3217, SH3142, SH3362, Yangambi Km5, and 
Kikundi were tested for resistance in a Fusarium-infested field. Of the nine genotypes, 
Pisang Awak and Fougamou showed typical symptoms of Fusarium wilt including leaf 
discoloration and wilting, pseudostem splitting, and browning of the vascular tissue. 
The seven other genotypes have not showed symptoms of Fusarium wilt after being in 
the infested field for 20 months. A new trial with 42 new genotypes is being screened for 
Fusarium resistance in the infested field. 
3.12 Musa improvement at the polyploid level 
by M. Pillay and A. Tenkouano 
Sixteen secondary triploids (Table 5) were selected for preliminary yield testing and evalu-
ation by farmers. Bunch weights of these hybrids ranged from 13 to 25 kg. Some hybrids 
were resistant to black Sigatoka while others were classified as tolerant. 
-25-
Table 5. Secondary triploids selected for preliminary yield trial (PYT). 
Hybrid Parentage BWT DFF NSL 
1. Opp30 1201K-l 1201K-l rpp~ 13 132 3 2. Opp79 1201K-l 1201K-l Opp 19 170 3 3. Opp81 1201K-l 1201K-l Opp 18 3 4. O~~85 120lK-l 1201K-l 0 14 152 4 5. 2 9K-5 660K-l x 8(f1l -7 16 170 3 6.2156K-41 1438K-l x 9128-3 22 122 3 7. 76898-20 1201K-l x 9128-3 12 164 2 8_ 7798~-2 917K-2 x 9128-3 21 lSI 3 9.8386 -15 917K-2 x 8H 3217 19 134 2 10. 83868-22 917K-2 x 8H 3217 20 2 11. 83868-50 917K-2 x 8H 3217 24 139 3 
12. 83868-19 917K-2 x 8H 3217 19 139 4 
13.91878-8 6OOK-l x 9128-3 20 131 4 
14. 9494S-1O 917K-2 x SH 3362 25 137 4 
15.97508-13 401K-l x 9128-3 19 139 5 
16_ 2409K-3 222K-l x 8075-7 17 150 4 
NSL - number of standing leaves at harvest; DFF = days to fruit filling; BWT = Bunch weight 
3.13 Evaluation of banana bacterial wilt resistance in Musa 
by M. Pillay, R. Bandyopadhyay, and W. Tushemereirwe 
Bacterial wilt, caused by Xanthomonas campestris pv. musacearum has been identified 
in Uganda. A task force was set up to evaluate the disease epidemiology and make recom-
mendations to the National Agricultural Research Organization. To enable screening of 
genotypes for bacterial wilt resistance, 41 genotypes (Table 6) were multiplied by micro-
and macropropagation. These genotypes have been established in the "hot spot" area in 
the Mukono district. 
Table 6. Genotypes used for banana bacterial wilt screening in Mukono, Uganda. 
PNG acees- TMPIB 2x F H I A TMHx TMH3x ITC accessions Introduced 
sions hybrids Hybrids Hybrids Hybrids landraces 
Ambiri 8075-7 8H 3217 660 K-l 83868-19 M. acuminata Yangambi M." Banksii Km5 
Dumingii 9128-3 SH 3362 917 K·2 91878-8 . acuminta ssp. Pisang lilin 
mallaccensis 
Kokopo 7197-2 SH 3142 1201 K-I 94948-10 M. acuminata cv. Mshale 
Merik 1968-2 FHlA23 1348 K-l 97508-13 
Pahang 
Pama 376K-7 U09K-3 
Pagatau 222K·l 77988-2 
Pora pora 365K-l 83868-15 
Yalim 401K-l 83868-22 
8386S-50 
3.14 Open Quarantine Facility (OQF) 
by E. Kanju, M. Hermes (ARI Kibaha, Tanzania), J. Whyte, and B. Khizzah 
The restricted movement of germplasm has been a major drawback in the expansion of 
production and utilization of cassava. It has not been possible for varieties or advanced 
genotypes developed in one country to benefit other countries in the region, which has 
impeded the spill over effects of cassava improvement activities in the region. The open 
quarantine system was established by plant quarantine services in the region to permit 
the transfer of a number of genotypes (in the form of cuttings) from the regional midalti-
tude germplasm enhancement program in eastern Uganda to western Kenya, Tanzania, 
and Rwanda specifically to combat the CMD pandemic. With the increasing devastation 
-26-
of cassava from cassava brown streak virus disease in the coastal lowlands of eastern 
and southern Africa, the open quarantine system was adopted to enable the exchange of 
tolerant/resistant materials in the region. 
In order to increase the number of resistant cultivars available to farmers, about 457 clones 
were introduced into the OQF established at Kibaha, Tanzania in March 2003, from the 
East African Root crops Research Network (EARRNET) breeding program at KARl Mtwapa, 
near Mombasa, Kenya. Mtwapa is a CBSD hot spot. Only clones, which did not show any 
CMD and CBSD symptoms, were introduced. The clones were also selected on the basis 
of their high yielding ability and high dry matter content. One stem per clone was taken. 
At Kibaha, each stem was cut into four cuttings and planted in the OQF. Planting was 
done on 4 May 2003. The spacing of 0.5 m x 0.5 m was used. After sprouting, the plants 
were closely inspected (weekly) for any disease symptoms. Plants showing any symptoms 
were recorded and then rouged and buried. After every two months, inspectors from the 
postentry quarantine station based at Arusha inspected the plants to make sure that only 
apparently disease free plants will be multiplied after the one·year compulsory confine-
ment in the OQF. The OQF is located in isolation (at least 200 m from the nearest cassava 
crop) and fenced and guarded. 
Plant establishment in the OQF was excellent due to watering. By the end of October 2003 
(5 MAP), 169 (37%) clones out of the 457 introduced into the OQF had all their plants 
rouged. Out of the 288 (63%) clones remaining, only 97 (21%) had all established plants 
free of disease symptoms. The use of the OQF has facilitated the quick and efficient intro-
duction of improved germplasm. 
4.15 Cassava Crossing Block 
by E. Kanju, M. Hermes (ARI Kibaha, Tanzania), J. Whyte, and B. Khizzah 
In order to generate new genotypes, which combine resistance to both CMD and CBSD, 
a crossing block was established at ARI Kibaha, Tanzania. The following cultivars were 
used: Kiroba, Arnani 46106/27, NDL 90/34, Namikonga, Kigoma Red, Kitumbua (CBSD 
tolerant/resistant); Kibaha, and TMS 4(2) 1425 (CMD resistant). Planting was done in 
April 2003. A randomized complete block design was used, with four replications. One row 
of ten plants constituted a plot. The spacing of 1.5m x 1.5 m was used. A full diallel mating 
system was used to generate F1 seeds for genetic studies. 
Due to severe drought experienced during the season (only 100 mm were received com-
pared to the normal expected amount of 900 mm), plant growth was adversely affected to 
the extent that no seeds were collected. Kibaha is not an ideal site for the establishment 
of crossing blocks due to poor soils and poor flowering of clones. Drought is not a common 
phenomenon at the site. In order to avoid the risk of poor seed set, two new sites (ARI 
Maruku, Bukoba and ARI Naliendele, Mtwara) were selected for the establishment of 
crossing blocks during the next planting season. 
3.16 Cassava Clonal Evaluation Trial (CET) 
by E. Kanju. M. Hermes (ARI Kibaha, Tanzania), H. Saleh. (ARI Kizimbani, 
Zanzibar), J. Whyte, and B. Khizzah 
In December 2002 a team of cassava breeders from eastern and southern Africa jointly 
met at ARI Kibaha, Tanzania for the harvesting of a cassava seedling trial. Their efforts 
-27-
resulted in the selection of about 513 seedlings that showed promising resistance to both 
cassava mosaic disease (CMD) and cassava brown streak disease (CBSD). The selected 
seedlings were cloned and planted in a clonal evaluation trial (CET) at three sites to evalu· 
ate their yielding potential and other quality attributes, and also to confirm their disease 
and pest reaction. 
Seedlings selected from the seedling trial were cloned and planted at three sites: Kibaha, 
a hot spot for both CMD and CBSD but poor soils (520 clones); Alavi estate, 10 km from 
Kibaba, where an entrepreneur (Mohamed Enterprises-MeTL) intends to establish a large 
cassava starch factory (480), and Zanzibar (170 clones). Planting was done in December 
2002. Three cuttings per clone were planted. A check plot design was used where a local 
check was planted after every ten clones. The spacing of 1.0 m x 1.0 m was used. Plant 
height and pest and disease scores were taken at three-month intervals up to harvesting. 
Diseases and pests were subjectively scored on a scale of 1-5, where class 1 indicated no 
symptoms and class 5 very severe symptoms. At harvest yield and yield component data 
were taken. The roots were also carefully observed for brown streak root necrosis. 
Table 7. Some characteristics of 41 cassava clones selected from the CET for advanced 
evaluation at Kibaha, Tanzania during the 200412005 seaSOD_ 
Clone no. Root weight No. of roots! Average root size Pedigree 
k!1/plant plant (kg) 
1. KBH 2000/010 1.27 3.0 0.42 71673-0P 
2. KBH 2000/016 4.47 6.0 0.74 71673-0P 
3. KBH 2000/017 1.90 5.5 0.34 71673-0P 
4. KBH 2000/018 2.40 6.7 0.36 7l673-0P 
5. KBH 2000/023 5.50 8.0 0.69 71673-0P 
7. KBH 2000/025 3.35 6.0 0.56 ~1425-0P 
8. KBH 2000/038 1.40 4.7 0.30 E 146-0P 
9. KBH 2000/049 1.83 3.3 0.55 88/00279-0P 
10. KBH 2000/057 1.73 5.0 0.35 91100459-0P 
11. KBH 2000/062 2.47 7.0 0.35 92/0325-0P 
12. KBH 2000/063 2.85 7.5 0.38 9210325-0P 
lB. KBH 2000/066 1.90 6.3 0.30 96/161B-OP 
14. KBH 2000/068 2.60 3.7 0.71 96/1613-0P 
15. KBH 2000/069 2.10 5.0 0.42 96/1613-0P 
16. KBH 2000/073 1.90 3.7 0.52 96/161B-OP 
17. KBH 20001100 3.55 5.5 0.64 88/00188-0P 
18. KBH 2000/111 1.73 6.7 0.26 88100188-0P 
19. KBH 2000/129 1.70 5.0 0.34 TME 139-0P 
20. KBH 20001164 1.50 3.7 0041 61677-0P-OP 
21. KBH 20001169 1.80 3.7 0.49 91102316-0P 
22. KBH 200011 75 2.10 7.3 0.29 TME 629-0P 
23. KBH 20001190 1.73 4.3 0.40 71762-0P 
24. KBH 2000/213 1.30 3.7 0.35 71762-0P 
25. KBH 2000/248 1.53 5.0 0.31 71762-0P 
26. KBH 20001257 1.87 4.0 0.47 88/00593-0 P 
27. KBH 2000/264 2.95 7.5 0.39 88/oo593-0P 
28. KBH 2000/280 2.50 6.0 0.42 88/01043-0 P 
29. KBH 2000/283 1.93 7.0 0.28 88/02122-0P 
30. KBH 2000/288 1.70 5.0 0.34 88/00417-0P 
31. KBH 2000/289 1.60 6.0 0.27 88/00417 -OP 
32. KBH 2000/306 1.83 5.0 0.37 88102334-0P 
33.KBH 2000/336 2.83 10.3 0.27 88/00022-0P 
34. KBH 2000/344 1.57 3.7 0043 91100424-0P 
35. KBH 2000/376 2.93 7.0 0042 84563-0P 
36. KBH 2000/417 2.40 4.7 0.51 84776-0P 
37. KBH 2000/418 3.07 7.7 0.40 84776-0P 
38. KBH 2000/453 1.77 4_3 0041 Mkondezi-OP 
39. KBH 2000/464 2.33 4_3 0.54 Mkondezi-OP 
40. KBH 2000/505 2.17 5_0 0.43 Kiroba-OP 
41. KBH 2000/513 1.63 4_3 0.38 Kiroba-OP 
Local check 0.55 2_0 0.27 
-28-
Plant establishment at all three sites was excellent. However, Kibaha and Alavi estate 
experienced unusual severe drought. The short rains expected in October-December also 
did not arrive, therefore, harvesting at these sites was postponed until the beginning of the 
2003/2004 rains (March 2004). Harvesting of the trial at Zanzibar was done in December 
2003. The number of clones selected for further evaluation and their characteristics are 
summarized in Tables 7, 8, and 9. 
A high genotype by environment interaction was observed. Different clones were selected 
at the three sites (only seven clones were common between Kibaha and Alavi estate, only 
one clone was common between Kibaha and Zanzibar, and only two clones were common 
between Alavi estate and Zanzibar). This highlights the importance of adopting breeding 
for specific adaptation approaches. Family performance in terms of both root yield and 
number of clones selected, seem to have also depended on the environment. Under poor 
soil and low moisture conditions (Kibaha environment) the following families seem to have 
performed well: 71673 and 96/1613; under good soil and low moisture conditions (Alavi 
estate) the following families seem to have performed well: 71762 and 88/00188 whereas, 
under good soils and good moisture conditions (Zanzibar), the following families seem to 
have performed well: 71762, 96/1632, 91100424, and Kiroba. 
Table 8. Some characteristics of 39 cassava clones selected from the CET for advanced 
evaluation at Alavi estate, near Kibaha, Tanzania during the 2004/2005 season. 
Clone no. Root weight No. of roots/ Average root size Pedigree 
1. KBH 2000/001 
2. KBH 2000/020 
3. KBH 2000/023 
4. KBH 2000/043 
5. KBH 2000/055 
7. KBH 2000/063 
8. KBH 2000/066 
9. KBH 2000/069 
10. KBH 2000/111 
11. KBH 2000/119 
12. KBH 2000/122 
13. KBH 2000/123 
14. KBH 2000/135 
15. KBH 2000/139 
16. KBH 2000/141 
17. KBH 2000/145 
18. KBH 2000/150 
19. KBH 2000/170 
20. KBH 2000/180 
21. KBH 2000/182 
22. KBH 20001188 
23. KBH 20001197 
24. KBH 2000/202 
25. KBH 2000/208 
26. KBH 2000/213 
27. KBH 2000/228 
28. KBH 2000/232 
29. KBH 2000/283 
30. KBH 2000/301 
31. KBH 2000/363 
32. KBH 2000/378 
33.KBH 2000/379 
34. KBH 2000/382 
35. KBH 2000/405 
36. KBH 2000/407 
37. KBH 2000/416 
38. KBH 2000/425 
39. KBH 2000/488 
Local check 
kg/plant plant (kg) 
2.60 6.3 0.41 
1.47 4.3 0.34 
1.67 5.3 0.31 
0.47 4.0 0.12 
0.87 3.0 0.29 
1.97 5.7 0.35 
3.60 8.5 0.42 
2.20 4.5 0.49 
5.20 6.5 0.80 
0.40 0.7 0.60 
5.35 11.0 0.49 
4.50 8.0 0.56 
1.50 2.7 0.56 
3.25 8.5 0.38 
4.15 9.0 0.46 
1.93 2.3 0.83 
1.37 5.0 0.27 
2.30 3.0 0.77 
1.13 3.7 0.31 
4.40 9.5 0.46 
3.00 7.0 0.43 
2.10 4.7 0.45 
1.57 5.0 0.31 
1.65 7.0 0.24 
0.87 1. 7 0.52 
5.85 7.5 0.78 
5.20 10.0 0.52 
2.23 4.0 0.56 
1.47 3.0 0.49 
3.00 2.7 1.13 
1.60 6.3 0.25 
2.70 5.0 0.54 
3.10 7.0 0.44 
1.50 6.7 0.23 
1.57 3.3 0.47 
1.50 5.0 0.30 
0.93 2.3 0.40 
1.57 5.3 0.29 
0.77 2.3 0.33 
-29-
96/1431·0P 
71673·0P 
71673·0P 
88/00623-0P 
88/00279-0P 
92/0325-0P 
96/1613-0P 
96/16138-0P 
88/00 188-0P 
88100 188-0P 
88100188-0P 
TME 130-0P 
TME 200-0P 
91/00438-0 P 
95/0947-0P 
TME 168-0P 
91100174-0P 
91102316-0P 
71762-0P 
71762-0P 
71762-0P 
71762-0P 
71762-0P 
71762-0P 
71762-0P 
71762-0P 
71762-0P 
88102122-0P 
88102334-0P 
9611632·0P 
84563-0P 
84563-0P 
84563-0P 
88102363-0P 
30211-0P 
84776-0P 
9610867-0P 
Kiroba-OP 
Table 9. Some characteristics of 26 cassava clones selected from the CET for advanced 
evaluation in Zanzibar during the 200412005 season. 
Clone no. Root weight No. of rootsl Average root size 
1. KBH 2000/64 
2. KBH 20001152 
3. KBH 20001191 
4. KBH 20001200 
5. KBH 2000/228 
6. KBH 2000/239 
7. KBH 20001251 
8. KBH 20001341 
9. KBH 2000/344 
10. KBH 20001352 
11. KBH 20001365 
12. KBH 20001367 
13. KBH 20001368 
14. KBH 20001383 
15. KBH 20001397 
16. KBH 20001462 
17. KBH 20001471 
18. KBH 2000/477 
19. KBH 20001480 
20. KBH 20001482 
21. KBH 20001488 
22. KBH 20001494 
23. KBH 20001500 
24. KBH 20001506 
25. KBH 20001513 
26. KBH 20001517 
Local Check 
kg/plant plant (kg) 
2.67 5.0 0.53 
1.67 3.0 O.M 
2.67 5.0 0.53 
2.67 4.3 0.62 
3.50 6.5 0.54 
2.33 5.3 0.44 
4.00 6.3 0.63 
3.00 4.5 0.67 
3.33 6.7 0.50 
4.67 5.3 0.88 
2.50 4.5 0.56 
2.50 3.5 0.71 
3.33 6.0 0.56 
1.17 4.7 0.25 
1.33 2.7 0.50 
1.83 5.3 0.34 
2.50 6.0 0.42 
1.00 3.0 0.33 
1.50 3.0 0.50 
2.00 10.0 0.20 
1.33 4.0 0.33 
1.17 4.0 0.29 
1.73 4.0 0,43 
0.83 3.7 0.23 
1.25 5.5 0.23 
1.50 5.0 0.30 
0.56 2.7 0.22 
Pedigree 
9210325·0P 
TME 117·0P 
71762·0P 
71762·0P 
71762·0P 
71762·0P 
71762·0P 
91100424·0P 
91/00424·0P 
91100424·0 P 
96/1632·0P 
96/1632·0P 
96/1632·0P 
84563·0P 
71173·0P 
Mkondezi·OP 
Kiroba·OP 
Kiroba·OP 
Kiroba·OP 
Kiroba·OP 
Kiroba·OP 
Kiroba·OP 
Kiroba·OP 
Kiroba·OP 
Kiroba·OP 
Kiroba·OP 
3.17 Genotype by environment interaction for native cassava (Manihot 
esculenta Crantz) starch quality and its use in the commercial sector 
by I.R.M. Benesi1, M. T. Labuschagne2, A.G.O. Dimn3, and N.M. Mahungu41Chitedze 
Research Station, PO Box 158, Lilongwe, Malawi; 2Department of Plant Sciences, 
University of the Free State, PO Box 339, Bloemfontein 9300, Republic of South 
Africa; 3 International Institute of Tropical Agriculture (IITA), Oyo Road, PMB 5320, 
Ibadan, Nigeria; 'Southern Africa Root Crops Research Network (SARRNET), PO 
Box 30258, Lilongwe 3, Malawi 
Genotype by environment study on interaction for native cassava was also initiated. Cas· 
sava starch is used directly in different ways or as a raw material for further processing 
like food, textile, paper, adhesives, chemicals, glucose, soap, detergent, laundry, ethanol, 
cosmetic powders, sausage making, pharmaceuticals, and insecticides. Dry matter content 
is closely related to starch content in cassava. This makes dry matter an important trait to 
cassava producers since it is a crop grown largely for its carbohydrate content. Dry matter 
varies very widely in cassava over years and environments. 
This study therefore was initiated to look into native cassava starch qualities from differ· 
ent Malawi cassava genotypes, determine the appropriate stability parameter to deal with 
G x E for starch quality traits, and also find out the possibility of use of cassava starch 
by the main industries in Malawi. Apart from starch quality parameters, root dry mater 
and starch extraction evaluation were also included. Various stability measures were used 
to deal with the problem of genotype by environment interaction. The results show that 
all the cassava genotypes produced starch with no protein just like the one used in the 
pharmaceutical industry. The moisture and ash content were much lower than the recom· 
mended allowable maximum (Table 10). 
- 30-
T
ab
le
 1
0.
 W
hi
te
ne
ss
, p
H
, m
o
is
tu
re
 c
o
n
te
n
t 
a
n
d 
a
s
h 
c
o
n
te
n
t 
fo
r 
s
ta
rc
h 
e
v
a
lu
at
ed
 in
 2
00
0/
20
01
 s
e
a
s
o
n
. 
W
hi
te
ne
ss
 o
f s
ta
rc
h 
(L
) 
pH
 o
f s
ta
rc
h 
St
ar
ch
 m
o
is
tu
re
 c
o
n
te
nt
 (%
) 
St
ar
ch
 a
sh
 c
o
n
te
nt
 (%
) 
G
en
ot
yp
e 
C
hi
te
dz
e 
M
ak
ok
a 
M
ea
n 
C
hi
t-
M
ak
ok
a 
M
ea
n 
C
hi
t-
M
ak
ok
a 
M
ea
n 
C
hi
t-
M
ak
ok
a 
M
ea
n 
ed
ze
 
ed
ze
 
e
dz
e 
Si
lir
a 
94
.7
5 
94
.7
5 
94
.7
5 
5.
5 
5.
5 
5.
5 
13
.
10
 
14
.
20
 
13
.6
5 
0.
13
 
0.
11
 
0.
12
 
Sa
ut
i 
96
.0
2 
96
.
02
 
96
.0
2 
5.
2 
5.
1 
5.
2 
12
.6
0 
13
.7
0 
13
.1
5 
0.
09
 
0.
11
 
0.
10
 
C
H
92
/0
82
 
94
.
94
 
94
.
94
 
94
.9
4 
5.
4 
5.5
 
5.
5 
12
.
20
 
14
.3
0 
13
.2
5 
0.
09
 
0.
15
 
0.
12
 
TM
S4
(2
)1
42
5 
95
.
36
 
95
.3
6 
95
.
36
 
5.
0 
5.
3 
5.
2 
11
.2
0 
14
.
20
 
12
.
70
 
0.
22
 
0.
13
 
0.
17
 
CH
92
/1
12
 
93
.4
9 
93
.4
9 
93
.4
9 
5.
1 
5.
5 
5.
3 
10
.
30
 
13
.
70
 
12
.
00
 
0.
15
 
0.
13
 
0.
14
 
C
H
92
/l0
5 
94
.5
6 
94
.
56
 
94
.
56
 
4.
8 
5.
5 
5.
2 
11
.4
0 
12
.3
0 
11
.8
5 
0.
13
 
0.
13
 
0.
13
 
LC
N
80
10
 
94
.6
3 
94
.6
3 
94
.6
3 
5.
1 
5.
7 
5.
4 
11
.8
0 
13
.4
0 
12
.
60
 
0.
20
 
0.
14
 
0.
15
 
30
78
6 
94
.9
4 
94
.9
4 
94
.9
4 
5.
2 
5.
6 
5.
4 
11
.8
0 
13
.5
0 
12
.4
5 
0.
10
 
0.
16
 
0.
13
 
83
35
0 
96
.
11
 
96
.1
1 
96
.
11
 
5.
2 
5.
1 
5.
2 
12
.5
0 
12
.
90
 
12
.
70
 
0.
25
 
0.
15
 
0.
20
 
T
M
E
1 
94
.5
1 
94
.5
1 
94
.5
1 
5.
3 
5.
3 
5.
3 
11
.6
0 
13
.7
0 
12
.
65
 
0.
20
 
0.
11
 
0.
15
 
81
10
00
15
 
94
.8
4 
94
.8
4 
94
.8
4 
5.
1 
5.
6 
5.
4 
11
.8
0 
14
.0
0 
12
.
90
 
0.
13
 
0.
11
 
0.
12
 
I 
C
H
92
/1
08
 
95
.4
2 
95
.4
2 
95
.
42
 
5.
5 
6.
2 
5.
9 
12
.2
0 
14
.0
0 
13
.1
0 
0.
22
 
0.
13
 
0.
17
 
~
 
M
K
95
/0
54
 
93
.
77
 
93
.7
7 
93
.
77
 
5.
0 
6.
0 
5.
5 
12
.7
0 
12
.9
0 
12
.8
0 
0.
11
 
0.
16
 
0.
13
 
~
 
M
bu
nd
um
al
i 
96
.2
8 
96
.2
8 
96
.2
8 
5.
7 
5.
3 
12
.6
0 
13
.
10
 
12
.8
5 
0.
12
 
0.
19
 
0.
15
 
I 
4.
8 
G
om
an
i 
96
.3
3 
96
.3
3 
96
.
33
 
5.
1 
5.
9 
5.
5 
12
.
80
 
12
.
60
 
12
.
70
 
0.
09
 
0.
19
 
0.
14
 
M
ko
nd
ez
i 
94
.
81
 
94
.
81
 
94
.8
1 
4.
9 
5.
7 
5.
3 
12
.6
0 
13
.2
0 
12
.
90
 
0.
12
 
0.
09
 
0.
11
 
TM
S6
01
21
 
95
.3
8 
95
.
38
 
95
.3
8 
5.
0 
5.
8 
5.
4 
12
.0
0 
13
.7
0 
12
.
85
 
0.
18
 
0.
09
 
0.
13
 
TM
S8
45
63
 
95
.2
2 
95
.
22
 
95
.2
2 
4.
7 
5.
8 
5.
3 
12
.1
0 
12
.0
0 
12
.
05
 
0.
12
 
0.
13
 
0.
12
 
M
au
nji
li 
95
.4
8 
95
.4
8 
95
.
48
 
4.
9 
5.
1 
5.
0 
12
.
80
 
13
.
90
 
13
.
35
 
0.
13
 
0.
10
 
0.
11
 
TM
S6
01
42
A
 
95
.
64
 
95
.
64
 
95
.6
4 
4.
9 
5.
7 
5.
3 
14
.
00
 
12
.7
0 
13
.3
5 
0.
14
 
0.
09
 
0.
11
 
N
 ze
ru
 c
o
rn
 s
ta
rc
h 
95
.2
0 
95
.2
0 
95
.2
0 
4.
7 
4.
7 
4.
7 
11
.2
0 
11
.
20
 
11
.2
0 
0.
15
 
0.
14
 
0.
14
 
PI
M
 c
o
rn
 s
ta
rc
h 
96
.4
6 
96
.4
6 
96
.4
6 
5.
1 
5.
0 
5.
0 
11
.8
0 
11
.8
0 
11
.8
0 
0.
18
 
0.
13
 
0.
15
 
M
PL
 c
o
rn
 s
ta
rc
h 
97
.
28
 
97
.2
8 
97
.2
8 
4.
1 
4.
1 
4.
1 
9.
90
 
9.
90
 
9.
90
 
0.
11
 
0.
14
 
0.
12
 
C
or
n 
flo
ur
 
94
.2
0 
94
.2
0 
94
.
20
 
3.
4 
3.
4 
3.
4 
10
.6
0 
10
.6
0 
10
.
60
 
0.
86
 
0.
70
 
0.
78
 
M
ea
n 
95
.3
3 
95
.3
3 
95
.3
3 
5.
0 
5.
4 
5.
2 
11
.9
7 
12
.
98
 
12
.4
7 
0.
17
 
0.
15
 
0.
16
 
C
V
(%
) 
0.
83
 
0.
83
 
0.
83
 
0.
19
 
0.
15
 
0.
17
 
5.
96
 
4.
54
 
5.
25
 
11
.4
0 
9.
36
 
12
.0
4 
LS
D
 fo
r G
 
1.
97
 
1.
97
 
1.
97
 
0.
02
 
0.
02
 
0.
02
 
1.
78
 
1.
47
 
1.
12
 
0.
05
 
0.
04
 
0.
03
 
LS
D
 fo
r L
 
0.
01
 
0.
32
 
0.
01
 
G
 =
 ge
no
ty
pe
; L
 :;;
: lo
ca
tio
n;
 M
PL
 :;;
: M
al
aw
i P
ha
rm
ac
ie
s L
im
ite
d;
 P
IM
 ==
 P
ac
ka
gi
ng
 In
du
st
ri
es
 M
al
aw
i L
im
ite
d;
 N
ze
ru
 =
 N
ze
ru
 R
ad
io
 C
om
p 
The pH for cassava starch was within the recommended range. Additive main effects and 
multiplicative interaction (AMMI) was strongly correlated with other stability parameters 
like Wi·ecovalence and stability variance-no covariate. AMMI is therefore recommended 
for use in the stability analysis of starch quality parameters since it provides additional 
information on appropriate environments for unstable genotypes. This study suggests 
that genotype has a greater influence on root dry matter (hence on the starch) than the 
environment. 
3.18 eSA Rural livelihoods project, biotechnology module 
by M. Ferguson 
A project entitled "Improving rural livelihoods in Southern Africa" was funded by USAID-
Regional Center for Southern Africa (RCSA), with a biotechnology component. Implemen-
tation of the biotechnology component began in October 2003 with the recruitment of a 
biotechnology coordinator, based at the NEPAD Biosciences Center in Nairobi, located 
at ILRI. The overall goal of the project is to improve the livelihoods of the rural poor in 
Southern Africa through an integrated approach to marketing and competitive production 
that links farmers and processors to growth, value· added markets. Countries participating 
within the project are Angola, Malawi, Mozambique, Tanzania, Zambia, and Zimbabwe. 
With the lack of a strong regional biotechnology forum in southern Africa the biotechnol-
ogy component proposes to build a regional platform, as part of a longer term strategy, for 
priority setting and technical capacity development in Southern Africa. This will evolve in 
close association with activities coordinated by the Program for Biosafety Systems Consor-
tium (PBS) aimed at strengthening biosafety policies and decision-making processes in the 
region. A stakeholders meeting was held in Malawi from 5-7 November 2003, to develop 
a workplan. Key elements of the strategy that evolved, that form the basis of this plan of 
work are (1) collating and disseminating baseline information for use as decision· making 
tools, (2) establishing a regional working group to initially define agricultural priorities and 
within these, biotechnology priorities (3) identify and facilitate opportunities for strategic 
private sector/public sector research partnerships with advanced (international) research 
institutions to transfer skills and develop policy and regulations to facilitate technology 
access, and (4) enhancing adaptation, application, and dissemination of biotechnology tools 
and products that address food and income needs. 
In this first year of the project, these key elements will be addressed in the following 
way: 
• Compile an inventory of ongoing and planned plant breeding and biotechnology activi-
ties in the region, with an indication of capacities. 
• Compile an inventory of biotechnology applications appropriate to Southern Africa. 
• Coordinate a regional biotechnology introductory and planning workshop to share 
baseline information and define the way forward; followed by a regional priority set-
ting workshop. The above two documents will provide valuable baseline information 
and facilitate the building of strategic partnerships. 
• Address constraints to cassava: a regionally important commodity particularly for 
agriculturally vulnerable households: 
Establish a tissue culture facility in Maputo, Mozambique to produce disease-free 
cassava planting material of disease resistant varieties. 
Support ARI-Tanzania at Mikocheni to assess the regional diversity of cassava 
brown streak resistant germplasm. 
-32-
Facilitate a strategic partnership between national institutes in Malawi and the 
Danforth Center, St. Louis, USA, where a transgenic cassava mosaic disease (CMD) 
resistant cassava has been developed. 
3.19 CSA Rural livelihoods project, legumes module 
by John Wendt, in collaboration with H. Soko (Malawi Agricultural Research), K. 
Kanenga (Zambia Agricultural Research), M. Amane, INEA·Mozambique, B. Mtonga 
and J. Hoopper (World Vision Zambia), D. Kanyerere (World Vision Malawi) 
Following a stakeholders meeting in Malawi in November 2003, the above stakehold· 
ers agreed to implement a program of soybean and cowpea selection. Best soybean and 
cowpea varieties were selected regionally from Malawi, Zambia, and Mozambique and were 
tested on· farm against two IITA soybean and one IITA cowpea variety. This activity was 
implemented by World Vision at over 70 farmer sites in Malawi and Zambia, and five in 
Mozambique. IITA soybean and cowpea germ plasm sets were planted in Malawi, Zambia, 
and Mozambique through their respective national programs in December 2003 through 
March 2004. Best varieties from these germplasm sets will be multiplied if found superior 
to already certified varieties for further testing. 
3.20 CSA Rural livelihoods project, banana module 
by M Pillay and C. Nankinga, in collaboration with B. Mwenebanda, J. Hoopper 
(World Vision), J. Heermans (CLUSA), and M. Mataa (University of Zambia) 
Stakeholders meetings were held in Malawi and project partners were identified for Malawi 
and Zambia. 
Weaning and hardening sheds were constructed in Malawi. The first phase of planting 
material was delivered to Malawi for weaning and hardening. A rapid appraisal of tissue 
culture facilities in Malawi was completed. 
3.21 Marker-assisted breeding in cassava 
by M. Ferguson 
'Ibtal funding of US$518 000 was received from The Rockefeller Foundation for a collab· 
orative project between ClAT, IITA, and The Agricultural Research Institute, Mikocheni, 
Tanzania entitled "Molecular marker·assisted and farmer participatory improvement of 
cassava germplasm for farmer/market preferred traits in Tanzania". The project will cross 
farmer· preferred germ plasm by agroecology to improved introductions that have resistance 
to cassava mosaic disease (CMD), cassava green mite (CGM), and cassava bacterial blight 
(CBB). Molecular markers associated with pest and disease resistance will be employed to 
reduce, in a logical manner, the number of progeny to a manageable number. The progeny 
selected by MAS will be evaluated in a single season in the corresponding agroecology and 
then evaluated over two cycles in collaboration with end-users (rural communities and other 
cassava producers). The project will be carried out in a total of six years divided into two 
three-year phases. A principal objective of the project is also the development of capacity 
for participatory plant breeding and marker-assisted breeding. This would be achieved by 
training two national program breeders at the MSc and PhD levels, and through two train-
ing workshops on participatory plant breeding and marker-assisted breeding. Activities, 
including molecular marker analysis, would be conducted at national program facilities, 
while ClAT and IITA that work on cassava will provide backstopping in conventional and 
modern methods of cassava breeding. A first work planning meeting was held in October 
2003. 
- 33-
4 Environmentally safe integrated plant protection technologies 
that reduce pre- and postharvest losses due to pests and 
diseases developed and applied 
4.1 Monitoring the CMD pandemic 
by J. Legg, in collaboration with S. Bigirimana, A. Jorge, G. Okao·Okuja, R. Obonyo, 
F. Ndjelassili, P. Ntawuruhunga, H. Obiero, S. Ajanga, 1. Ndyetabula, S. Jeremiah, 
L. Traoure, T. Hangy, J. Mabanza, and M. Toko 
In partnership with national research teams, an extensive monitoring and diagnostics 
program was implemented, covering more than 500 sites across East and Central Mrica 
including Kenya, Tanzania, the Democratic Republic of Congo (ORC), the Republic of Congo 
(ROC), Gabon, Cameroon, and Burundi. CMD surveys were also conducted in Senegal and 
Guinea-Conakry. These wide ranging surveys demonstrated the expansion of the pandemic 
into three new districts of Tanzania (Mwanza, Kahama, and Kasulu), the occurrence of 
severe CMD throughout ROC, and new pandemic spread into eastern Gabon, northeastern 
Burundi, and North Kivu Province in eastern DRC. CMD in both Senegal and Guinea was 
relatively mild and there was no evidence of the rapid spread characterizing the pandemic 
in East/Central Mrica. Similarly, there was no evidence for the occurrence of EACMV· UG 
and severe CMD in southern or southeastern Cameroon. This suggests that the western 
limit of the pandemic lies in eastern Gabon and has yet to reach Cameroon. GIS tech-
niques were used to analyze epidemiological factors contributing to CMD epidemics and 
the region currently most threatened by continued epidemic expansion is the eastern part 
of the Great Lakes region encompassing Rwanda, Burundi, northwestern Tanzania, and 
eastern DRC. 
A follow-up countrywide survey of cassava pests and diseases was completed in the Republic 
of Congo (ROC) and virus and virus-disease assessments were also made within a compre-
hensive pest and disease survey implemented in Mozambique. In ROC, CMD incidence was 
greater than 85%, an increase on the value of the previous year, and CMD continued to 
represent the most important biotic constraint to cassava in that country. In Mozambique, 
CBSD was confirmed as being the main constraint, although it was largely restricted to 
the northern provinces of Cabo del Gado, Nampula, and Zambezia. 
4.2 Characterization of cassava mosaic gerniniviruses 
by J. Legg, in collaboration with G. Okao-Okuja, R. Obonyo, B. Owor, P. Sseruwagi, 
J. Ndunguru, P. Ntawuruhunga, H. Obiero, S. Ajanga, 1. Ndyetabula, S. Jeremiah, 
T. Hangy, and J. Mabanza 
More than 2000 CMG diagnoses were completed during the year in the IITA-ESARC labora· 
tory with the primary aim of mapping the spread of the pandemic associated EACMV-UG 
(the Uganda variant). 
A survey of more than 100 sites in Uganda revealed the occurrence of EACMV·UG in 73% 
of severely and 54% of mildly diseased plants. ACMV was less frequent and occurred pri-
marily in mildly diseased plants. Three RFLP variants of EACMV-UG and two of ACMV 
were identified and their distributions mapped. An important finding from this study was 
that mixed EACMV-UG + ACMV infections occurred in both plants with mild and severe 
symptoms. During the epidemic period in Uganda, mixed infections had been uniquely 
associated with very severe symptoms arising from a synergistic interaction between the 
-34-
two viruses. The incongruence of these results suggests that different variants of the same 
virus may interact in different ways. Further investigation will be required to characterize 
these interactions and identify what triggers synergism. 
A survey in Tanzania which provided more than 500 virus-infected samples from all cas-
sava growing regions ofthe country revealed a unique level of cassava mosaic geminivirus 
diversity. In addition to the previously characterized EACMV- [Tanzania) (EACMV- [TZ)), 
eleven EACMV-like virus types, designated as EACMV- [TZl) to· [TZU) were identified 
following restriction analysis of PCR products using EcoRV and MiuI endonucleases. The 
Uganda variant was recorded from throughout the Lake Zone, in addition to the northern 
coastal zone in the Tanga area. ACMV was not detected in the coast region. Important 
differences in symtomatology were associated with the different virus types, highlighting 
the importance of further characterization studies. Fifteen virus isolates were cloned to 
produce partial and full length sequences of DNAA and B. Significantly, East African cas-
sava mosaic Cameroon virus was shown to occur in parts of Tanzania. This is a recombinant 
EACMV which has now been shown to occur throughout West Africa and now for the first 
time in a part of East Africa. 
4.3 Identification of novel satellite molecules associated with CMD in East 
Mrica 
by J. Legg, in collaboration with J. Ndunguru and C. Fauquet 
Two previously undescribed single-stranded DNA satellite molecules have been found in 
association with CMD in both Tanzania and Kenya. This represents the first record of such 
components in association with a bipartite begomovirus infection and may have significant 
implications for the epidemiology of CMD. The first satellite, SatDNA II (approx. 1030 
bp in length), when co-inoculated with EACMV-UG in the CMD resistant variety TME3 
seems to break the resistance, and SatDNA III (approx. 1209 bp long) enhances symptoms 
of EACMCV in the test plant, Nicotiana benthamiana. The satellites also appear to be 
associated with unusual CMD symptoms in the field, including a strap-like symptom in 
which cassava leaflets are sharply narrowed. The molecules occur widely in the pandemic 
affected areas of western Kenya and northwestern Tanzania, but also occur, albeit less 
frequently, in coastal Tanzania. Tests have yet to be carried out on CMD·diseased mate· 
rial from other geographical locations in the cassava belt of sub-Saharan Africa to verify 
whether or not the occurrence of satellites is a more general phenomenon or particularly 
associated with the pandemic affected zone. The ultimate, but as yet unachieved goal of this 
research is to develop a molecular marker that will facilitate the monitoring and tracking 
of the putative epidemic-associated genotype. 
4.4 Molecular markers to track CMD epidemic associated Bemisia tabaci bio-
types in East and Central Mrica 
by J. Legg, in collaboration with P. Sseruwagi and J. Brown 
Bemisia tabaci whiteflies collected in Uganda were characterized through sequencing a 
c. 850 bp portion of the mitochondrial cytochrome oxidase gene, in collaboration with the 
University of Arizona. Previous studies of whitefly populations collected at the time ofthe 
passage through Uganda of the CMD epidemic had shown the occurrence of two B. tabaci 
genotype clusters: one, designated the "invader", had been shown to occur at and behind 
the epidemic "front", whilst the other, the "local", occurred almost entirely ahead of the 
epidemic front. The 2003 analysis of whiteflies collected more widely through Uganda 
-35-
revealed the co-occurrence of the two genotypes in all regions surveyed, although the "local" 
was more frequent (83%) and widely distributed than the "invader" (17%). This apparently 
anomalous result could be explained by the fact that the two populations freely interbreed, 
and either the local haplotype confers a selective advantage on progeny or cytoplasmic 
incompatibility resulting from activity of endosymbiotic enterobacteria means that mating 
is unidirectional. Additional molecular characterization using sequences of the "kdr" gene 
is currently underway to test the hypothesis that "ocal" by "invader" hybrids occur. In 
addition to whitefly material from Uganda, samples are also being examined from other 
CMD pandemic affected countries, including Kenya, Tanzania, Rwanda, Burundi, DRC, 
and ROC. 
4.5 Interactions between cassava mosaic geminiviruses, effects on yield and 
putative cross protection 
by J. Legg, in collaboration with B. Owor, M. Latigo, S. Kyamanywa, and D. 
Osiru 
Previous studies have shown that cassava plants of CMD·susceptible varieties infected with 
mild strains of EACMV-UG have some protection against superinfection by severe strains 
of the same virus, and in the field situation perform better than initially CMD-free plants. 
Studies conducted to assess the sustainability of the benefit of this protection revealed that 
replanted cuttings from the initially mildly diseased treatment yielded almost 50% more 
than those from the initially CMD-free treatment. Similarly, mean severity of plants in the 
initially mildly affected treatment remained significantly lower than that of the initially 
CMD-free treatment. These results suggest that there are sustainable benefits accruing 
from the selection of plants with mild CMD symptoms under high inoculum conditions 
in Uganda. A proposal was developed for a PhD study which will examine the molecular 
mechanisms underlying these phenotypic effects. 
Molecular approaches have been used to examine the interactions between cassava mosaic 
geminiviruses occurring in Tanzania. A truncated DNAA component of EACMV has been 
identified from some CMD-affected plants in the Lake Victoria zone of northwest Tanzania. 
This appears to have a beneficial effect through modulating symptoms. The component 
is cotransmitted with the intact helper EACMV DNAA. Sub genomic satellites have been 
shown to have the converse effect, through enhancing symptoms of EACMCV in Nicotiana 
benthamiana and appearing to break resistance to EACMV-UG in the normally CMD-
resistant cassava variety, TME 3. Further interactions between Tanzania virus isolates 
are being studied following the production of infectious clones. These studies are of vital 
importance given both the complexity of the existing virus fauna in Tanzania and the epi· 
demic·associated gross geographical displacements of some species, such as the Uganda 
variant, EACMV·UG. 
4.6 Evaluating the efficiency of aphelinid parasitoids as biocontrol agents of 
whitefly vectors of cassava mosaic gemini viruses 
by J. Legg, in collaboration with M. Otim, P. Asiimwe, S. Kyamanywa, and D. 
Osiru 
Surveys of parasitoids of B. tabaci on sweetpotato in Uganda have been initiated and 
revealed the occurrence of two principal species, Eretmocerus mundus and Encarsia sophia. 
E. mundus was the more abundant of the two species. The level of parasitism was great-
-36-
est at the site at which whitefly populations were least. This mirrors the pattern recorded 
for parasitoids (the same two principal species) of B. tabaci on cassava and suggests that 
a negative density dependent relationship exists between the parasitoids and their host. 
Detailed assessments of the stratification of whitefly and whitefly parasitoid populations 
on sweetpotato were conducted to determine the optimal approach for sampling parasitoids 
on this crop. Fourth instar nymphs were most abundant on leaves 9-16 (counting from the 
growing tip) and parasitoid mummies on more mature leaves. Intercropping sweetpotato 
with cassava was shown to reduce the abundance of whitefly populations on cassava, and 
there was a concomitant reduction in CMD spread in the intercropped treatment. However, 
parasitism levels did not differ between the intercropped and sole crop treatments, and it 
is speculated that an augmentation of predator activity may have been responsible for the 
reduced whitefly populations on the intercropped cassava. 
4.7 Baseline survey of pests and diseases associated with cassava in the 
DRC 
by J. Legg, R. Hanna, M. Toko, D. Coyne, and A. Dixon in collaboration with 
SECID 
Further assessment of the Eastern region of the country has shown higher levels of root 
galling damage by Meloidogyne spp. are being observed than in most other parts of the 
country. Samples taken during the survey are currently being used for identification of 
species associated with cassava. 
4.8 Baseline survey of pests and diseases associated with cassava and sweet-
potato in Mozambique 
by J. Legg, R. Hanna, M Toko, D. Coyne, and A. Dixon collaboration with INERA 
Root-knot nematodes (Meloidogyne spp.) were present throughout the cassava belt of 
Mozambique. The severity of galling damage based on a 1 to 5 scale was slight, however, 
throughout most areas within about 90% of fields and moderate in about 10% of the fields 
sampled. Although the mean number of galls per 50 em of feeder roots was low in most 
sampled fields, the results indicate a consistent presence of root· knot nematodes throughout 
the cassava production areas. Results, however, from the soil and root samples collected 
during the survey, suggest that while Meloidogyne spp. consistently occur, Pratylenchus 
spp. (most likely P. brachyurus) are more prevalent and occur in greater abundance than 
Meloidogyne spp. Identification of samples are continuing. Few other nematode species 
have so far been observed. Details from the sweetpotato have yet to be assessed. However, 
observations of roots in the field suggest widespread damage to roots by nematodes. 
4.9 Importance of root-knot nematodes on cassava in yield loss assessed 
by D. Coyne, in collaboration with J. Whyte, A. Dixon, T. Munga, KARL Kenya; and 
M. Ogunlolu, University of lbadan, Nigeria 
Trials established in both West and East Mrica to assess the yield difference of cassava 
between nematicide treated or not treated in Meloidogyne spp. infested soils continue to 
provide limited insight into the relationship between the nematodes and crop production. 
A major factor appears to be the high levels of variability in crop growth within treat-
ments, but also the erratic occurrence of nematodes and lack of differences in nematode 
densities between treatments as the cropping season progresses. Regression of root galling 
- 37-
damage against yield parameters shows negative relationships however. In more controlled 
inoculation studies conducted in the field at ESARC-Uganda, cassava production yield 
was on average, 30% lower (P:S 0.05) in all inoculated treatments (100, 1000, and 10000 
nematodes per plant) than uninoculated for cv. Migyera (TMS30572) and 25% fewer small 
(unmarketable) tubers. There were no differences observed between inoculated treatments. 
Yield was 20% less for inoculated treatments compared with uninoculated for cv. SS4, but 
not significant (P:s 0.05). 
4.10 Maintain and further develop plant parasitic nematode cultures and col-
lections 
by D. Coyne, C. Dochez, and G. Georgen 
Plant parasitic nematode cultures are continuously being cultured and maintained in vitro 
and in vivo for use in nematode pathogenicity and biology studies at IITAand also for access 
by national programms. A reference collection of plant parasitic nematode microscope slides 
continues to be gathered and stored for training purposes from samples from diagnostic 
surveys. Photographic images of nematodes on microscope slides are being additionally 
taken for storage electronically at the Biodiversity Museum in Cotonou, and elsewhere. 
4.11 Country-wide survey of cassava pests and diseases in Mozambique 
by M. Taka', R. Hanna', M. Andrade', J. Legg', and D. Coyne', in collaboration with 
M. Otema', B. Agbonton', G. Okao-Okuja, R. Obonyo, A. Jone 2, EIMambo2 , and 
A. Sitole2 ' IITA; 2 Department of Sanidade Vegetal, Mozambique 
Upon request by IITA·PROAGRI (The National Programme for Agricultural Development) 
Project in Mozambique and with financial support by PROAGRI and the IITAAfrica-wide 
Cassava Green Mite (CGM) Project, a comprehensive survey was conducted in two stages 
in April-May 2003 to assess the importance of cassava pests and diseases throughout the 
cassava belt of Mozambique and in September--October 2003 to determine the impact of 
the most important pests and diseases on cassava production. Earlier surveys conducted 
in Mozambique were scanty and hence did not give an overall picture of the status of 
pests and diseases of cassava in the country. This activity reports on the survey done in 
April-May 2003. 
Through discussions with The National Institute for Agronomic Research INIA and SAR-
RNET-Mozambique and CGM Project, the itineraries of the survey were decided based 
on previous surveys conducted on cassava green mite in Mozambique. Eight provinces-
Maputo, Gaza, Inhambane, Sofala, Manica, Zambezia, Nampula, and Cabo Delgado-were 
retained for the survey. 
Frequency. The frequency was recorded to determine the importance of cassava in each 
of the provinces. It was determined by recording the presence/absence of cassava fields 
on the right and left side of the roads while driving. The distance covered per day was 
determined from the difference between the mileage recorded at the beginning of the day 
and the mileage at the stopping point of the day. The frequency per km was equal to the 
total number of cassava fields recorded on the left and right sides of the road. divided by 
the total mileage covered for the day. With this formula, one can calculate the frequency 
per district and per province as needed. 
- 38-
Field sampling. Young cassava fields (> 3 months and <10-12-months-old) were sampled 
at every 15-20 km interval along the itinerary of the day. Although most fields were selected 
along the main roads, others were selected from secondary roads when there were not 
enough fields along the main roads. Fields with less than 50 plants were not sampled as 30 
plants were the minimum for sampling. In addition, only fields near homes were sampled 
to obtain authorization from the owner of the field/farmer. When a field was identified at 
a given interval and the farmer was present, the provincial technician who accompanied 
the team explained to the farmer the purpose of the survey and asked him further ques-
tions such as the names of the location, district, varieties cultivated, age of the field etc. 
This information was recorded on a datasheet agreed upon by all scientists involved in 
the survey. Other relevant information recorded on the datasheet included the geographic 
position of the location (GPS), the cropping systems (monocrop vs intercrop), weed cover-
age (based on 1-5 scale where 1 = no weeds, 2 = < 25% of the field is weedy, 3 = 50% of the 
field weedy, 4 = 75% of the field is weedy, and 5 = 100% of the field is weedy), the soil type 
(classified as sandy, loamy, or clay), and the vegetation (classified as savanna, secondary 
forest, and primary forest). Thirty plants were then randomly selected in each field and 
were used to assess the abundance and damage severity of the pests and diseases. The 
most common pests evaluated included the cassava green mite (CGM) , red mites (RM), 
cassava mealybug (CM), whiteflies (WF), and nematodes while the main diseases were the 
cassava mosaic disease (CMD), cassava brown streak disease (CBSD), cassava bacterial 
blight (CBB), and cassava anthracnose disease (CAD). 
Pests. The damage severity of CGM, CM, and RM was evaluated using a rating scale from 
1 to 5, where 1 represented plants with no visible damage symptoms and 5 represented 
plants with very severe symptoms or death ofthe plant. The abundance of CGM was deter-
mined by directly counting the number of actives encountered on the first fully developed 
leaf while the abundance of CM was estimated by counting the number of actives in the 
shoot tips. The number of whiteflies was counted on the 5th leaf from the top of the plant. 
The presence/absence of the predatory mite, Typhlodromalus aripo was determined by 
examining the shoot tip of each sample plant. Nematodes were sampled on 10 of the 30 
plants randomly selected above, i.e., every other 3rd plant. A trowel and a knife were used 
to dig and loosen the soil around the selected plant to get access to the feeder roots without 
uprooting the plant. A 50-cm length of feeder roots was collected to count the number of 
galls to determine galling intensity. When less than 50 cm of feeder roots, the number of 
galls was counted on whatever length obtained. The root system was also scored for gall-
ing damage on a scale of 1-5 (where 1 + no galling, 5 = severe galling damage). All feeder 
roots collected from the plant with approximately 100 ml of soil dug around each sample 
plant were placed in a plastic bag and bulked for each field. These samples were stored in 
a cooler to prevent adverse heating and were later processed in the laboratorylhotel room. 
Nematodes were extracted within 24-48 hours from 100 ml soil and 5 g root fresh weight 
subsamples from each field, using a modified sieve system with plastic sieves and plates. 
Nematode suspensions were then reduced with a 38 micron sieve, killed with hot water, 
and fixed with formalin for later nematode identification and counting. 
Diseases. The severity of cassava mosaic disease (CMD), cassava brown streak disease 
(CBSD), cassava bacterial blight (eBB), and cassava anthracnose (CA) was evaluated 
using a similar rating scale from 1 to 5 as for the pests. Leaf samples from severely CMD 
infested plants were harvested and placed in vials with alcohol for DNA determination 
later in the hotel room. 
-39-
Frequency results. The highest frequency of cassava was found in Zambezia (8.1 fields! 
km), followed by Gaza and Inhambane with, respectively, 5.95 and 5.65 fieldslkm. Nam· 
pula occupied the fourth place with 4.24 fieldslkm and Cabo Delgado occupied the filth 
place with 2.45 fieldslkm, or about half of the frequency of Nampula. The lowest cassava 
frequencies were recorded in Sofala (0.56 fieldlkm), Manica (0.52 fieldlkm), and Maputo 
(0.49 fieldlkm). 
Pest results. Five pests were encountered during the survey. These were the cassava green 
mite (CGM), red mites (RM), cassava mealybug (CM), whiteflies (WF), and nematodes. 
However, only CGM infestations were severe in much of the fields sampled across the coun-
try (Figs 2a and b). The distribution and incidence of the predatory mite T. aripo coincided 
with that of CGM (Fig. 2c). Whitefly population densities were abundant particularly in 
Zambezia where surprisingly CMD severity was rather low, suggesting that WF may not 
only be vector of CMD virus but can also be an independent pest. Indeed, the leaves of the 
plants with high WF populations were black sooty which could seriously affect photosyn-
thesis. Others pests such as CM and RM were considered minor pests. Nematodes were 
present throughout the country. The most commonly observed plant parasitic nematodes 
associated with surveyed fields are root-knot nematodes (Meloidogyne spp.), population 
densities recovered have largely been low however. Other relatively common species include 
Pratylenchus spp. (lesion nematodes) and Helicotylenchus spp. (spiral nematodes), but also 
in low densities. Results suggest root-knot nematodes are widespread on cassava. 
Diseases results. Three of the traditional common diseases known to affect cassava pro-
duction throughout tropical Mrica were present in Mozambique. These were the cassava 
mosaic disease (CMD). However, based on the distribution and severity, only CMD was 
found as a threat to cassava production. CBB and CAD were found in very few locations 
and their severity was low. In addition to CMD, the cassava brown streak disease (CBSD), 
recently reported as devastating cassava in many countries in Eastern Mrica (Tanzania, 
Kenya), was also found in the three northern provinces of Mozambique-Cabo Delgado, 
Nampula, and Zambezia. Zambezia appeared as the most affected province. The incidence 
and damage severity for both CMD and CBSD are shown in Figures 3a and band 4a and 
b. 
CMD virus identification. The results of the identification are reported in Table 11. 
The Mrican cassava mosaic virus (ACMV) was the most commonly found in the samples 
collected. About 50% ofthe total fields with positive identification were infested with this 
type of virus. The virus was found particularly in Inhambane, Zambezia, and Nampula 
and was absent in Cabo Delgado and Gaza. The second most common virus was the mixed 
Mrican cassava mosaic virus (ACMV) and Eastern cassava mosaic virus (EACMV). About 
34% of the total fields were infected. The least common virus identified was the EastMrican 
cassava mosaic virus (EACMV) with only 17% of the fields infected. The devastating Eastern 
Mrica cassava mosaic virus -Ugandan variant (EACMV-UG2) was not found in any of the 
fields during the April-May survey. 
Conclusions and recommendations. Based on the frequency resuits, Zambezia, Gaza, 
Inhambane, and Nampula are in that order the greatest cassava producing provinces 
of Mozambique. However, this may change if more districts and more feeder roads were 
included during the survey. 
CMD and CBSD are the most threatening diseases of cassava in Mozambique. CBSD is still 
located in the Northern provinces (Cillelgado, Nampula, and Zambezia); however, urgent 
-40-
measures need to be taken now to prevent its spread southward. Although the Eastern 
African cassava mosaic-Ugandan variant was not found in the samples analyzed during 
this survey, all ingredients are present for the development of the disease. Preventive 
measures should be taken now. 
CGM remains a problem in areas where the crop is planted on poorly managed soils and 
where T. aripo has not established or spreads slowly as a result of cultivation of unfavorable 
varieties. Therefore, further redistribution/release of T. aripo should be made along with 
the introduction of varieties with favorable shoot characteristics. An integrated approach 
including soil management should be encouraged. In addition as vectors of CMD virus, 
whiteflies can be independent pests of cassava. There is need to determine whether the 
whiteflies on cassava are the same species. 
Table 11. Virus identification of samples collected from individual fields in different 
provinces in April-May 2003. 
Province Number of fields and types of virus identified 
ACMV 
Cabo Delgado 0 
Nampula 9 
Zambezia 12 
Sofala 7 
Manica 2 
Inhambane 15 
Gaza 0 
TotaUvirus 45 
CGM damages 
o No damage (1) 
• Slight damage (1-2) 
• Moderate (2-3) 
• Severe (~) 
• Very severe (>4) 
c=J Mozambique 
EACMV-UG2 EACMV ACMV+EACMV 
0 2 0 
0 2 0 
0 2 7 
0 0 15 
0 1 1 
0 4 6 
0 5 2 
0 16 31 
~ Carbo Delgado I 
~zambeZial 
~Sofalal 
"'---""""'===:;~ Inhambane 
Maputo 
TotaU province 
02 
11 
21 
22 
04 
25 
07 
92 
Figure 2a. Damage severity of CGM in different cassava growing provinces in Mozambique. 
-41-
CGM infestation 
o Low (0 -25) 
• Moderate (26 - 60) 
• High (61 - 100) 
• Very high (>100) o Mozambique 
-d Carbo Delgado I 
~zambeZial 
~Sofalal 
Ir---------~~~--~Inhambane 
Figure 2b. CGM population densities in different cassava growing provinces in Mozambique. 
o 
T. aripo distribution 
o Absence (0) 
• Low (0-25%) 
• Moderate (26-50%) 
• High (51-100-.4) o Mozambique 
-d Carbo Delgado ~ 
t""---=;;;;;;:;;;;;;;;;;;;;;;;d Inhambane 
Figure 2e. Percentage infestation of T. aripo in different cassava growing provinces in Mozambique. 
-42-
CMD Severity 
o No disease (1) 
• Slight damage (1-2) 
• Moderate (2-3) 
• Severe (34) 
• Very severe (>4) o Mozambique 
Figure 3a. Damage severity of CMD in different cassava growing provinces in Mozambique. 
CMD incidence 
o Low(~25%) 
• Moderate (26-50%) 
• High (51-100%) 
D Mozambique 
Figure 3b. Incidence of CMD in different cassava growing provinces in Mozambique. 
-43-
CBSD Severity 
o No damage (1) 
• Slight damage (1-2) 
• Moderate (2-3) 
• Severe (34) 
• Very severe (>4) o Mozambique 
~ Garbo Delgado ~ 
~zambezia~ 
~Sofalal 
~------~~~~~~Inhambane 
Figure 4a. Damage severity of cassava brown streak disease in different cassava growing provinces 
in Mozambique. 
CBSD incidence 
o Low (0-25%) 
• Moderate (26-50%) 
• High (51-100%) 
o Mozambique 
-====-l Ca rbo Delgado ~ 
~zambezial 
~Sofalal 
~------~~~~~~Inhambane 
Figure 4b. Incidence of cassava brown streak disease in different cassava growing provinces in 
Mozambique. 
-44-
4.12 Country-wide survey on the impact of cassava pests and diseases on cas-
sava production in Mozambique 
by M. Toko', R. Hanna', M. Andrade', J. Legg', and D. Coyne', in collaboration with 
M. Otema', R. Obonyo, A. Jone', and A. Sitole'. 'International Institute of Tropical 
Agriculture (IITA) 2Department of Sanidade Vegetal, Mozambique 
The survey in September-October 2003 was a follow-up of the survey carried out in 
April-May 20 03. This survey was mainly conducted to determine the impact of the most 
severe pests and diseases that were evaluated in April-May when cassava fields were still 
young (> 3 months and < 8 months) on cassava production when the fields had matured 
in September-October 2003. The same itinerary followed in April-May 2003 was also fol-
lowed in September-October 2003, except that in the September, the survey started from 
the Northern region (Cabo Delgado) moving down to Maputo. 
Field preselection was done from the office. The fields that were sampled in September-
October were preselected among the 202 fields sampled in April-May 2003. The fields were 
chosen on the basis of their mean damage severity recorded in April-May. Fields that had 
in April-May a mean severity above 2 (based on 1 to 5 scale) for one or more of the most 
important pests (CGM) orland diseases (CMD, CBSD) were selected for the evaluation of 
pests and diseases and yield parameters. The selection was based on the assumption that 
the fields that were severely infested/infected in April-May when they were young (> 3 
months and < 8 months) would have a greater negative impact on yield parameters. A total 
of 79 fields from Gaza, Inhambane, Sofala, Manica, Zambezia, Nampula, and Cabo Del-
gado provinces responded to these criteria. The fields were later identified using the GPS 
information and others recorded in April-May 2003. Fields in Gaza and Inhambane were 
particularly chosen because of their severe CMD infection and CGM attacks in April-May, 
while the fields in Nampula, Zambezia, and Cabo Delgado were chosen mainly because 
ofthe severity of attacks by CBSD. When the preselected field was not found or had been 
harvested, the next closest field (within 2-3 km) was selected. 
In each sample field, 30 cassava plants of the two most common varieties were randomly 
selected across the field and were evaluated for their severity of CMD and CBSD for the 
diseases, and CGM and the presence/absence of the predatory mite, Typhlodromalus aripo 
for the pests. The damage severity for all pests and diseases as well as the collection ofleaf 
samples for virus identification was done the same way it was done in April-May 2003. 
Nematodes were evaluated on the same 10 plants that were harvested for yield parameters 
(see yield parameters). 
Ten of the 30 plants sampled for pests and diseases were selected and tagged for harvest-
ing. Each target plant was uprooted using a hoe and/or a machete that served to dig the 
soil around the plant to facilitate the pulling of the whole plant. The harvested plants 
were left near the place they were uprooted from. Each uprooted plant was separated into 
stems and leaves, and tuber roots. All stems and leaves from the same plant were cut in 
pieces and placed in a bucket to facilitate weighing of the samples. The tuber roots were 
cut off from the base of the stems and separated into marketable (that can be accepted for 
sale at a market) and unmarketable tuber roots (that many buyers would not like to buy 
either because they are rotten or are too small). The damage caused by nematodes to tuber 
roots was determined by recording the presence/absence of the galling on a 1-5 scale. The 
number of tuber roots in each category (marketable and unmarketable) was then counted, 
placed in a bucket, and weighed using weighing balances of 5 to 25 kg to accommodate low 
- 45-
and heavy weights . Rotten tuber roots as a result of disease, physical damage, or rodents 
were not included among the marketable roots even when they were of a good size. They 
were considered unmarketable. After weighing the tuber roots, a few tubers from each 
category (marketable and nonmarketable) were cut in pieces with a machete to determine 
the presence/absence of root necrosis considered to be the symptoms of CBSD attack on 
the tuber roots. A rating scale from 1 to 5 developed by INIA was used. 
Pest and diseases. A total of 79 fields were sampled. The incidence and damage sever-
ity of CGM were low throughout the survey area, except in two locations in Cabo Delgado 
and Inhambane where the severity was moderate (Fig. 5a). As a result, the populations of 
the predatory mite Typhlodromalus aripo were equally low (Fig. 5b). The distribution and 
severity of CMD are shown in Figure 6. Except in Manica where all cassava fields sampled 
were CMD symptom free, CMD was found throughout the other provinces although the 
damage severity varied greatly. Figure 7 shows the distribution and severity ofCBSD, which 
was as in April-May only observed in Cabo Delgado, Nampula, and Zambezia. The sever-
ity, however, varied from slight to moderate, except in Nacala, Nampula province where 
the mean damage severity was as severe as (> 3.0 based on 1-5 scale). For the nematodes, 
the mean number of galls per plant varied between provinces. The highest infestation was 
recorded in Manica. 
Figure 5. Distribution and severity of cassava green mite (CGM) and incidence of the predatory mite, 
Typh/odroma/us aripo in September-October. 
- 46-
00 
CMD Severity 
o No damage (1) 
o Slight damage (1-2) 
• Moderate (2-3) 
• Severe (3-4) 
• Very severe (>4) 
Figure 6. Distribution and severity of cassava mosaic disease (CMD) in September-October survey. 
/ 
CBSO Severity 
o No damage (1) 
o Slight damage (1-2) 
Moderate (2-3) 
• Severe (3-4) 
• Very severe (4-5) 
Figure 7. Distribution and severity of cassava brown streak disease (CBSD) in September-October. 
- 47-
CMD virus iidentification. The virus identification of the samples collected in 
September-October is shown in Table 12. The majority of the samples collected (about 
80.5%) were infested by the Eastern Africa Cassava Mosaic Virus (EACMV) throughout the 
seven provinces surveyed. However, there were 11 fields (7.6%) infested with mixed African 
Cassava Mosaic (ACMV) and Eastern African Cassava Mosaic Virus (EACMV) and six 
fields (4.1 %) infested with the Mrican Cassava Mosaic Virus (ACMV). Unlike in April-May, 
the devastating Eastern Mrica cassava mosaic virus-Ugandan variant (EACMV-UG2) 
was identified in seven fields (4.8%) of the total samples. However, all the samples were 
identified from Nampula province. The strain was not identified in any other province. 
Overall yield parameters. The mean number and mean weight of marketable and 
unmarketable tuber roots and the weight of the stems and leaves are presented in Table 
13. The greatest mean number of tuber roots per plant (marketable and unmarketable) 
was recorded from Manica followed by Nampula and Zambezia provinces. The mean 
numbers were moderate in Sofala and Manica provinces while the lowest mean numbers 
were reported from Inhambane and Gaza. The mean weights of tuber roots and the mean 
weights of stems and leaves followed the same trend as those of tuber roots. The highest 
mean weight of tuber roots and of stems and leaves were found in Cabo Delgado followed 
by Nampula province. These means in the two provinces were greater than the means 
recorded in Zambezia and Sofala provinces. The lowest mean weights were reported from 
Manica, Inhamabane, and Gaza. 
Table 12. Virus identification of samples collected from individual fields in different 
provinces in September-October 2003. 
Province Number of fields and types of virus identified 
ACMV EACMV-UG2 EACMV ACMV+EACMV TotalJprovince 
Cabo Delgado 0 0 2 0 02 
Nampula 0 7 4 1 12 
Zambezia 0 0 47 2 49 
Sofala 0 0 47 1 48 
Manica 0 0 4 1 05 
Inhambane 6 0 13 1 24 
Gaza 0 0 4 5 05 
ThtalJv:irus 6 7 121 11 145 
Table 13. Weights oC stems and leaves per plant and number and weight of marketable 
and unmarketable tuber roots at harvest during September-October survey. 
Provinces NoMR MRWt NoUR URWt Wt of Stems + leaves 
Carbo del Gado 6.129 3.121 5.028 0.837 3.615 
Nampula 4.675 2.819 4.404 0.906 3.199 
Zambezia 4.379 1.836 3.532 0.622 2.123 
Sofala 3.075 1.701 3.025 0.523 1.855 
Manica 3.457 1.053 2.957 0.313 1.473 
Inhambane 2.812 1.329 1.935 0.385 1.141 
Gaza 2.888 1.373 2.158 0.407 1.124 
No MR = number of marketable tuber roots; MR Wt = marketsble tuber root weights 
No UR = number of unmarketsble tuber roots, UR Wt = unmarketsble tuber root weights 
Wt of stems + leaves = weight of stems and leaves 
-48-
Conclusions and recommendations. The results of September-October survey tor 
pests and diseases confIrmed the observation made in April-May that CGM, CMD, and 
CBSD are the most threatening cassava pest and diseases, respectively, in Mozambique. 
Although the incidence and severity of nematodes was not very severe in April-May, the 
damage caused to tuber roots in the form of quality and source of secondary infection may 
be important. There is need from the samples collected to identify the most common spe· 
cies found in Mozambique. 
There was no clear evidence of the impact of pests and diseases on cassava production. It is 
difficult to determine such an impact from a regular field survey as a single field was often 
infected or infested by more than one disease or pest. In addition to pests and diseases, 
agronomic and other soil or varieties etc. varied from province to province even within 
the same province. For instance, one would have expected lower yields in Cabo/delGado, 
Nampula, and Zambezia where CBSD was present in addition to other constraints, but 
this was not the case as reported in Table 13. 
4.13 Monitoring and diagnostic survey study in the Republic of Congo 
by P. Ntawuruhunga (consultant) under the supervision of J. Legg and J. Whyte 
Cassava is the main food staple and source of income of the population of the Republic of 
Congo. It constitutes the backbone of the farming system and the major commodity of the 
Congo population. However, since 199711998 CMD has become the most important con-
straint to cassava production in the Republic of Congo, threatening cassava production and 
the food security, having already caused food shortage in the country. Diagnostic tests of 
CMD·diseased samples collected in the Plateaux region in a 1999 survey in ROC revealed 
the presence of EACMV· UG, suggesting the possibility that the pandemic of severe CMD 
had expanded right across Central Mrica. The most severely diseased plants were infected 
with both EACMV-UG and ACMv, as observed for the severe CMD associated with the 
pandemic in East Africa. The principal objective of the study following that done in the 
previous year was to provide a comprehensive and detailed assessment of the status of 
CMD in ROC and to determine whether the disease was spreading, in the characteristic 
manner reported for the pandemic in East Mrica. 
An informal approach to sampling was used where stops were made at regulars intervals of 
at least 10 to 15 km along the main roads in the regions surveyed, occasionally penetrating 
areas away from the main road to try as much as possible to obtain all variation existing 
between fields in surveyed areas. 
In total, 163 fields were sampled in 35 districts in nine regions and Brazzaville Commune. 
They were 11 (6.7%) in Bouenza, 29 (17.8%) in Cuvette Centrale, 20 (12.3%) in Cuvette 
Ouest, 10 (6.1%) in Lekoumou, 12 (7.4%) in Likouala, 15 (9.2%) in Niari, 40 (24.6%) in 
Plateaux, 10 (6.1%) in Pool North, 10 (6.1%) in Sangha, and six (3.7%) in Brazzaville. 
Cassava fields aged between three and six months were sampled. Thirty plants in each 
field were selected along the two diagonals running across it. Crop field sizes, crop mixture, 
and the names of local cassava cultivars grown were recorded. Geographical data were 
recorded using GPS equipment, while biological assessments were made for CMD, and 
whitefly population, cassava green mite (CGM), cassava bacterial blight (CBB), cassava 
mealybug (CM), and Typhlodromalus aripo. 
-49-
The adult whitefly population was assessed on each plant sampled. Five apical open leaves 
of the tallest shoot of the plant were each observed taking it by the petiole turned gently 
upside down to expose the flies and count their number and the total flies on the five 
leaves was recorded. One to three CMD·diseased leaf samples were collected from each 
field for subsequent virus diagnosis. A two-stage process was used for this, in which DNA 
was extracted daily from leaf samples during the course of the survey in ROC and final 
diagnoses were done using polymerase chain reaction (PCR) techniques in the laboratory 
ofIITA-ESARC, Uganda. 
CMD was observed throughout the country and in all regions surveyed. The CMD 
situation in Cuvette Centrale and Plateaux had not changed compared to last year's obser-
vations, however it had deteriorated in Pool north where both incidence and severity had 
increased. %itefly populations were relatively low even in the forest areas compared to 
the observations made in 2002. More regular monitoring, however, would be required to 
gain a clearer understanding of whitefly population dynamics, since whitefly populations 
vary considerably from season to season. T. aripo, the introduced predator of CGM, was 
observed everywhere and with a greater abundance compared to 2002. Whilst this was 
true in all regions, it was particularly marked in Niari where it was first released (personal 
communication). Cassava bacterial blight was observed in all regions with a high incidence 
in Cuvette Centrale combined with high severity. The districts of Boundji and Gamboma 
were most severely affected by bacterial blight. 
Total incidence for the country was 86.2% with 81.7 % due to cutting infection diseased 
while only 4.3% was attributed to whitefly infection, indicating that the disease is primar-
ily being sustained through the planting of diseased cuttings (Table 14). These data show 
that CMD has increased and is characterized by high incidence and severity throughout 
the country. In 2002, total incidence was 80.3%. 
Table 14. Type of infection and total incidence of CMD and abundance of whitefly 
population by region in 2003, Republic of Congo. 
Cutting infec- Whitefly Without infec- Total Incidence Number of 
Region tion infection tion (%) B. tabacil 
plant 
Bouenza 145 ~80.8%) 8 f2.4%) 39 ~1l.8%~ 88.2 2.1 Brazzaville 145 80.6% 4 2.3% 31 17.2% 82.8 0.4 
Cuvette C. 667 76.70/U 65r5%~ 138 (15.9%) 84.1 3.4 Cuvette O. 513 85.55Vo) 13 2.2% 74 r2.3%) 87.7 1.2 Lekoumou M.Ml 8! .7O/~ 28 9.30/'12 90.7 1.6 Likouala 321 89.2% 36 10.0Vo) 90.0 2.1 Niari 400 88.9% ~312~~Vo! 37 8.2% 91.8 2.8 Plateaux 923 76.9% 67 5.6% 210 JI7.~%) 82.5 1.6 Pool North 206 68.7% 28 9.3% 65 ~ 1.4%) 78.0 0.6 Sangha 281 93.7% 3 (1.05) 16 5.3% 94.7 1.9 
Mean 4004 (81.7Vo) 212 (4.3%) 674 (13.g%) 86.2% 2.0 
Table 15. CMD severity scores for the regions oftbe Republic of Congo, 2003. 
Region CMD severity score 
1 2 3 4 5 
Bouenza 40 ll1.6%~ 68 ~19.7%l 109(31.6%) 84 ~24.3%l 44 12.8%) Brazzaville 31 17.2% 61 33.9% 57 (31.7%) 20 11.1% 
n ""I Cuvette Cs. 157 (18.0%) 175 (20.1%) 266ro.6%I 219 (25.2%) 53 6.1% Cuvette O. 74r2.3%) 80 ~13.3%l 198 33.0% 191(31.8%) 57 9.5% Lekoumou 28 9.3%) 98 32.7% 115 38.3% 45 ~15.0%l 14 4.7% Likouala 36 10.5%) 138 (40.4%) 115 33.6% 43 12.6% 102.9% 
Niari 37 8.2%) 43 (9.6%) 125l27.8%l 159 ~35.3%l 86 19.1%) Plateaux 229 (19.1%) 161(13.4%) 36230.2% 313 26.1% 135 (11.3%) 
Pool Noth 65 ~21.7%) 51 ~17.0%l 40 (13.3%) 78 i26.O%l 66 [22.0%) Sangha 16 5.3%) 32 10.7% 167 (55.7%) 74 24.7% 11 3.7%) 
Mean 13.3% 21.2% 32.6% 23.2% 9.8% 
-50-
CMD severity was much higher in Niari and Cuvette Ouest where more than 30% of fields 
had scores equal to or more than four (Table 15). The change in increasing incidence and 
severity observed in the Pool region compared to 2002 suggest continued epidemic expan-
sion in this region, although whitefly populations were relatively low. 
An overall percentage of more than 86.2 % of the cassava plants expressing CMD symptoms 
indicates clearly that CMD continues to be the dominant pest/disease constraint in the 
Republic of Congo. Furthermore, most stands were diseased through the use of infected 
planting materiaL An alarming situation therefore continues to persist in the country. 
Cassava mosaic virus types distribution in the Republic of Congo. The results (Table 
16) indicated that out of the 163 CMD infected samples 153 (82.2%) responded positively 
to the test while 28 (17.2%) did not yield a PCR product. More detailed characterization 
would be required to determine if this was due to the occurrence of a CMG not detected by 
the universal primers used, or deficiencies in the field-based DNA extraction. A totoal of 
49.6% of the positive samples were dual infections ofACMV and EACMV-UG2 [l]. Single 
infections of ACMV and EACMV-UG2 [1] made up 26_7% and 21.5%, respectively. Single 
EACMV was detected only in Sangha region and appeared to be a novel type unlike those 
occurring in EastAfrica. Dual infections ofACMV and EACMV were detected in two samples 
from Cuvette Centrale. The overall picture indicates that most of the plants (71%) were 
infected with EACMV-UG2 [1] in both single and dual infections. ACMV infected 77.7% of 
all plants sampled in both single and dual infections while EACMV infected 2.2% for both 
single and dual cases_ In conclusion, the present results show that the dominant viruses 
in the Republic of Congo are ACMV and EACMV-UG2 [1] and these results confirm what 
was reported previously from the 2002 survey. 
Gerrnplasrn development. Out of a total of 11 7 genotypes introduced into the Republic 
of Congo in 2002, 97 genotypes were developed and cloned in 2003 at Odziba research sta-
tion for further evaluation work. Evaluation of 17 UTA genotypes at three sites of CMD 
indicated that clone 92/277 was considered the best in all the three sites followed by clones 
921167 and 921 269_ Clones 92/106, 92/130, and 92/647 did not show any CMD symptoms, 
indicating resistance, but were poor in yield performance. 
Multiplication. The multiplication activity was conducted in collaboration withAgricongo 
on a contract basis whereby multiplication of clone MM 86 at three locations of three hect-
ares each (Odziba, Lekana, and Ewo) was done. In total 700,000 cuttings were distributed 
to farmers starting September 2003 by the FAO emergency program in collaboration with 
the Ministry of Agriculture. New multiplications on 9 ha of the clone MM 86 were estab-
lished again in June 2003 at the same three locations in collaboration with FAO. 
Training. Four cadres' scientists from Congo were trained in monitoring, sampling tech-
niques, and virus diagnostics at ESARC_ An in-country training course on Root crops 
production and CMD management of 30 technicians was organized in Brazzaville_ Five 
hundred farmers were trained in CMD management in Congo in four regions. 
- 51-
T
ab
le
 1
6.
 D
is
tr
ib
ut
io
n 
o
f v
ir
us
 s
pe
ci
es
 a
n
d 
s
tr
a
in
s 
in
 th
e 
di
ff
er
en
t r
e
gi
on
s 
o
f t
he
 R
ep
ub
li
c 
o
f C
on
go
, 2
00
3.
 
R
eg
io
n 
V
ir
us
 ty
pe
s 
A
C
M
V
 
E
A
C
M
V
-
E
A
C
M
V
-
A
C
M
V
 +
 E
A
CM
V
-U
g2
 [1
] 
A
C
M
V
 +
 E
A
C
M
V
-U
g1
 
T
ot
al
 
U
g2
[1
] 
C
ol
 
V
ir
us
 s
tr
ai
ns
 
A
k 
A
n 
EA
C
M
V
·U
g 
E
A
C
M
V
-
A
n+
U
I 
A
k+
U
l 
A
n+
U
2 
A
k+
U
2 
A
n+
E
l 
2[
1]
 
C
ol
 
B
ou
en
za
 
2 
1 
3 
0 
2 
1 
0 
0 
0 
9 
B
ra
zz
av
ill
e 
1 
0 
2 
0 
1 
0 
0 
0 
0 
4 
C
uv
et
te
 C
. 
0 
7 
5 
0 
14
 
0 
0 
0 
2 
28
 
C
uv
et
te
 O
. 
0 
0 
4 
0 
14
 
0 
0 
0 
0 
18
 
Q
1 
L
ek
ou
m
ou
 
0 
5 
2 
0 
0 
0 
0 
1 
0 
8 
~
 
,
 
L
ik
ou
al
a 
0 
0 
1 
0 
6 
0 
1 
0 
0 
8 
N
ia
ri
 
0 
1 
1 
0 
5 
4 
0 
0 
0 
11
 
Pl
at
ea
ux
 
0 
10
 
8 
0 
13
 
0 
0 
0 
0 
31
 
Po
ol
 N
or
th
 
3 
5 
1 
0 
0 
1 
0 
0 
0 
10
 
Sa
ng
ha
 
1 
0 
2 
0 
4 
0 
0 
0 
0 
8 
Tb
ta
l 
7 
29
 
29
 
1 
59
 
6 
1 
1 
2 
13
5 
%
 
5.
2 
21
.5
 
21
.5
 
0.
7 
43
.7
 
4.
4 
0.
7 
0.
7 
1.
5 
10
0 
W
he
re
 A
CM
V 
=
 A
fri
ca
n 
ca
ss
av
a 
m
o
sa
ic
 v
iru
s;
 E
AC
M
V·
 U
g2
 =
 U
ga
nd
a 
v
a
ria
nt
 o
f E
as
t A
fri
ca
n 
ca
ss
av
a 
m
o
sa
ic
 v
iru
s;
 E
A
CM
V
.U
gl
 =
 Pu
ta
tiv
e 
Ea
st
 A
fri
ca
n 
c
a
ss
a
v
a
 m
o
sa
ic
 
v
iru
a;
 A
k 
=
 A
CM
V 
K
en
ya
 s
tr
ai
n;
 A
n 
=
 A
CM
V 
N
ig
er
ia
 s
tr
ai
n;
 U
1 
=
 
EA
CM
V-
Ug
2 
[1)
; U
2 
=
 
EA
CM
V.
Ug
2 
[2]
; 
E1
 =
 
EA
CM
V-
Ug
1;
 E
AC
M
V-
Co
1 
=
 
U
nu
su
al
 E
A
CM
V 
st
ra
in
; 
A
n+
U
I =
A
CM
V
 N
ig
er
ia
 s
tr
ai
n 
+
 E
A
CM
V·
Ug
2 
[1)
; A
n+
U
2 
=
A
CM
V
 N
ig
er
ia
 s
tr
ai
n
 +
 E
A
CM
V.
Ug
2 
[2]
; A
k+
U
l =
A
CM
V
 K
en
ya
 s
tr
ai
n 
+
 E
A
CM
V-
U
g2
 [1
]; A
k+
U
2 
=
 
AC
M
V 
K
en
ya
 s
tr
ai
n 
+
 E
A
CM
V-
Ug
2 
[2]
; a
n
d 
A
n+
El
 =
A
CM
V
 N
ig
er
ia
 s
tr
ai
n 
+
 E
AC
M
V-
U
gl
 
Collaborators included: 
• Dr Armand Claude Mvila et Mr Albert Bembe, chercheurs au Center de Recherche en 
Amelioration Genetique des Plantes (CERAG), Ministere de I'Enseignement Superieur 
Charge de la Recherche Scientifique, Delegation Generale de la Recherche Scienti-
fique 
• Mr Maurice Obambi du Minisrere de I'Agriculture et de I'Elevage Charge de la 
Promotion de la Femme, Brazzaville, Republique du Congo 
• Mr Amadou Ouattara, the FAO representative in the Republic of Congo 
• Dr Majema J_M, the international consultant for the FAO cassava mosaic in 
Congo 
• Ministry of Agriculture 
• Ministry of Research and Technologies 
• NGO, Agricongo 
A project proposal entitled "Development of cassava technologies and enterprises for 
improved livelihood in the Republic of Congo" was developed and submitted to donors 
in the effort of mitigating the CMD effects_ Cassava may not only be described as food 
security for the population of Congo, rather it is more than that_ Each meal in Congo has 
cassava as major component either as chikwangue, leaves (sakasaka), or paste_ During the 
diagnostic surveys studies, it was observed that the daily work in all villages was related 
almost to cassa va_ 
4_14 Targeted multiplication of CMD resistant varieties in East and Central 
Mrica 
by J_ Legg., P. Ntawuruhunga, J _ Whyte_, B. Khizzah_, A Dixon, and B. Owor, in 
collaboration with H. Obiero, S. Akhwale, E. Kanju, S. Jeremiah, 1. Ndyetabura, R. 
Mayiga, S_ Bigirimana, A Ouattara, and J. Mabanza 
A program to mitigate the effects of the CMD pandemic in East Africa has been imple-
mented by UTA and it country partners since 1999, with the primary approach being the 
development, dissemination, and multiplication of CMD resistant germplasm. Through 
this program, more than 1000 clones with a background of CMD resistance have been 
introduced each to Kenya and Tanzania, and smaller but substantial numbers to Rwanda, 
Burundi, and eastern Democratic Republic of Congo. Following germplasm introduction 
through open quarantine arrangements, clones have been evaluated using a "fast-track" 
approach in which a small number of top performers (10-20) are identified from the open 
Oquarantine-based evaluation, multiplied up rapidly, then tested directly on-farm at mul-
tiple, agroecologically contrasting sites_ 
During 2003, highlights of the germplasm evaluation and exchange work included the 
completion of the evaluation of the second tranche of 17 fast-track clones in western Kenya, 
the initiation of multiplication of the best performing of the first group of 10 fast-track clones 
in Tanzania, the completion of the evaluation at three CMD hotspots of 17 UTA-derived 
clones in ROC and the identification of seven CMD-resistant clones from 15 tested under 
extreme CMD inoculum pressure conditions in northeastern Burundi. 
In partnership with NARS, NGOs, extension, and other stakeholders in Uganda, Kenya, 
Tanzania, ROC, and Burundi, more than 24 000 ha of CMD-resistant germplasm were 
multiplied, distributed, and planted in farmers' fields to mitigate the effects of the CMD 
pandemic_ Much of this multiplication and dissemination work was still based around 
-53-
the early TMS-type, CMD-resistant varieties, and in many areas these have been readily 
accepted_ However, increasingly, a greater diversity of germplasm, including TMS, TME, 
and TMS by TME cross material, is being introduced into the mass multiplication program. 
This provision of greater choice of germplasm with the whole range of quality character-
istics has had a marked beneficial effect on farmer adoption. 
A regional stakeholders workshop with the twin aims of reviewing project progress and 
planning for future work was held at Bukoba, Tanzania and involved participants from 
research, quarantine services, extension, and NGOs from Uganda, Kenya, Tanzania, 
Burundi, and the Republic of Congo. A proposal for a further one-year phase to run from 
October 2003 to September 2004 was approved by USAID, but will involve a smaller group 
of countries, comprising Kenya, Tanzania, and Burundi. 
4.15 Development of an action plan for casssava brown streak disease research 
in East and Southern Africa 
by J. Legg, M. Andrade, J. Whyte, B. Khizzah, E. Kanju, and R. Hanna, in collabo-
ration with R. Njeru, T. Munga, K. Mtunda, A. Zacarias, F. Chipungu, G. Kaitisha, 
R. Hillocks, M. Thresh, and D. Jennings 
Proceedings of the October 2002 cassava brown streak disease (CBSD) stakeholders work-
shop, held in Mombasa, Kenya, were edited and published through DFID's Crop Protection 
Programme. Six hundred copies were printed and these were distributed to participants, 
institutions, and other stakeholders, many of them in the different countries of Eastern 
and Southern Africa affected by CBSD. In order to strengthen UTA's role as a stake-
holder in CBSD research for development, a proposal was developed which targeted the 
strengthening of UTA fundamental research on CBSD. An important element of this was 
the collection of molecular data on isolates of Cassava brown streak virus from all areas 
affected by CBSD, so that new UTA plans to develop a transgene-mediated approach to 
resistance development would be fully supported by required baseline data. The strategic 
CBSD research program was to be financed through an extra core contribution to UTA 
from the Canadian International Development Agency. This program is expected to be 
initiated during 2004. Another of the main objectives of the new project will be to start to 
implement the action plan developed during the CBSD workshop and to strengthen UTA's 
coordinating role in this field of research. 
4.16 Assessment methods for banana weevil 
by C. Gold and P. Ragama, in collaboration with R. Coe and N.D.T.M Rukazam-
buga 
The banana weevil is an indirect pest on bananas and plantain. The larvae attack the 
corm, weakening stability of the plant and impeding nutrient uptake. Damage assess-
ment requires destructive sampling and is most often done on recent harvested plants. 
A wide range of methods have been developed to assess banana weevil damage. These 
include estimates of damage to the corm periphery (exposed by corm paring) and to the 
corm cortex and central cylinder (in cross cuts). The different damage estimates are not 
always well correlated with each other. In one screening trial, Nsowe had the highest level 
of damage to the corm periphery and the lowest level of damage to the central cylinder. 
Thus, interpretation of Nsowe's level of resistance to banana weevil depends upon the 
sampling parameter selected. 
-54-
At present there is no agreed sampling protocol. This is, in part, because there is no clear 
understanding on which damage parameters are most important, i.e., which types of 
damage have the greatest effect on banana yield. 
Preliminary analyses comparing assessment methods were reported in 2002. Further analy-
ses were conducted in 2003, incorporating additional data_ Multiple damage assessment 
parameters were employed in two long duration yield loss trials (cv_ Atwalira, Musa spp. 
AAA-EA) and a cultivar screening trial in Uganda_ Parameters included two estimates of 
peripheral damage on pared corms and estimates of damage to the central cylinder and 
cortex (plus a derived total damage score) observed in cross sections. In the first two trials, 
estimated yield losses to banana weevil exceeded 40% in latter cycles. 
Damage to the central cylinder had a greater effect on plant size and yield loss than damage 
to the cortex or corm periphery_ In some cases, a combined assessment of damage to the 
central cylinder and cortex showed a better relationship with yield loss than an assessment 
of the central cylinder alone. Correlation analysis showed weak to modest relationships 
between damage to the corm periphery and damage to the central cylinder_ Thus, damage 
to the corm periphery (less labor intensive to assess) is not a strong predictor of the more 
important damage to the central cylinder. Therefore, banana weevil damage assessment 
should be made for the central cylinder and cortex. 
4_17 Fungal endophytes for the microbial control of the banana weevil, nema-
todes and Fusarium wilt 
by B_ Niere, T. Dubois, C.S_ Gold, D. Coyne, and C. Nankinga, in collaboration with 
R. Sikora, E. Adipala, A. Viljoen, S. Kapindu*, N. Labauschagne, S_ Athman*, P_ 
Paparu*, P. Kilama* , and E_ Mukwaha* 
Banana weevil (Cosmopolites sordidus), nematodes (Radopholus similis, Helicotylenchus 
multicinctus, and others) and Fusarium wilt (Fusarium oxysporum f. sp_ cubense) are 
soilborne pests which attack the roots, corm, and vascular system weakening plant sup-
port and impeding nutrient uptake_ Clean planting propagules, like tissue-culture-derived 
plants, can be used to establish new fields, although pest (re)infestation remains a vital 
concern_ Microbial control offers excellent possibilities for controlling these pests_ The 
primary focus of this project is the use of endophytes for the management of banana pests 
and diseases. Fungal endophytes are microorganisms that colonize plant tissue internally 
for at least part of their life cycle without causing symptoms of disease_ Many endophytes 
have formed mutualistic relationships with their host plants and serve as antagonists 
to pests and diseases_ These endophytes can be inoculated into tissue culture plants and 
reduce pest and disease pressure. 
Screening of endophytes against banana weevils and nematodes_ Fifteen endophytic 
F. oxysporum isolates that were obtained from highland banana have been screened in vitro 
for effects of fungal culture filtrates on banana weevil eggs and nematode juvenile stages 
using a banana corm tissue medium and an artificial medium. In vitro screening protocols 
have been successfully established for weevil screening, and most isolates cause between 
40 and 50% egg mortality compared to 5% in the control. In vitro screening protocols using 
endophytic F. oxysporum against nematodes were improved by reducing mortality rates in 
the control from 30% to 5%. Fifteen in vivo screening experiments were conducted, screen-
ing a total of 14 endophytic F_ oxysporum isolates. The effects of several isolates on banana 
weevil or nematode infested plants were very promising_ For example, in banana weevil 
infested plants, inoculation of isolate Emb2_ 40 reduced inner corm damage by 45%, reduced 
- 55-
oviposition by 46%, and increased dry shoot weight by 39% compared to control plants. In 
nematode infested plants, isolate V5w2 reduced nematode population densities by 29%, 
decreased root necrosis by 11%, and increased fresh shoot weight by 11%. Interestingly, 
the data show that some isolates can contribute to the control of both pests. Isolate V5w2 
is now being tested in two on-station field experiments under different levels of nematode 
infestation. 
Studies on endophyte colonization, persistence and interactions. For an endophyte 
to provide sustained protection against banana pests, it is essential that the fungus suc-
cessfully colonizes and persists from the time of inoculation beyond the screenhouse period 
into the time of field planting. Also, interactions of endophytic fungi with their host plant, 
banana pests, and each other are not well understood. Two in vivo screening experiments 
that investigate combining different endophyte isolates for nematode control at different 
infestation levels revealed that endophyte cocktails (multiple isolate inoculation) might 
enhance their beneficial effect against nematodes and that endophytes retain their full 
beneficial effect even at increased nematode pressure. Three in vivo screenhouse and field 
experiments that investigate endophyte inoculation methods and subsequent persistence 
showed that endophytes can persist within the banana corm for at least 16 weeks, showing 
colonization rates of up to 81 %. 
Biodiversity and isolation offungal endophytes in highland banana. A biodiversity 
study at selected sites in Uganda on the occurrence and distribution offungal endophytes 
in banana yielded more than 400 new endophytic isolates. Results from morphological 
studies confirm earlier studies that F. oxysporum is the dominant fungal species grow-
ing endophytic ally in banana corms, roots, and pseudostem bases. Diversity among these 
endophytic strains of F. oxysporum and relatedness with pathogenic forms will be deter-
mined with molecular marker techniques. These isolates have been carefully documented 
and stored and are now also being used in new in vitro and in vivo screening for control of 
banana weevils and nematodes. 
Screening of endophytes against Fusarium wilt. Ten endophyte isolates were tested 
in in vivo screening experiments for activity against Fusarium wilt. A very aggressive 
Fusarium wilt (i.e., reflecting a highly virulent isolate of F. oxysporum f. sp. cubense) was 
used in these experiments and this explains in part why none of the isolates tested so far 
showed consistent activity against Fusarium wilt. 
Characterization of Fusarium wilt in Uganda. As a foundation for conducting endo-
phyte-Fusarium wilt studies, proper characterization of 108 Fusarium wilt isolates in 
Uganda, employing detailed morphological, molecular (AFLP), biochemical (VCG), and 
biological (pathogenicity tests) characteristics, is being compiled. The isolates were obtained 
from diseased banana plants in the major banana growing areas in Central and Southwest-
ern Uganda and as far north as Arua district. Fusarium wilt was, among other cultivars, 
also found on the banana cultivar Bluggoe, the differential host for race 2 of Fusarium 
wilt. Race 2 was up until now thought to be absent in Uganda. Since our preliminary data 
suggest that both race 1 and race 2 of the pathogen are present in Uganda, this finding 
might have profound implications in banana breeding programs. 
Field-testing of different banana planting materials. Although tissue culture technol· 
ogy is already accepted and used in many Central and East African countries, this technol-
ogy is currently not available to Ugandan farmers. Tissue-cultured banana is however an 
excellent source of pest- and disease-free planting material. The high cost of the tissue-
-56-
culture-derived planting material is seen as one of the reasons for slow dissemination and 
adoption of this technology in Uganda. 'Ib further the promotion and acceptance of tissue 
culture technology in Uganda, a field trial comparing different types of planting material 
of banana (tissue culture, hot-water treated suckers, and untreated suckers) of the widely 
grown East African highland banana cultivar Enyeru was planted in 2002. Harvest of the 
mother plants began in mid-2003 and will continue into mid-2004. In the ongoing field 
trial, each type of planting material is tested under high nematode pressure and under 
two different management practices, with and without mulch and manure. Economic 
analysis of the obtained data will help to determine the best type of planting material for 
farmers in Central Uganda. Preliminary results show that tissue culture banana plants 
produce heavier bunches and grow faster than sucker-derived plants. Since the present 
project depends on the promotion and acceptance of tissue culture plants, this finding is 
exciting because it highlights the potential of using tissue culture plants during the first 
crop cycle, even without endophytes, and will without doubt aid in the promotion of tissue 
culture in Uganda. 
Use of Paecilomyces spp. against banana nematodes. The use of Paecilomyces spp. for 
the control of R. similis was investigated in 12 in vitro bioassays and two in vivo screening 
experiments, using 30 isolates obtained from the banana rhizosphere. Bioassay protocols 
were not well established and as a result control mortalities were too high. Several improve-
ments, such as pH control, filtering of the fungal medium, rinsing of the bioassay medium 
and stimulation of nematodes, reduced control mortality from> 60% to ± 10% and yielded 
at least two promising Paecilomyces spp. isolates (23N5-2 and llNS-4). 
Tissue culture facility. A tissue culture lab is fully operational and routinely producing 
plants of six banana cultivars for all in vivo screenhouse and field experiments conducted 
at Sendusu, Uganda. Monthly output stands at 250-350 plants at the moment. A total of 
3750 plants were weaned in the reporting period. Output is currently maximized due to 
space limitations. 
4.18 Banana weevil response to synthetic pheromone lures and other infochemi-
cals 
by C. Gold and C. Nankinga, in collaboration with W. Tinzaara*, M. Dicke, A. van 
Huis, G. Kagezi, and W.K. Tushemereirwe 
Olfactory cues may be utilized by banana weevils in locating host plants, conspecifics, andior 
mates. A male-produced aggregation pheromone, sordidin, to which both female and male 
banana weevils respond, has been identified and synthesized. Chemtica International, Inc. 
in Costa Rica has begun commercial production of banana weevil pheromones in a product 
called Cosmolure+. These pheromones are distributed in small packets that are most often 
used as lures to pitfall traps. Studies in Costa Rica suggest that placement of only four 
traps per hectare, changed monthly and systematically moved through a banana stand, 
can reduce weevil populations and damage and lead to yield increases within one year. 
Trap effectiveness has not been well investigated under African agroecological conditions. 
Preliminary field tests have shown up to 18 times higher capture rates in pheromone traps 
than in traditional pseudostem traps. The attractiveness of the banana weevil to these 
pheromones might be enhanced by the addition of plant kairomones but further research 
is needed in this area. 
- 57-
The overall objective of this PhD research project is to develop an infochernical-based 
trapping system that is cost effective and can control the banana weevil under Ugandan 
conditions. 'Ib do this, the chemical and behavioral ecology of the banana weevil must 
be understood. The research focus of this study is therefore to elucidate pheromone and 
kairomone trap efficiency as related to weevil behavior and environmental conditions. 
This study is also investigating the impact of adult removal through trapping on weevil 
population dynamics. An effort will also be made to evaluate the potential of combining 
infochemical use with biological control methods (i.e., entomopathogens) for the manage-
ment of banana weevil. 
The relative attractivity of host plant volatiles and synthetic pheromone to 
banana weevil. Laboratory experiments were conducted in the dark room at Sendusu 
using a double pitfall olfactometer. The olfactometric results indicated that fermented 
banana pseudostem tissue (Cultivar Mbwazirume, Musa spp AAA-EA group) was equally 
attractive to banana weevil as pheromones but more attractive than fresh pseudostem 
tissue. Laboratory results demonstrated the additive effect of pseudostem tissue to the 
pheromone. Although all the four field experiments conducted showed an additive effect 
to the pheromone, a significant effect was only showed with two out of four experiments. 
Both results of laboratory and field experiments indicated an increased weevil attraction 
with an increase in pheromone release rate. A regression analysis should showed signifi-
cant relationship between the amount of fermented pseudostem tissue used and the weevil 
trap catches. 
The influence of female mating status, age and density on C. sordidus response 
to the pheromone. Laboratory activities were conducted in the dark room at Sendusu 
using a double pitfall olfactometer. Sexually immature and mature banana weevils (i.e., 
1-14 and 30-40 days after eclosion, respectively) collected from the culture were tested for 
their response to pheromone lures in a double pitfall olfactometer. 
Response by mated and unmated females to synthetic pheromones. Unmated wee-
vils responded significantly more (P-value = 0.031, x2-test, 2x2 contingency table test) to 
pheromones than mated ones. Previous weevil density appears not to have a significant 
effect on weevil response to pheromones. Field experiments were conducted at Kawanda 
and Sendusu. There were more weevils captured at higher weevil density (17.9 % of the 
released weevils) than at low weevil density (9.8% of the released weevils). More unmated 
weevils (24.1 % of the released) than mated (19.2%) were recaptured in pheromone baited-
traps. 
Factors influencing the effectiveness of the pheromone trap in attracting banana 
weevil. Experiments were conducted at Kawanda and Senge (5 km NW of Kawanda) to 
determine: (1) the effective radius of pheromone traps for attraction of banana weevil; 
(2) the sex and mating status of banana attracted to the pheromone traps in different 
locations relative to the host plant; (3) the upwind and downwind orientation of banana 
weevil in response to the pheromone source; (4) the effect of covering pheromone traps 
on weevil response; and (5) the effect of weevil behavior around the trap on catches. Trap 
performance decreased with increase in radius of a given trap area, and the effective trap 
radius tended to level off after 9 m. More weevils were recaptured within less than 5 m 
(15%) from the pheromone trap than from within 5-10 m (9%) and 10-15 m (5%). Relative 
humidity showed a significant positive relationship to banana weevil catches in pheromone 
traps while wind speed, temperature, and rainfall showed a weak relationship. Phero-
- 58-
mone-baited traps that were covered with banana leaves captured lower numbers of C_ 
sordidus compared to uncovered traps_ More weevils were captured in pheromone-baited 
traps placed in alleys than on mats. 
The effect of mulching on weevil movement relative to pheromone traps. The 
experiment was conducted in the field established at Sendusu to: (i) determine the effect of 
mulching on weevil catches in pheromone traps; (ii) to determine the distance and direction 
moved relative to the location of pheromone traps in mulched and nonmulched areas; and 
(iii) discriminate differences between sexes in terms of weevil response to pheromone traps 
in mulched and nonmulched plots. The treatments were: (i) heavy mulch, (ii) light mulch, 
and (iii) no mulch (trash). The percentage of weevils recaptured was higher in heavy mulch 
plots (10.9%) as compare to light mulch (8.1 %) and no mulch plots (6.9%). There were twice 
as many females than males recaptured and the mulch level had a limited effect on the sex 
ratio of recaptured weevils. Direction had a limited effect on weevils recaptured in mulched 
and nonmulched plots. This experiment is being repeated (December 2003-March 2004). 
Effect of pheromone-baited trap density on banana weevil populations. An on-farm 
study to evaluate the effect of pheromone trap density on the population of the banana 
weevil was conducted in Masaka District, Uganda. Forty-two farms were assigned to one 
of three treatments: control, four and eight traps/ha. The density of the traps depended on 
banana stand size. Lures were changed monthly at which time the traps were moved to a 
different location within the stand. Adult populations were estimated by using mark and 
recapture methods at 0, 6, 12, 18, and 21 months, while damage was evaluated at 0, 3,6, 
12, 18, and 21 months. Pheromone trap captures were generally low and not influenced by 
rainfall. Changes in banana weevil populations and corm damage were negligible. Doubling 
the number of traps did not increase catches. 
Response of banana weevil predators towards herbivore-induced plant volatiles 
and pheromones. Laboratory bioassays are being conducted to (i) investigate the effi-
cacy of different volatiles in attracting the banana weevil predators and (ii) determine the 
role of pheromones in enhancing the host finding by predators of the banana weevil. The 
predators that are being evaluated for response to host plant volatiles and the pheromone 
are (i) Dactylosternum abdominale (Hydrophilidae: Coleoptera) and (n) the ant Pheidole 
megacephala (Hymenoptera: Formicidae). 
The potential of using pheromone traps in disseminating B. bassiana for the con-
trol of banana weevil. The objectives of these trials were: (1) to determine the potential 
for the pheromone-baited traps to aggregate banana weevils around the trap mat; (2) to 
investigate the potential for banana weevils to become infected with pathogens in or around 
the trap; (3) to determine the rate of dissemination of the pathogens by infected individuals 
of banana weevil; and (4) to determine the effect of using B. bassiana in combination with 
pheromone traps in suppressing banana weevil populations and damage. 
Aggregation potential. The experiment was conducted at Kawanda and Sendusu. Twice 
as many C. sordidus were captured in pseudostem traps at the base of the trap mat (phero-
mone-baited mat) than at mats < 5 m from the trap mat, and four times as many at mats 
> 5 m away. The results suggest that many weevils may he attracted to pheromone lures 
but fail to enter the traps (only a small proportion are caught). This may result in weevil 
aggregation in the vicinity of pheromone traps. 
-59-
Field transmission. Of the weevils that were recovered by searching, 13.1 % that were 
released uninfected and 3.1% of the weevils unmarked died due pathogen infection after 
incubation in the laboratory. Of the weevils, which were recovered by pseudostem traps, 
7.1 % of the weevils released uninfected and 1.8% unmarked weevils died due to pathogen 
infection after incubation in the laboratory. The weevils (released uninfected and unmarked) 
that died due pathogen infection picked it from the weevils that were released infected. 
The results show that the pathogen could be transmitted from an infected weevil to a 
noninfected weevil in the field. 
Effects of Beauveria bassiana infection on weevil behavior. The experiments were 
conducted in the field and in open cages at Sendusu. The effect of Beauveria infection on 
weevil location and movement was evaluated. The effect of the level of infection with Beauve· 
ria on weevil behavior was also studied. Most dead weevils (>70%) due to pathogen infection 
were found in the leaf sheath and in soil by a mat in the field trial. In the open caged trial, 
most weevils (70.7%) were found on the corm surface and then leaf sheath (8.6%) and in 
soil by the plant (9.2%). About 30% of the infected weevils were found to move between 3 
and 12 m, suggesting that transmission of the fungus from a mat to another is possible. 
Delivery systems of Beauveria bassiana using pheromones. The experiment was 
conducted at Sendusu in the former yield loss trial. The treatments were: control (noth· 
ing), pathogen only applied to central mat, pheromone + pathogen inside trap, pheromone 
+ pathogen around the trap mat, and pheromone + pathogen on trap mat and mats less 5 
m. More weevils died due to B. bassiana infection in plots where B. bassiana was applied 
on the trap mat and four adjacent mats (14.2%) than where the pathogen was applied 
in the pheromone trap (4.1%) and around the pheromone trap (7.2%). There were more 
weevils that got infected per 200 g weight of B. bassiana for the pathogen applied around 
the pheromone trap (16.7 weevils/200 g B. bassiana) than when the pathogen was applied 
inside the pheromone-baited trap (9.0 weevilsl200 g B. bassiana) and on the trap mat and 
four adjacent mats (5.7 weevils/200 g B. bassiana). 
Impact of integrating B. bassiana with pheromones. The experiments are currently 
in progress at Kawanda and Sendusu to determine the impact of using B. bassiana in com-
bination with pheromone- baited traps on weevil population and damage. The treatments 
are: (i) control, (ii) B. bassiana on entire plot of 35 mats, and (iii) Pheromone + pathogen 
at trap mat and adjacent mats_ 
4.19 Diagnostics in banana-coffee systems 
by S. Okech, C. Gold, and S. Abele 
Banana and coffee form the economic base for the majority of the population of the Great 
Lakes region of Mrica. The two crops are grown in various patterns as intercrops or sepa-
rately as monocultures. However, the yield trends of the two crops over the last 20 years 
suggest that the system is declining, and this decline severely affects rural livelihoods in 
the region. It is important to generate information about the system on (i) the interaction 
(socioeconomic, agronomic, pests and diseases) of these two crops and (ii) how best these 
systems can be established, developed, and improved. 
A multidisciplinary team comprising specialists in Integrated Pest Management (IPM), 
agronomy, economics, and agricultural extension from the International Institute ofTropi-
cal Agriculture (UTA), National Agricultural Research Organization (NARO)--banana 
- 60-
and coffee research programs, and Mbarara District Department of agriculture conducted 
a survey in April and May 2003 with the following objectives: 
• Determine the importance of banana as a food security and cash crop vis-a.-vis coffee 
as a cash crop_ 
• Explain trends and shifts in production of banana and coffee, management practices, 
production constraints, market opportunities etc_ 
• Elucidate farmers' perceptions on production and management constraints with a view 
to developing collaborative an on-site research program on a sustainable banana---{!offee 
production system_ 
• Derive hypotheses for participatory research on the above-mentioned topics_ 
PRAs and diagnostic surveys were conducted at seven sites in Uganda during May/June 
2003 and a diagnostic survey at one site in December 2003 to answer the above objec-
tives. 
Economic aspects of the banana-coffee system. The most important economic rela-
tionship between the two crops is that they provide seasonal liquidity (especially coffee) 
for their mutual benefit on management. During the whole year, banana sales provide 
liquidity for hiring labor for banana and coffee management; coffee provides liquidity for 
banana management during the coffee-harvest season. 
Banana and coffee production systems. Three systems of production were reported 
across all the sites: (i) banana or coffee monoculture, (ii) banana, and (iii) banana + coffee 
+ annual crops (mainly beans). Perceived benefits from the banana coffee intercropping 
system were given by the farmers as: (1) to maximize on space and labor resources uti-
lization, all sites; (2) to maximize on limited land, all sites; (3) to replace an old banana 
plantation on a poor land, two sites; (4) bananas provide nutrients/moisture for young 
coffee, five sites; (5) banana provides shade and mulch for coffee, five sites; (6) coffee husk 
is used for soil fertility maintenance, six sites. The only constraint associated with the 
banana coffee intercropping system reported by the farmers was that coffee mines the 
soil and kills banana after some time--reported at all sites. Intercropped plots eventually 
turn into coffee monoculture. Time taken for bananas to die out in the intercrop depends 
on management and soil fertility but takes about five to more than 10 years. 
Constraints to the banana-coffee system. Market (low fluctuating prices) for both 
bananas and coffee was the common leading constraint at all sites followed by pests and 
diseases_ Declining soil fertility was ranked third or fourth at most sites. Poor management 
was noted at all sites and was associated with lack of resources including equipmentltools 
and labor. Shortage of land was ranked high among the constraints but it was given a low 
priority in this PRA because of the survey objectives. 
The key pest that farmers associated with banana in the system was banana weevil while 
coffee wilt was the main coffee disease. Farmers' observations on relationships between 
banana weevil and the production system adopted varied. Farmers at three sites believed 
that the weevil incidence was low in the banana-coffee intercrop while farmers at four 
sites reported that the incidence was high in the intercrop because of poor management 
of bananas usually associated with the intercropping. Coffee wilt was a new problem and 
farmers understood very little about it_ Diagnostic survey results showed that banana 
weevil incidence was consistently low in farms where coffee husk was used as mulch. 
-61-
The way forward. Both farming and marketing systems at the surveyed sites are complex 
and diverse and a lot of research is required to provide optimal support for the development 
of these systems. Both the economic and agronomic potential, especially of intercropping 
systems, is not yet fully exploited. Banana and coffee playa major economic role as cash 
crops and food crop (banana) at all the sites. Whether they are grown intercropped with each 
other or cultivated in pure stands, banana and coffee have many economic and agronomic 
relationships, both positive and negative. These relationships have to be further developed 
and improved. Behavior patterns of the farmers are not only profit orientation, but also 
liquidity and food security, and play an important role in economic decision.making. The 
major constraints to production are land scarcity, poor marketing options, and therefore 
a lack of liquidity, which leads to poor crop management and poor exploitation of the eco· 
nomic potential. 
The survey team proposed the following hypotheses to be tested during further research: 
(a) Optimizing banana--coffee intercropping management (proper spacing and adoption of 
improved technologies) will improve soil fertility, minimize pest problems, increase longev· 
ity of the plantations, and result in higher yields of both banana and coffee under the cur· 
rent key constraint of land and labor shortage. (b) In particular, application of coffee husk 
as mulch and soil fertility amendment in banana plantation reduces the banana weevil 
constraint and increases banana yield. (c) Improved marketing organization, by establish-
ing better market information, by taking over marketing functions through the farmer 
(replacing middlemen with farmers' organizations) will increase farmers' cash income. 
4.20 Banana weevil pest status: effects of NPK fertilizers on yields loss from 
banana weevil in Mbarara, Uganda 
by S. Okech and C. Gold 
In Mbarara, a long-term yield loss trial was planted to test the hypothesis that bananas 
growing under fertile soil conditions are more tolerant of banana weevil attack than those 
grown in less fertile soils. Amendments of NPK fertilizer to banana plots (cv. Enyeru) pro-
vided different levels of soil fertility in a trial planted at the Mbarara stock farm. Treat-
ments were: (1) fertilizer plus weevils; (2) no fertilizer plus weevils; (3) fertilizer without 
weevils; and (4) no fertilizer and no weevils. The study has been going on since 1996. The 
first five crop cycles (1996-2001) did not receive mulch as part of crop management. The 
subsequent crop cycles (6-8) received uniform application of mulch in all treatments. Plot 
size per treatment was 49 mats at a spacing of 3 x 3 m. Plots were separated by 20 m grass 
alleys. Fertilizers were applied at the rate of 100 kg Nlhalyearin a four-split application 
(Urea 46%), 50 kg Plhalyear in a two-split application (Triple Supper Phosphate 46%) and 
100 kg Klhal year (Muriate of Potash 52%) in a two-split application. Dursban was applied 
regularly to treatments 3 and 4 to maintain the plots free of weevils. 
Adding fertilizer provided minor yield advantages « 15%) during the first three cycles and 
a major yield advantage during the fourth cycle in both weevil free and weevil infested 
plots. In unfertilized plots, yield losses (of the cycles that did not receive mulch) increased 
from 5% in the plant crop to 40% in the 3rd ratoon. Weevil damage was 4.8%. The losses 
declined to 31% in the 7th ratoon after two years of application of mulch. Average weevil 
damage in the 7th ratoon was 3.5%. 
-62-
Whereas application of fertilizers provided a yield gain of up to 40% in the third ratoon 
(no mulch application phase) in both weevil free and weevil infested plots, a yield loss of 
14.9% was recorded in the 7th ratoon (after two years of mulch application) in the fertilizer 
with weevil plots. The yield gain from fertilizer without weevils was 7.8%. Application of 
mulch increased the plant vigor and growth of those plots that did not receive fertilizer. 
The results demonstrate that weevil damage, as low as 3%, can cause stunting of banana 
suckers and loss in yield when soil fertility is low. 
Banana weevil larvae feeding on the banana corm periphery can effect root detachment, 
while internal feeding is likely to influence root initiation and development. However, there 
are no available data quantifying the effect of weevil damage on banana root number. A 
study to investigate the effect of weevil attack on banana root number at flowering was 
conducted in three sites in Uganda (Kawanda, Masaka, and Ntungamo). Two hundred and 
ten recently flowered plants were uprooted, assessed for weevil and nematode damage, 
and for root number. All peripheral and internal weevil damage indicators were negatively 
correlated with the number of roots, whereas nematode root necrosis was not. In this study 
root number appeared to be more affected by damage to the corm interior than to the corm 
periphery (even though earlier observations suggested that root loss was proportional to 
the degree of attack on the corm surface). Heavily weevil-damaged plants had 30% fewer 
roots at flowering than lightly damaged plants. 
4.21 Studies on endemic ants and banana weevil in Uganda 
by C. Gold, in collaboration with A.M.K. Abera* and R. van Driesche 
The potential of indigenous natural enemies, especially ants, to control banana weevils in 
Uganda is being investigated through laboratory, screenhouse, and field trials as part of a 
PhD research program that will be completed in 2004. Ants are a major component of the 
arthropod community in the banana farming systems in the tropics. Surveys in Uganda 
and Tanzania have shown highly diversified ant fauna to be active in banana stands. 
Although ants are opportunistic predators, they are increasingly being used in biological 
control because they possess attributes that make them effective foragers. 
A survey conducted in major banana growing areas showed that banana-farming systems 
have a high species complex. Fifty species were encountered at the survey sites, while the 
number on individual farms ranged from 19 to 34. Many ofthese species could be considered 
as potential predators on banana weevil eggs. Myrmicine ants, which have been reported 
as effective predators on banana weevils in Cuba, were abundant on banana farms. For 
example, the Pheidole species complex comprised the predominant ant group on most farms. 
These were observed nesting in the soil, in standing plants, and in crop residues. Another 
important species, Odontomachus troglodytes, was considered by most farmers as a pest, 
but has also been reported as a predator. These were commonly found nesting at the base 
of banana mats or in crop residues. 
Both Pheidole sp. and O. troglodytes attacked and removed live baits of banana weevil eggs 
and larvae in the lab and in the field. This served as an indicator but could not be deemed 
conclusive, as many species of ants will remove readily available protein sources. We then 
tested the potential of these two ant species to extract weevil stages from living plants in 
microcosm experiments conducted within the laboratory. At low weevil density (i.e., two 
males and two females per plant), the consistent presence of ants in the experimental 
arena resulted in a 54% reduction in weevil eggs, 44% weevil larvae, and 100% reduction 
-63-
in weevil pupae. Weevil damage was 25% lower on the corm periphery and 57% lower in 
the corm interior (i.e., cross section measurement) than in controls. At higher weevil den· 
sities of five males and five females, a smaller proportion of weevil immature stages were 
removed. Proportionately more weevil stages escaped predation and damage reductions 
were 12% on the corm periphery and 38% in the corm interior. 
We also tested the ability of Plu!idole sp. and O. troglodytes to remove weevil stages from 
banana crop residues. In drum experiments, the two species reduced weevil egg numbers 
by 72% and 67%, respectively. Pheidole sp. also removed 16% of banana weevil larvae, 
while neither species removed banana weevil pupae. In four runs of a field experiment 
using larvae inserted into decomposing banana corm residues, neither species had much 
impact on larval number. In these experiments, it appeared that the ants were unable to 
extract larvae that had burrowed away from the corm's surface. 
Trials were planted at Sendusu in 2002 to determine the effects of Pheidole sp. and O. trog· 
lodytes on banana weevil damage under field conditions. This trial will continue through 
2004. 
4.22 Efficacy and persistence studies of the entomopathogen Beauveria bassi-
ana under different banana management practices 
by C. Nankinga, C. Gold, and T. Dubois, in collaboration with 1. Godonou, D. Moore, 
S. Gowen, w.K. Tushemereirwe, E. Magara*, V. Tumuhaise*, R. Kawuki, and S. 
Kyamanywa 
The effectiveness of a biopesticide depends not only on its initial capacity to kill the pest 
but also on its capacity to persist in the environment where it is applied and on its killing 
action. Some of the abiotic factors that can influence a biopesticide are pH, temperature, soil 
moisture, and UV light. The presence of antagonistic fungi, bacteria, and microorganisms 
that exist in soil with high organic matter have been reported to influence the efficacy and 
persistence of B. bassiana. Earlier, we observed that soil amendments such as ash, urine, 
tobacco, and pepper could influence the efficacy and persistence of the entomopathogenic 
fungi. Stable manure and coffee husks are common types of soil organic amendments that 
farmers apply in banana plantations to enhance soil fertility. These amendments can also 
influence the physical and chemical factors (e.g., pH, moisture content) of the soil. Studies 
were undertaken to investigate how such farmer practices like application of organic amend· 
ments and spacing of the banana plants can influence the performance of B. bassiana. 
In one experiment, maize formulated B. bassiana was applied at a rate of 200 g: (1) around 
the mats and then mulched; (2) around banana mats without mulch, and (3) applied 2-3 
cm below the soil surface. B. bassiana was applied on four randomly selected banana mats. 
The persistence of B. bassiana was assessed at two· week intervals by picking the Beauveria 
formulation from the field and exposing it to ten live adult weevils overnight. Microorgan· 
isms associated with B. bassiana formulations were also collected and identified. Results 
indicated that infectivity of B. bassiana gradually decreased after field application from 
90-99% after 14 days to 9.2 to 64.2% after 56 days. Beauveria formulation degradation 
was markedly faster when buried, as reflected by the low mortality «10%) caused to wee-
vils exposed to buried samples picked from the field after 56 days. Two microorganisms, 
Penicillium chrysogenum and Aspergillus flavus, were identified attacking the formulation 
substrate. Of these, P. chrysogenum more frequently colonized the maize-based formula-
tion of B. bassiana. 
-64-
A field trial on the effect of spacing and plant density on efficacy and persistence of B. bassi-
ana was established at Kawanda in October 2001. Banana plants were planted at spacings 
of 2 x2 m, 2.5 x 2.5 m, and 3 x 3 m. These spacings were chosen to simulate common plant 
densities in farmers' fields. It was hypothesized that different plant densities would affect 
the soil environment (sunlight penetration, soil moisture and temperature, microorganism 
activity) and growth characteristics of the plant, which will directly or indirectly influence 
the performance of the formulations or fungus applied in a banana field. The efficacy and 
persistence of B. bassiana is being evaluated with a solid maize-based formulation applied 
at two doses and two application frequencies (every 8 and 12 weeks). The two doses being 
tested are: dose 1 (high dose = 200 g of maize·based formulation, measuring to approxi-
mately 1015-16 conidialhectare) and dose 2 (low dose = 100 g of maize· based formulation, 
measuring to approximately lO" conidialhectare). Preliminary results showed low naturally 
occurring weevil populations and damage levels with no significant differences in the three 
plant spacings. Growth parameters show that plants planted at 2 x 2 m spacing show lower 
girth, height, and bunch weights. Data collection will continue for at least two more plant 
cycles. Beauveria bassiana samples have also been taken from the field to assess viability 
of the applied fungus. 
To investigate how organic materials applied in banana fields can influence the performance 
of B. bassi ana, pot experiments were conducted to evaluate the efficacy and persistence of 
B. bassiana applied as a maize· based formulation to soil amended with different levels of 
coffee husks, decomposed cow dung, and artificial fertilizers and planted with suckers of 
local cooking banana "Mpologoma". The treatments involved applying B. bassiana to 10 
combinations of soil and organic amendments (loam soil not mixed with any organic amend-
ments or soil mixed with coffee husks, cow dung, and inorganic fertilizers [NPK at rate of 
69 g/plant] in the ratios of 1:1, 1:2, 1:3). Coffee husks and cow dung were also evaluated 
as pure stands without soil. To check on the persistence of the fungus, 10 banana weevils 
were released into the pots (applied with 100 g of maize· based formulation of B. bassiana) 
at four-week intervals and monitored for mortality and Beauveria infection for 16 weeks. 
Generally B. bassiana applied to soil without any amendment caused higher mortality than 
all the other treatments and B. bassiana applied to pots with only coffee husks showed the 
lowest infectivity. Infectivity of the maize-based formulation of B. bassiana was observed 
to decrease in all soils mixed with the different organic and chemical amendments. Higher 
mortality (30-70%) was obtained in weevils released to the pots immediately after B. 
bassiana application. The mortality decreased below 50% in weevils released 60 days after 
applying B. bassiana. This decreased infectivity was attributed to the degraded maize· based 
formulation of B. bassiana that led to contamination ofthe fungus by other microorganisms 
in the soiL These results confirm earlier observations that suggested that the fungus on 
the maize-based solid substrate would need to be applied every 60-90 days to enhance its 
infectivity in the field. Otherwise, if longer persistence is required, more stable formula· 
tions will have to be developed that can stand this environmental degradation. 
Further evaluations were carried out from Kawanda on B. bassiana conidial yield from 
selected substrates, efficacy of different formulations against the banana weevil, and 
efficacy and persistence of the fungus under organic and inorganic soil amendments. The 
substrates evaluated were cracked maize, maize bran, "machicha", bagasse, cotton husks, 
maize bran + bagasse, maize bran + cotton husks, and bagasse + spent yeast and they 
were formulated with clay or loam soils. Evaluation of the substrates was based on conidia 
-65-
counts using a haemocytometer. Cracked maize grains were the best substrate (3.2 x 109 
conidia per gram), even though it was not significantly different from that of maize bran 
(3.1 x 109 conidia per gram). Cracked maize and cracked maize formulated with clay soil 
were the most effective with over 80% weevil mortality in 30 days. 
So far studies investigating abiotic and biotic factors affecting the efficacy and persistence 
of B. bassiana have focused on understanding how organic and inorganic amendments 
applied in the banana fields can influence the physical and chemical properties of the soil, 
and the possible effects of these changes on the performance of the fungus. The soil amend-
ments evaluated were unamended soils, coffee husks, decomposed cow dung manure, and 
inorganic fertilizers. During the study, weevil mortality was assessed on five-day intervals, 
soil temperature, and moisture at 10-day intervals, while soil fauna and antagonists were 
assessed at 30-day intervals. After 30 days, highest mortality (67%) was registered in the 
unamended soils, while in soils amended with coffee husks, decomposed cow dung manure, 
and artificial fertilizers caused respective mortalities of 30%, 40%, and 43%. The reduction 
in weevil mortality was attributed to the degradation of the B. bassiana by the microbial 
decomposers and colonizers that could have lead to first degradation of the maize-based 
B. bassiana formulation. Overall, the soil amendments significantly reduced the efficacy 
and persistence of the entomopathogenic fungus over time. We intend to build on these 
findings. Follow-up studies to address persistence and infectivity of B. bassiana under 
different banana agroecological zones are ongoing. 
4.23 The use of semiochemicals in delivery of B. bassiana 
by C. Gold and C. Nankinga, in collaboration with W. Tinzaara*, M. Dicke, and A. 
van Huis 
The use of semiochemicals in aggregating banana weevils was investigated by studying 
the insect's attractivity to traps made from various banana tissues that were pounded 
and placed in pots. Fresh pounded tissues of the susceptible highland banana cultivar 
"Atwalira" (genome AAA-EA) and the resistant banana cultivar "Keying" (genome ABB) 
were immediately placed at 0, 5, 10 and 15 cm below the soil surface in 6-liter buckets. 
Thus, we had eight kairomone treatments, i.e., two cultivars and four soil depths. The 
buckets were grouped into four subsets of three buckets per treatment. Ten adult field-
collected banana weevils were released per bucket in the subsets at 0, 1, 2, and 3 weeks, 
respectively, after placement of the banana tissues. There was a progressive decline in the 
percentage of weevils attracted to processed banana tissues with an increase in the depth 
at which the tissues were placed. Correlation analysis revealed a strong negative relation-
ship (r = -0.86) between the percentage of banana weevils recovered from pounded banana 
tissues and the depth at which the tissues were placed within the buckets. The declining 
trend was general for the fresh and one-, two-, or three-week-old processed banana tissues. 
Pounded banana tissues remained attractive to banana weevils up to three weeks, with 
one-week-old tissues attracting the highest number of weevils, though not significantly 
(P >0 .05) different from the fresh and two-week-old tissues_ However, a significant (P < 
0.05) difference was observed between the number of weevils attracted to one-week and 
three-week-old banana tissues. Correlation analysis revealed a weak negative relationship 
(r = -0.14) between the percentage of weevils in pounded tissues and age of the tissues. 
Analysis of weevil attraction to banana tissues grouped according to cuitivars showed no 
significant difference between the two cultivars tested. 
-66-
'l\vo field experiments were conducted to further test processed banana tissues as weevil 
trapping materials. The first experiment was conducted in May 2003 in which the trap· 
ping processed banana tissues were buried at 5 cm under the soil while the second one was 
conducted in December 2003 with the processed tissues placed at the soil surface. Tested 
banana tissues included fresh pounded corm, chopped corm, pounded pseudostem, and 
chopped pseudostem from the susceptible cultivar Mpologoma (AAA-EA) and the resistant 
cultivar Keying (ABB). Thus, a total of eight processed banana tissues were evaluated. 
Banana tissues placed at the soil surface caught more weevils (range 1.97-2.73 weevil/trap) 
than tissues buried at 5 cm below the soil surface (range 0.08-0.54 weevils/trap). In addi-
tion banana tissue traps at the soil surface were more attractive than split pseudostem 
traps (i.e., 1.97-2.73 vs 1.5--1.8 weevil/trap) as compared with buried tissues that were 
less attractive than split pseudostem traps (0.08-0.54 vs 1.0--1.7 weevil/trap). However, 
correlation analysis revealed a positive, though weak, relationship between the number of 
weevils captured by processed tissues both at 5 cm depth (r = 0.1) and at the soil surface 
(r = 0.29) and split pseudostem trap captures on the same banana mats. Further analysis 
revealed no significant (P> 0.05) difference in weevil catches among the various processed 
tissues within each of the two experiments. Similarly, no significant (p> 0.05) differences 
between the two cultivars tested, i.e., Keying (ABB) and Mpologoma (AAA·EA) and also 
between corm and pseudostem tissues. 
The next set of experiments concerned the use of semiochemicals in the delivery of B. 
bassiana. A laboratory study was conducted using an olfactometer in the darkroom at 
Sendusu to determine the quantity of B. bassiana conidia picked by a weevil from semio· 
chemical·baited traps treated with the fungus. At the same time, we determined the ability 
of weevils infected by B. bw;siana in semiochemical traps to transmit the fungus to other 
weevils. The following comparisons were made: (i) Pheromone + Beauveria vs Beauveria 
alone, (ii) pounded banana pseudostem + Beauveria vs Beauveria alone (iii) split pseudo 
ostem + Beauveria vs Beauveria alone. In all the three comparisons, pheromones attracted 
the highest number of weevils, followed by the split pseudo stem, while the fresh-pounded 
pseudostem (unprocessed kairomones) showed least attraction to the weevils. In the 
olfactometer, the weevils spent more time at pheromone sources, followed by the pounded 
pseudostem and finally under the pseudostem. Weevils picked up the most spores from the 
split pseudostem (perhaps because exudates wetted the formulation and facilitated adhe-
sion) although differences were not significant. This experiment suggested that pheromone 
and kairomone-based traps could facilitate the delivery of B. bassiana to banana weevils. 
The weevil behavior might also influence the amount of spores picked the by the insect. 
Some weevils walked over the surface of the fungus such that only their legs and abdomen 
were exposed to the inoculum. Other weevils burrowed into the fungus thus exposing their 
whole body to the spores. 
A field experiment was conducted to evaluate the ability of B. bw;siana treated semio-
chemical traps to attract banana weevils from various distances, and assess weevil infec-
tion from the traps. The following treatment traps were evaluated: (i) pheromone trap, (ii) 
covered pounded corm (covered with split pseudostem), (iii) exposed pounded corm, (iv) 
split pseudostem, and (v) control (empty pitfall trap). Ten banana weevils (five males: 5 five 
females) were released on mats at 0, 3, 6, and 9 m with respect from the trap mat. Weevils 
collected from the traps were taken to the laboratory and maintained in moist chambers 
to monitor for B. bassiana growth on dead weevils. Generally, percentage weevil recapture 
was very low ranging between 3 and 10% for the various treatment traps. Pounded corm 
-67-
tissues covered with split pseudostem apparently recaptured the highest number (10%) 
of the released weevils while the control had the lowest recapture rate of 0.5%. The high 
recapture value exhibited by a combination of pounded corm and split pseudostem could 
have been as a result of "double" quantity of the banana tissues (i.e., pounded corm tis-
sues + the split piece of pseudostem). The same value could have perhaps been obtained 
by doubling the number of split pseudostems (use two pieces) or doubling the quantity (by 
volume) of pounded corm materials. Lower weevil recaptures of the pheromone trap could 
have resulted from setting up the trap without water. No single trap recaptured weevils 
from all the weevil release positions. For all the traps, there was a higher recapture rate 
from the trap mat compared with other release positions. The pounded corm tissues covered 
with split pseudostem had the highest overall percentage recapture value and it recaptured 
weevils from all distances. Exposed corm tissues never attracted weevils beyond 3 m from 
the trap. A total of 227 weevils (including the recaptured and the natural population) were 
collected from all the traps over a one-month trapping period, and 56% of these weevils 
got infected with B. bassiana under moist chambers in the laboratory. This study further 
demonstrated that B. bassiana applied around semiochemical traps can successfully infect 
and kill weevils that are aggregated by the traps. 
Another field experiment was conducted to further evaluate banana weevil attractivity 
to kairomone traps made from banana tissues and treated with B. bassiana, and also to 
assess the dissemination of B. bassiana by weevils infected from the traps to healthy wee-
vils within the habitat. Pretreatment pseudostem trapping (two traps/mat) was conducted 
in all plots, targeting the selected trap mats and a sample of five mats around the each of 
the selected mats. Samples of the trapped weevils per plot were taken to the laboratory to 
monitor for weevil mortality and growth of B. bassiana due to natural infection from the 
field. Treatments included (i) disk·on-stump trap (with one disk) + B. bassiana, (ii) disk-
on·stump (with two disks) + B. bassiana, (iii) pounded pseudostem tissues + B. bassiana, 
(iv) split pseudostem trap + B. bassiana, and (v) B. bassiana alone. The number of weevils 
collected per trap was recorded and the weevils were taken to the laboratory for monitor-
ing mortality and B. bassiana growth. In addition, 200 g samples of B. bassiana were col-
lected from around each treatment trap and taken to the laboratory in dry Petri dishes. 
Ten adult weevils were released into the B. bassiana containing dishes and left in contact 
for 24 hours. The weevils were transferred to Petri dishes lined with moist tissue paper 
to monitor for mortality and mycosis. Posttreatment saturation trapping was conducted 
at two-week intervals for two months on the trap mat and four mats randomly selected 
around the trap mat. A record was taken of the total number of weevils trapped per mat, 
and weevils collected from the trap mat and from the neighboring mats were taken to the 
laboratory to monitor for mycosis. Detailed data analysis for this study has not yet been 
done. However, exploratory analysis indicates that the disk-an-stump trap (with two disks) 
attracted the highest number of weevils while pounded pseudostem tissues captured the 
lowest. Weevils captured at the various treatment trap mats exhibited higher rates of 
mycosis (range 9-22%) compared with weevils trapped from mats at 3 m (range 6-13%). 
Surprisingly, for the control trap, weevils trapped at the trap mat and those from 3 m 
showed as high as 16% and 13% infection, respectively, which were generally higher than 
the kairomone-based traps. 
-68-
4.24 Farmer perceptions and adoption of technologies for clean planting mate-
rial 
by C. Gold, S. Okech, and C. Nankinga, in collaboration with C. Kajumba* and 
W.K Tushemereirwe 
An MSc study on farmer adoption of clean planting material technologies was completed 
in 2003. Banana weevils and nematodes (especially the aggressive Radopholus similis) 
are often transferred to new fields through infested planting material. To address this 
constraint, clean planting material technologies, including corm paring and hot-water 
treatments were developed and disseminated to farmers in the districts of Ntungamo, 
Mpigi, and Luweero in Uganda, Oyuges and Kakamega in Western Kenya, and Bukoba 
in Tanzania. Approximately 1500 farmers in the five districts were trained using various 
approaches on the use of the clean planting material technology. Farmers then established 
banana fields using clean planting material. 
A survey was later done in these districts with 257 randomly selected farmers (trained and 
not trained by researchers) at training sites to gain insight into the factors that influenced 
adoption of these technologies. An interview guide was administered to the farmers. The 
results obtained revealed that farmers perceived corm paring as a simple technology that 
did not require a lot oflabor and was cost effective. On the other hand, hot-water treatment 
did not require a lot of labor, was not a simple technology, and required more additional 
costs in purchasing the thermometer, tank, and fuel. 
The farmers observed that the clean planting material led to improved sucker production, 
reduced pests, and better yields. Negative observations by the farmers included suscep-
tibility of the germinating suckers to drought and to livestock. Training and involvement 
of the farmers in the use of the technology helped improve their understanding of the 
banana pest status and also helped meet their expectations. Adoption of the technology 
was influenced by the farmers' characteristics including sex of the head of the household, 
farm size, land-tenure system, farmers' perception of the technology and its output, ability 
of the technology to met farmers' expectations, their understanding of the pests, and the 
approaches used in disseminating the technology. 
4.25 Effects of crop sanitation on banana weevil populations and damage 
by C. Gold, S. Okech, and P. Ragama, in collaboration with M. Masanza* and A 
van Huis 
Research on the effects of crop sanitation on banana weevil populations and damage were 
conducted through a PhD study that was completed in 2003_ Laboratory, on-station, and 
on-farm trials were conducted to determine relative banana weevil ovipositional prefer-
ences and egg and larval survivorship on crop residues vis-a-vis standing plants and the 
effects of crop sanitation on banana weevil populations and damage. Crop residues have 
been reported as an adult refuge and breeding ground for the banana weevil. It has been 
hypothesized that removal of these residues will reduce banana weevil numbers and damage 
to growing plants. An alternative hypothesis is that the residues draw ovipositing adults 
away from growing plants, thereby reducing damage. Little data has been available to 
support either of these hypotheses. 
Influence of different aged banana crop residues on attraction and host acceptance of the 
banana weevil. Two choice experiments were conducted in the laboratory at Kawanda to 
-69-
determine attraction and acceptance of crop residues of different ages to the banana weevil. 
In the first experiment, studies focused on different types and ages of residues of one sus-
ceptible highland banana clone "Nabusa" (genome groupAAA-EA). Corms attracted 65% of 
the test weevils, pseudostems 30%, and 5% were nonrespondents. Oviposition levels and the 
number of eggs per female were higher on young than old corms. In the second experiment, 
the same parameters were measured on banana residues of selected clones based on their 
levels of resistance and tolerance to banana weevil damage. Corms were more attractive 
to adults than pseudo stems and flower stalks except for fresh residues of resistant clones. 
Pseudostems were more attractive than flower stalks with a few exceptions. The number 
of eggs per female did not differ across clones, but varied with residue age. The number 
of eggs per female was highest on flower stalks, followed by corms and pseudostems. Old 
flower stalks were more acceptable for oviposition than any other residue part. 
Influence of plant and residue age on egg eclosion and larval survival of the banana weevil. 
We conducted laboratory trials to investigate banana weevil eclosion success and larval 
survivorship on banana residues of different ages at Kawanda in Uganda. When we inserted 
24-hour-old eggs into corm pieces of susceptible banana cultivar "Kisansa" (genome group 
AAA-EA) offour different ages, eclosion rates were 66% in fresh, 67% in moderately old, 
64% in old, and 58% in very old residues. To assess immature survival, 24-hour-old larvae 
were put on banana corms of suckers and crop residues of the same cultivar Kisansa in 
single rearing chambers. The number of surviving individuals was recorded at three-day 
intervals until adults emerged. The number of surviving individuals 48 days after eclosion 
was 12% on sword suckers, 10% on maiden suckers, and 7% on flowered plants, after 51 
days 12% on fresh, 8% on old, and 5% on very old corms. Larval duration and mean date 
of adult emergence increased with plant and crop residue age. Crop residue age did not 
affect adult weight, but the females were heavier than males. These results imply that 
fresh residues offer better nutrition for banana weevil than old residues. However, since 
all residues offered a successful breeding place for the weevil, they should all be destroyed 
to prevent population build-up of the banana weevil. 
Distribution, timing of attack and oviposition of the banana weevil on banana crop resi-
dues. On-station and on-farm trials were conducted in Uganda to investigate the effect of 
crop residue management on attack, oviposition, and distribution of the banana weevil on 
crop residues and growing plants. Oviposition and distribution were assessed on standing 
and prostrate residues of different ages by destructive sampling. Similar data were col-
lected from fields maintained at three sanitation levels. In the first experiment, oviposition 
occurred on residues as old as 120 days, but mainly between 0 and 30 days postharvest. 
Weevil infestation varied among banana clones. In the second experiment, oviposition 
levels on standing residues were not significantly affected by age. Oviposition levels on 
prostrate four-week-old residues were two times higher than those on two-week-old resi-
dues, while the number of larvae on eight-week-old residues was three times higher than 
on two-week-old residues. The number of pupae did not differ at all ages of prostrate and 
standing residues. The number of adults in standing and prostrate residues on 16-week-old 
residues was two times higher than that on two-week-old residues. In the third experiment, 
farmers' fields maintained at high sanitation level had 50% lower eggs per residue than 
those kept at low sanitation level. The number ofimmatures per residue was 50% higher 
on banana corms than on pseudostems. Larvae were three times more abundant at low 
than at high sanitation_ The number of pupae per residue at low sanitation was six times 
higher than at high sanitation leveL Residues in fields at high sanitation hosted 50% less 
-70-
adults per residue than in fields at low sanitation. The results suggest that removal and 
splitting of corms after harvest is effective and practical in destroying immature growth 
stages of the pest. 
Effects of covering highland banana stumps with soil on banana weevil oviposition. The 
effect of covering postharvest banana stumps with soil on banana weevil oviposition levels 
was investigated at Sendusu, Kawanda, and in Ntungamo district of southwestern Uganda, 
In the first experiment we assessed oviposition levels in a banana system comprising grow· 
ing plants and residues. Oviposition increased from sword suckers, reaching a peak one 
to seven days after harvest and decreased thereafter. In the second experiment conducted 
on farmers' fields, corms received 70% of the eggs and pseudostems 30%. The area 5--10 
cm below the collar received 27% of the eggs, the area 0-5 cm above the collar 30%, and 
the area 5--10 cm above the collar 0.3%. The remaining eggs (43%) were laid 0-5 cm below 
the collar. The effect of stump height and covering the stumps was evaluated in both the 
wet and dry seasons at Kawanda and Ntungamo. Cutting stumps to ground level alone 
had no effect on oviposition. Covering postharvest banana stumps reduced banana weevil 
oviposition in the wet but not in the dry season. 
Effect of crop sanitation on banana weevil Cosmopolites sordidus (Germar) (Coleoptera: 
Curculionidae) populations and crop damage in farmers' fields. An on·farm study of the 
effect of crop sanitation on the banana weevil populations and corm damage was conducted 
through farmer participatory trials in Ntungamo district, Uganda. Farmers practiced sani· 
tation levels that were broadly defined as low, moderate, and high, with most farmers at the 
lower end. During the study, some farmers were asked to maintain their original level of 
management, while others were asked to increase sanitation levels. However, farmers did 
not strictly adhere to prescribed treatments so that an incomplete block design was used in 
analysis of data. Increase in sanitation level from low to high significantly reduced adult 
banana weevil populations, lowered corm damage, increased plant maturation rates, and 
increased yields. The data suggest that improved sanitation management can contribute 
to control and improved banana productivity. 
Effect of crop sanitation on banana weevil and natural enemy populations in an on·sta· 
tion trial. We evaluated the effect of crop residue removal on weevil population, weevil 
damage (over four crop cycles), and nematode and arthropod natural enemy incidence in 
isolated young banana stands at Kawanda. A closed banana weevil population assumes 
no emigration or immigration between plots. We infested isolated banana plots with 5--10 
weevils and a complex of 3000 nematodes per plant. As harvesting ofthe plant crop started, 
we subjected the plots to low, moderate, and high crop sanitation levels. High sanitation 
levels reduced trap catch by up to 40%, and weevil population by up to 43%, and seemed 
to expose standing plants to increased weevil attack causing up to 34% reduction in yield. 
Crop sanitation seemed to reduce nematode populations but not damage. Plant growth 
was not affected until the fourth crop cycle when girth reduced by 10% and height by 6% 
in plots under high sanitation compared to plants in low sanitation fields. Plant girth and 
height increased with crop cycle. Complete removal of crop residues resulted in a three· fold 
reduction in arthropod natural enemies compared to leaving the residues intact, but did 
not affect total oviposition on growing plants. In young banana stands with closed weevil 
populations, removal of crop residues exposes growing plants to increased weevil attack. 
-71-
4.26 On farm validation of banana IPM metbods 
by C.S. Gold, C.M. Nankinga, P.E. Ragama, S.H. Okech, in collaboration with W. 
Tinzaara, A. Bareyke, W.K. Tushemereirwe, M. Byabachwezi, and S. Inzaule 
On·farm testing of banana IPM practices, including the use of clean planting material, 
application of neem and other biorationals, the use of crop sanitation, and testing of new 
cultivars being studied in ongoing on· farm trials in Masaka district, Uganda, Bukoba dis· 
trict, Tanzania, and Kakamega district, Kenya was completed in late 2003. Data analysis 
and write·up will be completed in 2004. 
4.27 Variability in reproductive fitness and virulence of four Radopholus similis 
populations from Uganda 
by C. Dochez, J. Dusabe, M. Pillay, and J. Whyte, in collaboration with D. De Waele 
(KULeuven, Belgium) 
Different populations of Radopholus similis from Uganda were compared for reproductive 
fitness and virulence. Different pathotypes might exist within a nematode species, and 
the resistance might only be effective against some pathotypes, but not all. If the target 
nematode species has a high level of genetic variability, the durability of the resistance 
might be affected. Therefore, populations of different locations within Uganda (Namulonge, 
Mbarara, Ikulwe, and Mukono) were collected and cultured monoxenically on carrot discs. 
Reproductive fitness of the four populations was compared on carrot discs in function of 
time and inoculum level. Pathogenicity experiments on host plants were carried out in pot 
trials. The population from Mbarara showed a higher reproduction ratio compared to the 
other populations, both on carrot discs and plants. Also the percentage root necrosis was 
highest on plants inoculated with the population from Mbarara. This population man· 
aged to break the resistance of Pisang J ari Buaya, worldwide known to be resistant to R. 
similis. The diploid hybrid TMB2x 9128·3 and Yangambi km5 showed resistance against 
the four R. similis populations. These results indicate that differences in pathogenicity 
among different R. similis populations exist and should be taken into consideration in the 
breeding program. 
4.28 Importance of yam nematodes in East Mrica established and genetic vari-
ability of Pratylenchus sudanensis compared with morphological variabil-
ity 
by D. Coyne, in collaboration with J. Mudiope, Makerere University, Uganda; and 
M. Moens, L. Waeyenberge, University of Gent, Belgium 
Pratylenchus sudanensis is the dominant nematode species associated with yam in Uganda, 
while Meloidogyne spp. commonly occur on yam across the country. P. sudanensis is strongly 
associated with cracking oftuber surfaces and Meloidogyne spp. with galled tubers. In pot 
studies, P. sudanensis was shown to cause cracking of tubers, increased root death, and 
reduced tuber production. Meloidogyne spp. caused galling of tubers and reduced tuber 
weight, which was more prominent on Dioscorea cayanensis (P ::: 0.05) than D. alata or 
D. rotundata. During storage, it was observed that galled D. rotundata tubers desiccated 
more quickly (P ::: 0.05) with greater weight loss than clean tubers. Molecular analysis of 
different isolates of P. sudanensis show that considerable variation exists between some 
isolates, some of which closely resemble sequences of Pratylenchus glutierrezi, P. loosi, P. 
coffeae, and P. pseudocoffeae, though none matched exactly. No reference sequence exists 
-72-
for P. sudanensis yet and therefore it was not possible to compare against a reference. Some 
isolates however had very little variation between them, while others had wide variation 
and are most probably different species. Assessment of the reproductive fitness of three 
isolates of P. sudanensis, isolated from yam, found that one isolate (from Jinja) had a lower 
(p:",: 0.05) reproductive fitness (and therefore most likely pathogenicity) than the other two 
isolates (from Rakai and Masaka). 
4.29 Comparative efficiencies of mulching quantity of Tithonia for nematode 
management and Musa crop improvement assessed 
by D. Coyne, in collaboration A Tenkuano, lITA 
Use of Tithonia diversifolia applied as fresh mulch, was assessed for its affect on plantain 
crop growth and production and on nematode activity, in a microplot experiment and field 
experiment conducted in Nigeria. In microplots, the experiment, which is currently being 
repeated, was terminated after nine months growth. Application of mulch at the equiva· 
lent of 2 tlha dry weight gave higher plantain leaf dry weight (p:",: 0.001) compared with 
no mulch and higher stem (P:"': 0.01) and leaf (p:",: 0.001) dry weight at 4 tlha dry weight 
application. No differences in root weight were observed between the treatments (P :"': 
0.05) at nine months. Nematode densities of Helicotylenchus multicinctus and Radopholus 
similis appear to be suppressed at high application rates of mulch (> 8 tlha dry weight 
application) and resulting in lower root damage indices. In the field experiment mulch 
was applied annually at 0, 6 (in one application), 6 (in 2 x split application), and 6 tlha dry 
weight application (split in between 5 x monthly application). In all mulched plots, plant 
height was greater than without mulch (p:"': 0.01) but increased the number of days to 
flowering (P:"': 0.05). Yields were higher in the single application and 5 x split application 
(p :"': 0.05) than the remaining two treatments to date, although harvests are continuing 
for the mother crop and therefore not finalized. The effect of mulching with T. diversifolia 
on plantain nematode activity and multiplication is so far not conclusive, but appears to 
have a suppressive effect. Both microplot and field trials are continuing. 
5 Crop and natural resource management practices for sustain-
able, competitive, and commercially based production systems 
developed and disseminated 
5.1 Literature review on soil quality constraints in EA highland bananas 
by Piet van Asten 
Banana research in the East Mrican highlands has traditionally focused on pest· and 
disease·related issues and the introduction of improved exotic cultivars. However, soil 
quality constraints are generally considered to be one of the primary reasons for the low 
and declining banana yields in this region. IITA has assigned a VVOB associate scientist 
to conduct research on this topic. Before starting on·station and on· farm activities, there 
is a great need to get an overview of what has been done already in the field of soil quality 
constraints in banana cropping systems, both in the East Mrican Highlands and elsewhere 
in the world. The literature study was conducted in Uganda, sourcing key informants, 
personal libraries of banana scientists in the region, and the Internet (a.o. INIBAP's 
MUSALIT banana library). 
The review has shown that: (i) there is scanty evidence for soil fertility decline in banana 
systems, (ii) that banana soils have relative good soil fertility compared to annual cropped 
-73-
fields and grassland, (iii) that, nonetheless, nutrient deficiencies (mainly K, N, Mg) are fre-
quently detected in banana fields, (iv) that yields can increase substantially (> 50%) when 
fertilizers are applied, but (v) that fertilizers are only efficient when pest pressure is low, 
(vi) that there is great risk of accelerated soil fertility decline due to the increasing export 
of banana bunches to the farmers, in combination with the reduced availability of organic 
matter and fallowing to restore soil nutrient stocks, (vii) that there are clear indications 
that balanced plant nutrition can improve the plants' resistance against pest attacks and 
that micronutrients might play an important role in this_ The review will be published in 
the African Crop Science Journal as part of a series of papers that were presented at the 
IPM conference in Kampala, Uganda, in December 2003_ 
5.2 Cassava soil fertility program Uganda and Kenya 
by A. Fermont, J_ Wendt, Yona Baguma (NARO Uganda), and Hannington Obiero 
(KARl Kenya) 
The cassava mosaic disease (CMD) epidemic that started in Uganda in the early nineties 
has presently reached most of eastern and central Africa with devastating effects on cas-
sava production. IITA and its national partners are successfully developing and releasing 
CMD resistant cassava varieties to counteract its impact. The latest introductions have a 
high yield potential characterized by multiple resistance to the major biotic stresses, drought 
tolerance, earliness, and higher dry matter content. Yield levels up to 50 tJha in advanced 
yield trials have been obtained in Uganda and western Kenya under nonfertilized conditions. 
These yield levels have raised concerns about the impact of the new varieties on soil-fertility 
depletion. 
Soil-fertility depletion has been described as one of the most important constraints to food 
security in sub-Saharan Africa. Nutrients are commonly not replaced to the degree that 
they are removed in crop harvesting and other losses, resulting in highly negative nutrient 
balances. Soils in western Kenya and Uganda are predominantly Ferrasols, Acrisols, and 
Nitisols; old weathered soils that contain predominantly kaolinite and are virtually free 
of weathering minerals. The general trend of soil-fertility decline that has been noted by 
many authors, particularly in western Kenya, reinforces the concern on the impact of the 
introduction of high yielding cassava varieties on soil fertility. 
Recently, cassava has no longer seen as a pure food security crop. Instead it has been iden-
tified as an important commodity for both income generation and food security in Africa 
in the future. There is a careful trend to commercialization of cassava production, both 
for the local market (fresh roots and dried chips or flour) as well as for industrial usage 
(animal feed, flour, starch). In Kenya, small-scale pilot processing units are being put in 
place by IITA and KARl to stimulate local commercialization of cassava products, while 
in Uganda, Ugachick, a big animal-feed industry, has recently started using cassava roots 
in the production of animal feed. 
Cassava soil-fertility issues in Eastern Africa have been largely neglected as a research 
theme due to the image of cassava as a food security crop, which grows reasonably well on 
marginal soils. However, the release of high yielding cassava varieties in combination with 
the general trend of soil fertility decline in this part of Africa requires that more attention 
is given to these issues. Specifically there is a need to understand and quantify the impact 
the introduction of improved cassava varieties will have on soil fertility levels and the per-
formance of cassava-based cropping systems, and develop cropping practices for various 
-74-
types of cassava cropping systems (low, medium, high input levels) which maintain both 
yield and soil-fertility levels_ 
Activities undertaken_ In collaboration with researchers from NARO and KARl three trial 
protocols were developed to start answering the above questions: 
• Variety x fertilizer trial. Objectives: (a) Compare the impact on soil-fertility depletion 
of three local cassava varieties versus three improved varieties under both fertilized 
and nonfertilized conditions using three methods; and (b) screen these varieties for 
their response to fertilizer. 
• Intercropping trial: Objectives: (a) Compare the impact on soil-fertility depletion of 
two local versus two improved varieties in the three most common cropping systems 
(sole/cereallbeans) using three methods, and (b) evaluate the impact of the improved 
varieties on the performance of the intercropping system as a whole_ 
• Nutrient response curve trial: Objective: Determine the limiting macro and/or meso 
elements for cassava production. 
All trials have a CRBD and four repetitions. Plot size is 8 x 9 m. Harvest area is 20 plants. 
The trials were installed at one location in central Uganda (N amulonge) and at one location 
in western Kenya (Alupe) in April 2003. Soil-fertility samples using the mass sampling 
method were taking before planting in the harvest areas of all plots. Agronomic data (plant 
height, pest, and disease scoring) are taken at 3, 6, 9, and 12 MAP. Harvesting will be done 
in March 2004. 
Results available from 2003 are preliminary as the trials will only be harvested in March 
2004. In general aboveground biomass production is much better in Namulonge than in 
Alupe. Fertilizer use seems to promote aboveground biomass production in Namulonge, 
both in improved and in local varieties. In Alupe, none of the varieties seems to develop 
additional aboveground biomass due to fertilizer usage. Fertilizer use seems to have no 
negative impact on the occurrence of pests and diseases. Intercropping reduces plant height 
in similar percentages in improved and in local varieties. 
Trials will be continued for 2-3 years before final conclusions can be drawn. This work will 
be part ofthe PhD work of Ann eke Fermont with Wageningen University in the Netherlands 
(promoter: Ken Giller). On-farm work will start in 2004 in both Kenya and Uganda. 
5.3 Soil sampling methods for detecting changes in soil properties 
by John Wendt 
Sustaining soil fertility is one of the most pressing issues facing African farmers. Conti-
nent-wide, nutrient removal from crops considerably exceeds nutrient addition through 
inorganic and organic sources. As a result, soil fertility decline is widespread. 
Integrated natural resource management (INRM) strategies involve combinations of 
improved germ plasm, crop arrangement/rotation, inorganic and organic inputs, residue 
management, and weeding strategies. These elements, individually or in combination, 
work to achieve sustainability of the soil resource. UTA as a leader in sustainable INRM 
management must therefore have the necessary tools to evaluate whether its INRM strate· 
gies degrade, sustain, or enhance soil nutrient status. 
Measuring soil changes requires special soil sampling procedures based on soil mass rather 
than soil depth. Many researchers, including soil scientists, still use soil depth sampling 
-75-
procedures, which result in errors and consequent data misinterpretation and erroneous 
conclusions regarding soil change. Soil mass sampling procedures are not new, but the 
necessity of using soil mass to precisely calculate nutrient changes has until now been 
difficult. Previous procedures required soil bulk density determination, which is laborious 
and time-consuming. 
Soil mass sampling procedures were developed that do not require a bulk density deter-
mination. This greatly speeds up soil sampling. Working with a soil sampling equipment 
company, AMS Soil Samplers, modified soil sampling equipment was developed to rapidly 
take samples for nutrient determinations on a soil mass basis. The equipment modifications 
include a modified soil probe tip made of special hardened steel that lasts much longer, 
and a modified tip design that accelerates removal of the soil sample from the probe, which 
speeds up the soil sampling process. 
An equation was developed that relates soil nutrient content to soil mass, as follows: 
M 1AlU = aQvt sOli r 
where Mout, is the nutrient mass, M~i' is the soil mass, and a and n are constants that 
describe nutrient distribution, and are calculated from the soil mass and nutrient mass 
from two or more depths in a sample core. Using the soil sampling equipment developed 
in collaboration with AMS Soil Samplers, it is possible to calculate the nutrient mass to 
a high degree of accuracy for any chosen soil mass in below the first depth increment in a 
soil core using interpolation procedures. 
The calculation procedure is complicated and tedious. Therefore, an Excel spreadsheet was 
developed to perform all of the necessary calculations from raw data input. The inputs 
include the diameter of the soil sample probe, the laboratory analysis, the number of cores 
taken for the sample, the weight of the soil, and the soil laboratory analysis. Any reference 
soil mass can then be chosen by the researcher. The calculation procedure also corrects for 
large inputs of organic matter which may occur when mulches are applied. 
These new soil sampling and calculation techniques were taught at trainings in Yaounde, 
Ibadan, and Kano in 2003. Approximately 20 technicians from Yaounde, 40 from Ibadan, 
and 15 from Kana participated in these half-day trainings. Soil sampling equipment from 
AMS was purchased for all three of these locations. The spreadsheets were distributed to 
soil scientists to facilitate the calculations. 
A publication of this work will be completed in 2004. Already, the methods are being 
employed to evaluate changes in soil fertility in Cameroon, Kenya, Mozambique, Nigeria, 
and Uganda. These procedures should greatly enhance UTA's abilities to evaluate the 
effects of INRM strategies on soil fertility. 
5_4 Improved cassava distribution and fertilization trials 
by N. Mahungu 
Four cassava varieties have been selected for distribution and multiplication in Malawi. 
These are Mbundumali (for fresh market and dried chips), Maunjili (for silage produc-
tion), and Sauti and Yizaso (for dried chips and flour). Over 6 500 000 one-meter cuttings 
have been distributed to farmers across the project sites. This quantity can plant about 
2000 ha. In Tanzania, 23 000 stems were used to establish mUltiplication plots at Alavi 
and Kibaha. 
-76-
Table 17. Effect of fertilizer application on cassava storage root yield (tJha) at 10 months 
after planting, (Alavi estate, Kibaha, 2003). 
Clone 
TMS4 (2) 1425 
Kiroba 
Kalolo 
NDL90/034 
Kibaha 
Bola Kupata 
Mean 
S.E. + 
C.V. (%) 
No fertilizer 
1B.2 
14.B 
21.5 
15.4 
15.2 
14.4 
1B.2 
1.81 
43.4 
35 kg N, 15 35 kg N, 15 0 kg N, 0 kg 
kg P,O" 30 kg 
K.,0/ha 
22.9 
31.2 
35.2 
21.3 
22.5 
19.0 
25.3 
2.55 
kg P,O" 15 kg 
K.,O Iha 
22.1 
23.8 
32.6 
21.2 
12.0 
14.6 
21.0 
2.08 
P,O" 15 kg 
K.,0 Iha 
18.3 
15.9 
30.7 
15.4 
18.7 
14.7 
18.9 
5.11 
Mean 
19.9 
21.3 
29.9 
18.3 
17.1 
15.7 
Table 18. Effect of fertilizer application on cassava storage root dry matter content (%) 
at 10 months after planting (Alavi estate, Kibaha, 2003). 
35kgN,15 35 kgN, 15 Okg N,Okg 
Clone No fertilizer kg P,O" 30 kg kg P,O" 15 P,O" 15 kg Mean 
K.,O Iha kg K.,0 Iha K.,O Iha 
TMS4 (2) 1425 33.2 29.8 35.4 35.5 33.5 
Kiroba 36.B 35.1 41.4 36.6 37.4 
Kalolo 34.9 35.3 32.9 36.9 35.0 
NDL90/034 39.2 42.2 40.6 43.3 41.3 
Kibaha 37.3 38.8 36.3 39.3 37.9 
Bola Kupata 42.1 38.8 37.1 38.1 39.0 
Mean 37.2 36.7 37.3 38.3 
S.E. + 0.58 0.82 0.67 1.64 
C.v.~ 7.64 
Trials aimed at evaluating breeders' lines for responsiveness to fertilizers are also being 
conducted in Malawi and Tanzania. In Tanzania, six cassava varieties, TMS4 (2) 1425, 
Kiroba, Kalolo, NDL901034, Kibaha, and Bora Kupata, are being evaluated with fertilizer 
inputs at Alavi estate (METL) for their suitability for starch production after 10, 12, 14, 
and 16 months of growth. Three harvests have been made and root yield and dry matter 
content for the first harvest are presented in Tables 17 and 18, respectively. All agronomic 
parameters and starch content are being determined. 
6 Public and private sector partnerships that promote 
information and technology exchange with emphasis on 
scaling up strategies for increased trade developed and/or 
strengthened 
6.1 Strengthening regional trade intelligence and trade information ser-
vices 
by Shaun Ferris*, Martin Nahamya*, Luwandagga David*, Charles Lwanga*, Nick 
Maunder", Thomas Awuor, Andrew Mutengu", Harriet Nsubuga''', Mark Wood"', 
and Stephen Njukia' (FoodNetT", FEWSNET-, ADC/IDEA"', RATES') 
The Regional Agricultural Trade Intelligence Network (RATIN) was designed in early 2003 
and kicked off in mid-2003 with the launch of an interactive website: www.ratin.netwhich 
is perceived as a platform for exchange of information on trade for major food commodi-
ties in Kenya, Tanzania, and Uganda. The information exchanged included trade news, 
cross-border trade information, trade balance sheets, weather forecasts, food aid updates, 
millers' prices, and crop forecasts. Through this project, monthly trade meetings have been 
-77-
held in Kenya, Tanzania, and Uganda and this has created a forum where traders, market 
researchers, food aid agencies, and development NGO partners interact and discuss trade· 
related issues. A couple of regional monthly bulletins have been produced and disseminated 
in two languages: Swahili and English, since April 2003. Links with world space satellite 
have been initiated and information will be syndicated to several radio stations within the 
region for faster dissemination. Under this same activity, personnel have been stationed 
on border posts ofthe three countries to monitor volumes traded across the borders. 
6.2 Market Information Services 
by Shaun Ferris*, Geoffrey Okoboi*, and David Luwandagga* (FoodNet*) 
Provision of market information is crucial for improving market access and efficiency. This 
particular component of FoodN et is funded by ACDI· VOCA under a five· year arrangement, 
which started late last year. The governments, donors, and several development partners 
have acknowledged the importance of market information. FoodNet has continued to excel 
in collection, processing, and dissemination of market information in Uganda and this is 
used as a model for other countries to emulate. Market agents record daily wholesale and 
retail price data for major markets in Kampala and also for upcountry towns covering 28 
commodities in 19 districts of Uganda. This information is mainly disseminated using FM 
radio stations and is broadcast in eight local languages. Data is also sent to policy makers, 
food aid agencies, development partners, research institutions, and private sector members 
via email, internet SMS, newspapers, and through world space satellites. 
6.3 The NAADS local Marketing Information Service 
by Dennis Bisase* and Emily Arayo* (FoodNet*) 
The NAADS/FoodNet market information project was designed to provide a micro·market 
information service for producers at the subcounty level in six Districts of Uganda where 
the project is being implemented. These districts include Arua, Mukono, Kibaale, Kabale, 
Tororo, and Soroti. At present, 20·minute market information programs run on six com· 
munity radios twice daily in respective local languages. A "market-market" serial radio 
drama program has been developed in English to run on the radio stations that are under 
the project arrangement. 
Client· based trainings on how to utilize market information have been conducted with 
district market officers and government officials in 22 subcounties of the six target dis· 
tricts in Uganda. Over 600 farmers have been trained how to interpret and utilize market 
information. 
A market information dissemination method at village level has been developed and this 
involves pinning price bulletins at some subcounties' headquarters of the local markets 
covering both low·value crops and high·value priority commodities such as livestock, Mor· 
inga, and Vanilla. 
6.4 Market analysis 
by Shaun Ferris*, John Jagwe*, Robinah Nyapendi*, Kelly Wanda*, John Spi/s· 
buryA, James Mbwika A, Jackson Nkuba+, and Samson Makworo+ (FoodNet*, Private 
consultant A, NARS') 
A number of market studies have been conducted in the past year, the summary of which 
is shown in Table 19. Most of them have been submitted for review before they can be 
-78-
Table 19. Summary of market studies conducted by FoodNet in 2003. 
Study Partners Coverage Status 
Sesame eRS, FoodNet Uganda, Kenya, Ready for publica-
Tanzania tion 
Banana ARDI-Maruku Uganda, Kenya, Ready for publica-
INIBAP, KARl, Food-Net, Tanzania, tion 
CIAT-ATDT-ISAR Rwanda 
Maize Makerere University, Uni- Uganda, Kenya, Almost complete 
versity of Nairobi, Sokoine Tanzania, 
Beans 
univers~ EARO 
ECABR , CIAT, Food-
Ethio~a 
Ugan ,Kenya, Ready for publica-
Net Tanzania EARO tion 
Rice FoodNet, NARS Uganda, kenya, Almost complete 
Tanzania 
Soybean FoodNet, NARS Uganda, Kenya, Initial stages 
Tanzania 
Market opportunities CRS, SSARP-USAID, Southern Sudan Almost complete 
studies on Agric, for- FoodNet 
estry, wildlife sectors 
Maize, beans, vegeta- EARO, FoodNet Ethiopia Initial stages 
bles, livestock market 
study 
Dried fruit ¥rocessin~ 
Marketing 0 Vitamin 
NARO, FIT(U) 
KARl 
Uganda 
Kenya 
Almost complete 
Complete and ready 
rich r.;tatoes in Kenya 
Mar et opportunity FoadNet, eIAT, Urban Uganda 
for publication 
Complete and ready 
identification for urban Harvest far publication 
& peri urban farmers 
in Kampala 
market opportunity FaodNet, NAADS-Uganda Uganda Study ongoing 
identification farmers Gov't 
six districts of Uganda 
published with IITA The methodologies used in these studies were documented and shared 
with socioeconomists in research institutions equipping them with skills to conduct similar 
studies in the future. The draft reports of the above mentioned studies are available on the 
FoodNet website: www.FoodNet.cgiar.organd the publications will be availed to interested 
parties on request. 
6.5 Agroenterprise development 
by Robinah Nyapendi*, Rupert Best, and Shaun Ferris*(FoodNet*, ClA~) 
A ''learning alliance" has been developed between Catholic Relief Services (CRS) and 
FoodNet aimed at improving the capacity of CRS personnel and those of its partner insti-
tutions to support the development of sustainable links between their target farmers and 
markets_ The 'territorial' approach to agroenterprise development is what is being used in 
the alliance. Four major workshops have been conducted for the alliance in phases and the 
participating countries (Burundi, Ethiopia Kenya, Madagascar, Malawi, Rwanda, Sudan, 
Tanzania, and Uganda) facilitated in the course of their work. A report of the process 
undertaken is yet to be produced. 
Within this learning alliance, two manuals have been developed and utilized in the market 
opportunity identification process and improvement of competitiveness of market chains in 
small-scale producers. These manuals were developed in collaboration with CIAT's Carlos 
Ostertag and Rupert Best with funding from the International Development Research 
Center (IDRC), the Swedish Development Corporation (SDC), and the Inter-America 
Development Bank (IDB). 
-79-
6.6 Training 
by Shaun Ferris*, John Jagwe*, Robinah Nyapendi*, Rupert &s~, Andrew Temu', 
Hugo De Groote', Carlos Ostertag", and Charles Lwanga* (FoodNet* CIAT', 
CYMMIT',IFPRI') 
Since capacity building is one of the priority areas of FoodNet, a number of trainings have 
been conducted by FoodNet with an aim of building capacity of partners and stakeholders 
in dealing with market challenges in their domains of operations. Table 20 is a summary 
of trainings conducted by FoodNet in 2003- mid·2004. High profile agricultural economists 
and agroenterprise experts from IFPRI, CIAT, and CYMMIT conducted the agroenter· 
prise and marketing course. The course material included Marketing basics, Agricultural 
Reforms, and Designing and executing a market opportunity identification survey together 
with data analysis and utilization of statistical packages such as SPSS. 
The website development course was conducted by IITA·FoodNet webmaster to a group 
of private sector partners working with a market· oriented NGO (Faidha Mali) operating 
in the northeastern part of Tanzania. The participants acquired skills on how to develop 
and maintain simple websites using user friendly website development software such as 
Dreamweaver. 
Table 20. Sununary of trainings conducted by FoodNet, 2003 to mid-2004. 
Course title Participants Partners Location Dates 
Market analysis 31 Socioeconomists IFPRI, CIAT, Nazareth, March 
training course from Ethiopia Agric. EARO, CYMMIT, Ethiopia 2004 
Research Org. FoodNet 
(EARO) 
Market analysis 42 CRS staff and Catholic Relief Ser· Dire Dawa, March 
training course partners in agric vices (CRS) Ethiopia 2004 
extension activities 
Website development 20 Faidha Mali (pri· Private sector Arusha, Tan· April 
course vate sector partners) zania 2004 
6.7 Website development 
by Charles Lwanga* and Shaun Ferris* (FoodNet*) 
Under a bilateral arrangement made between FAO and FoodNet, a webmaster was hired 
two years ago with the aim of developing and maintaining websites for FoodNet and 
its partners since leTs were perceived as crucial media channels for both now and the 
future. FoodNet was to avail the same services and skills to partners. Under this arrange· 
ment, the FoodNet Web master has been able to design seven websites for partners both 
in research institutions and the private sector. Some of the websites developed under this 
arrangement include the potato research network website: www.asareca.org/prapace, the 
bean research network website: www.asareca.org/ecabren, the root crop research network: 
www.asareca.org/earrnet, the regional trade network website: www.ratin.net, and the 
IFPRI scrip project: http://www.FoodNet.cgiar.org/scrip. 
7 Capacities of NARS and other stakeholders to generate, evalu-
ate and disseminate appropriate knowledge and intervention 
technologies for target subsectors enhanced 
-80-
7.1 Backstopping of Ugandan NARO national banana research program 
by C. Nankinga, C. Gold, S. Okech, and P. Ragama, in collaboration with W. Tush· 
emereirwe 
During 2002, IITA helped coordinate on·farm banana research activities in 10 subcounties 
in Masaka and Luwero districts, Uganda on banana integrated pest management and germ· 
plasm evaluation. More than 26 000 improved and East Mrican highland banana clones 
were disseminated to over 1000 farmers. In collaboration with NARO, IITAconducted GIS 
mapping of banana o·farm sites in seven subcounties in central Uganda and selected sites 
for multilocation banana germplasm evalaution. 
IITA also worked with NARO in developing technology disseminination tools with NARO 
through the production of posters on (1) Banana germplasm improvement, (2) Conven· 
tional breeding of bananas, (3) A cell suspension system for highland bananas, (4) Banana 
micropropagation, (5) Postharvest processing of bananas to chips and flour, and (6) Clean 
planting material: a key to a successful banana plantation. 
In collaboration with NARO· National Banana Research Program, IITA produced brochures 
on Postharvest processing of bananas to chips and flour and clean planting material: a key 
to a successful banana plantation. These brochures were distributed to over 2000 partici-
pants of agricultural shows conducted in Masaka district, Jinja district and the Buganda 
Kingdom. 
Jointly with NARO-National Banana Research Program, IITA produced a summary of 
Highlights of banana research activities and Highlights of banana germ plasm improve-
ment in Uganda that have been distributed to over 500 visitors that visited the National 
Banana Program. 
IITA and NARO-National Banana Research Program conducted four stakehoders' meet-
ings (farmers, agrictural extension, NGOs, donors, community leaders, opnion leaders) at 
two banana benchmark sites, reaching over 1000 stakeholders. IITAjointIy organized with 
NARO-National Banana Research Program an open day at Bamunanika benchmark site 
that attracted more than 4000 show participants. This open day was intended to transfer 
banana management and utilization technologies being evaluated and disseminanated at 
the site. 
7.2 Student training 
by C. Gold, B. Niere, C. Nankinga, T. Dubois, S. Okech, and P. Ragama 
PhD project formulation and research was supported in host plant resistance to banana 
weevil through conventional and nonconventional methods (three students); biological 
control of banana weevil using ants; the effects of crop sanitation on banana weevil; the use 
of info chemicals in banana weevil control; and the use of endophytes for control of banana 
weevil and nematodes (two students). MSc research supported by IITAincluded the use of 
Beauveria bassiana against banana weevil (two students); the use of endophytes against 
banana weevil and nematodes (two students); technology transfer of hot-water treatment 
for control of weevils and nematodes; and data base management. Research partners 
include Wageningen University, the University of Massachusetts, Cornell University, the 
University of Pretoria, and Makerere University. 
- 81-
7.3 Postgraduate training 
by J.P. Legg, in collaboration with J. Brown, J. Colvin, D. Gerling, C. Rey, T. Avel· 
ing, G. Thompson, C. Fauquet, S. Kyaanywa, M. Latigo, D. Osiru, and E. Adipala 
A male Ugandan PhD student, registered with the University of Greenwich, UK, completed 
a PhD study of Bemisia tabaci, as vector of cassava mosaic geminiviruses. His research 
aimed to identify the factors leading to the superabundance of B. tabaci populations in CMD 
pandemic affected areas of East Africa. A second male Ugandan PhD student, registered 
at the University of Witwatersrand, South Africa, conducted molecular characterization 
studies of cassava mosaic geminiviruses and their associated whitefly vectors at the Uni· 
versity of Arizona, USA, and began to examine the bionomics, mating compatibility, and 
genetics of different cassava B. tabaci genotypes at the Natural Resources Institute, UK 
Two other Ugandans completed a Rockefeller Forum Makerere University based master's 
program project on sustainable IPM for CMD and whitefly management. One of the two, 
a male student, examined the potential for using aphelinid parasitoids to manage whitefly 
populations, whilst the second, a female, investigated interactions between cassava mosaic 
geminiviruses, their effects on yield, and putative cross protection. A further Ugandan, 
doing an MSc at Makerere, was in the second year of a program examining the parasitoid 
diversity in B. tabaci on sweetpotato and assessing the potential for augmenting parasitism 
through using sweetpotato and cassava intercrops. A male Tanzanian finished the second 
year of a PhD program on cassava mosaic virus variability and interactions in Tanzania. 
In 2002 he carried out molecular characterization studies at the Donald Danforth Plant 
Science Center, St. Louis, USA. A second Tanzanian began the second year at Makerere 
University, Uganda on the topic ofphytosanitation versus resistance in the management 
of CMD and sweetpotato virus disease. A second female student began the second year of a 
masters program with Egerton University, Kenya, developing elements of a CMD manage· 
ment program for Siaya district, western Kenya. 
7.4 TARGET project 
by M. Pillay and Guy Blomme (IPGRI-INIBAP) 
Project partners were identified in Tanzania and Mozambique. Four main partners were 
identified in Tanzania: Adventist Development and Relief Agency (ADRA), the Agricul-
tural Research and Development Institute (ARDI), a marketing institute (FAlDA), and the 
Kagera Community Development project (KCDP). The main project partners in Mozam-
bique were: Instituto Nacional de Investigacao Agronomica (INIA), University Eduardo 
Mondlane, and the Institute of Tropical and subtropical Crops (ITS C) in South Africa. A 
training program was held in both countries. Weaning sheds were constructed and dem-
onstration plots were set up. The demonstration plots were used to train the extension 
workers and farmers in banana field establishment, banana agronomy, pest and disease 
management, and postharvest techniques. 
The planting material (FHIA 17, FHIA 21, FHIA 23, FHIA 25, SH3436-9, BITA-3 and 
Yangambi km 5) has been established in farmers' fields in both countries. 
-82-
7.5 Training in cassava production and processing technologies in Malawi 
by N. Mahungu 
More than 250 individuals (farmers and extension staff) in Malawi have been trained in 
cassava production and processing technologies in collaboration with partners such as 
NGOs, the private sector, and the Ministry of Agriculture, Irrigation and Food Security. 
In addition to farmers, SARRNET hosted four students from Bunda College of Agriculture, 
a constituent College of the University of Malawi and two students based in France on 
attachment (Table 21). These were trained in various activities of SARRNET. 
A book entitled ''Use of cassava in animal feed" has been published with a joint effort from 
CLAYUCA. 
Table 21. Students on attachment to SARRNET in 2003. 
Name of student 
Mr Nicolas Leley 
Mr Eric Nonry 
Betina Eliphas (Ms) 
Harvey Tchale 
Wilford Phiri 
Patrick Chiwaya 
Wongani Chisala 
Tione Kaonga 
College 
Ecole Superieure 
D'Agriculture D'Angers, 
France 
Ecole Superieure 
D'Agriculture D'Angers, 
France 
Bunda College of Agriculture, 
Malawi 
Bunda C()lIege of Agriculture, 
Malawi 
Bunda College of Agriculture, 
Malawi 
Bunda College of agriculture, 
Malawi 
Bunda College of agriculture, 
Malawi 
Bunda College of Agriculture, 
Malawi 
Field of specialization 
BSc in Agriculture 
BSc in Agriculture 
BSc in Agriculture (Crop Science) 
BSc in Agriculture (Crop Science) 
BSc in Agriculture (Crop Science) 
BSc in Agriculture (Crop Science) 
BSc in Agriculture (Agriculture 
economics) 
BSc in Agriculture (Agriculture 
economics) 
7.6 Statistical and Geographic Information Systems training 
by P. Ragama, C. Legg, S. Korie, S. Ofodile, H. Kisingo, and T. Alabi 
A course in SAS/GIS was conducted for 2\1, weeks between 15 September and 1 October 
2003, where 12 UTA employees and 13 collaborators from NARO attended. The course 
was funded jointly between Biometric/GIS unit of IITA·Ibadan and IITAIESARC as our 
colleagues from West Africa accepted to come and backstop us in conducting the course. 
Several topics in statistical models using SAS package were covered in a period of two weeks 
(Table 22). The GIS course was conducted in three days after concluding Statistics/SAS. 
Overall, participants rated the course above average and the general feeling was that the 
frequency of conducting such courses should be increased. 
The course was designed for persons involved with handling experimental data from the 
fieldllaboratory and surveys that involve ordinal and/or continuous data. Participants 
were encouraged to come with their own data so that during practical sessions they would 
be assisted in analyzing their data. SAS package was used for statistical analyses while 
ArcView was used for GIS. The venue was the National Banana Resource Center, Kawan 
da. 
The objectives of the course among others were to empower scientists and research 
associates/scholars in analyzing data collected efficiently in order to meet the mandate 
- 83-
of IITA-ESARC in the region. The course compared different alternatives for obtaining 
modern analyses of key statistical designs in use. An attempt was made to review spatial 
methods and the linkage of geophysical and biological attributes through GIS_ About 75% 
of the course was practical while 25% was theory. 
Table 22. Timetable of statistical computing and data analysis with SAS course, 15 
September to 1 October 2003. 
Date 
15--16109/2003 
17109/2003 
18109/2003 
19/09/2003 
20/09/2003 
22/09/2003 
23/09/2003 
24/09/2003 
25/08/2002 
2610912003 
27109/2003 
29/09-1110/2003 
Publications 
Monographs 
'!bpic 
SAS basics and data management 
Data entry through Access and linkage to excel 
and basic analysis 
Introductory statistics with SAS 
Desigo concepts and ANOYA derivatives 
Weekend practical session (morning only) 
Two-way ANOYA (CRBD, factorial, split-plot, 
etc.) 
Mixed models 
Longitudinal data analysis Irepeated measures 
desiplfanalysis (with mixed models) 
Ordmary least squares and dummy variable 
regression, trend/partial 
Multivariate analysis 
On-farm design/analytical strategies and surveys 
GIS course 
Resource person(s) 
HusseinlRa~ama 
OfodilelKone 
OfodilelRagama 
KorielRagama 
RagamalKorie/Ofodile 
KorielRagama 
Ragama/Korie 
KorielRagama 
KorielRagama 
KorielRagama 
Ragama/Korie 
Chris Legg & Alabi 
Collinson C.D., K. Wanda, A_ Muganga, and R-S.B. Ferris. 2003 Cassava market-
ing in Uganda: constraints and opportunities for growth and development_ 
ASARECAIlITA, Monograph 2 IITA, Ibadan, Nigeria. 
Summary. This report presents research into constraints and opportunities within tradi-
tional fresh and dried cassava marketing in Uganda. DFID's Crop Post Harvest Program, 
DFID's bilateral aid section in Kampala, and the East Mrican Research Network funded 
the research. 
Key findings 
• Dried cassava flour is an important food staple for particular groups of poor consumers 
within Kampala. Real cassava flour prices in Kampala are volatile and increasing in 
the long term, suggesting that food security among sections of the urban poor may be 
threatened. 
• Improving the flow of market information would increase dried cassava marketing effi-
ciency. This would lead to better spatial integration between urban markets (thereby 
reducing the absolute levels and volatility of consumer prices, and increasing market-
ing opportunities for farmers), and to a reduction of transaction costs throughout the 
marketing chain, particularly to the benefit of farmers. 
• Efficiency could also be improved by increasing liquidity within the dried cassava trad-
ing chain. Greater capital availability would allow wholesalers and retailers to expand 
the scale of their businesses and spread their overhead costs over a greater volume of 
trade. Food retailing in general has poor cost efficiency, and consequently consumer 
food prices are much higher than is necessary_ In the case of cassava flour, retailers add 
approximately 36% to the price of flour in between buying and selling, yet the greatest 
-84-
part of this margin is accounted for by overhead costs. Retailing is a simple service and 
should not cost consumers so much. 
• Product quality is not a problem during the dry season, when optimal drying conditions 
permit sufficient dried cassava production that is of adequate quality and affordable to 
urban consumers. However, during the wet season, good quality flour becomes expen-
sive, not only because oflow levels of in terseasona I storage but also because farmers are 
unable to dry cassava roots effectively. Appropriate drying technologies might improve 
the situation, but they would have to be both nonlabor and noncapital intensive to find 
favor within prevailing farming systems. 
• Current market facilities constrain wholesaling efficiency. Competition with alternative 
land uses limits business expansion, maintains high rents, and creates congestion and 
pollution. 
• There is no evidence that any group within the dried cassava trading chain makes 
unjustiflably large proflts. On the contrary, excessive competition at the retail level 
appears to squeeze profits to the extent that individual retail outlets operate at an 
inefficiently small scale. 
• Fresh cassava trading is driven by the high perishability of the roots and by the price 
premia that consumers are willing to pay for freshness. 
• Large marketing cost savings could be realized if current methods for reducing the 
perishability of cassava roots can be adapted to meet the requirements of Uganda's 
fresh cassava trade. 
• Recent cassava breeding research has neglected widespread consumer preferences for 
medium sized, sweet varieties. 
• Transport charges between farms and assembly points are more than ten times more 
expensive than interurban transport charges. Although this finding was made in the 
context of dried cassava trading, it probably applies to most domestically marketed 
farm produce, and may indicate the importance of public investment in reducing the 
costs of community level transport. 
• Lending to the food marketing sector is not only constrained by high transaction costs 
and poor loan security but also by conservative attitudes among potential borrowers. 
Recommendations 
Many of the constraints that apply to fresh and dried cassava trading are common to other 
types of food marketing in Uganda. Examining cassava marketing in isolation can there· 
fore lead to less than optimal research impact. The following are suggested as avenues of 
general food marketing research and technical assistance. 
• Formal lending to the food marketing sector: This research should concentrate on find-
ing innovative lending practices that will reduce costs and risks to both lenders and 
borrowers. Ideally, it should involve traders, commercial lenders, government policy 
personnel, and legal experts. 
• Capturing economies of scale within food wholesaling and retailing: Building on the 
work into improved lending, this research should be broad ranging and include inves-
tigation into capital, business skill, entrepreneurial, cultural, and infra structural 
- 85-
constraints. An understanding of the relevant constraints will allow policy makers to 
develop strategies to encourage larger scale wholesaling and retailing. 
• Improved market facility planning: If Kampala's urban authority is willing to cooper· 
ate, we recommend assistance aimed at improving the planning of Kampala's markets, 
with the ultimate aim of reducing marketing costs, improving sanitation, and reducing 
congestion in the city center. The work should engage urban planners and traders in 
a collaborative approach. 
The final two recommendations apply specifically to technical innovation within cassava 
marketing systems. 
• Drying technology: We do not recommend spending research funds on developing ways 
for drying cassava roots at the farm level. New drying techniques would only find 
relevance during the three months of the wet season. Furthermore, in order to find 
favor among farmers, drying innovations would have to satisfy the impossible combina-
tion of being nonlabor and noncapital intensive. We do, however, recommend further 
research on the financial viability of drying innovations at the village assembler level 
of the marketing chain, where capital constraints are lower. Village assemblers handle 
greater volumes of chips and may therefore be able to exploit economies of scale in the 
drying process. 
• Fresh cassava storage: We recommend research into the feasibility of adapting CIATI 
NRI fresh cassava storage technology for use by Ugandan traders. Costs and benefits 
to the traders should form the central theme of the research. 
Robbins, P. and R.S.B. Ferris. 2003. The impact of globalization on the 
agricultural sectors of East and Central African countries. ASARECAlIITA, 
Monograph. IITA, Ihadan, Nigeria. 124 pp. 
Executive summary. Globalization is the term used to describe the recent impact of 
innovations in communications and transport systems on trade and the growing interde· 
pendence of countries due to economic sophistication and burgeoning output. These inno-
vations have encouraged nations to reduce the high levels of protection between trading 
blocks of countries and to adopt policies to liberalize their economies in order to increase 
their volume of trade, including trade in agricultural products. 
It has been proved that, for many countries, increased economic liberalization and open-
ness leads to growth. It has also been recognized, however, that for some countries and for 
some communities within countries the transition from a protected, centrally controlled 
economy may bring with it serious, negative, short-, and medium-term consequences. 
Some Eastern and Central Mrican countries have recognized the importance of striving to 
increase their role in the international economy and have, over the last two decades, adopted 
appropriate economic measures--others have done so more recently. These measures have 
resulted in benefits to ECA countries including the stimulation of private sector trading 
networks needed in a modern economy. However, the risks associated with adopting a more 
exposed position in a highly competitive global agricultural market have presented these 
countries with some serious difficulties. A combination of the impact of structural adjust-
ment programs and partial reform of the rules governing international trade has reduced 
the prices of primary commodities exported by ECA countries and caused an increase in 
- 86-
imports of agricultural products from more competitive producers, some of which remain 
highly subsidized in their country of origin. 
The result of oversupply and weakening demand due to the current recession, has led to 
commodity prices falling to a 40-year low and analysts suggest that commodity prices are 
likely to remain at these low levels for the foreseeable future. This bleak outlook is reflected 
in the dramatically falling terms oftrade for many ECAcountries and suggests a profound 
downturn in their economic outlook and performance. The international community has 
recognized some of these difficulties and has made some effort to assist these countries to 
overcome them. Much remains to be done by these countries themselves, however, to take 
advantage of the opportunities offered by globalization and to ameliorate the negative 
impacts of the process. 
In the next round of WTO talks, the radical reform of the trading relationship between 
ACP countries and the EU and the establishment of closer regional economic cooperation 
will have further major implications for agriculture. 
In the opinion of the authors of this paper, ECA countries have not appreciated the scale 
and implications of these changes and that, without urgent action on their part, they may 
seriously weaken their economies in the years ahead. Measures need to be adopted by a 
very wide range of agencies in both the public and private sectors. These range from a major 
effort to increase the understanding of issues in multilateral trade negotiations, urgent 
efforts to devise strategies to reduce economic dependence on primary commodities, and 
major reforms of agricultural development and research strategies. 
This paper attempts to set these issues in an historical context, to highlight the main 
issues that need to be addressed, and to list important questions that need to be asked of 
policy·makers throughout the agricultural industry. The conclusions of this study are that 
decision·makers should give urgent consideration to the following suggestions: 
1. Strengthening negotiating capacity in trade talks. Mrican countries have been disap· 
pointed by the effects of decisions made in previous WTO and ACP-EU negotiations. 
ECA countries are poorly represented in these and other multilateral and bilateral 
talks and they lack capacity to analyze important and highly complex issues, to develop 
negotiating positions, and to respond quickly and effectively to their various negotiat. 
ing teams. Consideration should be given to establishing national and regional teams 
of experts with the necessary authority to analyze the interests of their stakeholder 
groups and to establish appropriate negotiating positions. Negotiating teams should 
be significantly strengthened in Brussels and Geneva especially. Resources should be 
made available by cutting diplomatic expenditure in other countries where necessary. 
The negotiators need to be directly linked to the policy analysis groups and to the line 
Ministries of Trade, Agriculture, and Finance, such that informed decisions can be made 
rapidly and effectively. Such reform will be particularly necessary in the forthcoming 
WTO trade round which will focus on greater inclusion of developing country interests 
and may take into account proposals associated with the "Development Box" and other 
nontrade issues proposed by developing countries. 
2. Managing the oversupply of primary product exports. Over the last two decades the 
adoption of internal and international market liberalization polices have led to a cata· 
strophic fall in the prices of many of the agricultural products exported by ECA coun· 
tries. The plunge in prices has been caused by systemic overproduction stimulated by 
- 87-
components of structural adjustment programs. Economists call this phenomenon the 
fallacy of composition, i.e., less income is earned as more commodities are produced. 
ECA countries are highly dependent on the production of cash·crop commodities for 
employment, economic growth, and export revenue. 
Countries that produce and export raw commodities such as coffee, sugar, tea, cotton 
etc. through small·scale production systems are unable to create new jobs or re·invest 
into alternative market sectors. Countries and individual farmers, who rely on cash· 
crop production for revenue, are obliged to continue to grow and sell these commodities, 
no matter how low prices fall. 
To address this issue, efforts should be made to find a common cause with other produc. 
ers of these commodities in Africa and in other continents to bring some order into these 
markets and to devise strategies that involve donors and support agencies such as the 
IMF and the World Bank to bring supply of these products in line with demand. 
3. Enforcing existing trade protection. ECA WTO members have agreed to limit the pro· 
tection given to domestic farming. Fixed import tariffs still apply in many categories, 
however. Greater efforts should be made to increase control of porous borders to discour· 
age unwanted imports and to collect excise revenue. The dumping of heavily subsidized 
agricultural commodities from developed countries should be actively opposed where 
such imports disrupt local farming economies. These efforts need to be pursued within 
the WTO mechanism and in bilateral exchanges. 
Efforts should also be made to analyze the impact of imports of food aid and food mon-
etization schemes on domestic and regional farming. Such imports should be controlled 
with the objective of meeting relief needs whilst avoiding the undermining of local and 
regional production. 
4. Stimulating production of added· value products. Most analysts believe that the prices of 
primary agricultural commodities will continue to fall in the foreseeable future. Unless 
the mix of industrial activity is changed, economic growth will not occur. 
The "Everything But Arms' initiative", the "Africa Growth Opportunity Act", and other 
similar market·access measures now offer LDC countries in ECA the opportunity to 
attract investment into the region to improve the quality and range of products and, 
more importantly, to produce added·value products made from locally produced raw 
materials. Every effort should be made to capitalize on these opportunities by promoting 
inward investment now that many tariff barriers to added·value products have heen 
removed in the main consuming markets. (Kenya should seriously consider applying 
to be reclassified as a Least Developed Country for this reason.) 
Consideration should be given to strengthening the role of existing export and invest· 
ment promotion organizations to include the preparation of detailed investment plans 
and packages in added·value products that will attract greater foreign direct invest· 
ment (FDI). Tax regimes should be modified where necessary to encourage this form of 
investment. Vertical diversification may represent the only option for ECA countries 
to avoid the economic damage caused by falling raw commodity prices. 
5. Establishing an Agricultural market analysis unit. An Agricultural market analysis 
unit should be established in each ECA country. This unit would be concerned with 
coordinating and developing policy on the development of a market·orientated strategy 
- 88-
in agriculture and setting policy guidelines for agricultural research. The Unit should 
also coordinate its activities with relevant regional bodies. It should be staffed with 
appropriately qualified economists and market experts. The Unit should work closely 
with the private sector and, especially with those private·sector support groups work-
ing to stimulate production for growth markets. 
6. Establishing a National market education progmm. Many actors in the agricultural 
sectors in ECA countries are still not familiar with the idea of competitive markets. A 
National market education program should be established targeted, primarily, at farm· 
ers, traders, and agricultural product processors. Such a program needs to be linked 
to the Agricultural market analysis unit (see above) and Market information services 
(see below) and run in conjunction with other stakeholders including Ministries of 
Agriculture, Education, and Trade, farmers' and traders' associations and other private 
sector actors, and with extension services. 
The program needs to set targets for training farmers to understand how competitive 
markets work, to take advantage of market information and to inform them of the dif-
ficulties and opportunities associated with market conditions. Issues addressed need 
to include the stimulation of collective activity to improve economies of scale, linking 
supply variety and quality to market needs, negotiation of sales and inputs, and the 
use of credit and business management. 
The program should have a limited duration and should be administered efficiently as 
a separate unit within a national agricultural development reform program. 
7. Establishing a Market information service. Many typical, small- and medium-scale 
farmers, traders, and processors in ECAcountries are very poorly informed about prices 
and market conditions of the commodities they produce. Farmers find themselves in a 
weak bargaining position with traders which results in lower-than-market farm·gate 
prices, high transaction costs, and wastage. Market nformation services need to be 
established at local, national, and regional level to gather, process, and disseminate 
market information in the appropriate language of intended recipients. Such services 
need to be fully coordinated with each other and involve full participation of stakehold· 
ers. 
The aim of these services should be to stimulate more competitive markets. They are 
likely to be supported by the agricultural industry itself as they are in more developed 
countries, once competitive markets become more established. 
8. Strengthening agricultural research and extension and services. Research and exten· 
sion services need to continue with their vital role in controlling plant and animal 
diseases and pests, discovering and distributing new varieties, training farmers to 
improve their technical abilities, etc. If ECA countries wish to compete successfully in 
the world economy, however, these institutions need to develop or acquire new skills 
and expertise in market analysis and market linkage. Producers need to ensure that 
there are viable markets for any existing or new products. They need to ensure that 
the quality and packaging of those products meet the requirements of customers both 
on the domestic and export market. Research and extension services have a vital role 
to play in this effort and must be prepared to reform quickly to meet the challenges of 
globalization. 
-89-
In many respects national research programs have succeeded in their goal to achieve 
food security; the current emphasis should now be to develop dynamic and commer· 
cially orientated research that supports improved market analysis, market access, and 
added value processing. Extension services should now focus on assisting producers to 
trade more effectively within a liberalized market. Special attention should be given 
to aspects such as linkage of production to markets, access to credit, and collective 
marketing which will enable the millions of atomized, small·scale farmers to gain from 
economies of scale in their input and output markets. 
Government research services need to work closely with the private sector which is 
increasingly developing its own research capacity, particularly in regard to higher value 
commodities and research related to issues and problems further up the value chain. 
9. Reducing imports of goods that can be competitively produced domestically. Many ECA 
countries import fruit juices, soluble coffee, cooking oils, etc. when they are rich in all 
the raw materials needed to make these products and have low labor costs. An effort 
should be made to examine import data and to analyze the prospects for developing the 
local manufacture of such products and to encourage investment in the production of 
such goods but only if this can be done profitably without resort to market protection. 
It should be remembered that savings on imports are as valuable as export revenue. 
10. Strengthening the legal framework for market activity. Market manipulation and 
collusion among traders to the detriment of farmers, consumers, and exporters are 
widespread practices in ECA countries. In some countries, road tolls and taxes are 
arbitrarily applied and often restrict trade and increase transaction costs. Where nec· 
essary, governments must institute a program to reform the legal framework within 
which agricultural product transactions take place, establish or reform laws of contract, 
outlaw restrictive practices, and regulate a competitive market in agricultural goods. 
In addition, governments must ensure that these laws are properly enforced. 
Refereed journal articles 
Bigirimana, S., P. Barumbanze, R. Obonyo, and J.P. Legg. 2003. First evidence 
for the spread of East African cassava mosaic virus-Uganda (EACMV-UG) and 
the pandemic of severe cassava mosaic disease to Burundi. New Disease Reports 
[http://www.bspp.org.uklndrl] Volume 8. 
Chiwona-Karltum, L., L. Brimer, J.D. Kalenga,A Mhone, J. Mkumbira, L. Johans-
son, M. Bokanga, N.M. Mabungu, and H. Rosling. 2004. Bitter taste in cassava 
roots correlates with cyanogenic glucoside levels. Journal of the Science of Food 
and Agriculture 84: 581-590. 
Coyne, D.L., K.L. Sahrawat, and RA Plowright. 2003. The influence of mineral 
fertilizer application and plant nutrition on plant-parasitic nematodes in upland 
and lowland rice in Cote d'Ivoire and its implications in long-term experiments. 
Experimental Agriculture 20 (in press). 
Abstract. Mineral fertilizer application and consequent plant nutrition has long been 
observed to influence associated plant.parasitic nematode population densities, offering 
the potential as a nematode management option. Observations were made on the influence 
of mineral fertilizer application on nematode populations on three separate long·term rice 
-90-
experiments, (differential mineral application on upland and on lowland rice and P applica-
tion on upland rice) undertaken between 1994 and 1997 in Cote d'Ivoire. On upland rice, 
treatments with K or N withheld from the comprehensive mineral application treatment 
(range of elements including N, P, K, Ca, Mg and Zn) led to lower densities (P S 0.05) of 
Pratylenchus zeae at harvest than the comprehensive mineral application, in 1995. By with-
holding K or Mg, Helicotylenchus psuedorobustus densities were greater than with either 
the control (no mineral application) or comprehensive mineral application in the same 
year. No differences were observed between treatments in 1994, or between treatments for 
densities of other nematode species present (Meloidogyne incognita, Criconemella tescorum) 
or for total nematode density. In the lowland rice trial, no treatment effects on nematode 
species (Hirschmanniella oryzae and miginotylenchus palustris) were observed. In the P 
application trial on a P-deficient mtisol, Heterodera sacchari densities were lower (P s 0.05) 
in treatments receiving 180 kglha P, than untreated in 1995; in 1996 no differences were 
observed between untreated and 135 kglha P, while in 1997 higher densities of H. sacchari 
were present in 135 kglha P than untreated. Regression analysis of nematode densities 
against mineral straw content in the P application trial revealed a negative correlation 
between M. incognita and Mn (P S 0.001) and Ca (P S 0.05), and between P. zeae and Zn or 
Fe (P S 0.05). A positive correlation was observed between Helicotylenchus spp. and Mg (P 
s 0.05). This study provides strong arguments for the assessment of nematode constraints 
in long-term research trials, the omission of which may severely limit the validity of the 
primary experimental objectives. 
Coyne, D. L., H.A.L. Talwana, and N.R. Maslen. 2003. Plant-parasitic nematodes 
associated with root and tuber crops in Uganda. Mrican Crop Protection 9: 
87-98. 
Abstract. In a nematode survey of eight commonly grown root and tuber crops (cassava, 
sweetpotato, potato, yam, tannia, taro, carrot, and turmeric) from 4303 fields in Uganda, 
68 species of plant-parasitic nematodes from 28 genera were extracted from soil and roots. 
About twice as many nematode species were recovered from soil (64) as from roots (36), 
while 32 species were found only in soil and four species only in roots, usually in mixed 
populations. Meloidogyne spp. (root· knot nematodes) were the most frequently recovered 
nematodes across crops, with the major species (M. arenaria, M. hapia, M incognita, and 
M.javanica) observed on cassava. Some nematodes were recovered only from specific crops. 
Sweetpotato was associated with the greatest diversity of species (55 species in 25 genera), 
followed by cassava (40 species in 19 genera) and yam, which was sampled principally in 
the northern and eastern regions of Uganda (39 species in 14 genera). The study examined 
the distribution of the plant·parasitic nematodes, which provides information for more 
detailed local population studies and nematode pathogenicity evaluation on important 
root and tuber crops. 
DixonA.G.O., R. Bandyopadhyay, D. Coyne, M. Ferguson, R.S.B Ferris, R. Hanna, 
J. Hughes, I. Ingelbrecht, J. Legg, N. Mahungu, V. Manyong, D. Mowbray, P. 
Neuenschwander, J. Whyte, P. Hartmann, and R. Ortiz. 2003. Cassava: from poor 
farmers' crop to pacesetter of Mrican rural development. Chronica 43: 8-15. 
Legg, J.P. and C.M. Fauquet. (2003). Cassava mosaic geminiviruses in Mrica. 
Plant Molecular Biology (in press). 
Mkumbira, J., N.M. Mahungu, and U. Gullberg. 2003. Grouping locations for effi-
cient cassava evaluation in Malawi. Experimental Agriculture 39: 167-179. 
- 91-
Mkumbira, J., L. Chiwona-Karltun, U. Lagercrantz, N. Mahungu, J. Saka, A. 
Mhone, M. Bokanga, L. Brimer, U. Gullberg, and H. Rosling. 2003. Classification 
of cassava into "bitter" and "cool" in Malawi: from farmers' perception to char-
acterization by molecular markers. Euphytica 132 (1): 7-22. 
Nwakanma, D.C., M. Pillay B.E. Okoli, and A. Tenkouano. 2003. PCR-RFLP of the 
ribosomal DNA internal transcribed spacers (ITS) provides markers for the A 
and B genomes in Musa L. Theor. Appl. Genet 108: 154-159. 
Abstract. Musa acuminata Colla (AAgenomes) and M. balbisiana Colla (BB genomes) are 
the diploid ancestors of modern bananas that are mostly diploid or triploid cultivars with 
various combinations of the A and B genomes, including AA, AAA, BB, AAB, and ABB. The 
objective of this study was to identify molecular markers that will facilitate discrimination 
of the A and B genomes, based on restriction site variations in the internal transcribed 
spacers (ITS) of the nuclear ribosomal RNA genes. The ITS regions of seven M. acuminata 
accessions and five M. balbisiana were each amplified by PCR using specific primers. All 
accessions produced a 700 bp fragment that is equivalent in size to the ITS of most plants. 
This fragment was then digested with 10 restriction enzymes (AZul, Cto1, Dde1, Haell1, 
Hin{l, Hpa11 , Msp1, Rsa1, Sau3A1, and Taq1) and fractionated in 2% agarose gels, 
stained with ethidium bromide and visualized under UV light. The Rsa1 digest revealed 
a single 530 bp fragment unique to the A genome and two fragments of 350 bp and 180 bp 
that were specific to the B genome. A further 56 accessions representing AA, AAA, AAB, 
AB, and ABB cultivars and synthetic hybrids were amplified and screened with Rsal. All 
accessions with exclusively A genome showed only the 530 bp fragment while accessions 
having only the B genome lacked the 530bp fragment but had the 350 bp and 180 bp frag-
ments. Interspecific cultivars possessed all three fragments. The staining intensity of the 
B genome markers increased with the number of B genome complements. These markers 
can be used to determine the genome constitution of Musa accessions and hybrids at the 
nursery stage and, therefore, greatly facilitate genome classification in Musa breeding. 
Keywords Musa, genome markers, internal transcribed spacers (ITS), PCR-RFLP. 
Nwakanma, D.C., M. Pillay, B.E. Okoli, and A. Tenkouano. 2003. Sectional relation-
ships in the genus Musa inferred from PCR-RFLP of organelle DNA sequences. 
Theor. Appl Genet 107: 850-856. 
Abstract. The genus Musa is divided into the sections Eumusa, Rhodochlamys, Australi-
musa, and Callimusa on the basis of morphological characters and differences in basic 
chromosome number. Information on the evolutionary relationships in the genus would 
improve our understanding of taxonomic delimitations in Musa and assist in the choice of 
compatible sources of useful variation for the improvement of the cultivated bananas. The 
objective of this study was to construct a molecular phylogeny of the genus Musa using 
restriction site polymorphisms of the chloroplast and mitochondrial DNA. Six chloroplast 
and two mitochondrial DNA sequences were amplified individually through peR. The 
amplified products were digested with 10 restriction endonucleases. A total of 79 restric-
tion site changes were scored among 13 species sampled from the four sections of Musa. 
Wagner parsimony using the branch and bound option, with Ensete ventricosum as the 
outgroup, defined two lines of evolution in Musa_ One lineage comprised species of the sec-
tions Australimusa and Callimusa which have a basic chromosome number ofx = 10 while 
the species of section Rhodochlamys (x = 11) were in the other lineage. Species of Eumusa 
(x = 11) were distributed into both lineages. Section Rhodochlamys appeared as a sister 
-92-
group of section Eumusa with M. laterita Cheesman having identical organelle genome 
type as some subspecies of the M. acuminata complex. The progenitors of the cultivated 
bananas, M. balbisiana Colla and M. acuminata, were evolutionarily distant from each 
other. M. balbisiana occupied a basal position in the cladogram indicating an evolutionarily 
primitive status. The close phylogenetic relationship between M. laterita and M. acuminata 
suggests that species of the section Rhodochlamys may constitute a secondary gene pool 
for the improvement of cultivated bananas. 
Okao-Okuja, G., J.P. Legg, L. Traore, and M. Alexandra Jorge. 2003. Viruses asso-
ciated with cassava mosaic disease in Senegal and Guinea Conakry. Journal of 
Phytopathology (in press). 
Pillay, M., C. Dimkpa, G. Ude, D. Makumbi, and W. Ushemereirwe. 2003. Genetic 
variation in the banana cultivar "Sukali Ndizi" grown in different regions of 
Uganda based on RAPD and AFLP markers. Mrican Crop Science Journal 11: 
87-95. 
Abstract. The objective of this study was to assess genetic relationships among "Sukali 
Ndizi" clones collected from 16 different localities in Uganda using the RAPD (random 
amplified polymorphic DNA) and AFLP (amplified fragment length polymorphism) tech-
niques. Thirty-four RAPD primers used singly and in combination produced 234 unambigu-
ous bands_ The RAPD primers produced identical banding patterns in all the samples and 
were, therefore, not useful in differentiating the clones. On the contrary, the nine AFLP 
primer pairs produced a total of 554 fragments, 17 of which were polymorphic. Genetic 
relationships were established from theM LP data by UPGMAciuster analysis. 'Two groups 
of clones were clearly dermed. It was interesting to find that while clones from some con-
tiguous districts such as Lira and Soroti, Bushenyi, and Kasese were highly similar, other 
closely related clones were from disjunctive localities_ The similarity ofthe clones in adjacent 
districts could be explained by local exchange of germ plasm while similarity of clones in 
noncontiguous districts is probably the result of more purposeful and selective transfer of 
planting material. If we assume that a single clone of "Sukali Ndizi" was introduced into 
Uganda, then the observed variation is possibly due to somatic mutations. 
Pillay, M., J. Hartman, C. Dimkpa, and D. Makumbi. 2003. Sukali Ndizi is a triploid 
with an AAB genome composition. Mrican Crop Science Journal 11: 119-124. 
Abstract: Cultivated bananas originated from interspecific hybridization oftwo wild diploid 
(2n = 2x = 22) species, Musa acuminata and M. balbisiana that were the donors of the A 
and B genomes, respectively. Most cultivated bananas are triploids with 2n = 3x = 33 chro-
mosomes. Banana improvement programs make use of interspecific hybridization for gene 
introgression. Consequently knowledge of the exact ploidy level and genome composition of 
a plant is important for breeding purposes. The ploidy level of a plant can be determined 
in two ways: (1) by a physical count of its chromosomes, and (2) by flow cytometry. In this 
study, the ploidy level of the dessert banana cultivar, Sukali Ndizi, was determined by con-
ventional chromosome analysis and flow cytometry. Our results showed that Sukali Ndizi 
is a triploid. In the past, Sukali Ndizi was considered to be a diploid with anAB genome 
composition. We also determined the genome composition using a set of RAPD (random 
amplified polymorphic DNA) markers and found that it has anAAB genome composition. 
-93-
Sseruwagi, P., W.S. Sserubombwe, J.P. Legg, J. Ndunguru, and J.M. Thresh. 2003. 
Methods of surveying the incidence and severity of cassava mosaic disease and 
whitefly vector populations on cassava in Africa: a review. Virus Research (in 
press). 
Ude, G., M. PiIlay, E. Ogundiwin, and A. Tenkouano A. 2003. Genetic diversity in 
a core African plantain collection using AFLP and RAPD markers. Theor. Appl 
Genet 107: 248--255. 
Abstract. FifteenAFLP primer pairs (EcoRI +3 and MseI +3) and sixty 1 D-mer RAPD prim-
ers were used to detect polymorphisms and assess genetic relationships in a sample of 25 
plantains from diverse parts of Western and Central Mrica .. The discriminatory power of 
the AFLP technique was greater than that of the RAPD technique, since the former pro-
duced markers with greater polymorphic information content (PIC) than the latter. Hence, 
AFLP analysis appeared to be a more powerful approach for identifying genetic differences 
among plantain accessions. In this regard, significant genetic diversity within the plantains 
was shown by unweighted pair-group method of arithmetic averages (UPGMA) and mul-
tidimensional principal coordinate (PCO) analyses. The AFLP derived clusters indicated 
closer relationships between similar inflorescence types than the RAPD derived clusters. 
A small group of cultivars mainly from Cameroon were separated from the bulk of other 
plantains, suggesting that Cameroon may harbor accessions with useful or rare genes for 
widening the genetic base of breeding populations derived from plantains. A greater effort 
should be directed at collecting and characterizing plantain cultivars from Cameroon. 
Book chapter, peer-reviewed 
Legg, J.P., D.Gerling, and P.N. Neuenschwander. 2003. Biological control of white-
flies in sub-Saharan Africa. Pages 87-100 in Biological Control in IPM Systems in 
Africa, edited by P. Neuenschwander, C. Borgemeister, and J. Langewald. CABI, 
Wallingford, UK. 
Conference papers, workshop proceedings, abstracts, newsletters 
Abele, S., B. Bashaasha, and C. Gold. 2003. The need for improved marketing 
and processing to enhance rural livelihoods in Africa. Paper presented at the 
Workshop on Improved Banana Marketing and Utilization, 2-3 October 2003, 
Kampala, Uganda. 
Abstract. Sub-Saharan Mrica has the world's weakest agricultural performance, includ-
ing lowest per capita calorie supply, and trade participation. Future prospects on world 
markets are bleak, as the 2002 US farm bill and the 2003 EU CAP reform are unlikely 
to reduce surplus pressure and thus will continue to hamper Mrican agricultural exports 
and finally the commercialization of Mrica's agriculture. 
But problems are not only external in origin. Efficiency in Mrican agriculture is low, pro-
cessing is poor, and trade is hampered by poorly functioning institutions. Studies have 
proved that the removal of collusion, corruption, and poor infrastructure are more effective 
in stimulating agricultural trade than lowering tariffs. Public investments in infrastructure 
and market information systems are required. Competition must be fostered and institu-
tional failures rectified. Promising new agrobusinesses are often hampered by poor access 
to credit and finance due to poorly developed capital markets. Increasing rural investment 
by combining public and private finance schemes could help foster competition by assisting 
new competitors through their "infant years". 
-94-
Abele, S., U. Fiege, and K. Reinsberg. 2003. The impact of agricultural development 
on agricultural employment and rural labor markets: evidence from Eastern 
Europe and Mrica. Poster presented at the International Agricultural Research 
for Development Conference, 8-10 October 2003, Goettingen, Germany: http:// 
www.tropentag.de/abstractslfuI1l354.pdf. 
Abstract. Agricultural employment is the decisive factor for rural well-being in develop-
ing countries as well as in many eastern European transition countries. This is because, 
differently from Western Europe, there are, in many regions, almost no other employment 
opportunities than farming. Therefore, the decrease of agricultural employment opportuni-
ties yields social and economic distress, especially for the most vulnerable, the landless, 
and the small-scale farmers who have to seek supplementary income from working casu-
ally on other farms. 
But agricultural employment is subject to many factors. Labor hiring is strongly affected 
by short-term productivity developments. Structural adjustment in the agricultural sector 
decreases labor by substituting it with capital. 
This contribution analyses the effects of agricultural development on agricultural labor 
markets and thus on rural employment. It gives evidence from two regions that may seem 
different at first glance, but show surprising similarities: Eastern Europe and Western 
Africa, especially Niger. Using the method of calculating short-term and long-term labor 
demand elasticity for the agricultural sector, effects of both short-term distortions (e.g., 
price or yield declines) and long-term adjustments are assessed. 
For both regions, it can be shown that short-term distortions in agricultural productivity, 
mainly caused by price decline, lead to a sharp decrease in hired labor employment. The 
following structural adjustment mostly leads to a substitution of labor by capital. Con-
sequently, when the initial productivity losses are compensated, less labor than before is 
employed at the same productivity level so that the net effect is a decline in agricultural 
employment over time. 
Based on this knowledge, recommendations are given for both Eastern Europe and West 
Mrica. It is most important to establish a balanced structural policy that aims at both 
improving agricultural efficiency and creating off-farm labor, e.g., in the downstream 
sector. 
Benesi, I.R.M., M.T. Labuschagne, N.M. Mahungu, A.G.O. Dixon, and C.D. Vil-
joen. 2003. Genetic distance analysis of elite Malawi cassava (Manihot esculenta 
Crantz) genotypes using morphological and AFLP marker techniques. Page 166 
in Book of Abstracts. Arnel R. Hallauer International Symposium on Plant Breed-
ing, 17-22 August 2003, Mexico City. 
Ferris, S., S. Abele, and P. Robbins. 2003. Developing market information services 
as a first step in meeting the challenges of globalization in the small-scale farm-
ing sector of sub-Saharan Mrica. Presentation on the Mini-Symposium "Market 
research for the development of commercialized agriculture in Sub-Saharan 
Mrica", 25th IAAE-conference, 16-22 August 2003, Durban, South Mrica. 
Abstract. Globalization is the term used to describe the recent impact of innovations in 
communications and transport systems on trade and the increasing integration of world 
markets. Whilst it has been proved that increased economic liberalization and openness 
-95-
leads to growth, it is also recognized that for many developing countries, liberalization 
has caused serious economic difficulties. A combination of structural adjustment programs 
and partial reform of international trade policies has had the overall effect of reducing the 
prices of primary commodities exported by many African countries. These effects combined 
with localized insecurity have also led to significantly increased imports of agricultural 
products, from more competitive producers, some of which remain highly subsidized. The 
result of oversupply of the commodities markets has led to commodity prices falling to their 
lowest levels in 40 years. This bleak situation is reflected in the dramatically falling terms 
of trade for many African countries and suggests a profound downturn in their economic 
outlook. At the same time, the profits of international trading processing houses and retail 
companies in developed countries have shown year on increases. The international com-
munity has recognized these marketing difficulbes and is designing tariff reducing trading 
opportunities for LDCs and is attempting to reform agricultural support programs in a 
package that will assist developing countries to overcome them. 
Despite these new trading initiatives, it is the view of the authors that many African 
countries have not fully appreciated the scale and implications of globalization on their 
futures and that, without urgent action on their part; they may seriously weaken their 
economies in the years ahead. Options for improving the competibveness of small-scale 
agricultural producers in SSA, include (i) Improving the negotiating powers of LDCs for 
global trade, (ii) Managing the oversupply of primary product exports, (iii) Stimulating 
production for exports of added-value products, (iv) Strengthening internal institutions, 
especially markets. It can be shown that African trade is not only hampered from outside, 
but there are also severe impediments arising from inefficient structures inside African 
trade systems, such as collusion, taxation and trade policies, poor infrastructure, and poor 
market information. It is therefore vital to improve African markets by providing access for 
new actors such as businessmen and farmers' trade cooperatives, and to improve market 
information services, and (v) Reducing imports of goods that can be competitively produced 
domestically. 
To make gains in any of these areas, African governments need to initiate practical, stepwise 
processes, in partnership with donors, NGOs, and private sector associations, to gradually 
shift their agricultural sectors to a more competitive basis. The way this reform process is 
both planned and implemented is particularly important in those countries that rely upon 
many millions of small-scale producers. As a starting point, this paper focuses on the oppor-
tunities for assisting small-scale farmers and small- to medium-scale traders to improve 
their understanding of market conditions and signals, through the provision of integrated 
market information services that target the specific needs of producers, processors, and 
traders. This educational/information based approach to market reform is considered a 
relatively low cost and practical means of supporting the different types of actors along 
the market chain, to gain a better understanding of the marketing dynamics that affect 
their ability to trade profitably, and to provide them with the necessary decision support 
tools for them to engage more effectively in the marketplace. 
Gondwe, F.M.T., N.M. Mahungu, R.J. Hillocks, M.D. Raya, C.C. Moyo, M.M. Soko, 
F.P. Chipungu, and I.R.M. Benesi. 2003. Economic losses experienced by smaII-
scale farmers in Malawi due to cassava brown streak virus disease. Pages 28-38 
in Cassava Brown Streak Virus Disease: Past, Present and Future, edited by J.P. 
-96-
Legg and R.J. Hillocks. Proceedings of an International Workshop, 27-30 October 
2002, Mombasa, Kenya. 
Kajumba, C., C. Gold, P.R. Speijer, W. Tushemereirwe, D. Coyne,A.R. Semana, and 
J. Mudiope. 2003. Farmers' perception on use of banana clean planting material 
for pest management in Uganda. In Integrated Pest Management Conference Pro-
ceedings, IPM Conference, edited by J.S. Tenywa, M.P. Nampala, S. Kyamanywa, 
and M. Osiru, 8--12 September 2002, Kampala, Uganda (in press). 
The East Mrican highland banana (Musa genome group AAA-EA) is a staple crop in the 
Ugandan diet consumed by a large proportion of the population. Banana weevils (Cos· 
mopolites sordidus) and nematodes (especially the aggressive Radopholus similis) have 
been observed to reduce production. These are often and easily transferred to new fields 
through infested planting material. To address this constraint, an inexpensive and afford· 
able method using clean planting material was developed and disseminated to farmers 
in three districts, i.e., Ntungamo, Mpigi, and Luweero. In these districts, approximately 
1000 farmers were trained on the use of clean planting material technology using a farmer 
participatory approach. Farmers then established banana fields using clean planting mate-
rial. From the study, a total of 180 farmers were interviewed on their opinions on the 
technology. Among the positive observations, 13% and 34% observed that the technology 
led to reduced pests and improved yields, respectively. Some of the negative observations 
by the farmers included drought susceptibility (19%) and 6% observed that they were 
more susceptible to livestock. According to the farmers' perceptions, the technology was 
appropriate and worthwhile under their circumstances, managed pests, improved banana 
yields, and proved profitable. 
Kanju, E.E., E. Masumba, M.Masawe, S. Tollano, B. Muli, A. Zacarias, N. Mahungu, 
B. Khizzah, J. Whyte, and A. Dixon. Breeding cassava for brown streak resistance: 
regional cassava variety development strategy based on farmers'/consumers' 
preferences. Proceedings of the 13th symposium of the International Society 
for Tropical Root Crops, 10--14 November 2003, Arusha, Tanzania (in press). 
Abstract: Cassava is an important food crop in the EastMrican coastal lowlands. However, 
yields are low due to various reasons amongst which are pests and diseases. Cassava brown 
streak virus disease (CBSD) is one ofthe major diseases of economic importance in Kenya, 
Malawi, Mozambique, and Tanzania. Yield losses attributed to eBSD ranging from 49 to 
74% have recently been reported in some coastal areas of Tanzania. It is estimated that 
in Tanzania alone, CBSD causes economic losses of more than USD 16000000 annually. 
The wide adoption of resistant varieties is the best sustainable control strategy. However, 
only a handfull of resistant/tolerant varieties is available in each of the severely affected 
countries. In order to increase the number of resistant varieties available to farmers, a 
seedling nursery was raised at Kibaha, Tanzania. About 513 seedlings were selected and 
cloned for further evaluation. At Kibaha, eBSD disease pressure and spread was very high 
therefore, 88% of the harvested seedlings were infected. This indicated that this site was 
ideal for screening cassava genotypes for resistance to CBSD. Although the vector is still 
unknown, it was very active and infective at this site. Because the seedlings were raised 
from botanical seeds, ruling out the chance of seedborne transmission, these results prove 
that vector transmission can account for a large percentage of plants infected in a field. 
Families with a high percentage of number of seedlings selected (PNSS) included Kiroba, 
71762, and 188100188. The importance of marking diseased seedlings during the grow· 
-97-
ing season is discussed. Furthermore, massive introduction of germplasm through open 
quarantine, the establishment of crossing blocks to generate improved genotypes, and the 
use of a decentralized participatory evaluation procedure have been initiated as important 
strategies towards availing farmers of a wider choice of improved varieties. 
Kanju, E.E., J. Whyte, and A. Dixon. 2003. Is resistance/tolerance of cassava to 
brown streak disease associated with the zigzag stem trait? ROOTS Newsletter 
8(2): 15-19. 
Abstract. Cassava brown streak disease (CBSD) first reported from Amani, Tanzania 
in 1936, is now considered to be one of the major biotic constraints to cassava production 
especially in the coastal low lands of East and Southern Africa. Efforts to develop resistant 
varieties as the major control measure for CBSD started way back in the 1940s including 
intraspecific crosses, which, however, resulted only in few varieties with field tolerance to 
brown streak. In efforts to improve breeding efficiency, some researchers have investigated 
the correlation between morphological traits and resistance to CBSD. The zigzag stem 
growth habit is one of the marker genes identified in cassava. Marker genes control the 
expression of traits which can be easily classified into distinct states, each controlled by a 
different allele, and whose expression is little influenced by the environment. There is evi-
dence to suggest that a single recessive gene (zz) controls the zigzag stem. Field observations 
have started to reveal that most ofthe cultivars, which are tolerant/resistant to CBSD, are 
heterozygous (Zz) for the zigzag stem characteristics. In Tanzania, the following cultivars 
that are tolerant to CBSD are all heterozygotes: Kigoma Mafia, Nanchinyaya, Kiroba, 
TMS 30001, and Amani 46106127. In Kenya, two cultivars that are tolerant to CBSD are 
also heterozygotes: Kaleso (Amani 46106127) and Kahoteli. Furthermore, in Mozambique 
two cultivars that are tolerant to CBSD are heterozygotes: Mulaleia and Macia 1. These 
observations imply that there is a strong possibility that the gene conferring resistance 
to CBSD is linked to the "z" allele. However, the heterozygote in most cases has a normal 
stem growth habit. Therefore, if it is true that the "z" allele confers resistanceftolerance to 
CBSD, it may prove difficult to use this correlation by visual scoring of clones. 
Further evidence is needed to prove the association between CBSD resistanceltolerance 
and the zigzag stem trait. Rapid genetic progress will be made in breeding for resistance 
to CBSD if this association is proven. 
Mahungu, N.M., M. Bidiaka, H. Tat, S. Lukombo, and S. N'luta. 2003. Cassava 
brown streak disease-like symptoms in Democratic Republic of Congo. ROOTS 
Newsletter 8(2): 8-10. 
Niere, B., C. Gold, and D. Coyne. 2003. Endophytes and banana IPM: Current 
status and future prospects. Pages 171-178 in Integrated Pest Management Con-
ference Proceedings, IPM Conference, edited by J.S. Tenywa, M.P. Nampala, S. 
Kyamanywa, and M. Osiru, 8-12 September 2002, Kampala, Uganda. 
Endophytic fungi colonize living plant tissue without causing symptoms of damage to their 
hosts. Endophytic fungi have been isolated from all plants investigated and mutualistic 
relationships of fungal endophytes with their host plant could be demonstrated. Studies 
on banana material originating from Uganda have shown that banana corm and root 
tissue hosts a variety of endophytic fungi. Preliminary results in Uganda have shown that 
indigenous endophytic fungi from banana can effectively colonize tissue-cultured banana 
when plants are inoculated at an early stage. Inoculation could lead, depending on fungal 
-98-
isolate and banana clone, to increased plant growth, reduced nematode multiplication, 
and reduced weevil damage. Control levels vary between fungal species and strains and 
host-€ndophyte interactions influence chances of success. Interactions of the fungi with 
their host plant and weevils and nematodes, respectively, and the mechanisms leading to 
effective control are not well understood. Similarly, the duration of protection provided 
by the endophyte to the plant is not known at present. Currently, research focuses on the 
inoculation of single strains onto endophyte-free planting material, such as tissue-cultured 
plants, and targets the initial period of plant establishment in the field. As our knowledge 
of the microbial biodiversity inside the plant increases, combinations of different fungal 
strains or species might be used in the future. This will help to achieve different levels of 
control for a number of pests and also to target specific periods. Long-term control might 
involve persistence of the fungal endophyte or the activation of effective and durable plant 
defense responses to pests. Ideally, fungal endophytes are applied only once and persist 
within the host plant to ensure protection for a considerable or the entire time span of the 
crop. This is a major advantage to biocontrol agentsibiopesticides, which are more costly 
to produce, have to be applied more frequently, and are often highly influenced by envi· 
ronmental factors. The use of fungal endophytes also reduces the risk of side effects on 
nontarget crops and humans as no field applications are required. The use of indigenous 
strains further limits environmental risks. Acceptability of endophytes for pest control is 
therefore considered high from an environmental and economic point of view. Research at 
the International Institute of Tropical Agriculture in Uganda focuses on the identification of 
fungal endophytes that are effective against a number of pests and diseases of banana and 
the mechanisms leading to biocontrol. One of our prime objectives is the optimization and 
integration of this new technology into agricultural systems in East Africa. Management 
of endophytes must be seen as one component in IPM strategies and has to be combined 
with other control options such as tissue·cultured planting material. 
Ntawuruhunga, P., J.B.A. Whyte, and P. Rubaihayo. 2003. Effects of altitude and 
plant age on cyanogenic potential, dry matter, starch and sugar content in cas-
sava genotypes. Presented at the 13th Symposium of the International Society 
for Tropical Root Crops (ISTRC), 10-14 November 2003, Arusha Tanzania. 
Abstract. A study was conducted at three locations in Uganda (Bulisa: 650 m asl; Namu-
longe: 1150 m asl, and Kapchorwa: 1750 m asl) to evaluate the effect of altitude on dry 
matter, starch and sugar content, and cyanogenic potential in cassava. Ten clones from five 
different sources were used in a randomized complete block design with three replications, 
and data were repeatedly measured at three monthly intervals up to 15 months after plant-
ing. The mixed model analysis indicated significant (P < 0.001) differences among genotypes 
at each elevation for dry matter production. Significant interaction between plant age and 
source and between genotypes within source and plant age at all elevations indicated geno-
typic differences among sources a well as among genotypes within sources in production 
and partitioning of assimilates. The study indicated that dry matter content was higher 
at high altitude than at low and midaltitude. Highly significant (P < 0.001) differences for 
CNP were also observed between sources and among genotypes within source at midalti-
tude. Plant age effect was significant for CNP. Significant (p < 0.001) interactions between 
plant age with source and with genotypes nested within source indicated that genotypes 
as well sources produced different amount of CNP during plant growth cycle, obtaining 
high levels at all elevations at around six months after planting. The study also showed 
that starch and sugar contents were high at higher altitude than at low and midaltitude, 
-99-
a similar trend observed with dry matter content. This study showed that there is a need 
of further physiological investigation for a better explanation. 
Ntawuruhunga, P., J.P. Legg, J. Mabanza, and A.J.C. Mvila. 2004. UTA's efforts 
in mitigating cassava mosaic disease crisis in the Republic of Congo-Brazzaville: 
A poster presented at the Sixth International Scientific Meeting of the Cassava 
Biotechnology Network (CBN), 8--14 March 2004, CIAT, Cali, Colombia. 
Abstract. Cassava (Manihot esculenta Crantz) is the main food staple and source of income 
of the population of the Republic of Congo. It constitutes the backbone of the farming 
system and the main commodity of the Congo population. Its consumption per capita is the 
highest on the continent. However, since 1997/98 the cassava mosaic virus disease (CMD) 
has become the most important constraint to production following a period of unrests that 
complicated food security causing food shortage in the country. 
The USA Embassy in Brazzaville requested an urgent intervention from the Office of United 
States Foreign Disaster Assistance (OFDA) through the Cassava Mosaic project managed 
by the International Institute of Tropical Agriculture (IITA) in East Mrica. Subsequently, 
IITAconducted two follow· up diagnostic surveys in 2002 and 2003 that have shown that 
severe virus strain (EACMV·Ug or Uganda Variant) associated with the East African pan· 
demic is the most important devastating disease constraint and had spread throughout the 
country. With limited funds, mitigation efforts were deployed through different activities: 
introductions ofIITAresistant germplasm, collection of national germplasm, its evaluation 
under hotspots of the disease, and multiplication of tolerant cultivars accompanied with 
activities of sensitization through training programs of personnel and farmers. 
Ntawuruhunga, P., J.B.A. Whyte, and P. Rubaihayo. 2003. Correlation and prin-
cipal components studies of some plants attributes at early plant growth in cas-
sava. Presented at the 6th Africa Crop Science Society Conference, 12-17 October 
2003, Nairobi, Kenya. 
Abstract. A potted experiment was conducted at Sendusu farm of ESARC to determine 
morphological and physiological determinants of yield performance using two genotypes 
from five sources of origin viz. East African lowland, midaltitude, highland, and West 
African lowland and midaltitude. A randomized block design with three replications was 
used. The results showed variability among genotypes in all the characteristics evaluated. 
Significant differences (P < 0.001) among sources for sprouting, plant height, leaf area, 
potential storage root, storage root formation, and differentiation were observed. The geno· 
types within sources did not differ significantly, indicating the importance of environmental 
source in guiding genotype development and germplasm exchange. Simple correlation and 
principal component analyses revealed that the meristem width played a key role in con-
trolling plant development and storage root differentiation. Principal component analysis 
grouped the different cassava plant traits in three categories, (1) plant biomass (plant 
height, leaf area, and leaf dry matter), (2) plant physiology (meristem width, length, and 
volume) and (3) plant underground (total number of root and potential storage root). The 
findings will be used in guiding cassava breeding and selection at early stages. 
PilIay, M., A. Tenkouano D. Makumbi, and P. Ragama. 2003. Genotype x Environ-
ment interactions in muItilocational testing of Musa hybrids in three sites in 
Uganda. 6th African Crop Science Society Meeting, 12-17 October 2003, Nairobi, 
Kenya. 
-100-
Abstract. Banana is an important starchy staple and income· earning crop in the highlands 
of East Mrica. The crop is affected by a number of diseases and pests that are responsible 
for declining yields. The development and deployment of resistant hybrid cultivars is 
considered the most economically feasible way of enhancing banana production in East 
Mrica. The challenge facing breeders is to select hybrids that will perform well in various 
agroecologies. Multilocational evaluation trials (MET) were set up to evaluate the stability 
of hybrids. The objective of this study was to determine genotype x environment interac· 
tions for bunch weights, fruit filling time, and black Sigatoka resistance in Musa. Fourteen 
genotypes were studied in three different ecoregions in Uganda. The experimental layout 
was a randomized complete block design with three replications and 10 plants per genotype. 
Data on various growth, disease, and yield components were collected over three cycles. 
Bunch weights varied amongst the hybrids within each site. Bunch weights were influenced 
by altitude and were greater at higher elevations. Fruit filling times also varied across 
the three locations and seemed to be dependent on altitude. Plants grown at higher eleva· 
tions were observed to have longer fruit filling times. All the hybrids took longer to reach 
maturity than the controllandrace in all three locations. The index of nonspotted leaves 
at flowering showed that all the hybrids were less susceptible to black Sigatoka than the 
controllandrace. This was not unexpected since most of the hybrids had Sigatoka· resistant 
genes introgressed from a wild diploid species. 
Talengera, D., M. Pillay, and C. Kasozi. 2003. Genetic diversity studies in East 
African highland bananas using RAPD markers. 6th African Crop Science Society 
Meeting, 12-17 October 2003, Nairobi, Kenya. 
Abstract. Banana is a major staple food crop in the highlands of East Africa. The genetic 
diversity in the crop has not been investigated in detail. In this study we used the RAPD 
(random amplified polymorphic DNA) technique to assess the genetic diversity and rela· 
tionships in banana germplasm from Uganda. Genetic relationships among the cultivars 
are useful to determine parents that could be used in a program breeding for superior 
hybrids with disease and pest resistance. DNA was extracted from 33 landraces using a 
modified eTAB procedure and amplified by PCR (polymerase chain reaction) using XX 10 
base primers from OPERON Technologies. Gels were stained in ethidium bromide and 
photographed under UV light. The results showed that all the cultivars produced identical 
banding patterns with all the primers used in this study. These results suggest that the 
East Mrican highland bananas in Uganda have a very narrow genetic base that may be 
due to domestication or genetic bottleneck effects. The results also showed that the DNA 
markers were not able to separate the "beer" and "matooke" bananas. Other multiplex 
DNA marker systems such as AFLP (amplified fragment length polymorphism) may have 
to be used to resolve differences in these unique bananas. 
Bigirimana, S. and J.P. Legg. 2003. The threat to cassava production on Burundi 
posed by the CMD Ppndemic and its likely impact. Page 69 in Abstract, 13th Sym-
posium of the International Society for Tropical Root Crops, 10-14 November, 
Arusha, Tanzania. 
Jeremiah, S.C., I. Ndyetabura, J.P. Legg, and H. Kulembeka. 2003. Monitoring 
the spread of severe mosaic diseases in the Lake zone of Tanzania. Page 185 in 
Abstract, 13th Symposium of the International Society for Tropical Root Crops, 
10-14 November, Arusha, Tanzania. 
-101-
Legg, J.P and R.J. Hillocks (eds). 2003. Cassava brown streak virus disease: past, 
present and future. Proceedings of an International Workshop, 27-30 October 
2002, Mombasa, Kenya. Natural Resources International Limited, Aylesford, 
UK. 100 p. 
Legg, J.P. 2003. Cassava brown streak virus characterization and diagnostics. 
Pages 41-45 in Proceedings of an International Workshop, 27-30 October 2002, 
Mombasa, Kenya. Natural Resources International Limited, Aylesford, UK. 
Legg, J.P., F. Ndjelassili, and G. Okao-Okuja. 2003. First report of cassava mosaic 
disease and cassava mosaic geminiviruses in Gabon. New Disease Reports [http: 
flwww.bspp.org.uklndrf] Volume 8. 
Legg, J.P. and B. Owor B. 2003. Cassava mosaic disease in Africa: where are the 
epidemics? 6th African Crop Science Society Conference, 12-17 October, 2003, 
Nairobi, Kenya. 
Legg, J.P. and Owor. 2003. Cassava mosaic disease in Africa: epidemics and con-
trol priorities. Joint IITAlIFAD Emergency Workshop on the Preemptive Manage-
ment of the Virulent form of Cassava Mosaic Disease in West and Central Africa, 
9-10 October 2003, Yaounde, Cameroon. 
Legg, J.P., R.W. Gibson, and J. d'A Hughes. 2003. Virus diseases of root crops 
in Africa: An Overview. Page 67 in Abstract, 13th Symposium of ISTRC, 10-14 
November 2003, Arusha, Tanzania. 
Legg, J.P., S. Mallowa, and P. Sseruwagi. 2003. First report of physical damage to 
cassava caused by the whitefly Bemisia tabaci (Gennadius) (Hemiptera: Sternor-
rhyncha:Aleyrodidae). Page 41 in Abstract, 3rd International Bemisia Workshop, 
17-20 March 2003, Barcelona, Spain. 
Legg, J.P., M. Otim, B. Owor, P. Ntawuruhunga, I. Ndyetabura, H. Obiero, S. Kya-
manywa, J. Colvin, and D. Gerling. 2003. Page 131 in Managing cassava mosaic 
gemini viruses and their Bemisia tabaci vector in Africa: current practice and 
future opportunities. Abstract, 3rd International Bemisia Workshop, 17-20 Marcb, 
Barcelona, Spain. 
Legg, J.P., B. Owor, P. Ntawuruhunga, J. Ndunguru, and P. Sseruwagi. 2003. 
Cassava mosaic geminiviruses, Bemisia whiteflies, and the African pandemic of 
cassava mosaic disease. Page 50 in Abstract, 3rd International Bemisia Workshop, 
17-20 March, Barcelona, Spain. 
Mabanza, J., A.M. Makiza, and O. Ntawuruhunga. 2003. Reprise de la recherche 
sur Ie manioc au Congo suite a la pandemie de la mosaique africaine du manioc. 
Presented at the 13th Symposium oflSTRC, 10-14 November 2003, Arusha Tan-
zania. 
Mallowa, S.O., J.P. Legg, D.K. Isutsa, and A.W. Kamau. 2003. Use of cultural meth-
ods for the management of cassava mosaic disease in Western Kenya. Abstract. 
6th African Crop Science Society Conference, 12-17 October, Nairobi, Kenya. 
Mvila, A.C., P. Ntawuruhunga, P.A. Bembe, and M. Obambi. 2003. La mosaique 
africaine du manioc en Republique du Congo: Distribution et importance de la 
maladie. Presented at the 13th Symposium of the International Society for Tropi-
cal Root Crops (ISTRC), 10-14 November 2003, Arusha Tanzania. 
-102-
Ndyetabura, I., J.P. Legg, and E. Adipala. 2003. Reaction of local and research 
released sweetpotato varieties to sweetpotato virus disease (SPVD) and yield 
performance in the Lake Zone of Tanzania. Page 153 in Abstract, 13th Symposium 
of the International Society for Tropical Root Crops, 1-14 November, Arusha, 
Tanzania. 
Obiero, H.M., M.S. Akhwale, B.W. Khizzah, P. Ntawuruhunga, J.B.A. Whyte, and 
J. DeVries. 2003. Participatory evaluation and selection of improved cassava 
varieties in Western Kenya. Presented at the 13th Symposium of the Interna-
tional Society for Tropical Root Crops (lSTRC), 10-14 November 2003, Arusha 
Tanzania. 
Obiero, H., J.B. Whyte, J.P. Legg, M.S. Akhwale, J. Malinga, and T. Magut. 2003. 
Successful restoration of cassava production on western Kenya. Page 189 in 
Abstract, 13th Symposium of the International Society for Tropical Root Crops, 
10-14 November, Arusha, Tanzania. 
Omongo, C.A., J. Colvin, T. Alicai, J.P. Legg, J.M. Thresh, and G.W. Otim-Nape. 
2003. Field colonization by B. tabaci of CMD-affected and unaffected cassava 
plants and its relationship to spread of cassava mosaic disease. Abstract, 6th 
Mrican Crop Science Society Conference, 12-17 October, Nairobi, Kenya. 
Omongo, C.A., J. Colvin, T. Alicai, W. Sserubombwe, J.P. Legg, and R.W. Gibson. 
2003. Host selection by Bemisia tabaci of CMD-symptomatic and non-symptomatic 
cassava plants. Page 199 in Abstract, 13th Symposium ofISTRC, 10-14 November, 
Arusha, Tanzania. 
Otim, M., J.P. Legg, S. Kyamanywa, A. Polaszek, and D. Gerling. 2003. Potential 
for biological control of the whitefly, Bemisia tabaci, vector of cassava mosaic 
geminiviruses using paras ito ids in Uganda. Page 74 in Abstract, 3rd International 
Bemisia Workshop, 17-20 March, Barcelona, Spain. 
Otim, M.H., J.P. Legg, S. Kyamanywa, and D. Gerling. 2003. Occurrence of para-
sitoids of Bemisia tabaci on cassava in Mrica: a review. Abstract, 6th Mrican 
Crop Science Society Conference, 12-17 October 2003, Nairobi, Kenya. 
Owor, B., J.P. Legg, G. Okao-Okuja, R. Obonyo, P. Ntawuruhunga, H. Obiero, I. 
Ndyetabura, E. Marandu, J. Sato, R. Mayiga, S. Bigirimana, and J. Mabanza. 
2003. Novel approaches to controlling cassava mosaic disease in East Mrica. 
Abstract, 6th Mrican Crop Science Society Conference, 12-17 October 2003, Nai-
robi, Kenya. 
Owor, B., J.P. Legg, G. Okao-Okuja, R. Obonyo, and M. Ogenga-Latigo. 2003. Inves-
tigation of virus interactions and putative cross-protection through grafting and 
transmission studies with Bemisia tabaci. Page 62 in Abstract, 3rd International 
Bemisia Workshop, 17-20 March 2003, Barcelona, Spain. 
Owor, B., J.P. Legg, G. Okao-Okuja, and R. Obonyo. 2003. Diagnostic techniques 
for cassava mosaic begomoviruses in Mrica. Page 190 in Abstract, 13th symposium 
of the International Society for Tropical Root Crops, 10-14 November, Arusha, 
Tanzania. 
Semaganda, R., D. Gerling, J.P. Legg, S. Kyamanywa, and E. Adipala. 2003. Aug-
menting the activity of cassava whitefly parasitoids by intercropping cassava 
-103-
with sweetpotato. Page 39 in Abstract. 3rd International Bemisia Workshop, 17-20 
March 2003, Barcelona, Spain. 
Semaganda, R., J.P. Legg, D.Gerling, S. Kyamanywa, and E. Adipala. 2003. Aug-
menting the activity of Bemisia tabaci parasitoids on cassava by intercropping 
cassava with sweetpotato. Abstract, 6th African Crop Science Society Conference, 
12-17 October 2003, Nairobi, Kenya. 
Sseruwagi, P., J.P. Legg, D. Rogan, and J.K. Brown. 2003. Molecular character-
ization and geographical distribution of cassava Bemisia tabaci (Gennadius) 
(Hemiptera: Aleyrodidae) genotypes in Uganda. Abstract, 6th African Crop Sci-
ence Society Conference, 12-17 October 2003, Nairobi, Kenya. 
Sseruwagi, P.P., J.P. Legg, C.M.E. Rey, J. Colvin, and J.K. Brown. 2003. Molecular 
variability of cassava Bemisia tabaci genotypes in East and Central Africa. Page 
18 in Abstract, 3rd International Bemisia Workshop, 17-20 March, Barcelona, 
Spain. 
Oral presentations 
Coyne, D. 2003. Integrated nematode management in developing countries. First 
PINCNET-Africa Workshop, 21-26 September 2003, Ibadan, Nigeria. Oral presen-
tation. 
Coyne, D. 2003. Nematology research for development at UTA. First PINCNET-
Africa Workshop, 21-26 September 2003, Ibadan, Nigeria. Oral presentation. 
Coyne, D., J. Mudiope, and B. Hugues. 2003. Advances in yam and cassava nematol-
ogy research at UTA. 16th Nematological Society of Southern Africa Symposium, 
1-4 July 2003, Strand, South Africa. Oral presentation. 
Dochez, C., J.B.A. Whyte, M. Pillay, D. Coyne, and D. De Waele. 2003. Reproductive 
fitness and pathogenic potential on Musa of different Radophilis similis popula-
tions from Uganda. Sixteenth Nematological Society of Southern Africa Sympo-
sium, 1-4 July 2003, Strand, South Africa. Oral paper presented by Dochez. 
Mudiope, J., D. Coyne, E. Adipala, and R. Sikora. 2003. Pathogenicity of the lesion 
nematode, Pratylenchus sudanensis (Loof & Yassin) on Yam variety Dioscorea 
rotundata. Proceedings IV African Crop Science Conference, 12-17 October 2003, 
Nairobi, Kenya. Oral presentation by Mudiope. 
Niere, B., C. Gold, and D. Coyne. 2003. Can fungal endophytes control soilborne 
pests in banana? IOBC, Bonn, Germany, 1-4 June 2003. Oral paper presented by 
Niere. 
Paparu, P., N. Bjoern, C. Gold, D. Coyne, and E. Adipala. 2003. Colonization of 
banana by fungal endophytes. Proceedings IV African Crop Science Conference, 
12-17 October, Nairobi, Kenya. Oral presentation by Paparu. 
Extension/promotional material 
Nematology at ITTA: research awareness poster. Coyne, D. 
Nematology in Musa production: awareness poster. Coyne, D. 
Pest identification sheet: insects on yam. Coyne, D. and L. Kenyon 
Pest identification sheet: nematodes on yam. Coyne, D. and L. Kenyon 
Pest identification sheet: fungal diseases on yam. Coyne, D. and L. Kenyon 
-104-
Pest identification sheet: viruses on yam. Coyne, D., J. Hughes, and L. Kenyon 
Yam pest identification poster: awareness poster. Coyne, D. and L. Kenyon 
Train ing/ meeting/workshops 
International (Mrican) nematology workshop organized and conducted at UTA, 
Ibadan, 19-23 September 2003. Coorganized with University Gent, Belgium for 
Postgraduate International Nematology Course (PIN C) Mrican Alumni. 
Mozambique (13-21 April). Training in nematology sampling and damage assess-
ment for diagnostic survey on cassava for INIA technician, 13-21 April 2003, 
Maputo, Mozambique. 
-105-