Coffee Wilt Disease

Coffee Wilt Disease
This page intentionally left blank
Coffee Wilt Disease
Edited by
Julie Flood
CABI
Bakeham Lane
Egham
Surrey
TW20 9TY
UK
CABI is a trading name of CAB International

CABI Head Office
Nosworthy Way
Wallingford
Oxfordshire OX10 8DE
UK
Tel: +44 (0)1491 832111
Fax: +44 (0)1491 833508
Email: cabi@cabi.org
Web site: www.cabi.org
CABI North American Office

875 Massachusetts Avenue
7th Floor
Cambridge, MA 02139
USA
Tel: +1 617 395 4056
Fax: +1 617 354 6875
Email: cabi-nao@cabi.org
CAB International 2009. All rights reserved. No part of this publication may be reproduced
in any form or by any means, electronically, mechanically, by photocopying, recording or
otherwise, without the prior permission of the copyright owners.
ISBN: 978-1-84593-641-9
Typeset by MTC, Manila, Philippines

Printed and bound in the UK
Contents
Contributors
vii
Preface
C. Dengu
ix
Acknowledgements
xi
1.
Introduction
J. Flood
2.
Coffee Wilt Disease in Democratic Republic of Congo

A. Kalonji-Mbuyi, P. Tshilenge Djim and N.T. Saiba
3.
Coffee Wilt Disease in Uganda

G.J. Hakiza, D.T. Kyetere, P. Musoli, P. Wetala, J. Njuki,
P. Kucel, P. Aluka, A. Kangire and J. Ogwang
28
4.
Coffee Wilt Disease in Ethiopia

A. Girma, A. Million, H. Hindorf, Z. Arega, D. Teferi and
C. Jefuka
50
5.
Status of Coffee Wilt Disease in Tanzania

D.L. Kilambo, N.M. Nghoma, J.M. Teri and L. Masumbuko
69
6.
Socio-Economic Impact of Coffee Wilt Disease

R.O. Musebe, J. Njuki, S. Mdemu, G. Lukwago, A. Shibru and
T. Saiba
83
7.
Biology, Taxonomy and Epidemiology of the Coffee Wilt

Pathogen Gibberella xylarioides sensu lato
M.A. Rutherford, D. Bieysse, P. Lepoint and H.M.M. Maraite
8.
9.
HostPathogen Interactions in CoffeaGibberella

xylarioides Pathosystem
A. Girma, D. Bieysse and P. Musoli
Management of Coffee Wilt Disease
N. Phiri, M. Kimani, N. Efa, S. Simons and G. Oduor
CAB International 2009. Coffee Wilt Disease
99
120
137
10. Breeding for Resistance Against Coffee Wilt Disease

P.C. Musoli, A. Girma, G.J. Hakiza, A. Kangire, F. Pinard,
C. Agwanda and D. Bieysse
11. Extension Approaches and Information Dissemination for
Coffee Wilt Disease Management in Africa: Experiences
From Ethiopia
E. Negussie, M. Kimani, A. Girma, N. Phiri and D. Teshome
155
176
12. Concluding Remarks

J. Flood
196
Index
201
The colour plate section can be found after page 84.
vi
Contributors
S. Admasu, Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research, PO

Box 192, Jimma, Ethiopia
C. Agwanda, CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya.
E-mail: c.agwanda@cabi.org
P. Aluka, Coffee Research Centre, PO Box 185, Mukono, Uganda. E-mail: cori@africaonline.co.ug
Z. Arega, Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research, PO Box
192, Jimma, Ethiopia.
D. Bieysse, Centre de Coopration Internationale en Recherche Agronomique pour le Dveloppement
(CIRAD), TA41/K , Campus International Baillarguet, 34398 Montpellier, Cedex 5, France.
A. Girma, Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research, PO Box
192, Jimma, Ethiopia. E-mail: girma.adugna@yahoo.com
G. J. Hakiza, Coffee Research Centre, PO Box 185, Mukono, Uganda. E-mail: cori@africaonline.
co.ug
H. Hindorf, INRES-Phytomedizin, University of Bonn, Nussallee 9, 53115 Bonn, Germany.
C. Jefuka, Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research, PO Box
A. Kalonji-Mbuyi, Universit de Kinshasa, BP 117, Kinshasa XI, DRC and Centre Rgional dEtudes
Nuclaires de Kinshasa (CREN-K), BP868, Kinshasa XI, DRC. E-mail: adrienkalonji@yahoo.fr
A. Kangire, Coffee Research Centre, PO Box 185, Mukono, Uganda. E-mail: cori@africaonline.co.ug
D. L. Kilambo, Tanzania Coffee Research Institute (TaCRI), Lyamungu, PO Box 3004, Moshi, Tanzania. E-mail: www.tacri.org
M. Kimani, CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya. E-Mail:
m.kimani@cabi.org
P. Kucel, Coffee Research Centre, PO Box 185, Mukono, Uganda. E-email: cori@africaonline.co.ug
D.T. Kyetere, Coffee Research Centre, P.O.Box 185, Mukono, Uganda, email: cori@africaonline.
co.ug
P. Lepoint, Unit de Phytopathologie, Universit Catholique de Louvain, Croix du Sud 2/3, B-1348
Louvain-la-Neuve, Belgium.
G. Lukwago, National Agricultural Research Organization, PO Box 421, Kabale, Uganda.
H. M. M. Maraite, Unit de Phytopathologie, Universit Catholique de Louvain, Croix du Sud 2/3,
B-1348 Louvain-la-Neuve, Belgium.
L. Masumbuko, Tanzania Coffee Research Institute (TaCRI), Lyamungu, PO Box 3004, Moshi,
Tanzania. E-mail: www.tacri.org
S. Mdemu, Tanzania Coffee Research Institute (TaCRI), Lyamungu, PO Box 3004, Moshi, Tanzania.
A. Million, Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research ,PO
Box 192, Jimma, Ethiopia. E-mail: girma.adugna@yahoo.com
R. O. Musebe, CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya.
E-mail: r.musebe@cabi.org
P. C. Musoli, Coffee Research Centre, PO Box 185, Mukono, Uganda. E-mail: cori@africaonline.
co.ug
E. Negussie, CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya.
N. M. Nghoma, Tanzania Coffee Research Institute (TaCRI), Lyamungu, PO Box 3004, Moshi,
Tanzania. E-mail: www.tacri.org
vii
J. Njuki, CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya. Present
address: ILRI, PO Box 30709, Nairobi 00100, Kenya.
G. Oduor, CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya.
J. Ogwang, Coffee Research Centre, PO Box 185, Mukono, Uganda. E-mail: cori@africaonline.co.ug
N. Phiri, CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya. E-mail:
n.phiri@cabi.org.
F. Pinard, Centre de Coopration Internationale en Recherche Agronomique pour le Dveloppement
(CIRAD), TA41/K, Campus International Baillarguet, 34398 Montpellier, Cedex 5, France.
M. A. Rutherford, CAB International (CABI), Europe-UK, Bakeham Lane, Egham, Surrey, TW20
9TY, UK. E-mail: m.rutherford@cabi.org
N.T. Saiba, Office National du Caf, BP 8931, Kinshasa Kingabwa, DRC.
A. Shibru, Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research, PO Box
S. Simons, CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya. Present
address: Global Invasive Species Programme (GISP), UN Avenue, PO Box 633-00621, Nairobi, Kenya.
E-mail: s.simons@gisp.org
D. Teferi, Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research, PO Box
J.M. Teri, Tanzania Coffee Research Institute (TaCRI), Lyamungu, PO Box 3004, Moshi, Tanzania.
E-mail: www.tacri.org
D. Teshome, Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research ,PO
Box 192, Jimma, Ethiopia.
P. Tshilenge Djim, Universit de Kinshasa, BP 117, Kinshasa XI, DRC.
P. Wetala, Coffee Research Centre, PO Box 185, Mukono, Uganda. E-mail: cori@africaonline.co.ug
viii
Preface
This is story of an African coffee farmer in Uganda who I interviewed at the start of the project at
the end of 1999. This farmer had 1000 coffee trees on his 1-ha farm. He was earning US$1000 per
annum from approximately 1000 kg of green coffee he produced from his field. He had a family
of seven that included five children who ranged in age between 5 and 15 years. He did not
hire any outside labour to assist with crop harvesting and processing, and he supplemented his
income by intercropping his coffee field with bananas, which are stable food crop in the area.
In 1997/1998, when the war started in the Democratic Republic of Congo (DRC), there was
a lot of coffee trafficking through illegal boarder crossings from DRC to Uganda. The military
personnel operating in the area were bringing into Uganda ripe and unripe coffee cherries
for processing and sale or export from Uganda. During this time, a disease called coffee wilt
disease became evident to the coffee-growing public along the routes used by the military
personnel. Before this period, the disease had been identified in very limited areas and was a
minor threat to the coffee sector. After this period in the first year alone, the disease affected
more than 600 trees from the farmers field. In that year, the farmers production yield was
reduced by half, and by the second year, these 600 trees were dead and needed uprooting.
By 1999, the yield from the coffee farm had reduced from a peak of 1000 kg of green
beans to 400 kg, representing a 60% decline in income for the farmer. In absolute terms, this
represents a decline of income from US$1000 to US$400 per annum for the family of seven.
The Common Fund for Commodities has the farmer as its main client; therefore, we
have every reason to worry when we witness such disasters. This project was designed to try
to mitigate some of the problems by assisting the farmer described above through teaching the
farmer different methods of controlling the spread of the coffee wilt disease and introducing
to the farmer new disease-resistant varieties of coffee. The Common Fund has walked this
journey with farmers in DRC, Uganda, Tanzania and Ethiopia and stands witness to the
challenges faced by these farmers. I am happy to say that most farmers now know how to
identify the disease in its early stages and how to stop its rapid spread to other trees in the
field. If the farmers had the knowledge in the early 1990s, which they now have today, I am
confident that at least 50% of the coffee crops loses could have been avoided.
Just to put the effects of the coffee wilt disease into context in terms of the value of
coffee lost by the affected countries, you may wish to note that the affected countries export
over US$1 billion worth of coffee every year. Thus, a 30% crop loss due to coffee wilt disease
represents US$300 million lost income per year. In real terms, this amount represents 1 million
children not having access to education, health care and sometimes basic food for survival.
This project is a development lesson for policy makers who sometimes fail to appreciate the
challenges encountered by farmers in the agricultural business. As Project Manager, I would
like to sincerely thank all the coffee scientists, farmer group leaders and Extension Service staff
who guided the smooth implementation of this project.
Caleb Dengu
First Project Manager
Common Fund for Commodities
ix
Acknowledgements
I would like to thank all the contributors to this book- without their patience, dedication and
hard work, this book would not have been possible. I would also like to thank our many collaborators including scientists and technicians from national and international research institutes
and others who work in the coffee sector in the participating countries such as farmers and
extensionists who participated in and helped with, many of the activities that are reported in
the book. They have to live with the reality of CWD on a daily basis. Acknowledgement is also
due to the funders of the Regional Coffee Wilt Programme (RCWP) including the Common
Fund for Commodities (CFC), the EU through its INCO-DEV Programme as well as through
the Coffee Research Network (CORNET) of the Association for Strengthening Agricultural
Research in East and Central Africa (ASARECA) as well as the UK Department for International Development (DFID) through its Crop Protection Programme. I would particularly like
to thank Mr Caleb Dengu (CFC) for his support and encouragement from the original development of the RCWP through to its conclusion and the final outputs (this volume and the final
technical report). The ICO (International Coffee Organization) are also acknowledged as the
Supervisory Body for the CFC component.
Lastly I would like to thank my colleagues in CABI notably Dennis Rangi, Mike Bodinham,
Peter Baker, Noah Phiri and Sarah Cranney for their inputs and advice.
xi
Introduction
Julie Flood
CABI, Bakeham Lane, Egham, Surrey, TW20 9TY
During the 20th century, a devastating disease (coffee wilt disease [CWD])
spread across Africa reducing yields, destroying millions of coffee trees in affected countries and costing hundreds of millions of dollars in lost earnings
to farmers. The disease has also contributed to a decline in revenue for several
African nations due to reduced coffee production and will be a contributory
factor in any attempt at revitalization of the African coffee sector in the future.
CWD is of particular significance because unlike many other coffee diseases,
it kills trees and the farmer is often unable to replace his crop easily and faces
major short falls in his cash income. The disease has reached epidemic proportions twice during the last century, becoming a serious constraint during
the 1930s, 1940s and 1950s and was responsible for the complete failure of
excelsa coffee commercially. However, following the systematic introduction of sanitation methods (uprooting and burning) plus comprehensive
breeding programmes in many of the affected African countries, CWD was
considered a minor problem. Nevertheless, the disease re-emerged again on
robusta coffee in Central Africa with sporadic outbreaks observed during
the 1970s and building to another extensive epidemic during the 1980s and
1990s. CWD has become an increasingly important production constraint on
arabica in Ethiopia. In the intervening time between the two main epidemics, the African coffee sector had changed significantly. During the second
half of the 20th century, there has been a shift from being predominantly
plantation cultivation to being predominantly smallholder production, and
consequently, management of the disease required reappraisal. The Regional
Coffee Wilt Programme (RCWP) was a multi-country, multi-donor project
initiated at the start of the 21st century to examine the disease critically and
to identify suitable management options for smallholder farmers in order to
reduce the incidence of the disease, minimize coffee losses and lessen impact
on farmer income.
CWD, which is also called tracheomycosis, is a fungal disease. The causal
agent, Fusarium xylarioides (sexual form, Gibberella xylarioides), is a vascular
wilt pathogen. The fungus invades the coffee tree and colonizes the xylem
Flood
system (water-conducting elements of the tree). External symptoms exhibited

by coffee plants affected by CWD are generally similar to those affected with
other vascular wilt pathogens. Colonization of the vascular system induces
host responses, which disrupt water conduction, and this is manifested as
wilting and desiccation of leaves followed by defoliation and dieback of affected branches. Symptoms may appear at any stage of crop growth, and the
rate at which they develop varies. Once affected, death of the plant is inevitable, and in mature trees, it usually occurs between 3 and 15 months after the
first appearance of symptoms. Young plants, however, may be killed within
a matter of a few weeks of infection. Symptoms can be more pronounced
on one part of the tree, a likely consequence of initial infection occurring on
one of several main stems, but defoliation gradually extends to the entire
plant. Coffee berries that would normally be green may redden as if ripening
prematurely but often remain intact on shoots following defoliation. Other
external symptoms include the swelling of the trunk and the appearance of
vertical or spiralling cracks in the bark of mature trees. Small blackish-brown
perithecia (sexual stage) of the fungus, similar in appearance to dark soil particles, may be produced in the cracks of the bark. Characteristic bluish-black
staining of the wood can be observed directly beneath the bark.
The disease was first observed in 1927 in a plantation of coffee excelsa
near Bangui in the Central African Republic (CAR), which was then known
as Oubangui-Chari (Figueres, 1940). By 1945, the disease had destroyed most
of the countrys Coffea excelsa plantations (Saccas, 1951). Later, the disease
was found on other coffee species (Coffea canephora and Coffea neo-arnoldiana)
in parts of the CAR. It was similarly reported in several other countries in
West and Central Africa around the 1930s and 1940s. In 1939, Steyaert observed the disease on C. excelsa plantations in the Democratic Republic of
Congo (DRC; formerly known as Zaire and the Belgian Congo), close to the
border with Sudan at Aba (Steyaert, 1948; Fraselle, 1950). Initially, it caused
a few problems, but, as in CAR, it later reached epidemic proportions when
it spread to robusta coffee. Around the same time (19381945), the disease
also became established on excelsa in Cameroon, causing complete destruction of plantations in the east of that country. In 1947, it was discovered on
C. canephora in Cte dIvoire, and major losses occurred there in the 1950s.
More than 50% of the coffee-producing areas were destroyed in both Cte
dIvoire and in DRC, and the kouillou line of robusta was completely wiped
out in Cote dIvoire. By 1958, the disease was reported in Guinea, and it
spread quickly to most of the coffee areas, causing coffee production to fall
by nearly 50% (Chiarappa, 1969). In East Africa, in Ethiopia, symptoms similar to those of CWD were documented on Coffea arabica for the first time by
Stewart (1957) and by Lejeune (1958).
At an international conference held in 1956, recommendations were
made for a management programme to include (i) systematic sanitation over
vast areas where affected coffee plants were to be uprooted and destroyed;
(ii) where possible, relocation of coffee production to new locations; and
(iii) replanting with resistant germplasm. Implementation of these recommendations reduced the impact of CWD, and literature produced during the
Introduction
1970s and 1980s referred to this as a minor disease, of little importance for
arabica and robusta coffee production.
However, in remote areas of DRC, the disease continued to be observed
especially around abandoned plantations in the north-east of the country
(Chapter 2, this volume). As early as the 1970s, farmers in Aketi (76 km from
Isiro) had observed the disease in abandoned plantations, and during a survey conducted from 1974 to 1975, a number of Institut National pour ltude
et la Recherche Agronomique (INERA) fields around Yangambi were also
reported to be affected. Throughout the 1980s, reports persisted of a wiltlike disease affecting coffee around the town of Isiro, and surveys were conducted (Chapter 2, this volume). There is a considerable distance between
these areas, and this suggests that the disease was very widespread in that
region during this time. In 1995, CAB Internationals (CABI) plant clinic laboratory received samples of diseased coffee plants from the Managing Director of Esco Zaire sprl (Mr. Philip Betts), and F. xylarioides was isolated. These
samples had been collected from robusta coffee in Beni and Rutchuru, again
indicating that the disease was widespread in the region. In 1995, Office Zarois du Caf (OZACAF), now called Office National du Caf, prepared a
detailed report for the International Coffee Organisation (ICO) in which they
outlined the extent of the problem and its serious effects on the economy of
north-east DRC.
In March 1996, ICO facilitated contact between OZACAF and CABI because of CABIs experience with diseases of perennial crops including coffee.
This was the start of my own personal involvement with CWD. I was invited
by OZACAF to go to DRC and prepare an independent report on the nature
and extent of the problem. I travelled to the north-east of DRC in July 1996
and conducted surveys of plantations and smallholder farms as well as interviewed farmers about the disease, which they called coffee AIDS. Unlike
many other coffee diseases where the yield can be reduced (for example, if
the disease affects the berries) with CWD, once affected, the coffee trees died.
One agriculturist I interviewed told me that he had observed the disease first
in the 1970s on a large abandoned plantation on the road to Aketi, which
confirms other reports that the disease had re-emerged as a serious problem
in this part of Africa at that time just when most scientific authorities considered it as a minor problem.
The diagnostic symptom of CWD blue black staining below the bark of
the coffee bush was commonly seen, and the pathogen was isolated (Flood,
1996). The despair of coffee growers in this region was palpable. They were
angry, and they wanted action. They wanted to know how to manage the
problem. Unfortunately, the only recommendation then available was the
same as had been given in the 1950s sanitation, i.e. uprooting and burning coupled with long-term breeding programmes. Sanitation was being
conducted on plantations, but the disease was spreading faster than plantation workers could undertake the work. For smallholders, removal of very
old coffee bushes using hand tools was impossible; even where a plant had
been dead for a year or more, its root system was virtually impossible to
excavate.
Flood
I presented my report to the Executive Board of the ICO in September

1996 (Flood, 1996), and I was formally invited by Mr. William Naggaga (then
Secretary of the Ugandan Coffee Development Authority) to make a similar
fact-finding visit to Uganda because similar symptoms of a coffee disease
had been observed in some of the districts bordering DRC and in Mukono.
Several surveys had been undertaken by Ugandan scientists, and the pathogen had been identified (Chapter 3, this volume). It had also been noted in
Uganda that the disease was only affecting robusta coffee; arabica coffee was
not affected there, although the disease has continued to be a problem in
Ethiopia on C. arabica (Chapter 4, this volume).
Following presentation of my report on the Ugandan situation to the ICO
(Flood, 1997) and in recognizing the severity of the situation, an international
workshop was held in Kampala, Uganda (July 28th30th, 1997). A proposal
for a funded programme of work to alleviate the problem was developed
and submitted to several donor agencies including the Common Fund for
Commodities (CFC), the European Union (EU) and the UK Department of
International Development. Subsequently, the proposal had to be developed
as a fully integrated programme of activities addressing different aspects
of the disease and its management to facilitate funding of discrete activities
by the various agencies, in line with their priorities and timelines, and to
make tenable financial arrangements. The revisions resulted in the establishment of five interrelated and interdependent projects, each dealing with specific components of the disease. The collective name for these projects was
the RCWP, and its overall coordination (by CABI) was funded by the CFC.
Biological and socio-economic surveys were conducted in-country to ascertain the extent of the disease and its impact on livelihoods. These surveys
were funded by the EU through the Association for Strengthening Agricultural Research in Eastern and Central Africa/Coffee Research Network. The
RCWP also conducted research on remote sensing (funded by the EU), on
durable resistance to the pathogen in Africa (funded by EU-INCO-DEV), as
well examining the epidemiology and pathogen variation (funded by the
UK Department of International DevelopmentCrop Protection Programme
and EU-INCO-DEV). In addition, considerable attention was made to the
training of extensionists and farmers and to the dissemination of information
about CWD to facilitate surveillance and to limit spread. This component of
the RCWP was funded by CFC. For further details of the individual projects
including details of implementing agencies, partners, sponsors and results,
see the Final Technical Report of the RCWP (Phiri & Baker, 2009).
This book represents one output of the RCWP and, in addition to Phiri &
Baker (2009), represents a compilation of the current knowledge of various
aspects of CWD including information about the pathogen and its management. Although much has been achieved, considerable work remains. Over
the 13 years since I conducted surveys in DRC, the disease has spread much
more extensively within DRC (Chapter 2, this volume) and has been reported
in the province of Equateur, threatening coffee production in Western Africa.
It has continued to spread throughout all coffee-growing districts in Uganda
(Chapter 3, this volume) and into Tanzania (Chapter 5, this volume). The
Introduction
disease continues to be a problem of C. arabica in Ethiopia (Chapter 4, this

volume). Severe impacts on the incomes and livelihoods of coffee farmers in
affected areas of DRC, Uganda, Ethiopia and Tanzania have been observed,
and producers have had to diversify away from coffee production by either
growing other crops or engaging in non-agricultural activities (Chapter 6,
this volume). Early observations suggested that in areas where both coffee
species existed, only one species was affected, and artificial inoculation has
confirmed that there is host specificity of the strains from C. arabica and C.
canephora (Chapter 8, this volume). In addition, strains from arabica isolated
from Ethiopia are pathogenic to C. arabica germplasm from other coffeeproducing areas (Chapter 8, this volume). These two distinct populations
were also discernable using various molecular methods, although the overall
variability appears to be low despite the production of perithecia in nature
(Chapter 7, this volume). As part of the RCWP, participatory trials on managing the disease were undertaken in many locations throughout affected
countries (Chapter 9, this volume) so that farmers could search for ways of
reducing the impact of CWD. The effect of wounding coffee trees on increasing the incidence of the disease was quickly appreciated. These participatory
trials were also part of a need to raise awareness of the disease across the
region so producers and others, such as extensionists and researchers, are
able to quickly recognize its symptoms. If the disease spread to new areas,
then for eradication to have any chance of being effective, early detection
is critical. Thus, the RCWP undertook extensive training of farmers and of
extensionists (as trainers) in participating countries using the Farmer Field
School approach, which was adapted for coffee (Chapter 11, this volume).
The Farmer Field School approach plus field days and workshops proved
to be effective tools for facilitating farmers learning and in enhancing their
critical thinking and problem-solving skills. Mass communication methods
such as print materials and use of radio and television were also used and
discussed (Chapter 11, this volume). Farmers in affected countries are now
much more familiar with the disease and its management. Breeding for resistance remains the best option for management of this disease (and was
successful in earlier epidemics in the 20th century). Programmes for screening for resistance have been initiated in participating countries, and resistant
material has been identified. In Uganda, several resistant genotypes have
been selected and are currently being multiplied for distribution to farmers
(Chapter 10, this volume).
References
Chiarappa, L. (1969) International assistance in plant pathology in developing countries. With particular reference to FAO programmes. FAO Plant Protection Bulletin
17 78.
Figueres, R. (1940) Sur une maladie trs grave du cafier en Oubangui. Rapport. Ministre des Colonies, Paris, France.
Flood, J. (1996) A study of tracheomycosis or vascular wilt disease of coffee in Zaire.
Report presented to Zairean Coffee Organization (OZACAF). August 1996. 13 pp.
Flood
Flood, J. (1997) Tracheomycosis or vascular wilt disease of coffee in Uganda. Report
presented to Ugandan Coffee Development Authority (UCDA), Uganda. 12 pp.
Fraselle, J. (1950) Observations prliminaires sur une trachomycose de Coffea robusta.
Bulletin Agricole du Congo Belge XLI, 361372.
Lejeune, J.B.H. (1958) Rapport au Gouvernement Imperial dEthiopie sur la production
cafire. Rapport du la FAO, Rome FAO158/3/1881.
Phiri, N. and Baker, P.S. (2009) Coffee wilt disease in Africa; a synthesis of the work
for the Regional Coffee Wilt Programme (RCWP). Final Technical Report. CABI.
200 pp.
Saccas, A.M. (1951) La trachomycose (carbunculariose) des Coffea excelsa, neoarnoldiana et robusta en Oubangui-Chari. Agronomie Tropicale 6, 453506.
Stewart, R.B. (1957) Some diseases occurring in Kaffa Province, Ethiopia. In: Imperial Ethiopian College of Agriculture and Mechanical Arts. Alemaya, Ethiopia,
pp. 1516.
Steyaert, R.L. (1948) Contribution ltude des parasites des vgtaux du Congo Belge.
Bulletin de la Societe Royale de Botanique de Belgique 80, 1158.
2
1
Coffee Wilt Disease in Democratic

Republic of Congo
A. Kalonji-Mbuyi,1,2 P. Tshilenge Djim1 and N.T. Saiba3
Universit de Kinshasa, BP117, Kinshasa XI, Democratic Republic of the Congo

Centre Rgional dEtudes Nuclaires de Kinshasa (CREN-K), BP 868, Kinshasa XI,
Democratic Republic of the Congo
e-mail: adrienkalonji@yahoo.fr
3
Office National du Caf, BP 8931, Kinshasa, Democratic Republic of the Congo
2.1. Introduction
Coffee is one of the key cash crops in the Democratic Republic of Congo
(DRC). Several Coffea species are found across the country. The first coffee
plantations grown in DRC were from spontaneous species identified in some
provinces. The main exotic species include Coffea abeokuta, Coffea excelsa, Coffea myrtifolia, Coffea stenophylla and some varieties of Coffea arabica imported
from Belgian Congo (now DRC) by the colonial garden (Kinds, 1930). Coffea
liberica from Liberia was the first introduced Coffea species in DRC in 1881
(Jagoret and Decroix, 2002). However, commercial coffee production relies
mainly on two species, C. arabica L. (70%) and Coffea canephora Pierre (30%).
The main development of the coffee crop in DRC took place between 1924
and 1930. This was made possible because of an important collection of local
C. canephora accessions assembled mostly from prospection in the forests of
the Yangambi and Lula regions, as well as in other parts of Congolese and
foreign zones.
Discovered in 1898 in what is now DRC, Coffea robusta is the most commonly grown variety of C. canephora, accounting for up to 95% of C. canephora
populations worldwide (Miny, 1930). Unlike C. arabica plants, C. robusta does
not need to grow at high altitude and requires less care to grow because
it is hardier and tends to be less susceptible to pests. Its area of distribution is variable and corresponds to hot and humid climatic regions. It is
found in low- and middle-altitude areas in Africa (Ivory Coast, DRC, Cameroon, Uganda, Angola, Ghana, Togo, Madagascar, Tanzania and Republic of
Kalonji-Mbuyi et al.
Central Africa), in the Far East (India, Indonesia, Philippines, etc.) and in
Oceania, for example, New Caledonia.
Most of the robusta coffee plantations in DRC derive from progenies of
a few elite local clones from the Yangambi Research Stations, and one introduced clone (SA 34) from Indonesia, that were previously selected for yield.
No specific breeding programmes were developed to improve the agronomic
traits of genotypes, but coffee selection based on yields and morphological
traits was carried on by the Institut National dEtudes et de Recherches
Agronomiques (INERA). Six elite local clones including L-147, L-36, L-251,
L-215, L-93 and L-48 were selected from monoclonal isolate plots (Drachoussof et al., 1956). This standard mixing continues to be highly productive and
provides financial sustainability to coffee producers.
Robusta coffee is of paramount economic importance in DRC. It represents almost 90% of all the plantations across the country and is DRCs
biggest cash crop. It is with maize and rice as the top leading crops. Robusta
coffee is the third exported product in DRC after copper and cobalt, and it
generates nearly 9% of the total revenue from exportation for the country.
This represents almost 60% to 65% of the total value of exported agricultural
products (Banque Centrale du Congo, 2005). Coffee is produced at village and
agro-industrial levels (Petites et Moyennes Entreprises Agricoles). At village
level, coffee generates for local producers considerable cash flow estimated at
2 billion Congolese franks (Office National du Caf [ONC]). This represents
on average nearly half of the income generated from coffee exportation zones
(National Agricultural Statistic Service, 2006, unpublished report).
2.2. Importance of Coffee in DRC

Coffee is an important commodity in the economy of DRC. The coffee species
grown in DRC are C. arabica, C. canephora var. robusta and C. canephora var.
kouillou. The spatial distribution of coffee farming is dependent on ecological
conditions that are favourable for their cultivation. The current total area under coffee covers nearly 193,433 ha. The estimated area under robusta, including the Kouillou, is about 164,533 ha (85. 1%), and arabica coffee covers 28,900
ha (14.9%). Most coffee farmers are smallholders, which represent around 86%
of the total production with average fields per farmer of 1.3 ha for robusta and
0.8 ha for arabica. From the agricultural census, it has been estimated that a
third of the coffee plantations are now abandoned.
Production of coffee has been fluctuating for many years. For the 2006
2007 season, production was estimated at 34,553 t, with 27,007 t for robusta
and 7546 t for arabica (ONC, 2007, unpublished data). The national average coffee production varies between 150 and 200 kg/ha for robusta and
between 150 and 300 kg/ha for arabica. This is produced on 193,433 ha of
land, so the productivity is low. At the time of its independence in 1960, the
country was the leading African coffee exporter, with a production of 52,000 t
of coffee annually. Coffee production subsequently increased, reaching a
40000
35000
30000
Tonnes
25000
20000
15000
North Kivu
10000
Orientale
Equateur
5000
0
1997
1998
1999
2000 2001
2002 2003
2004 2005
2006
Year
Fig. 2.1.
Annual coffee production (var. robusta) in the infected provinces by CWD in DRC.
record level of 130,000 t from 1989 to 1990, comprising 25,600 t of arabica

and 104,400 t of robusta, including Kouillou (Conjoncture Economique, 1991;
ONC, 1990, unpublished data).
The data of the average robusta coffee production in Equateur, North
Kivu and Orientale province areas for the period 19972006 are summarized
in Fig. 2.1.
Coffee production decreased over the last decade, although there was
some increase in production in 2003 and 2004. Production then decreased
again rapidly, presumably due to the continuing impact of the coffee wilt disease (CWD).
2.3. Constraints to Coffee Production

The coffee sector in DRC faces a number of challenges of diverse nature. Notable challenges include the degeneration of the existing planting material,
the aging of the coffee gardens, the poor management of plantations, the
lack of skills and knowledge of coffee production by smallholder producers
and an inadequate, poor and dilapidated road and port infrastructure. In addition to these conditions, the potential of the sector has been progressively
reduced due to the following: the effects of nationalization of the coffee production sector in 1974; the continuing low farm gate price of coffee, which is a
worldwide phenomenon; the drift of people from the land into the cities; and
10
for those that remain on the land and try to farm, they are faced by a lack of
credit.
2.4. Most Important Diseases and Pests

Although socioeconomic and political conditions have changed, environmental conditions (climate, soil, etc.) remain very favourable for coffee production. However, the absence of any plant health surveys as a surveillance
mechanism for alerting the authorities to the presence and subsequent upsurges in plant diseases and the absence of management methods have favoured the resurgence of numerous diseases in the Congolese coffee culture;
some of which have reached epidemic proportions. Among these diseases
is CWD or tracheomycosis caused by Gibberella (Fusarium) xylarioides Heim
& Saccas. When surveys were subsequently undertaken, the presence of anthracnose (Colletotrichum sp.) and a root rot caused by Armillaria mellea mainly
in Bas-Congo province in the Bas-Fleuve district (A. Kalonji-Mbuyi and P.M.L.
Onyembe, 1996, unpublished data) were also detected. The anthracnose induced large brown spots on the leaves with or sometimes without a yellow
halo. Leaves fall, and the young branches wither, causing the whole plant
to dieback. In some instances, the berries are also attacked. This defoliation
was observed at all stages, even at the nursery. The root rot was frequently
observed in the Seke-Banza zone in the Bas-Congo province. The symptoms
are essentially the sudden and general wilting of the aerial part of the plant
immediately followed by plant desiccation. This disease presents very similar
external symptoms to those associated with CWD. The diagnostic root rot
symptom is the presence of splits on the lateral roots that show the fungal mycelial net characteristic of Armillaria. When the root rot becomes very severe,
the plant falls down if shaken or pushed.
2.5. A Brief History of CWD in DRC

CWD was first observed in 1939 by Steyaert on C. excelsa materials collected
in plantations located at the borders between DRC and Sudan (most precisely
at Aba) and others from Bangui in the Central African Republic (Fraselle,
1950).
In the beginning, the disease was minor with limited economic impact.
However, around 1949, an epidemic was reported on thousands of robustatype plants in Yangambi plantations (Heim and Saccas, 1950). The disease
was later observed in Haut-Ul, North Kasai and Katanga. Its massive
reappearance during the following year illustrated its danger not only because of the considerable damage it causes but also because of its epidemic
character (Fraselle, 1950). Initial CWD symptoms seen included a generalized chlorosis of the leaves, which become flaccid and curl inwards, but on
other occasions, the leaves do not become yellow but wilted while remaining green. The infected tree quickly becomes defoliated. Prior to their fall,
11
the leaves often darken, especially those on the seedlings and the nursery
stage.
A number of measures to contain the disease were initiated in Yangambi
City and Yangambi station, including uprooting and burning of infected coffee trees on the spot. These practices were followed, and the disease was successfully controlled. CWD was subsequently forgotten and lost its economic
importance because of the development of resistant robusta varieties (Fraselle and Geortay, 1952). Successful adoption of sanitation practices and resistant varieties led to the decline of incidence of CWD from the late 1950s.
In the late 1970s, a farmer in the north-east of DRC observed a wilt-like
disease of coffee in abandoned plantations around Aketi about 76 km from
Isiro (Pochet, 1986, personal communication). Later, the disease was observed on coffee plants during a survey conducted from 1974 to 1975 in a
number of INERA fields at Yangambi (A. Kalonji-Mbuyi, 1975, unpublished
data). Studies of the samples collected by Kalonji-Mbuyi recovered a strain
of F. xylarioides from infected plant branches, stems and roots. The identity
of the fungus was later confirmed by Professor Felix Pierre-Louis to be F.
xylarioides. Because the disease was not observed in the surrounding farmers
plantations, it has been supposed that it was due to the heterogeneity of the
material used by farmers, some of which may be resistant to this fungus,
while INERAs genetically uniform and homogenous plantations were considered to be more likely to be affected.
Throughout the years 1980 to 1981, national reports indicated the presence
of some source of CWD around the town of Isiro in the region of Haut-Ul,
Orientale province (ONC, 1982, unpublished report). These reports indicated
that the disease attacked only robusta coffee plantations and that both young
and old established plantations were affected. In addition, the disease occurred
both in plantations planted on rich soil and in localities with less-fertile soils.
The disease became widespread in all the Haut-Ul plantations, progressively spreading to all the surrounding territories. More plantations became abandoned, and smallholders became desperate as they saw their only
source of income disappear. The existence of a very marked and decreasing
gradient of infection as one moved away from the triangle formed by Isiro
territories, Wamba and Mungbere in the valley of Nepoko led to the conclusion that this area is the primary source of the infection.
Between 1985 and 1989, ONC initiated surveys at Ituri, Bas Uel and
Haut-Ul districts with the objective of obtaining information on the disease and of the environmental factors that influence the disease. Disease surveys revealed that the disease was also widespread in Orientale province.
According to estimates determined in 1987, its incidence was 19.3% of these
coffee areas (M.K. Katenga, 1987, unpublished data). Two years later (1989),
it reached 30.6% mainly in the district of Haut-Ul (A. Kalonji-Mbuyi, K.
Mukuna, R. Masozera and N. Isungu, 1990, unpublished data ).
The progression of CWD from Haut-Ul has been very irregular.
The disease continued to spread down through a corridor of robusta coffee, which stretches from Mambassa to Irumu in the Ituri district (Fig. 2.2).
The incidence of infection was reported then as being around 30% in 1992
12
Fig. 2.2. Primary source of CWD Infection in DRC (Isiro-Wamba-Mungbere) and progress to
North Kivu.
13
(Kalonji-Mbuyi et al., 1990; Pronalutra, 1992, unpublished report). The disease then spread towards the North Kivu province, and over the following 3
to 4 years, it was consistently observed in that area from 1995; its appearance
was always accompanied with considerable damage. The incidence of infection in 1995 was estimated to range from 13% to 30% of plants in plantations
located in Mavivi, Mbawu and Mutwanga. However, the disease remained
less serious in the surrounding areas of Butembo-Muhangi and Musienene
in the Lubero district of Lubero and did not exceed 3%.
In 1996, ONC, then known as OZACAF, invited Dr Julie Flood (CABI)
to assess the epidemical situation in eastern DRC, and she confirmed the
presence of CWD caused by F. xylarioides (Flood, 1996). In 1997, assessments
indicated that the incidence exceeded 50% in the coffee in the Haut-Ul district (Tshilenge-Djim et al., 1998).
Currently, CWD is a serious constraint that has caused considerable
damage to coffee production in DRC, and its effects compromise the very
existence of coffee growers. The disease is very widespread in the country,
and its spectacular progress has significantly disrupted production levels of
robusta. In its epidemic phase, the disease has manifested itself as destroying many of the coffee plantations in the greater production zone of DRC. In
certain places, it appears to be endemic with little infection sources, too scattered and difficult to detect by the planters. Nevertheless, these small groups
(disease foci) may represent a future threat to any future coffee material.
2.6. CWD Distribution in DRC

In DRC, CWD attacks robusta coffee and other coffee species such as C. liberica (C. liberica var. liberica, C. liberica var abeokutae, C. liberica var. dewverii) in
the Botanical Garden collections at Yangambi station (INERA). Several visits
to the arabica coffee area showed that this species is not affected.
A biological survey was carried out between 1991 and 2004 in different plantations across the seven coffee-producing provinces in DRC: BasCongo, Bandundu, Equateur, Kasai Occidental, Kasai Oriental, Nord-Kivu
and Orientale province. The general and specific objective of the survey was
to determine disease incidence on the coffee trees in the plantations within
the production area. The survey involved controlled sampling of coffee plantations to collect information on plant health, state of the coffee trees, the
presence or absence of CWD, and in the case of the disease being present, to
determine the incidence, severity and relationships with the ecoclimatic and
environmental conditions so as to contribute to the design of control strategies adapted to the various conditions. The methodology used to collect
data was based on the identification of infected plants in the plantations by
taking into account the external symptom expression (yellowing, drying of
the crown) and the internal characteristics (presence of the blueblack bands
on the level of the collar of the plant or in the aerial parts). Principal sampling of the inspected plantations was undertaken from the plantations chosen randomly in groupings, traditional entities gathering several villages.
14
Fig. 2.3.
The coffee wilt disease distribution in DRC.
15
Large commercial plantations were rare, and those still existing were subdivided into several groups, in which three to six were chosen at random for
surveying.
The ONC coordinated the survey, and scientists who took part included
plant pathologists from the University of Kinshasa and from the INERA;
staff from CABI Africa were the enumerators. The team members visited the
farms selected, with farmers doing the data collection. In each farm visited,
30 trees in a randomly selected diagonal or median direction were scored.
Biophysical factors such as production system, age of the coffee, altitude,
topography and information about the cultural practices (shade, fertilizer
use, intercropping, mulching, etc.) were also collected to assess their possible
influence on CWD incidence and severity.
The surveys revealed that CWD is present in North Kivu, Orientale and
Equateur provinces (Fig. 2.3). The disease was not observed in the provinces
of Bandundu, Bas Congo, Kasai Occidental, Kasai Oriental and South Kivu.
2.7. Importance of CWD in DRC

Results obtained in the infected zones for the periods 19911996 and 20022006
showed that a lack of an adequate plant health service is a major contributor
to the establishment of CWD in epidemic proportions in these areas. CWD
has increased its frequency, forming epidemic levels because the symptoms
on a few isolated diseased plants were not recognized rapidly enough, and
even when diagnosed, the management techniques were not implemented
quickly enough.
The data obtained (Fig. 2.4) varied from site to site, although similar rates
of infection were observed. For all the observations throughout the period
19911996, the highest values in all the sites were registered in 1996.
In Orientale province, Isiro had 90% of the plantations affected, Mambassa had 36%, Poko had 34%, Bafwasende had 28%, Opala had 29% and
Banalia had 27%. In this period, the Yangambi station was disease-free. For
the same period, in the province of North Kivu, the disease incidence varied
according to sites, with 46% of the plantations in Mangina affected, 41% at
Ocha, 38% at Muhangi and 37% at Mutwanga.
Two distinct periods are observed when considering CWD evolution.
The evolution of the disease between 1994 and 2002 has followed the normal
epidemic disease progress expanding to reach the maximum values for all
the investigated sites. The observations carried out in 2004 revealed a revival
of the infection cycle for all sites, and this revival was, in fact, faster, for example, 73% infection was observed at Ocha. Such type of epidemiologic evolution confirms the polycyclic feature of tracheomycosis on coffee in DRC.
When ground surveys could not be conducted due to periods of political instability (19962006), a simulation of the disease progression was
made by applying logistical models so as to forecast the disease situation.
Comparisons between the disease incidence forecasted with the actual observed values in 1995, 2002 and 2004 indicated some deviations in the actual
16
3000E
Yandongi
2004: 2,9
Poko
2004: 2,6
Mondjamboli
2002: 3,7
Buta
Yandongi
Yamongili
Mondjamboli
Yandombo
Yangambi
1991: 0
1992: 0
1993: 0
1993: 0
1994: 0
1995: 0
1996: 0
2002: 81,8
2004: 100
Yandombo
2004: 26,6
Banalia
Yangambi
000
Equateur
Banalia
1991: 0
1992: 2,7
1993: 1,1,2
1994: 1,3,4
1995: 23,7
1996: 26,9
2002: 91,4
2004: 7,4
Opala
1991: 0
1992: 2,8
1993: 12,8
1194: 17,6
1995: 27,8
1996: 28,9
2002: 96
2004: 13,3
Bafwasende
Bafwasende
1991: 0
1992: 3,7
1993: 10,6
1994: 15,2
1995: 25,3
1996: 27,8
2002: 100
2004: 5,2
Opala
Isiro
Isiro
1991: 37,7
1992: 40,3
1993: 43,4
1994: 47,2
1995: 50,6
1996: 90,6
2002: 100
2004: 3,5
Mambasa
1991: 36,7
1992: 18,5
1993: 20,5
1994: 22,3
1995: 33,6
1996: 36,7
2002: 80,6
2004: 24,5
500N
Oicha
1991: 14,8
1992: 16,9
1993: 22,9
1994: 29,7
1995: 36,3
1996: 41,3
2002: 32,3
2004: 73
Mangina
1919: 13
1992: 15,2
1993: 20,6
1994: 25,2
1995: 30,6
1996: 46,2
2002: 88,8
2004: 53,7
Mambasa
Mangina
Mutwanga
Nord-Kivu
Muhangi
Muhangi
1991: 3,7
1992: 4,6
1993: 10,6
1995: 9
1996: 38,2
2002: 100
2004: 64,5
Mutwanga
1991: 11,2
1992: 13,5
1993: 18,3
1994: 20,7
1995: 24,9
1996: 36,5
2002: 93,3
2004: 53
Maniema
No CWD
Sud-Kivu
Kasai
Oriental
CWD obseved
No survey
Kasai
Occidental
Data source:
Proesseur Adrien KALONJI MBUYI,
Universit de Kinshasa
Datum WGS 1984
Systme des coordonnes gographiques
2500 E
Fig. 2.4.
000
Province-Orientale
SOUDAN
Poko
1991: 21,1
1992: 26,1
1993: 27,4
1994: 27,9
1995: 28,6
1996: 33,6
2002: 80,6
2004: 26,6
OUGANDA
500N
2500 E
50
100
Km
3000E
Evolution of CWD in the infected zone between 19912004
observations compared with the forecasts at some sites. Those deviations

are due to various reasons according to the sites. Sites such as Bafwasende,
Opala and Banalia (Fig. 2.5ad), which are located in Central Basin in the
Orientale province, were characterized by low altitude (less than 400 m) and
heavy annual rains (superior to 188 mm), and there, the observed and simulated evolution of CWD are similar. However, in Mambassa, the logistical
model did not accurately predict the situation of sigmoid development of the
CWD epidemic when we consider data for the 19951996 period. This was
probably because of plantation abandonment.
At Isiro (Fig. 2.5e), the development of CWD shows a difference between
the values observed in 1996 and the forecast, with the model reproducing
a disease level for the site, which was already observed around 1983 (M.K.
Katenga, 1987, unpublished data).
In the province of North Kivu (Fig. 2.6ad), there are two situations. At
Muhangi, the development of CWD was very different between observed
and simulated values. It is worth noting for this site that the farmer gave up
maintaining plantations of robusta coffee in 1996 and adopted arabica coffee.
The Muhangi site is located in a high altitude of more than 1300 m, which is
favourable for arabica culture. The development of CWD, as revealed by the
model, suggested that with routine maintenance practices and uprooting of
diseased plants, the source of inoculum for the subsequent infection would
have been around 10.5% in 1996 (according to the model), but 38.2% infection
17
60
40
0
1990
2000
1980
1990
2000
Year
Year
Opala
Mambasa
60
(d)
20
20
40
(c)
40
% infected plant
60
80
80
1980
% infected plant
(b)
20
% infected plant
60
20
40
(a)
% infected plant
80
Bafwasende
80
Banalia
1980
1990
2000
1980
Year
1990
Year
2000
60
20
40
(e)
% infected plant
80
Isiro
1980
1990
Year
2000
Fig. 2.5. Evolution of CWD in DRC by applying logistical models: (a) Banalia, (b) Bafwasende,
(c) Opala, (d) Mambada and (e) Isiro.
18
80
60
40
% infected plant
1990
2000
1980
1990
Year
Year
Muhangi
Oicha
2000
60
40
% infected plant
(d)
20
40
(c)
20
60
80
80
1980
% infected plant
(b)
20
40
(a)
20
60
80
Mutwanga
% infected plant
Mangina
1980
1990
Year
2000
1980
1990
2000
Year
Fig. 2.6. Evolution of CWD in DRC by applying logistical models: (a) Mangina, (b) Mutwanga,
(c) Muhangi and (d) Oicha.
was actually observed. The sites of Ocha, Mutwanga and Mangina demonstrated similarities between the model and observed values with slight gaps
in data for the period 19951996. Any large deviations are attributable to the
abandonment of plantations due to civil wars in those regions.
From the simulation studies, it seems that if all the factors required to
stop the development of CWD were maintained in all the sites throughout
the period (19911996), then the development of the disease would have
reached maximal level in the years 1995 to 1996. However, in reality, in most
of the cases, after 1996, CWD reached a peak in north-east DRC, which corresponded to its maximal level, and a new disease cycle started on young
replantings in 2004.
The abandonment of almost all the plantations due to civil wars that
have occurred in that part of the republic seems to have favoured the de-
19
velopment of the disease. The case of Yangambi might serve as an example,

where CWD was almost absent until 1996 and its sudden reappearance in
2002 with an increased severity (81.8%) and incidence (100%).
Presently, CWD has infected plantations in the province of Equateur
through the district of Mongala on the border with Orientale province, and
it has progressed into the province to where the latest observations (Kalonji,
2007, unpublished data) indicate its presence in Gemena, a town located in
the north-west.
2.8. Factors Affecting Severity of CWD in DRC

There are many factors that influence the incidence and severity of CWD in
DRC.
2.8.1. Agronomic and environmental factors

Analysis of some agronomical and environmental conditions has shown that
CWD is a disease that manifests itself in many diversified environments including valleys, forest zones and high and low locations. The effects of human
influence through agronomic practices vary from province to province. The
age of plantation varies between 5 and 15 years in North Kivu (88%), whereas
in Orientale and Equateur provinces, most of the plantations are 15 years old
(75% and 85%, respectively). Age is a factor connected to the presence of CWD
in the three provinces where this disease exists. The presence of CWD on all
plantations of over 15 years in Bas-Congo (81.6%) and in Bandundu (64.6%)
shows that age could be a major risk factor, and the probability of infection is
increased each time the coffee trees are pruned. Fraselle and Geortay (1952)
reported that the period of the first production would be a favourable circumstance for the extensive development of the disease, although it can also affect
the young plantlets in the nursery.
The advanced age of the plantations necessitates replanting to improve
the productivity. Most farmers use locally produced seedlings to plant new
fields. Ninety percent of farmers in Bas-Congo, 80% in Bandundu, 82% in Kasai Occidental and 53% in Kasai Oriental use this practice. In North Kivu and
Orientale province, the local germplasm is the most frequently used (89.8%
and 65%, respectively). Generally, the use of seedling can have an important
influence on the development of CWD. Seedlings normally have a high level
of heterogeneity, and such diverse germplasm could include those genotypes
that have some resistance to the disease. However, the local material used in
the affected provinces could have a restricted genetic base due to selections
being constantly taken from the farmers own fields or plantations.
Furthermore, in the three provinces where CWD is present, the production
system is mainly of the large-scale plantation system (78.5% in Orientale province and 59.5% in North Kivu) followed by the semiplantation system (54.6%
in Equateur province). These two production systems favour the disease.
20
Weed management is another factor affecting CWD; most of the plantations do not have weed management (81.4% in Equateur and 44% in Orientale). This could have an inverse relationship with the development of
the disease, because in the provinces where CWD is absent, such as Kasai
Oriental, Kasai Occidental and Bas-Congo, this agronomic practice is very
frequent (54%, 44.3% and 43.6%, respectively).
With regard to the influence of pruning on the incidence of CWD, the
results are contradictory. For instance, in Orientale province, pruning, which
is practiced by 80% of the farmers, would have favoured the spread of the
disease, whereas in North Kivu, where 84% of the farmers do not practice
pruning, the incidence of CWD is higher (mean of 90% and 94% of infection,
respectively, for Orientale province and North Kivu). Similar results have
been reported by Fraselle and Geortay (1952), who indicated that the pruning practice does not appear to significantly influence disease incidence. The
authors however have observed that at Yangambi, where coffee bushes have
two or three stems (multicauline pruning system), an accelerated spread
of CWD was observed. This might be due to the many wounds created by
this pruning system, which enhance the pathogen penetration of the coffee
tree. The practice can be safe if applied using sharp and disinfected tools.
Other agronomic practices such as shade, fertilizer use and mulching were
similar in all the surveyed regions, so any differential effects on CWD were
not apparent.
It is not obvious to find a link between intercropping and CWD. Intercropping is more frequent in North Kivu (71.4%), in Kasai Occidental (53.3%)
and in Orientale province (52.5%), whereas it is less frequent in Kasai Oriental (27.8%), in Bas-Congo (20.4%) and in Bandundu (12.9%). Other factors
such as environmental conditions (valley, forest zones, hilltop and flat landscape) do not seem to have a particular influence on CWD because of the
presence of the disease in these various conditions.
2.9. Country-Specific Management Practices in DRC

2.9.1. Cultural practices and chemical control
The main strategy used in the management of CWD in DRC involved the systematic elimination of all the infected coffee trees once the most typical external symptoms of CWD (yellowing, browning unilateral or general of young
shoots, fading and wilting) are observed. If conducted judiciously, then the
method has the advantage of reducing pathogen inoculum at the primary
centre of infection, thus preventing the spread to neighbouring healthy trees.
For the method to be effective against CWD, the following observations and
operations are recommended (Fraselle et al., 1953):
For each field, a detailed plan locating each coffee tree is required.
Regular monitoring is required to detect any coffee bush showing even
partial dieback symptoms on the aerial parts of the plant; all doubtful
21
plants are clearly identified using visible signs (string, color material or
any other marking).
After identifying a plant suspected to be infected by CWD, it is advisable
to work in group to confirm whether or not that coffee tree is actually
infected.
The disease can be confirmed if the superficial bark is removed on all
sides of the stem, and blueblack bands are seen below the bark; this is
diagnostic of CWD.
If the suspect tree does not have these characteristic blueblack staining
below the bark, then the wounds need to be disinfected and the knife
used treated with a chemical or heat treated.
If the suspect tree is infected with CWD, then the plant needs to be clearly
marked to identify its location. A spray of 10% carbolineum preparation
in water solution (about 2 l of solution) should be applied to the aerial
parts of the infected coffee bush.
The infected plants should then be dug out (including the root system).
The tree should be cut into sections and incinerated in situ. If the wood is
dry, then this can be done immediately, but if the wood is still green, then
a few days may be needed to allow the wood to dry to facilitate burning.
Trials undertaken in situ where the plantations are characterized by the absence of CWD symptoms, but close to infected plantations, allowed an assessment of the impact of some agronomic practices and management methods
on both the expression and the development of the disease. Protected plants
were left to natural contamination, and the assessments consisted of recording the numbers of infected plants on a monthly basis.
The agronomic practices tested consisted of clean weeding, slashing and
pruning. The clean weeding was obtained by spraying herbicides glyphosate at 0.1%. The spraying was done in such a way that the herbicide did not
touch the coffee tree, that is, by keeping a distance of about 30 cm from the
tree. Slashing, which is a common practice of weeding, was made by using
the machete. Pruning by means of shears or saw was undertaken. Chemical
control was undertaken by monthly spraying of Cupravit (copper oxychloride), a fungicide, at 0.5% (wt/vol) on the base of the coffee trunk and by
painting the stem with the same fungicide (0.3% [wt/vol] solution) applied
to the trunk at a height of 0.5 m from the soil.
Monthly spraying of fungicide was more effective in reducing the attack of CWD than other treatments (Fig. 2.7). This was followed by slashing
and clean weeding, which also seem to significantly reduce CWD incidence.
Treatment using stem painting and pruning did not seem to be so effective.
The high prevalence of infected trees, scored in the case of the plots under
pruning (6%) and the control plots (7%), demonstrates the danger that might
be associated with wounding coffee trees during routine maintenance, and
the lack of maintenance or phytosanitary practices in the abandoned plantations is also a key cause of the rapid spread of CWD.
The foregoing results obtained in DRC show that besides fungicide spraying, agronomic practices that are aimed at maintaining a clean plantation
22
9
Clean weeding
% infected plant
6.5
5.75
6
5
Pruning
4
4
3
Slashing
2.5
Fungicide spraying
2.25
2
1
0
cultural practices and control methods
Stem paint fungicide
Control
Fig. 2.7. Incidence of CWD in the robusta coffee recorded after 18 months of application of
different cultural practices and control methods.
such as clean weeding and slashing would, if conducted efficiently, significantly contribute to reducing CWD spread. Alternatively, any practice that
involves wounding, such as pruning, is likely to increase the rate of spread
of the disease. Phiri (2005, personal communication) reported the results registered in Ethiopia, Uganda and Tanzania, where the same study was conducted at the same period as in DRC, and concluded that the prevention of
any wound contributes to reducing and slowing down the spread of CWD in
plantations. However, the use of chemical treatment cannot be recommended
for our farmers fields, given the cost and the danger that it represents for
humans and the environment. Agronomic practices such as slashing, if conducted appropriately, for example, by avoiding wounds on coffee trees or
ensuring that plantations are well maintained, will remain the main method
for the management of CWD in DRC and can be easily adopted by farmers.
2.9.2. Availability of the improved material for replanting

The current situation in DRC with respect to CWD requires rehabilitation of
existing production areas and the extension of these production areas using
planting material of desirable characteristics. Currently, in DRC, when starting new plantations, farmers seldom use seeds from the research centres that
have guaranteed characteristics for good production. Generally, the vegetative material used is from cuttings, and often, this material performs poorly.
However, the elite clones selected at Yangambi, which include seven clones of
robusta (L 36, L 48, L 93, L 147, L 215, L 251, SA 158) and five clones of small
Kouillou (LAF 93, LAF 159, S 9, S 19, S 23), are available, and those can be
used preferentially by the farmers.
One of the priorities of the DRC government, as described in the Strategic Action Plan, is the revival of the coffee production by the systematic
23
replanting of the old plantations. The government proposes to rehabilitate all

the centres of production and the multiplication of the elite clones of INERA.
The plan also includes the reorganization of the supply chains of quality coffee material by the creation of multiplication centres and the dissemination
of improved material in each of the four principal zones selected above. This
will require the involvement of the Ministry of Agriculture (technical and financial support) and the ONC as partner already on the ground (Beni/NordKivu and soon in Boma and Isiro).
The ONC and the farmer trade union (Syndicat pour la Dfense des Intrts Paysans) have set up seven centres for the multiplication of rooted cuttings
at Beni, Ocha, Irango, Bingo, Mutwanga, Vuyinga and Mayi Moya in North
Kivu. The Environment and Natural Resource Accounting Company has
recently acquired an in vitro propagation unit to enhance the multiplication.
2.9.3. Use of variety resistance

The production of disease-resistant coffee plants is the most desirable method
of achieving control of pathogens. It appears that the coffee varieties being
grown currently are very susceptible to CWD. The revival of coffee research
is necessary to increase productivity and offer coffee growers the option of
using high-producing vegetative material. The effort has to be undertaken
to obtain new CWD-resistant clones. The research task will have to be focused mainly on determining the behaviour and the level of resistance of the
robusta clones (INERA elite clones and other materials introduced) towards
CWD, as well as on trying to investigate the adaptation and introduction of
highly productive material from foreign origins, such as clones S/2, 1S/3,
1S/6, 223/32, 257S/53 and 258S/24 from Uganda.
Researchers should also seek sources of resistance by prospecting in infected coffee plantations and obtaining germplasm that shows some level of
natural resistance to the pathogen. Any promising material identified in the
farmers fields will need to be multiplied and tested to ensure that this material has a good degree of tolerance or resistance.
At the University of Kinshasa, coffee research is undertaken, and the
University has taken part in various projects that have contributed to the
improvement of an integrated management strategy for CWD. The study
about varietal resistance was carried out in natural conditions of CWD at
Beni in North Kivu.
Evaluation of resistance of these accessions under natural conditions has
revealed differences in reaction in these accessions. One possibility now is to
operate a preselection test for genetic resistance and to clone interesting material for the evaluation of resistance in the field. A future research activity
is to study the level of resistance of the seedling material taken as survivors
from devastated plantations.
To date, the results obtained (based on the mortality rate recorded on different genotypes 5 mo after inoculation) have shown that a certain level of variability exists (P < 0.05) in the coffee genotypes studied for their resistance to CWD
24
(Table 2.1). Different groups were identified by using Tukeys test. The materials
tested were classified in four different groups: the first is composed of genotypes
where the level of mortality ranged between 6.6% and 7.6%; the second, between
7.8% and 8.4%; the third, between 9% and 12.8%; and the fourth, between 13.3%
and 16.1%. Considering these different groups, the first is represented by the
genotype KR 8/10, where no infection was recorded, and the fourth is represented by the genotype KR 2/5, which was totally destroyed.
As part of the breeding for resistance studies, preliminary investigations
were also made on the genetic diversity of the F. xylarioides population. An
Table 2.1. Varietal behaviour (percentage of mortality of inoculated plants) of different
genotypes of robusta coffee tested for their CWD resistance.
No.
Genotype
% of Plant mortality
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
KR16/13A
KR19/1B
KR8/10
KR19/11
KR20/51
KR19/18B
KR17/55
KR10/7A
KR18/10
KR19/28
KR19/26
KR6/6
KR18/30
KR19/1A
KR19/55
KR20/50
KR19/31
KRA/6
KR16/55
KR17/47
KR1/3
KR16/13B
KR18/10A
KR1/1
KR19/18A
KR10/7B
KR19/12
KR20/31
KRC/3
KR3/5
KR12/6A
KR8/8
KR20/10
KR9/8
KR2/5
6.6
7.5
7.5
7.6
7.8
8.1
8. 2
8.4
8.49
8.9
9.1
9.3
9.3
9.5
9.7
10.2
10.3
10.6
11.0
11.2
11.2
11.2
11.4
11.5
11.8
12.1
12.3
12.4
12.5
12.6
12.9
13.3
14.4
15.0
16.1
Group
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
25
absence of genetic diversity within the population of F. xylarioides was revealed by molecular analysis using random amplified polymorphic DNA
markers.
2.9.4. Sensitization of stakeholders

Extension is an essential part of management of the disease because farmers
have to be informed of the results of research in order for them to be sensitized
on the use of innovations. The structure of the Congolese coffee plantations
must be taken into account when developing new strategies of extension and
training of the farmers (e.g., the Farmer Field School approach). Financial,
logistic and human means must also be taken into account, not only with the
vast areas to be supervised but also, more especially, with the extent of the
tasks to be achieved. Under the Common Fund for Commodities programme,
ONC has conducted a training programme on CWD diagnosis and identification as well as training on management of the disease. We have initiated radio
programmes in Beni and Isiro for farmers and all stakeholders, which cover
all aspects of coffee production and CWD.
2.10. Conclusion
Coffee production in DRC should be flourishing because the ecoclimatic conditions in the country are favourable. However, coffee production is in decline
because of the multiple factors that we have previously mentioned. Within
the framework of revitalization, measures likely to improve the various levels
of the sector such as production, processing, distribution and logistics must
be considered.
The government of DRC, under the auspices of the multisector programme of rebuilding and rehabilitation, has initiated studies on the coffee sector. The ONC also has a plan of development for the coffee sector,
which remains to be updated. It is important to say that even with limited support, the rural communities readily commit to the programme of
development and take responsibility for themselves, often by organizing
themselves into basic groups with the material and technical support of
development non-government organizations or private companies; up to now,
this has been largely independent of the government of DRC. Nevertheless, the
government of DRC will have to continue to play an essential part in the development of the sector so that an effective development strategy for the revival
of the coffee sector is produced. The government of DRC must encourage the
private sector to intervene, restructure and regulate coffee-buying campaigns
and support the local industrialization of finished coffee products.
According to all the studies made in DRC, it is important to continue
to emphasize the threat and danger that CWD presents for the regions that
remain currently free of CWD (provinces of Bandundu, Bas-Congo, Kasai
Occidental and Kasai Oriental). Understanding the risk factors involved and
26
alerting farmers to these risk factors, for example, rapid diagnosis of infected
trees and that older plantations are more susceptible, as well as alerting them
to the fact that improved propagation materials are available, will lessen the
chance of CWD having such an impact in these provinces. In addition, it is
important to put an emphasis on the risk of introduction of the disease into
currently wilt-free areas by scientific staff in their survey inspection visits.
Their equipment could be changed for every visit to reduce the chances of
introduction.
Acknowledgements
We are grateful to the ONC coffee authority in DRC, the Kinshasa University,
the Common Fund for Commodities and the EU for the generous financial
support to the Laboratory of Phytopathology of the University of Kinshasa
(DRC) during the implementation of this work. We would like to particularly thank Professor Kizungu Vumilia (Biometrics Unity of University of
Kinshasa) for the contribution in the model logistic analysis. We also wish to
express our gratitude to the Congolese stakeholders.
References
Banque Centrale Du Congo (BCC) (19952007) Rapports annuels dactivits. Service
de Documentation.
Conjoncture Economique (2001) Rapports annuals (19917). Ministre de IIndustrie,
Commerce, Petites et Moyenne Entreprises et Artisanat, Rpublique Democratique
du Congo.
Drachoussof, V., Focan, A. and Hecq, J. (1956) Le dveloppement rural en Afrique Centrale 1908-1960/196fs2. Synthse et Rflexions 1, 254260.
Report presented to Zairean Coffee Organization (OZACAF), August 1996, 13 pp.
Bulletin Agricole du Congo Belge II 2, 361372.
Fraselle, J.V. and Geortay, G. (1952) Une grave maladie du cafier Robusta La trachomycose. Avertissement et conseils aux planteurs. Bulletin dInformation de
lINEAC I 12, 87102.
Fraselle, J.V., Vallaeys, G. and De Knop, O. (1953) La lutte contre la trachomycose du
cafier Yangambi et le problme que pose actuellement cette maladie au Congo
belge. Bulletin Agricole du Congo Belge II 6, 373394.
Heim, R. and Saccas, A. (1950) La trachomycose des Coffea excelsa et robusta des
plantations de lOubangui-Chari. Comptes Rendus des Sances de lAcadmie des
Sciences Paris CCXXXI 11, 536538.
Jagoret, P. and Descroix, F. (2002) Evolution de la culture de Coffea canephora en Afrique et problmatique de dveloppement. Recherche et Caficulture 9, 4459.
Kinds, A. (1930) Introduction despces de Cafier Qu Congo Belge, Bulletin Agricole
du Congo Belge XXI, 915.
Miny, M.P. (1930) La culture du Caf Robusta au Congo Belge. Bulletin Agricole du
Congo Belge XXI, 924929.
27
Tshilenge-Djim, P., Kalonji-Mbuyi, A., Onyembe, P.M.L., Mukuna, K., Dibwe, M. and
Oripale M. (1998) Caractristique et volution spatio-temporelle de la trachomycose fusarienne du cafier robusta en Rpublique Dmocratique du Congo (RDC).
Rvue Congolaise des Sciences Nuclaires 14, 132140.
G.J. Hakiza,1 D.T. Kyetere,1 P. Musoli,1 P. Wetala,1 J. Njuki,2

P. Kucel,1 P. Aluka,1 A. Kangire1 and J. Ogwang1
1
Coffee Research Centre, PO Box 185, Mukono, Uganda

CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621,
Nairobi, Kenya
3.1. Introduction
Ugandas economy depends largely on agriculture with coffee as the main
foreign exchange earner. The contribution of agriculture to the gross domestic product has been as high as 51% in 1991 (Anon., 1993, 1994), but there has
been a steady decline in the contribution to 36.3% in 2006 (Anon., 2006). This
decline is not an indication of loss of significance of agriculture in Ugandas
economy, but rather, it is due to the development of other sectors. Both arabica
(Coffea arabica L.) and robusta coffee (Coffea canephora Pierre) are produced in
Uganda. The distribution of coffee farming in Uganda is highly influenced by
the agro-ecological adaptations of the two commercial species. Robusta coffee
is grown at warmer and lower elevations, which vary in altitude from 800 to
1500 m in the central and southern parts of the country, covering total acreage of 240,000 ha. The annual rainfall in the robusta areas ranges from 1000
to 2000 mm. Arabica was first introduced in Uganda in 1900 from Malawi
(then Nyasaland) for cultivation at all altitudes (Thomas, 1935; Leakey, 1970;
Jameson, 1970), but these initial introductions failed to adapt to low altitudes
<1400 m because of coffee leaf rust (CLR; Hemileia vastatrix Berk. and Br.) and
climatic and soil conditions. Arabica cultivation is currently limited to areas
within 1500 to 2300 m to the east, western and north-western Uganda, with an
estimated total acreage of 37,000 ha. Rainfall in the arabica areas varies 1500 to
2000 mm annually. Both crops do well on deep fertile soils.
Most coffee farmers are smallholders (>90%) with an average hectarage
per farmer of 0.23 ha for robusta and 0.36 ha for arabica. Both crops are grown
predominantly mixed with food crops such as bananas and beans. The recommended spacing is 3 3 m for robusta and 2.74 2.74 m for arabica coffee. The
28
29
national average coffee production varies between 400 and 500 kg/ha clean coffee (0.40.5 t/ha) with average weight ratio of cherries to clean beans of about
4.5:1 for robusta, and 300 to 400 kg/ha clean coffee (0.30.4 t/ha) with average
weight ratio of 5.0:1 for arabica (UCDA, 2006). These yield levels are attributed
to the fact that Uganda coffee is produced under low-cost systems. All arabica
coffee is wet processed/washed, whereas robusta is dry processed.
3.2. Importance of Coffee in Uganda

The major export commodity for Uganda for many years has been coffee.
Coffee contributes about 20% to 25% of the national annual foreign currency
earnings from exports valued between US$84 million and US$456 million
during the period from 1996 to 2005. More than 3 million Ugandans derive
their livelihood directly from coffee as farmers, processors, exporters, transporters, traders etc. Arabica contributes 15% to 20% of coffee production by
volume and 20% to 35% of the earnings from coffee exports (UCDA, 2005).
Although robusta is rated second to arabica coffee, it is an important
commodity in the economy of many African countries including Uganda.
It has a high extracting rate and therefore is best suited to processing into
instant and decaffeinated coffee (De Graaf, 1986). In addition, robusta coffee
generally has better resistance to CLR than arabica coffee. The bulk of Ugandas coffee is robusta (grown on 242,000 ha) and accounts for about 80% to
85% of coffee production by volume and raises 65% to 80% of revenue from
coffee exports (UCDA, 2005). Robusta is most widely cultivated at altitudes
ranging from about 800 m in Bundibugyo district (rift valley area) to about
1500 m, e.g. in Bushenyi district.
Robusta coffee (C. canephora Pierre) is indigenous to Uganda and is embodied in cultural and socio-economic norms of the Ugandan population.
The Baganda used coffee beans in their ritual of blood brotherhood, and
chewing of dried beans is still practised. Long before the arrival of the British
explorers Speke and Grant in Uganda in 1862, the Baganda were growing robusta coffee around their homesteads (Thomas, 1947; Purseglove, 1968). The
first coffee (robusta) exports from Uganda, mainly from Sese Islands (now
Kalangala district) in Lake Victoria, were in 1878 (Thomas, 1947). From this
humble beginning, robusta coffee cultivation expanded and thrived to become the major foreign exchange earner for the country for several decades.
The importance of coffee in terms of export volume/quantity and value in
US dollars, in comparison to other export commodities, is presented in Table
3.1. Although the volume of coffee exports has declined, it retains the highest
foreign exchange compared to a wide range of commodities (Table 3.2).

There are a number of production constraints that confront coffee farmers.
Among these constraints is a lack of varieties with resistance to coffee wilt
30
Hakiza et al.
Table 3.1. Quantities of various commodities exported 20022006.

Commodity
Coffee
Cotton
Tea
Tobacco
Maize
Beans and other
legumes
Fish and fish
products
Cattle hides
Electric current
Unit
2002
2003
2004
2005
2006
t
t
t
t
t
t
201,591
12,322
30,400
23,266
59,642
10,753
146,299
16,762
36,669
24,669
60,298
18,070
159,983
29,762
36,874
27,843
90,576
26,233
142,513
30,403
36,532
23,730
92,794
28,332
126,887
18,480
30,584
15,794
115,259
27,087
25,525
26,422
31,808
39,201
36,461
t
000 kW h
20,049
264,685
18,565
217,486
18,502
193,104
25,349
62,577
22,214
53,019
Source: Uganda Export Promotion Board 2006.
disease (CWD), drought, other pests and diseases. Other constraints include
low farm gate prices paid to farmers by coffee buyers or middlemen in addition to low and fluctuating world coffee price, which results in very small
returns to the farmers. In the absence of stabilisation funds, many farmers
have found coffee production increasingly less profitable. Moreover, there is
no incentive to produce good quality coffee because there is no grading system to reward those who produce good coffee. Poor agronomic practises by
farmers contribute to reduced productivity. For instance, stumping of coffee
is one of the recommended practises for rejuvenation of coffee, but very few
farmers practise it. The low adoption of recommended agronomic practises
can be attributed to limited access to information/technologies due to inadequate extension services and to limited access to inputs. Inputs may also
be available but are too expensive for most smallholders. The declining soil
fertility is another concern because farmers do not use fertilizer or organic
manure and is considered as the main cause of low productivity.
Table 3.2. Export commodities and their value (000 US$) 20022006.
Commodity
Coffee
Cotton
Tea
Tobacco
Maize
Beans and other
legumes
Fish and fish
products
Cattle hides
Electric current
2002
2003
2004
2005
2006
96,626
9519
31,293
45,262
10,609
3284
100,233
17,755
38,314
43,042
13,724
5,235
124,237
42,758
37,258
40,702
17,896
8,968
172,942
28,821
34,274
31,486
21,261
8,693
189,830
20,474
50,873
26,964
24,114
8,162
87,945
88,113
103,309
142,691
145,837
9810
15,645
4,925
13,778
5,409
12,075
7,064
4,465
8,030
4,855
Source: Uganda Export Promotion Board 2006.
31
3.4. Most Important Diseases and Pests

Arabica coffee generally has more diseases and pests than robusta coffee,
particularly at low altitude. Even at higher altitudes, arabica coffee suffers
serious attacks of pests and diseases. Insect pests such as Antestiopsis spp.,
stem borers (Bixadus sierricola White) and leaf-sucking insects including leaf
miners (Leucoptera spp.) and lace bugs (Habrochila spp.) can reduce yield and
quality of arabica coffee. Incidences of coffee berry borer (Hypothenemus hampei Ferrari) have been observed to increase on arabica coffee at higher altitude
than was previously seen. Root mealybugs (Planococcus spp.) are becoming
increasingly destructive and difficult to control. The most damaging diseases
are CLR (H. vastatrix Berk. and Br.) and coffee berry disease (CBD; Colletotrichum kahawae Waller and Bridge), which can cause yield losses of more than
50% under conditions favouring the disease. Bacterial blight (Pseudomonas/
Ralstonia syringae van Hall) occurs in all arabica areas, but its occurrence is
sporadic. Use of copper fungicides (50%) for the control of CBD and CLR at
high altitude also keeps bacterial blight under control.
For many decades, robusta coffee cultivation in Uganda experienced
only minor crop protection problems. The main pest on robusta has been
coffee berry borer, which can be controlled by good cultural practises such
as regular picking, stripping off and burning old and dry cherries (mbuni)
that remain on the trees at the end of the picking season. These practises remove sources of potential infestation and destroy the insects within the berries as well as their food source. Pruning is essential to keep the canopy more
open, less humid and unattractive to the pest. Root mealybugs have been
causing serious but localised damage on robusta coffee, and the symptoms
have sometimes been confused with CWD. Other species of mealybugs and
scale insects occur on aerial plant parts and can be a problem particularly on
young plants just transplanted in the field or those still in the nurseries.
Diseases on robusta had often been minor and insignificant, for instance
the berry red blister disease caused by Cercospora coffeicola (Cooke and Berk)
has been present on robusta since the 1960s with occasional and sporadic
incidences. Severe incidences of red blister disease causing berry fall have recently been reported in Rakai and several other districts. Research is needed
on this disease to find effective management measures. Some robusta clones
or varieties appear to have a lower resistance to CLR. Root rot caused by
Armillaria mellea Vahl. occurs often on coffee established on land previously
under forest but can also appear when shade trees such as Ficus spp. are suddenly cut down. Yellowing of affected plants followed by wilting and death
can be mistaken for CWD.
The emergence of the highly destructive CWD in the early 1990s was a
challenge to the whole coffee industry in Uganda (Gibberella xylariodes Heim
and Sacc), of which Fusarium xylarioides Steyaert in the anamorphic phase is
a highly aggressive disease that spreads rapidly, and all infected plants are
killed. Whole or partial coffee fields were wiped out and, almost overnight,
most farmers lost their only means of livelihood. Plates 1 and 2 illustrate the
effect of the disease. This field was one of the most productive coffee fields
32
Hakiza et al.
at the Coffee Research Institute (CORI, formerly COREC), and within 3 to 4

years, it was reduced to a few unproductive trees.
More than 120 million robusta trees (44.5%) have been destroyed by
the disease since 1993. In other words, 80,000 ha of robusta, 1.2 million bags
(60 kg bags) of coffee and US$100 million lost to CWD (UCDA, 2004). In addition, many coffee farmers and other stakeholders who were dependent on
robusta coffee have lost their livelihoods.
3.5. Brief History of CWD in Uganda

An account of the spread of the disease from the time it was first reported
and the actions that followed are provided here and are meant to provide lessons learnt for the whole coffee sector and should allow for a better response
to future epidemics.
In 1992, information was received from John Schluter, a businessman on
a visit to the Democratic Republic of Congo (DRC), of a devastating robusta
coffee disease in the Beni and Isiro areas of DRC and warned of its consequences if allowed to cross into Uganda. To find out as much information as
possible on the new disease, its symptoms, causal organism, epidemiology
and control, the senior author turned to literature to be able to recognize the
disease and what to do about it should it appear. Recommendations for control of the disease were obtained from literature (Wrigley, 1988; Coste, 1992).
Uprooting and burning of infected plants at the earliest symptoms and other
sanitary measures were intended to destroy sources of infection/inoculum to
reduce the rate of spread within the field and outside the field. Infected plant
parts carried by people as firewood was a mechanism for disease spread
as were agencies such as wind or rain. Spread of the disease through these
agencies ensured dissemination far from the original affected farms.
Ironically, in September 1993, wilting and death of a few robusta coffee
trees were observed in a 2.8-ha experimental plot at CORI, Kituza, in Mukono district of central Uganda. The plant pathologist (main author) recovered F. xylarioides, the anamorphic form of Gibberella xylarioides, from infected
plant parts (stems and roots). Identification of the pathogen was based on
cultural and spore morphological characters as described by Booth (1971)
and Gerlach and Nirenberg (1982). Reproduction of symptoms on seedlings
was done by transplanting 6-month-old seedlings/clones in sterile soil infected with spores of the pathogen and maintaining the inoculated plants
in the screen house until symptoms appeared. Re-isolation of the pathogen
from infected seedlings confirmed the pathogenicity of the fungus causing
the wilt.
The recommended standard control method for CWD is uprooting
and burning of the diseased plants to eliminate potential inoculum sources
for further infection and was maintained in this particular plot from 1993
to 2000. At first, the method appeared effective. However, with time, the
method proved unsustainable due to the high cost of labour for uprooting
the diseased plants. It appeared that inoculum sources were not only from
33
within the same field but also from external sources. It transpired that coffee
trees in the adjacent forest were also infected by wilt, and it was suspected
that this was the source of inoculum. As long as other sources of inoculum
exist, from outside the targeted field, infection continued. Farmers should
therefore work in conjunction with their neighbours to minimize the transfer
of inoculum from one farm to another.
Plants can be affected at any stage of development, from seedlings to
adults. The first symptom exhibited from the top of infected plants is a flaccidity of the leaves on one side of the affected stem. Chlorosis may or may
not occur. Wilting, rapid defoliation and dieback soon follow (Plate 3). During the rainy season, leaf wilting may not be readily observed, and the only
indication of wilt is rapid defoliation (noticed as carpet of leaves at the base
of trees) and dieback, which is the progressive death of branches or shoots
beginning at the tips leaving behind bare dry twigs. Dieback at the tips of the
main stem spreads down to all primary branches resulting in death of stem.
Dark brown necrosis occurs on stems of young suckers and branches, which
leads to death of those parts. Leaf veins show dark brown necrosis. On a
multistemmed coffee plant, stems die in sequence, one by one until the whole
bush dries up completely. Coffee berries on the affected tree ripen prematurely (turn red) and dry up but remain attached to the primary branches.
CWD is confirmed by the presence of blueblack streaks or bands in
the wood when a bark of the affected stem is peeled off. This distinguishes
CWD from all other wilting, which could be due to root mealybugs A. mellea and other fusaria such as Fusarium oxysporum, Fusarium solani, Fusarium
lateritium and Fusarium decemcelulare. At the advanced stages, cracks may occur on the stem from around the collar region and up along the stem. Within
these cracks in the bark are embedded brown to black perithecia, which bear
numerous ascospores. Ascospores are the sexual or teleomorph phase of the
fungus. These are spread by wind or washed by rain into soil to become soil
inoculum, then spread by runoff to other plants in the vicinity.
Infected plants do not recover even after pruning and/or stumping.
When pruned or stumped, any suckers that may sprout develop dark brown
necrosis and eventually wilt, die and dry up.
The affected and dried up plants remain firmly rooted in the ground,
unlike those affected by fungi such as Armillaria or root mealybugs, which
easily topple over when pushed.
Following the outbreak in Mukono, recommendations were made based
on literature (Coste, 1992) as an emergency measure in the absence of any
other information and were not then based on research findings done in
Uganda at the time.
Thus, regular field inspection to detect early symptoms of the disease
was recommended to farmers. Any plants confirmed as infected were to be
uprooted and burnt on the spot. Farmers were advised to avoid using coffee
husks as mulch in their coffee fields because these were strongly suspected
to be contaminated with the pathogen. Some districts, e.g. Masaka, passed
a by-law prohibiting milling of coffee from outside the district. Restriction
on the movement of infected plant materials, coffee husks and soil was also
34
Hakiza et al.
recommended. The berries on infected plants were to be separately picked

from coffee on healthy plants. Farmers were also advised to plant coffee on
new land, preferably some distance from the infected field. Replanting in the
affected fields was recommended not be done immediately after uprooting
but to delay planting for at least 1 year and preferably longer; other crops
can be grown. Tools were to be cleaned by flaming or use of disinfectant to
avoid passing infection from an infected plant to another. When pruning or
handling plants, always start with healthy-looking plants first and sterilize
tools after each plant. Wounding the stems during cultivation (such as weeding) of the coffee should be avoided because wounds are likely sites for entry
of the pathogen into the plants. These recommendations are now backed by
research findings that have been conducted in the past 4 to 5 years during the
project period, and therefore, the above management measures remain.
The first report of wilt outside Mukono district was received in October 1993 from Bundibugyo district in south-western Uganda bordering the
DRC concerning the presence of a destructive wilt disease of robusta coffee.
The Uganda Coffee Development Authority (UCDA) sponsored a multidisciplinary team of researchers composed of a pathologist, an entomologist, an
agronomist and an economist to travel to Bundibugyo and identify the problem. The main author was part of the expedition. The affected robusta areas were Bwamba, Bubukwanga and Bubandi counties. In this district, both
robusta and arabica are cultivated. Arabica coffee occupies higher grounds
in Nyankonja, Busaru and parts of Bubukwanga counties. The disease was
reported only on robusta coffee. During the survey, two farms in each of the
three subcounties were visited, and samples were collected. All the fields visited had some dead plants and partially diseased trees. Symptoms of the disease were not typical of the wilt disease caused by F. xylarioides as described
in literature (Coste, 1992) and observed at Kituza hardly a month earlier. The
dead and partially diseased plants had dieback, some leaf fall occurred but
a lot of dry leaves remained attached to the dead branches. When bark was
scraped off from the stem of the affected plants, there was no blueblack
streak characteristic of F. xylarioides, instead only brown to dark brown disintegrating/rotting tissues were observed. A whitish powder covered most of
the dead and blackened suckers and branches in many trees. From the samples collected, the fungi recovered in the laboratory were Fusarium stilboides,
F. lateritium, F. solani and F. oxysporum, which can also cause wilting and
death of plants under stress. In addition, there was a high incidence of stem
borer attacks leading to wilting and death of trees in the same fields. During
this first visit, the presence of CWD was somehow missed probably due to
the high incidence of other fungi and stem borers. It was later that these other
fungi became less frequent, and F. xylarioides became most dominant from
specimens received from this area (Hakiza, 1998, unpublished report).
More reports and samples were received in 1994 from Kanungu district
(formerly part of Rukungiri) in western Uganda, and the CWD pathogen
was consistently recovered from samples of roots, stems and branches/primaries. Specimens from the Bundibugyo district, received about the same
time, also clearly revealed the presence of the pathogen.
35
The worsening situation prompted UCDA in 1996 to organize and fund

a joint survey conducted by a multidisciplinary team from the Ministry of
Agriculture Animal Industry and Fisheries (MAAIF), Coffee Research Centre (COREC) and UCDA to establish the distribution/extent and damage
caused by the disease in ten major coffee-growing districts of Uganda. It
revealed the presence of the disease in ten districts: Mukono, Mubende,
Mpigi, Ntungamo, Kasese, Kabarole, Rukungiri, Bundibugyo, Kiboga and
Masaka.
In the same year, UCDA invited Dr Julie Flood of CABI to assess the situation, as she had just conducted a survey in Eastern DRC (Isiro area) on the
same problem (Flood, 1996), and she made an assessment in January 1997.
Her findings confirmed the presence of CWD as had been reported by the national scientists. She recommended regional collaboration to enable Uganda
to benefit from the past experiences of African countries that had previously
successfully managed outbreaks of CWD in earlier decades.
A regional workshop on CWD was held in the International Conference
Centre, Kampala, from July 28 to 30, 1997. Representatives from the International Coffee Organization (ICO), CABI and CIRAD attended the meeting, as
well as participants from African countries. The main outcome of the meeting was the initiation of a regional research project proposal (Regional Coffee
Wilt Programme) to contain the situation. Uganda and DRC, which were the
worst affected countries, were to play leading roles in research and development on CWD. This was a major landmark in the struggle against CWD in
the region. The project was finally initiated in 2000.
However, the incidence of the disease had continued to increase through
19952000, and samples and reports were received from other parts of Mukono
district, Kiboga, Mpigi, Jinja, Iganga and Mayuge. In 2000, reports and samples
were received for robusta coffee from the West Nile region where both coffee
species are grown; the pathogen was reisolated. By the end of 2000, all robusta
coffee districts in Uganda had been affected. Consequently, during the intervening time between the research proposal (Regional Coffee Wilt Programme)
being written and obtaining the funding, the National Agricultural Research
Organisation (NARO) of Uganda became so concerned that they convened a
technical workshop on CWD in 1999 to examine research progress and define
action points to accelerate research activities toward finding a solution. Following this meeting, a task force composed of plant pathologists, plant breeders and socio-economists from NARO and Makerere University was formed to
formulate a research programme on CWD and to decide on its implementation.
A comprehensive research programme was drawn up for a period of 5 years
together with its budget. CORI is continuing to implement the programme.
Many reports of wilt on arabica coffee and samples received at different
times from all the arabica areas in the eastern, western and north-western/
West Nile region of Uganda revealed other Fusaria but not G. xylarioides.
Many follow-up visits to those areas have also been made to investigate and
verify the reports. In all cases, G. xylarioides has not been recovered from
samples or observed in any of the field visits made. The fungi that have been
consistently recovered from samples were F. solani and F. oxysporum and
36
Hakiza et al.
were confirmed as the cause of localised wilting and death of arabica coffee
in those districts (Arua, Nebbi, Bushenyi, Kasese, Ibanda, Rukungiri, Mbale,
Sironko, Bududa and Manafa districts). Root rot caused by A. mellea Vahl has
also caused concern and has been mistaken for CWD. Insect attacks, the most
common of these encountered in the arabica areas, that induce wilt symptoms that can be confused with CWD are coffee stem borers (B. sierricola),
coffee root mealybugs (Planococcus ireneus) and twig borers. In essence, to
date, CWD infects only robusta coffee, even in districts where the two crops
are grown side by side.
The absence of the disease in arabica coffee in Uganda shows that the
pathogen is currently host specific. It is not known if with time the pathogen
will change and infect arabica as well, particularly in areas where the two
species are grown side by side. G. xylarioides remains a potential threat to
arabica coffee in Uganda.
3.6. Coffee Wilt Distribution in Uganda

In Uganda, CWD infects robusta coffee (commercial species), but it also affects
Coffea liberica and Coffea kapagota in the germplasm collections at Kawanda
Agricultural Research Institute. The disease manifests itself as classical vascular wilt with wilting leaves sometimes preceded by chlorosis followed by
dieback. A characteristic blueblack staining of the wood is seen under the
bark of affected trees.
After 1993, the incidence and severity of CWD continued to rise causing decline in yields in all affected districts. The worst affected districts were
Bundibugyo, Kyenjojo, Hoima, Kabale and Kayunga, where it was estimated
that 40% to 50% of the coffee fields were completely destroyed or abandoned
(CORI, 19961997; Lukwago and Birikunzira, 1997; UCDA, 1999). An estimate
by UCDA in 1999 indicated that 14.5 million robusta coffee trees (4.8% of the
coffee) were destroyed countrywide. However, these estimates were based on
inadequate information from an earlier survey conducted in 1996 in only ten
districts. The data capture tools used in the 1996 survey were not comprehensive. Moreover, there were no follow-up farm visits to monitor and check the
seemingly rapidly increasing rate of disease spread and incidence levels. Consequently, a more accurate estimate of current magnitude of the CWD problem
was needed to provide quantitative information on the current status of CWD
in the country. Various surveys (biological, remote sensing and socio-economic
surveys) were therefore conducted to provide comprehensive and up-to-date
information on the status of CWD in Uganda in 2002 and again in 2004.
The first biological survey was carried out in Uganda in March to April
2002 with the following objectives:
To establish CWD prevalence and severity at farm, district and national

levels
To quantify the impact of CWD on coffee production at farm, district
and national levels
37
To assess the influence of human and biophysical factors within the coffee
farming system on incidence, spread and severity of CWD in Uganda
To develop a survey tool for biological surveys in African coffeegrowing countries participating in the Regional Coffee Wilt Programme
(Improvement of Coffee Production in Africa through the Management
of CWD).
The survey was conducted in ten districts in the central, six districts in the
western and five districts in the eastern part of the country. The districts
surveyed in the central zone included Mukono, Kayunga, Wakiso, Mpigi,
Luwero, Masaka, Rakai, Sembabule, Mubende and Kiboga. The coverage
in central Uganda included almost all coffee-producing districts except the
Lake Victoria district of Kalangala and the marginal rainfall district of Nakasongola, which are minor producers. The ten districts selected from the
Central region produce 80% to 90% of robusta coffee in the country. In the
West, the districts surveyed included Hoima, Bundibugyo, Kibale, Kyenjojo,
Bushenyi and Rukungiri. Similarly, in the east, throughout the period
from 1995 to 2000, reports of the disease were received from other parts of
Mukono district, Kiboga, Mpigi, Jinja, Iganga, Mayuge and Iganga; the
West Nile region was affected in 2000. Thus, by the end of 2000, all robustagrowing coffee districts in Uganda had been affected (Fig. 3.1).
In each selected district in the central zone, a number of robusta-coffeegrowing parishes were randomly selected. The number of parishes selected
from each of the districts was proportional to the total no of parishes in the
districts. In the western and eastern zones, two parishes were selected from
each selected district. Three villages were randomly selected from each of
the selected parishes except in Mukono and Kayunga where one village was
randomly selected. In each village, three farms were randomly selected except in Mukono and Kayunga where nine farms were selected. Data were
collected from each of the selected farms using a data input form.
Survey results indicated that CWD was present in all 21 districts, although at varying incidences. All farms surveyed in Bushenyi, Hoima, Jinja,
Kamuli, Kibale, Kiboga, Kayunga, Luwero, Mukono and Rukungiri districts
had CWD. Rakai district had the highest percentage of farms without CWD.
These results were based on the assumption that the missing coffee trees on
the farms were assumed to have been infected by CWD and destroyed by the
farmers as recommended.
Only Iganga had a low mean CWD incidence of 3.5%. Bugiri, Rakai, Mayuge, Sembabule, Masaka and Bundibugyo had a moderate percentage of incidence of CWD. Bundibugyo is where CWD was first reported, and farmers
there had been advised to uproot all affected trees as soon as symptoms appeared. CWD incidence in the remaining districts was very high. Kayunga,
Luwero, Kibale, Mubende, Hoima, Mukono and Kiboga were most affected
with more than 60% of the trees infected (sick or dead-stumped or deadstanding or uprooted due to CWD). The severity of CWD in farms was high,
ranging from 3.5 in Iganga to 5.6 in Bundibugyo on a 16 scale and ranged
from 1.1 in Iganga to 3.3 in Kibale.
38
Hakiza et al.
MOYO
KITGUM
YUMBE
AN
JU
KOTIDO
AD
ARUA
PADER
GULU
MOROTO
NEBBI
LIRA
KATAKWI
APAC
AI
MASINDI
NAK
L. KYOGA
ASO
OLA
BA
KA
WAKISO
SE
L. GEDARGE
HO
RW
A
SIRO
NKO
MBALE
TORORO
IG
MUKONO
KAMPALA
MPIGI
KA
PC
G
AN
MAYUGE
GE
EN
PALLISA
JINJA
MUBENDE
KYENJOJO
KA
SIA
YO
G
U
IB
D
N
LE
KAMULI
RO
BU
GA
LUWERO
KAYUNGA
BO
KI
KIBALE
KASESE
KUMI
NG
HOIMA
BUGIRI
AL
NA
SOROTI
SU
L.
KA
IR
AP
BE
RT
BE
RIT
IPI
AM
BA
BU
LE
L. EDWARD
MASAKA
BUSHENYI
IRI
UNG
RUK
U
O
UNG
OR
KAN
KIS
KALANG
MBARARA
AL
A
L. VICTORIA
RAKAI
NTUNGAMO
KABALE
Key:
0-5%
= low
>5-20% = medium
>20%
= high
Arabica coffee growing areas
Fig. 3.1.
Map of Uganda showing the distribution of CWD of robusta coffee (not to scale).
3.7. Importance of CWD in the Country

Robusta coffee is a key commodity in household incomes and Ugandas main
revenue earner. The disease has serious implications to the national economy
as a whole but even of more concern to the household income of the farmers.
For example, the farmers visited in 1999 in the Mukono, Rakai and Mpigi

Table 3.3.
39
Loss of income incurred by two farmers due to CWD in Mukono district.

Case 1
Period
Before CWD
1 year after CWD
2 years after CWD
Case 2
Period
Case 1
Case 2
Income (gross revenue

in Uganda shillings)
5600
2800
0
3,920,000
1,960,000
0
2100
420
0
1,470,000
294,000
0
Source: Field visits to selected farmers in Mukono district (CWD Task Force Report, 1999).
Table 3.4. Yield loss in 60-kg bags of dry cherry (kiboko) due to CWD (inclusive of missing
trees).
District
Bugiri
Bundibugyo
Bushenyi
Hoima
Iganga
Jinja
Kamuli
Kayunga
Kibaale
Kiboga
Kyenjojo
Luwero
Masaka
Mayuge
Mpigi
Mubende
Mukono
Rakai
Rukungiri
Sembabule
Wakiso
Grand total
Mean
Average
annual
yield (kg)
before
CWD (B)
18.0
16.1
22.5
16.1
12.9
18.6
20.0
46.5
25.0
41.6
16.3
32.4
17.4
7.5
26.8
27.2
40.0
14.9
15.6
33.5
17.7
26.2
No. of
farms
giving
value
for yield
before
Average
current
annual
yield (kg)
(after
CWD) (C)
17
18
17
18
17
18
17
63
18
50
18
134
122
15
134
140
140
115
18
31
139
1259
7.4
2.3
2.6
0.9
3.4
4.3
6.1
3.4
2.4
4
2.1
4.1
7
3.7
4.5
2.9
3
9.8
3.3
19.1
2.2
4.7
Source: Report on CWD Biological Survey, 2002.
Farms
giving
value of
current
yield
(after
CWD) (C)
17
15
16
18
15
17
17
57
18
50
18
133
118
15
128
133
141
109
17
30
128
1210
Average
% yield
loss
(current
before)
Farms
where %
loss could
be
calculated
58.9
85.7
88.4
94.4
73.6
76.9
69.5
92.7
90.4
90.4
87.1
87.3
59.8
50.7
83.2
89.3
92.5
34.2
78.8
43.0
87.6
17
15
15
18
15
17
17
57
18
49
18
130
114
15
123
129
128
106
17
30
127
1175
76.9
40
Hakiza et al.
districts reported that the disease had adversely affected their earnings and
subsequently their ability to meet household cash requirements such as school
fees, domestic needs, hired labour, taxes and investment (CWD Task Force
Report, 1999). According to the same report, the estimated average loss in
income due to CWD per household was 362,973 Uganda shillings per annum, which was well above the average per capita income of Uganda estimated at US$250. Members of the task force visited some farmers in Mukono
district who had incurred losses due to CWD. Two of the farmers who kept
records revealed their losses as indicated in the table below.
Within a very short time, the disease caused a total loss of the crop. Many
farmers, particularly those who invested substantial amounts of money to
improve their ten or more hectares of coffee, did not only lose their crop and
income (Table 3.3) but also their health.
Yield loss at a district level (Table 3.4) showed that CWD was causing
77% loss in yield of robusta coffee at the national level. The mean yield loss
across districts ranged from 34.2% in Rakai district to 94% in Hoima. Even
among districts with the least losses in yield such as Rakai, some individual
farms incurred losses of more than 50%.
The current situation indicates the presence of the disease in all robusta
districts (Fig. 3.1). Its presence in some non-traditional robusta-growing areas, e.g., Busia and Arua districts, has also been confirmed. The incidence
and severity, however, have markedly declined in all the districts probably
because the most susceptible plants were attacked first and those that have
escaped or have some resistance are currently being infected. Farms that had
low incidences during the 2002 survey, e.g. Rakai and Iganga, are currently
experiencing some increase in incidence of CWD. Despite the low incidence
of the disease in some districts, farmers and all stakeholders should maintain vigilance and conduct disease management practises as has been recommended. Under favourable conditions, CWD could re-emerge.
3.8. Factors Affecting Severity or Incidence of CWD in Uganda

There are probably many factors, some known and many still largely unknown, that influence the incidence and severity of CWD. Some of the factors are examined here.
3.8.1. Lack of information

One of the factors that could have accelerated the spread of the disease in
Uganda was lack of information/knowledge on all aspects of the disease.
Research, extension and the Ministry of Agriculture had no useful information to give to farmers to slow down or stop the epidemic. In addition, it was
widely believed that robusta coffee was resistant to most pests and diseases.
This state of uncertainty gave time for the disease to continue to spread to all
districts within a relatively short time.
41
3.8.2. Varietal susceptibility to CWD

Most robusta plantings in Uganda originated from seeds selected by farmers
from their own coffee fields. Farmers preferred their own seeds/seedlings,
although there are recommended clones or clonal seeds available that have
better yields, resistance to some diseases and good bean and cup qualities.
The dependence on self-saved seeds/seedlings was largely because they incur no cost in obtaining them whereas the recommended clones/varieties
were very expensive. The cost of recommended robusta cuttings was 500
Uganda shillings for each single clone, and seedlings cost 300 Uganda shillings each. Clonal seeds from the research station cost 2000 per kilogram. In
addition to cost, the source of good planting materials may be far removed
from the farmers, so further costs in transportation are incurred; all these
costs encourage farmers to use their own materials. This seedling coffee was
very susceptible to CWD and encouraged the spread of the disease. Although
CWD can also infect the recommended clones, partial resistance is evident as
compared to seedlings.
3.8.3. Environmental factors

At the time of the onset of the epidemic, rainfall was observed to be higher
than normal for a prolonged period of time. Favourable environmental conditions, in the presence of susceptible germplasm and an aggressive pathogen, led to the disease spreading rapidly throughout the country. Two rainy
seasons normally occur in Uganda. The first rainy season (MarchJune) is
normally heavier than the second rains (SeptemberNovember). Recent field
observation trials at Kituza revealed an increase of CWD and disease progress in infected plants during the rainy season and a drop in number of new
infections during the dry season. This can be attributed to increased plant
activities during the wet season, which enhances fungal infection and colonization of the host as well as rain being involved in the spread of the disease.
In addition to climatic factors, soil type seems to have an influence in
that farms on predominantly sandy soils had more wilt-infected coffee trees
than farms on predominantly loam, sandyclay or clay soil types.
3.8.4. Lack of preparedness and resources

Generally, the response to the epidemic was slow for various reasons. Because
it was new, no one could predict the magnitude or gravity of the problem.
Lack of resources and lack of preparedness to mitigate emergency situations
such as this could not allow a quick response. Despite recommendations for
uprooting and burning, as well as other sanitary measures, farmers lacked
resources to uproot and burn infected plants. The exercise proved too expensive and too demanding on the ageing farmers who are the majority. Lack of
resources also delayed the start of serious research by the national research
42
Hakiza et al.
institutions. Research priorities were based on constraints with measurable

impact on the farmers and national economy but could not readily accommodate emerging problems. Lessons learned from these experiences probably
caused a more positive action when bacterial wilt of bananas (Xanthomonas
musacearum) emerged in Mukono district in 2002 (Banana Bacterial Wilt Task
Force Report, 2002, 2003). Mobilization of farmers for uprooting and burying infected banana stools was quickly done and a research agenda defined
and funded within the same year the report was received, which was in total
contrast to what happened when CWD was first observed.
3.8.5. Human activities

Unknowingly, farmers, farm workers and others spread the disease within the
farms and out of the farms to their neighbours and to other districts. Within
the farms, farmers and others cut the dry trees for use as firewood at the homestead. The mode of transport of the dry plants was by dragging them through
the coffee field and taking the shortest route to the homestead, to where these
were stacked outside the kitchen house. Shortly afterwards, more infected
trees were observed along the routes taken through the coffee field. The infected stems were also transported on bicycles and through other means of
transport to distant places where new infections could have been started.
G. xylarioides is considered a wound pathogen, which can only invade
the plant through wounds or damaged tissue (Muller, 1997). In many normal farming practises, such as weeding using hoes or slashers or machetes/
pangas, wounds can be created on roots and stems. A higher disease incidence was evident when coffee was hoe weeded as compared with other
weed management practises. Pruning and desuckering also create wounds,
which provide entry points for the pathogen, and the incidence was greater
in coffee where the canopy was unpruned as compared to coffee under canopy management (pruned). High disease incidence was also observed on
farms that did not use fertilizer or manure. Wounds created close to stem
base are more likely to result in systemic infection than wounds further up
the plant (Hakiza et al., 2004). Under favourable conditions of temperature
and moisture, spores of the pathogen deposited on the wounds germinate
and penetrate the host resulting in infection, and later visual symptoms of
the disease would appear. Shading of the coffee increased the probability of
disease incidence, and coffee farms with shade trees had significantly more
infected trees than farms without shade.
Increased CWD incidence was also observed on farms where livestock
were tethered to coffee trees or animals were allowed to graze in coffee fields,
as compared to farms where livestock (cattle, goats, sheep and pigs) were not
tethered or allowed to graze in coffee fields. Extensive wounding of stems
was observed as well on big roots where pigs had been digging up soil. Most
probably these wounds led to increased infection.
Infected planting material/seedlings could have played a role in the dissemination of the wilt pathogen. It is widely believed that CWD crossed to
43
Jinja, Mayuge and other districts in Eastern Uganda, mostly through volunteer seedlings collected from forests in Mukono district where CWD was also
rampant on forest coffee. The farmers are attracted by the low cost of these
seedlings compared to good seedlings from known sources. The seedlings
may appear healthy but could be infected or contaminated with G. xylarioides.
Symptoms may develop later when transported to other areas for planting.
Abundant inoculum in the form of perithecia is readily produced around the
collar region and a few centimetres below the soil surface following the death
of the seedlings. This inoculum is washed by rain into soil or spread by wind
to initiate infection to healthy seedlings planted in the same field or further.
As a soil-borne pathogen, any material that carries soil is suspect. Nursery operators often collect topsoil from the forests, as this soil is considered
very fertile. It has been demonstrated that unsterilized soil collected from
forests may have enough contamination with the disease pathogen to cause
infection to seedlings (Flood, 2005). The seedlings/clones raised in soil contaminated by the wilt pathogen can be transported far and wide and cause
new infections when they die, and perithecia are formed abundantly around
the collar and just below the soil surface.
Coffee plantations had significantly more CWD than other production
systems, and coffee husks were used for mulching in many coffee plantations. A number of farmers who lost their coffee have attributed this to
mulching with coffee husks. It was also claimed that the disease got to Mukono in coffee brought from the infected districts in western Uganda and
hulled in Mukono district. Many farmers took the husks for mulching their
coffee fields. Mukono district is among the districts that had a high incidence
of CWD. Because the coffee husks were suspected carriers for the pathogen,
its use for mulching was stopped by the local leaders. The recommendation
is maintained to date.
3.9. Country-Specific Management Practices in Uganda

3.9.1. Farmer support and other establishments
Control of CWD requires the participation of all coffee stakeholders from
village level to policy makers. Each stakeholder has a role to play to ensure
the effectiveness and success of control programmes. On-farm measures to
prevent spread of the disease are done by the farmer himself, ideally in conjunction with his immediate neighbours, and good sanitation measures can
minimize the transfer of inoculum from one field to another.
Key coffee stakeholders spearheading the fight against CWD comprise
MAAIF, which is responsible for providing extension services to farmers,
policy issues and provision of funds. Non-government organizations, such
as World Vision, also do some extension work, as well as various farmer organizations, among them the National Union of Coffee Agribusinesses and
Farm Enterprises. The UCDA is a body created by the government to oversee
the coffee subsector and to provide extension services to farmers through its
44
Hakiza et al.
extension agents posted to major coffee production districts. Research services are provided mainly through NARO, universities and other private
organizations. NARO, through its CORI, conducts research to solve priority
constraints of which CWD has been of major economic concern.
Coffee research in Uganda has been directed towards the control of CWD
to increase the knowledge base of the disease and its pathogen. The formulation of effective means of control/disease management and breeding for
durable resistance to CWD depends on detailed knowledge of the pathogen.
Concurrently, research is also directed to finding agronomic measures for
management of the disease. Use of resistant cultivars is the most effective longterm measure against CWD. In addition, resistance can eliminate or reduce the
need for other measures. General selection and breeding for resistance have
been making use of available germplasm in the coffee collection at Kawanda
Agricultural Research Institute and from survivors of CWD in the farmers
fields. The goals of these include improvement of yield and vigour, better local adaptation (resistance to other pests and diseases) and better bean quality.
There are high expectations from the selected CWD-resistant robusta coffee
lines, which are undergoing final field evaluations on farms in several districts
before release to farmers on a large scale (see Chapter 9).
3.9.2. Replanting programme in areas infected by CWD

Coffee production in Uganda had been on the decline for over a decade as
evidenced by the reduced volume and value of the most important export
commodity (Table 3.1). Among the reasons for the decline was low productivity due to old trees (more than 50 years old). To revitalize production, a
gradual replacement of old robusta trees with new plantations, using the six
high-yielding clonal robusta, was planned before the emergence of CWD. The
replanting programme was initiated in 1990 by UCDA, and when the disease
emerged, this operation continued and has been effective in maintaining coffee stands and ensured that there was some production of robusta coffee.
Replanting has been based on the recommended clones in the absence of
better varieties that could be used. As a precaution against CWD, all the nursery
operators were trained in nursery management, CWD control and general coffee field management including control of other pests and diseases. The planting
materials were purchased by UCDA and distributed free of charge to farmers to
encourage them to replant and also to ensure they get the recommended planting materials. It has been estimated that 100 million clonal seedlings were supplied to farmers since the start of the programme in 1990 (UCDA).
3.9.3. Introduction of robusta coffee in non-traditional areas free from CWD

To avoid/escape CWD, the production of coffee in selected areas suitable
for robusta coffee, outside the areas where robusta had been traditionally
grown, was encouraged and supported. Some areas in the north and east
45
of the country were already growing some coffee. With UCDA support and
technical input from the Coffee Research Centre, training of extension workers and nursery operators was undertaken. Training included general coffee management with an emphasis on CWD especially to avoid spread of
the disease from infected districts to new areas. The programme of capacity
building and raising awareness is ongoing.
3.9.4. Promotion of arabica production at lower altitudes (14001500 m)

Because CWD has not been observed on arabica coffee and also confirmed
through inoculations that it is not infected by the pathogenic strain present in
Uganda, it seemed logical to promote the cultivation of arabica in the robusta
areas where CWD is prevalent. Trials were initiated using three Catimor varieties, which are resistant to CLR but highly susceptible to CBD and cannot
be grown successfully at high altitude. The quality profiles of the varieties at
various altitudes have been assessed, and results show that some areas 1400
to 1500 m could produce good-quality arabica.
3.9.5. Training of trainers

The need for information on the managements of CWD and the dissemination of this information has been high. In order for farmers and extensionists
to manage the disease, they needed information from research. Lack of information at the beginning allowed the disease to spread rapidly. As information became available, there was need to increase information flow between
farmers, researchers and the extension systems to manage the disease more
effectively.
Under the CFC-funded project, CORI at Kituza together with their
partners CABI conducted a Training of Trainers in some disease recognition and management based on available information. CORI, in conjunction
with CABI Africa, trained 75 extension staff of the MAAIF and an additional
25 extensionists belonging to the National Union of Coffee Agribusinesses
and Farm Enterprises coming from all coffee regions including arabica areas and from those areas still free from the disease These later groups were
trained in the recognition of disease symptoms to allow early detection of
the disease and hence improved surveillance. This is crucial for a quick response to enable effective control to be applied before the disease is widespread. These extensionists now form a basis for farmer training on CWD
management.
3.9.6. Participation of leaders in the CWD control programme

Under the decentralized system of governance, the districts operate as small
units with power to formulate and implement policies affecting their areas
46
Hakiza et al.
and to recommend to the central government issues which require new legislations. It was observed that leaders in the districts would be useful partners in the campaign against CWD and therefore should be sensitized on the
problem. The aim of sensitization was for them to appreciate the magnitude
of the problem and its effect on the coffee farmers and the local communities
in general. It was also essential to enlist their participation in sensitization
and mobilization of farmers to try to manage the disease particularly where
by-laws are required. Successful sensitization workshops for local leaders
have been carried out in ten districts to date and are ongoing and have allowed them to influence their respective communities and mobilize and sensitize farmers effectively.
3.9.7. Information dissemination

Through mass media
Radio programmes in Luganda for farmers and other stakeholders go on air
every Sunday morning at 8:30 AM by UCDA on Radio Simba/CBS. Information on all aspects of coffee production, crop protection in particular CWD
and coffee trade are covered in the programme. The programmes have been
sustained for more than 2 years now, covering about 30,000 listeners each
time the programme goes on air.
Farmer field schools approach to information dissemination
The traditional extension service is where the service provider/extension
worker is the messenger between research and farmer (Ladela, 2001). The
farmer is the passive recipient of this information of which he has never
taken part in its generation; consequently, many good recommendations
have never been adopted. However, farmers are known to possess a lot of
information, which they have acquired over time through experience (Marseden, 1994). Farmers too have the ability to analyse situations and make
rational decisions.
Information dissemination through the traditional extension system is
very slow because the number of extension workers is quite low compared
to the number of farmers, whom they deal with one by one. To enhance the
farmers ability and to accelerate information dissemination, a farmer field
school (FFS) approach was initiated. Through participatory approaches,
farmers work in groups and have the advantage of learning from each other
as they share information. In an FFS, farmers make regular field observations, relate their observations to the ecosystem and apply their previous experience and any new information to make management decisions on their
crop guided by the extension worker.
A coffee FFS follows the production cycle of the crop. It consists of a
group, usually 20 to 30 farmers, who set up a study field. The group is responsible for the care and maintenance of the study field from soil preparation to harvesting/post harvest. Such schools have been started in
47
Bugiri, Kayunga, Mukono, Kiboga, Masaka and Rakai districts. Each of

these districts has at least two active field schools. Information dissemination through FFS has been so successful through open field days hosted by
these farmers, so that in all the six districts, extension workers who have
been on this coffee programme are now using the approach for other crops/
commodities.
The farmers are involved in training other farmers, for instance, the field
school in Rakai (Jjongeza Coffee Farmer Field School) trained more than 500
out-grower farmers of Kaweri Coffee Company in CWD management and
other good agricultural practises in 2006.
The creation of more FFS will greatly enhance the farmers knowledge
and give a chance for most farmers to get the much needed production information within relatively short time.
Coffee production campaign
Coffee stakeholders showed concern about the declining volume of coffee
exports from Uganda. Consequently, a number of meetings/workshops
were convened in Kampala during 2006, with the goal of improving coffee
productivity, so as to raise the volume of coffee from 2 million bags to 4 million bags by the year 2015. A plan of action was drawn up for the various
stakeholders, which included research, extension, input stockists and others. Control of CWD is essential if the programme is to succeed. Resistant
varieties would greatly impact on the progress to increase production. The
role of research is to ensure available resistant materials are evaluated, multiplied and distributed to farmers. At the same time, information dissemination through various print materials in English and main local languages on
CWD management and other agronomic and crop protection problems is
emphasized. The campaign began in 2006, and this is the third year.
3.10. Conclusion
Until resistant planting materials become widely available, farmers are
strongly advised to continue to manage CWD through sanitary and cultural
measures. These include regular inspection of their fields to identify trees
with early symptoms, followed by uprooting and burning of the affected
bushes in situ. Ensuring there is adequate soil fertility through addition of
fertilizers or manure is also advocated. Preventing introduction of inoculum
into the field or at least delaying introduction of inoculum into their fields
through infected planting materials or contaminated soil is recommended,
and thus, the source of planting materials should be from well-run, reputable
nurseries where the soil is sterilized. Farmers should avoid use of volunteer
seedlings from forests that could be harbouring the disease without showing
symptoms. More than 40,000 farmers have already been trained and sensitized to the disease, and many are currently being trained in FFS in Rakai,
Bugiri and Mukono.
48
Hakiza et al.
The information gap that existed at the beginning of the CWD epidemic
has currently been bridged. Participatory approaches to technology generation and dissemination have played a major role and are being adopted by
extension workers. Local leaders have been sensitized about CWD and their
pivotal role in the control of the disease. These leaders have realized that
without coffee, their revenue base would remain small, and they are showing more interest in the crop than before, particularly in Masaka district and
a number of others.
Recommendations for CWD management are now based on research
findings. There are sound scientific reasons why uprooting and maintaining
other sanitary practises are essential and are being communicated to farmers and extension staff. It is essential that information dissemination should
be maintained to integrate new information from research into management
strategies and keep farmers and extensionists updated.
Breeding work should continue to establish the nature of resistance in
the current selected CWD-resistant cultivars to provide a basis for planning
breeding strategies for durable resistance to CWD. Attempts to have DNA
finger printing of the selected resistant clones for identifying them and for
patent rights should continue to logical conclusions.
Acknowledgements
We gratefully acknowledge all the farmers and extension staff that participated in the programme with us; without their cooperation, patience, time
and interest, no useful results would have been achieved. We would also like
to thank UCDA for its commitment to this commodity and to all the development partners (CFC, ICO, EU, DFID, CIRAD etc.) for providing funds at the
most needed time. In addition, we would like to acknowledge CORI staff for
their dedication, in particular Mr Sammy Olal for his hard work in taking
photographs and processing data.
References
Anon (1993) Background to the Budget for the Financial year 1993/1994. Ministry of
Finance and Economic Planning. Uganda government.
Banana Bacterial Wilt Task Force Report (2002) Report for the Eradication of Banana
Bacterial Wilt Disease in Kayunga and Mukona Districts. National Agricultural Research Organisation, Uganda, pp. 2021.
Banana Bacterial Wilt Task Force Report, (2003) The National Research and Development Strategy and Action Plan for Control of Banana Bacterial Wilt. National Agricultural Research Organisation (NARO) Uganda.
Booth, C. (1971) The Genus Fusarium. CMI, Kew, Surrey. Pp 28, 107, 118121.
49
Coste, R. (1992) Coffee. The Plant and the Product. Macmillan Press Lt., London,
Pp 58.
Coffee Research Institute (CORI) Annual Report for 19961997.
CWD Task Force Report (1999) The National Research and Development Strategy and
Action Plan for Control of Coffee Wilt Disease NARO, pp. 78.
De Graaf, J. (1986) The economics of coffee: economics of crops in developing countries, No 1. Pudoc, Wageningen, The Netherlands, pp. 2126.
Flood, J. (1996) A study of the tracheomycosis or vascular wilt disease of coffee in Zaire
(IMI). Report presented to Zairian Coffee Organisation (OZACAF). Pp 13.
Flood, J. (2005) Epidemiology and variability of Gibberella xylarioides, the coffee wilt
pathogen. Final Technical Report. CABI Bioscience, p. 32.
Gerlach, W. and Nirenberg, H. (1982) The genus Fusarium a pictorial Atlas, Berlin,
Pp 79, 297300.
Hakiza, G.J., Kyetere, D.T. and Olal, S (2004) Mode of penetration and symptom expression in Robusta coffee seedlings inoculated with Gibberella xylarioides, the cause
of coffee wilt disease in Uganda. Proceedings of the 20th International Conference
on Coffee Science. Bangalore, India. pp. 12311234.
Jameson, J.D. (Ed) 1970 Agriculture in Uganda. Oxford University Press, London.
Ladela, A.A. (2001) Beyond training and visit: a sustainable extension approach for
Africa through a phased participatory extension education system. African Crop
Science Conference Proceedings 5, 805810.
Leakey, C.L.A. (1970) The improvement of Robusta coffee in East Africa. In: Leakey,
C.L.A. (Ed) Crop Improvement in East Africa. CAB. pp. 250277.
Lukwago, G. and Birikunzira, J.B. (1997) Coffee wilt disease (tracheomycosis) and its
implications on Ugandas economy. African Crop Science Conference Proceedings. Pretoria. pp 969974.
Marseden, D. (1994) Indigenous management and improvement of indigenous knowledge. In: Scoones and Thompson (Eds) Beyond Farmer First. Intermediate Technology Publications. London. pp. 5256.
Muller, R. (1997) Some aspects of past studies conducted in Western and Central Francophonw Africa on Tracheomycosis (Cote dIvoire, Cameroon and Central African
Republic). Proceedings of the first Regional Workshop on Coffee Wilt Disease (Tracheomycosis). International Conference Centre, Kampala, Uganda. pp. 1526.
Purseglove, J.W. (1968) Tropical Crops. Dicotyledons 2. Longmans Green and Co. Ltd
London, Harlow, pp. 458491.
Thomas, A.S. (1935) Types of coffee and their selection in Uganda. Empire Journal of
Experimental Agriculture 12, 112.
Thomas, A.S. (1947) The cultivation and selection of Robusta coffee in Uganda. Empire
Journal of Experimental Agriculture 15, 6581.
Uganda Coffee Development Authority (UCDA). Annual Reports for 1999.
Uganda Coffee Development Authority (UCDA). Annual Report for 2004.
Uganda Coffee Development Authority (UCDA). Annual report for 2005.
Uganda Coffee Development Authority (UCDA). Annual report for 2006.
Wrigley, G. (1988) Coffee. Tropical Agriculture Series. Longmans, London.
4
1
A. Girma,1 A. Million,1 H. Hindorf,2 Z. Arega,1

D. Teferi1 and C. Jefuka1
Jimma Agricultural Research Center, Ethiopian Institute of Agricultural Research,

PO Box 192, Jimma, Ethiopia
e-mail: girma.adugna@yahoo.com
2
INRES-Phytomedizin, University of Bonn, Nussallee 9,
53115 Bonn, Germany
4.1. Introduction
Unlike Carl Linnaeus, contemporary scholars and writers about coffee authentically confirm that the primary centre of origin and genetic diversity
of Coffea arabica L. (arabica coffee) is Ethiopia, and some of them wish to
rename this species as Coffea abyssinica. Wild forest coffee is still found today in the south-western parts of the country. The intact afromontane rainforests of that region are centres of biodiversity of flora and fauna. Arabica
coffee has been disseminated from this original place to many other countries like Brazil, Colombia and Vietnam, which are the leading producers
and suppliers of coffee beans to the world market. Coffee has rapidly become one of the prominent commodity crops in global transactions, and
it stands first in earning foreign currency for many countries including
Ethiopia.
Ethiopia has the longest tradition of coffee production and consumption
in the world with a traditional way of cultivation and the performance of
inimitable coffee ceremony. The country is also the oldest exporter of the
worlds finest original arabica coffees, some of which are unique and worldrenowned speciality coffees, such as Harrar, Jimma, Limu, Sidamo, Yirgacheffe,
Ghimbi and Lekempt, which receive premium prices. Coffee is crucial to the
Ethiopian economy because it contributes 10% of the countrys gross domestic product and generates more than 40% foreign exchange earnings.
Coffee remains crucial to the biological, social and economical values of
the country, but despite being the birthplace of coffee, Ethiopia has not ex-
50
51
ploited and benefited from the crop to the best of its genetic and ecological
potential. Coffee production systems remain predominantly traditional, and
diseases and insect pests greatly reduce the productivity and quality of the
produce.
This chapter reviews the status of coffee wilt disease (CWD), including
its occurrence, distribution and importance on arabica coffee in Ethiopia, and
highlights some of the factors accelerating the disease and efforts made to
contain the problem.
4.2. The Importance of Coffee in Ethiopia

Ethiopia stands first in Africa and sixth in the world with regard to production of arabica coffee, and substantial increases have been achieved from
245,650 t in 2000 to 320,000 t in 2007 (Fig. 4.1). More than 90% of the production is from the garden, semi-forest and forest coffee systems of small-scale
farmers, whereas the remaining 10% comes from large-scale plantation coffee. Coffee is by far the number one export crop for Ethiopia, contributing
about 40% of the countrys foreign currency income. The total export revenue from coffee was previously much higher (ranging from 60% to 70%), but
since 2000, the proportion of exchange earnings of the crop has declined to
around 40%. This decline is due to substantial price increases and increased
production of other export commodities such as oil crops, pulses, leather and
leather products, fruits and flowers for the international market (Kassahun
and Getnet, 2008). Coffee is directly or indirectly a source of livelihood for
more than about 25 million people engaged in production, processing and
marketing of the crop.
In addition, the potential economic value of the Ethiopian wild coffee
gene pool as a resource for international coffee breeding programmes has
350,000
308,300
285,890
267,300
300,000
Production (tonnes)
246,653
320,000
301,500
250,000
248,890
200,000
150,000
192,896
100,000
50,000
0
2000
2001
2002
2003
2004
2005
2006
2007
Fig. 4.1. Coffee production status over the past 7 years in Ethiopia (20002007) (Kassahun and
Getnet, 2008).
52
Girma et al.
been estimated to amount to c. US$0.5 and US$1.5 billion per year (Hein and
Gatzweiler, 2006). This genetic resource provided a unique opportunity to
develop more than 15 coffee berry disease (CBD)-resistant cultivars through
rigorous research within a 5-year period. The immediate provision of seeds
from these resistant materials to coffee farmers in the early 1970s saved the
countrys economy from catastrophe. Coffee yields have steadily increased
over the last four decades from 260 kg/ha in the early 1960s to more than
900 kg/ha during the late 1990s. Total Ethiopian production has risen to
more than 3.8 million bags (over 2 million bags for export) with a stable overall quality; the amount of wet processed mild coffee has increased since liberalization. Admasu (2006) estimated that a compound value amounting to
US$270,317,647 (2297.7 million birr) was gained in the period 19722001 due
to the increase in coffee yield by planting the improved cultivars on about
20% of land area.

Although there has been a considerable increase in coffee yields and overall
production, the average coffee productivity in Ethiopia remains low (about
500 kg/ha per year) as compared to the world standard and to other coffeeproducing countries, namely, Brazil, Vietnam, India and Kenya. Coffee is
largely produced by small-scale subsistence farmers and is a low input
output crop in Ethiopia. It grows under age-old traditional production systems (forest, semi-forest and garden). Agronomic practices and crop husbandry remain conventional, and the adoption and diffusion of improved
technologies, e.g. high-yielding cultivars and better pre- and postharvest
management activities, have been slow. Coffee diseases such as CBD, CWD
and coffee leaf rust (CLR) cause severe crop losses. Coffee insect pests, mainly
antestia, leaf miners and coffee berry borer, cause damage in most coffeegrowing regions of the country. Volatility and fluctuation in coffee prices, be
it at local or international markets, can have enormous impacts to farmers
livelihoods and to investment decisions in the coffee industry as a whole.
4.4. Most Significant Disease Problems and Pests

Fifteen diseases (Eshetu et al., 2000) and more than 45 insect pests (Million,
2000) have been recorded on C. arabica in Ethiopia, and among others, CBD,
CWD and antestia are foremost in reducing the quantity and quality of coffee in the country. CBD (Colletotrichum kahawae) can cause up to 100% yield
loss, with national average losses varying between 25% and 30% (Van der
Graaff, 1983; Merdassa, 1986; Eshetu et al., 2000). CWD is prevalent in almost
all coffee-growing regions, with national average incidence and severity of
28% and 5%, respectively. CLR (Hemileia vastatrix) occurs in most coffee areas with varying intensities but is more prominent on Harrar coffee, and as
high as 27% severity was reported (Eshetu et al., 2000). Antestia (Antestiopsis
53
intricata) is the major coffee pest inflicting considerable damage, amounting

9% berry fall and 48% darkened coffee beans (Million, 2000).
4.5. Brief History of CWD in Ethiopia

Historically, CWD on C. arabica was first observed in Ethiopia (Keffa province) by Stewart (1957), who described the wilting symptom and also identified the causal organism to be Fusarium oxysporum f.sp. coffeae. Lejeune (1958)
also noted the presence of this disease on arabica coffee. Later, based on comparative studies of the isolates collected from dying arabica coffee trees from
different origins and different Coffea spp., the causal was confirmed to be
Gibberella xylarioides Heim & Saccas, of which Fusarium xylarioides Steyaert
is the imperfect (conidial) state (Kranz and Mogk, 1973). Van der Graaff and
Pieters (1978) reported that this pathogen caused a typical vascular wilt disease and was the main factor of coffee tree death in Ethiopia.
Subsequent surveys accompanied by isolation and identification demonstrated occurrence of G. xylarioides (F. xylarioides) in major coffee-growing
regions of south and south-west Ethiopia (Van der Graaff and Pieters, 1978;
Merdassa, 1986; Girma, 1997; Eshetu et al., 2000). Even in some localities like
Bebeka and Teppi, CWD outbreaks were noticed in large-scale plantation coffee (Girma, 1997, Eshetu et al., 2000). During recent years, the prevalence
and importance of CWD have been markedly increasing throughout coffeeproducing areas of the country (Girma et al., 2001; Girma and Hindorf, 2001;
CABI, 2003; Girma, 2004; Oduor et al., 2005).
CWD (also known as tracheomycosis) kills the whole coffee tree within a
short period. Infected arabica coffee plants (of any age) usually occur singly
or at random in groups in the affected fields. The earliest symptom of infection on both mature and young coffee trees is epinasty of leaves on some
branches in the lower tree canopy. The leaves become brown or dark brown
within two or more weeks and finally drop off the branches. These external symptoms most frequently start on one side of an infected coffee tree
(unilateral or partial wilting) but eventually advance so that the whole plant
is affected. Later in the season, completely wilted trees become desiccated
and severely defoliated (Plate 4). These trees cannot be easily pushed over
and uprooted as opposed to coffee trees that have died from root rot disease (Armillaria mellea). Internally, brown or blueblack discoloured bands
are seen on the exposed wood of the stem (Plate 8). Dark fruiting bodies
(perithecia) of the pathogen can be observed on the bark of stems especially
around the collar region (Plate 7) and occasionally on branches of dead coffee
trees (Girma et al., 2001). Similar symptom description was given by Van der
Graaff and Pieters (1978) and Flood (1997).
The characteristic partial wilting symptom accompanied by discoloured
internal tissues effectively facilitates diagnosis and recognition of infected
coffee trees in the field. This early detection allows rouging out of the infected trees early in the season before fungal sporulation at the advanced
stage of pathogenesis (Girma and Hindorf, 2001).
54
Girma et al.
4.6. Coffee Wilt Distribution in Ethiopia

Coffee production in Ethiopia is broadly grouped into four systems on the
basis of biological diversity of the species and level of management, namely,
forest, semi-forest, garden and plantation coffee (Meyer, 1965; Paulos and
Demel, 2000). Forest coffee (sometimes referred to as wild coffee) regenerates spontaneously from self-sown seedlings as in the understorey of intact multilayered tropical rainforests. The system is characterized by a very
rich genetic diversity of both coffee and other flora and fauna. This coffeegrowing system is situated in the west and the south-west of the country.
Semi-forest or semi-domesticated coffee is simply derived from the wild
forest coffee through human intervention and domestication, i.e. by thinning the dense overstorey of the forest trees and slashing the understorey
bushes and shrubs. The open areas are filled mostly by transplanting naturally propagated coffee seedlings from under the mother trees resulting in
irregularly spaced trees with a high population density. Slashing is practiced
once a year just before or during coffee picking season. Semi-forest coffee
also contains diverse populations of coffee and contributes 35% to the total
coffee production of the country, although its productivity is low (Paulos
and Demil, 2000).
The garden coffee system is predominant in the southern and southeastern regions, with plots of varying size (usually less than 0.5 ha) around
farmers dwellings and predominantly intercropped with a variety of fruit,
root and cereal crops. The coffee population is less diverse consisting of
many landraces. This system involves more intensive management practices
such as slashing and hoeing (three to five times per year), and to some extent,
the coffee trees are pruned, mulched and fertilized with organic materials
(Workafes and Kassu, 2000).
Plantation coffee is a relatively new farming system for coffee cultivation in Ethiopia. The system involves monoculture of plants derived from
CBD-resistant and high-yielding cultivars in nurseries and transplanting
them into well-prepared land. Improved agronomic practices such as row
planting, correct spacing, intercropping, mulching, pruning, shade regulation and to some extent advanced postharvest technologies are being applied. Higher yield and good-quality coffee are obtained from plantation
coffee.
CWD occurs in all of the above coffee production systems (Fig. 4.2)
to varying extent of damage among and within coffee fields and districts
(weredas) depending on different interacting factors, mainly susceptibility
of coffee trees, intensity of cultural practices and environmental conditions
(Merdassa, 1986; Girma and Hindorf, 2001; CABI, 2003; Girma, 2004).
4.6.1. CWD in the forest and semi-forest coffee
The occurrence of CWD was reported after assessment in four forest coffee
areas in south-west and south-east afromontane rainforests with incidence
55
Areas not visited

(low or no coffee production)
No CWD found
< 5% Infected trees on farm
5 - 15% Infected trees on farm
> 15% & <30% Infected trees on farm
30 - 50% Infected trees on farm
>50% Infected trees on farm
Fig. 4.2. Average severity (% trees infected per farm) of CWD in different zones in Ethiopia
(CABI, 2003).
ranging between 5% at Sheko and 30% at Yayu. Although it was indicated

that the damage was minimal in the dense stands of coffee (Van der Graaff,
1983; Merdassa, 1986), this was the first documented report that showed
presence of CWD on forest coffee trees. Arega (2006) also demonstrated
increasing occurrence of CWD in some forest areas like in Harenna (Bale)
and Bonga (Keffa). The mean incidence in semi-forest coffee ranged from
3.6% at Mettu to 15.5% at Gera situated in south-west coffee-producing areas
(Fig. 4.3), and the severity varied between 18.6% and 25.4% in some coffee
fields at Yirgacheffe (Fig. 4.4) (Girma, 2004). A similar situation was observed
in Bale, Jimma, Ilubabor, and West Wellega zones (CABI, 2003).
4.6.2. CWD in garden coffee

CWD is prevalent in the southern region, specifically in the three major
quality-coffee-producing districts, namely, Wonago, Kochore and Yirgacheffe
of Sidama and Gedeo zones, with highest incidence in Yirgacheffe followed
by Kochore and Wonago. The severity of wilting in the sample fields in
Yirgacheffe varied between 27.2% and 43.5% in the garden coffee as compared to that of the semi-forest coffee (Girma, 2004) (Fig. 4.4). Although the
disease was not evenly distributed in most coffee-growing areas of Southern
56
Girma et al.
Incidence (%)
100
90
Semiforest
80
Plantation
70
60
50
40
30
20
10
0
a
Jim
ich
ira
Ch
om
a
bb
To
hi
ec
yu
Ya
tu
et
pi
p
Te
ek
b
Be
ea
Localities (District)
Fig. 4.3. Incidence of CWD in the semi-forest and plantation coffee production systems under
farmers conditions in south-western Ethiopia (error bars indicate standard deviations).
Nation, Nationalities and Peoples region, the average incidence (35%) and
severity (5.0%) was significantly (P < 0.001) higher than in other regions. It
was particularly high in the Sidama and Gedeo zones, with an incidence over
90% and severity of 25% (Fig. 4.5). The incidence of CWD was also above
35% in garden coffee of West Gojam zone of Amhara regional state, but it
was very low in Wolaita (Southern Nation, Nationalities and Peoples) and
West Harerghe (Oromiya) (CABI, 2003).
60
Garden
Incidence (%)
50
Semiforest
40
30
20
10
0
Konga
Sede-Menitu
Sede-Wonata
Coffee fields
Fig. 4.4. CWD incidence (%) in the garden and semi-forest coffee fields in Yirgacheffe (southern region) (error bars indicate standard deviations).
57
Incidence
Gambella Oromia
Burji
Konso
Walita
Amaro
S. Omo
Sidama
Sheka
KAT
Kefa
Gedeo
Hadiya
Bench Maji
W.Wellega
Jimma
W.Hararghe
Bale
Illubabor
Godere
S.Wello
Amhara
W.Gojam
Oromia
N.Shewa
Severity
E.Gojam
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
Agew-Awi
Incidence & Severity
SNNP
Fig. 4.5. Incidence and severity (%) of CWD in different regions and zones of Ethiopia in 2002
(CABI, 2003).
4.6.3. CWD in plantation coffee

The disease incidence is more severe in plantation coffee such as at research
centres, on larger farmer holdings (1 to 5 ha) and in large estate commercial farms. CWD is commonly encountered in the research plots at Gera and
Jimma amounting 42.5% and 48.2%, respectively (Table 4.1). It is serious in
the farmers coffee plantations at the Gera, Chira and Gechi districts, with respective mean incidence ranging from 21.7% to 25.5%, from 32.3% to 77% and
from 35% to 60%, respectively (Table 4.2). The overall mean coffee tree loss in
the farmers plantation was more than 30%, and in total, about 10 ha of planTable 4.1. Prevalence and incidence of CWD in various experimental plots of coffee
research centres in Ethiopia.
Research
centres/stations
Number of
fields (n)
Jimma
Agaro
Gera
Mettu
Teppi
Wenago
Mean
a
SD
= standard deviations.
10
3
15
3
3
3
Incidence (%)
Range
Mean and SDa
Altitude (m)
19.882.0
5.212.1
21.061.1
23.330.9
6.513.4
5.714.6
5.282.0
48.2 23.1
8.7 3.4
42.5 18.7
27.1 5.4
10.0 4.9
9.8 4.5
24.4 17.7
1750
1650
2000
1550
1200
1850
58
Girma et al.
Table 4.2. Incidence (%) of CWD plantation coffee under farmers condition in south-west
Ethiopia.
Location
Gera
Chira
Tobba
Gomma
Gechi
Yayo
Mettu
Mean
Incidence (%)
Field
Estimated area
(ha)
Range
Mean
Gicho 1
Gicho 2
Sedi-Loya
Gure-Genji
Chira 1
Chira 2
Yachi
Kilole
Ageyu
Shashamene
Echemo
Sombo
Camp
Mine-kobba
Asendabo
Jitto
Sor
Total = 17
1.0
1.5
1.0
5.2
4.5
1.5
0.3
0.4
0.2
0.5
0.3
0.2
0.5
5.0
5.0
1.0
0.5
Total = 28.6 ha
11.535.0
8.738.0
23.927.1
38.075.0
55.089.0
14.042.0
12.120.8
14.623.9
8.327.0
12.719.4
12.515.5
25.834.2
25.070.0
15.055.0
37.778.6
11.034.0
8.033.3
8.389.0
24.5
21.7
25.5
51.5
77.0
32.3
16.5
19.3
16.1
10.8
13.6
29.2
48.9
35.0
59.7
22.5
20.4
30.9 18.2
tation coffee had been abandoned completely. The lowest percentage of the
disease was recorded in the farmers plantation at Tobba (17.3%), whereas
the highest was at Bebeka (65.2%) (Fig. 4.3). Girma et al. (2001) confirmed that
the disease was more severe in plantation coffee at Bebeka, Teppi, Gera and
Jimma. Van der Graaff (1979) remarked that some spectacular failures of the
modern plantations system could be due to G. xylarioides, and when comparisons are made across production systems, the disease is more destructive in
garden and plantation coffees than in forest and semi-forest coffee systems.
The latter two systems are composed of heterogeneous coffee populations
possessing varying levels of resistance and less human interference. However, in the former systems, characterized by relatively homogenous coffee
trees and high levels of intervention, the disease spreads from tree to tree,
from row to row and from one block to the other developing throughout the
field (Girma, 2004). A remarkable increase in CWD severity of (11.5%) was
recorded over a 6-month period in nine districts (weredas) of Gedeo and
Sidama zones of Ethiopia (CABI, 2003).
4.7. Importance of CWD in Ethiopia

For many decades, CWD was considered as a minor problem in Ethiopia, and
its impact therefore largely remained unnoticed and its effect underestimated,
59
but the losses incurred due to the disease are comparable to those caused by
CBD. With CWD, the whole tree dies and all neighbouring coffee trees die,
so there is a loss of capital to the farmer; CBD affects only cherries. In addition, CBD can be controlled relatively easily with fungicides, but CWD is a
soil-borne pathogen and this presents difficulties in the application of chemical treatments; affected fields may need to be left as fallow for some years or
other crops planted (Girma, 2004).
Coffee production (yield) at the farm level decreased by 37% (from 1482
to 932 kg per sample farm), and this led to a decline in income of 67% (from
5038 to 1651 birr). The annual national crop losses attributed to CWD was
3360 t amounting to US$3,750,976 in Ethiopia (CABI, 2003). This economic
loss coupled with difficulty to manage the disease indicates that CWD is the
second leading disease of coffee, after CBD in Ethiopia.
4.8. Factors Influencing Severity of CWD in Ethiopia

Some preliminary observations indicate that temperature, rainfall, topography, coffee tree age, shade, soil type and weeding methods have significant
effects on CWD. The incidence of CWD is higher on coffee trees that are
older, shaded, planted on loamy soil and weeded by slashing (CABI, 2003).
Above all, susceptibility of coffee cultivars and agronomic/cultural practices
have consistent effect on the occurrence and severity of the disease.
4.8.1. Susceptibility of coffee cultivars in the field

A number of researchers have reported existence of marked differences in
resistance levels in arabica coffee populations to CWD under field conditions
at various locations (Van der Graaff and Pieters, 1978; Merdassa, 1986; Girma
and Hindorf, 2001; Girma, 2004). Merdassa (1986) assessed the incidence of
the disease in single-tree progenies of different coffee accessions for 6 years
(19791984) at Gera and obtained tree loss ranging from 0.3% to 87%.
In a field at Bebeka, 23 cultivars (including four introduced Catimor
lines) were planted in a completely randomized block design with three replications and 90 trees per plot. Cultivar 785, 1185, 1785 and 4485 were uniformly attacked in all plots and showed significantly high mean death rates
of 80.0%, 72.9%, 83.4%, and 97.4%, respectively, indicating their susceptibility to coffee wilt. In contrast, the Catimor lines (1579, 1779, 1979 and 2179)
and some French collections (F-15, F-27 and F-59) had the lowest infection
levels of less than 10% (Girma, 2004). The introduced coffee lines such as
Caturra Rojo, Caturra Amerello and Catuai showed significantly (P < 0.05)
higher mean incidences of 83.0%, 80.5% and 80.0%, respectively, and were
more susceptible to CWD than the indigenous cultivars 7454, 74110, 74112,
74140 and 74165 at Teppi (Girma, 2004). In all cases, the disease developed
dramatically into large foci starting from a single tree, and spread is faster in
plots composed of susceptible trees.
60
Girma et al.
4.8.2. Agronomic/cultural practices

Besides the inherent genetic factors of coffee cultivars, certain agronomic
and routine cultural practices have long been postulated to aggravate CWD.
There is a strong association with wounding and dissemination of the fungal fruiting bodies (perithecia and ascospores) from a single infected tree to
other disease-free trees/plots. The predominant disease-spread mechanisms
mainly involve human activities including pruning, stumping (to rejuvenate
old and unproductive trees), slashing and hoeing to control weeds and transporting infected trees from one field to the other (long-distance spread). The
use of contaminated farm implements across various plots/fields also plays
a significant role in spreading the fungus inoculum. A common practice in
Ethiopia is to cut wilted trees, store them somewhere in the field or near
the houses and use for various purposes such as for firewood, for fencing
around dwelling houses or coffee farms and as a support for climbing beans.
These trees and remaining stumps harbour the fungal fruiting bodies (perithecia with ascospores) that serve as inoculum source for further infection
and initiate disease epidemics (Girma, 2004). It has been estimated that 60%
of the farmers in Ethiopia used the wood for fencing, 26% for constructing
houses and animal sheds, 10% gave surplus wilted trees to their neighbours
for firewood and 2% sold the trees (CABI, 2003).
4.8.3. Spacing and population density of coffee trees

The severity of CWD seems to be affected by spacing between coffee trees
and population densities. The results of disease assessments in a population
density trial at Gera, consisting of eight cultivars planted in a circular fashion
fan design with each circle as a spacing treatment, showed that the mean incidence of disease linearly increased and significantly rose from 16.0% for 2 2
m to 28.5% for 1 1 m spacing between trees (Table 4.3). The disease outbreak
Table 4.3. Severities (%) of CWD in arabica coffee population density trial at Gera.
Coffee
cultivars
741
744
7440
7454
74148
74158
74165
75227
Meana
a
Spacing between coffee trees (m)

2.0 2.0
1.75 1.75
1.5 1.5
1.25 1.25
1.0 1.0
Meana
29.2
12.5
13.9
30.6
8.3
19.4
4.2
9.7
16.0 a
22.2
18.1
9.7
37.5
5.6
12.5
4.2
2.8
14.1 a
41.6
25.0
18.1
50.0
12.5
11.1
15.3
9.7
22.9 ab
44.4
15.3
22.2
45.8
9.7
16.7
11.1
22.2
23.4 ab
47.2
34.7
20.9
38.9
20.8
26.4
8.4
30.6
28.5 b
36.9 B
21.1 A
16.9 A
40.6 B
11.4 A
17.2 A
8.6 A
15.0 A
Means followed with the same letter(s) are not significantly different, least significant difference values (P = 0.05) for cultivars and spacing treatments are 10.9 and 8.6, respectively.
61
started mostly at the centre of the plot where trees were planted very close
to each other (1 1 m) and radiated outwards as spacing distances between
trees increased (2 2 m). This fact may be ascribed to root-to-root contact via
which the fungus can be transmitted from infected to nearby healthy trees.
The lateral and feeder roots of arabica and robusta coffees can spread in the
soil for a distance of 1.2 to 1.8 m from the trunk (Wrigley, 1988). There is
ample evidence that G. xylarioides is abundantly recovered from root parts of
symptomatic and asymptomatic trees (Girma et al., 2001).
In addition, closely spaced trees are more liable to wounding, and cross
inoculation from diseased to healthy trees can occur during slashing or hoeing coffee fields. Almost all stems of coffee trees have wounds at the crown
level or a few centimetres above, and on average, 98% of the diseased and
95% of healthy trees were noted to have one to three wounds per coffee stem
(Girma, 2004). The wounds arose practically from the intensive slashing of
weeds in coffee fields by machetes, which is the most common method of
weed control in coffee in Ethiopia (Getachew, 1991; Tadesse, 2001). Getachew
(1991) noted that weeds are slashed frequently, sometimes more than ten
times a year, depending on the dominating weed flora in plantation coffee,
and most of the coffee trees were found to have at least one wound.
The effect of wounding was further demonstrated in young coffee
seedlings in the greenhouse. When seedlings with healthy roots were transplanted into either naturally or artificially infested soils, no wilting symptoms appeared. Infections were exhibited when the tap roots were injured and
transplanted into artificially or naturally infested soils. In addition, only
those seedlings inoculated by stem wounding through nicking with G. xylarioides-infested scalpels or by injecting conidial suspensions into the stems
became infected. The stem-nicking inoculation method also illustrates the
role of contaminated farm implements in cross-inoculating coffee trees and
disseminating the pathogen across coffee fields (Girma, 2004).
4.8.4. Replanting into the fields infested with the pathogen

Replanting susceptible coffee cultivars into already contaminated fields increases fungal inoculum, and further infection occurs in the second planting,
e.g. at Gera, where about 57.8% and 36.8% infection were recorded during the
first and second plantings in the same field (Girma, 2004; Girma et al., 2007).
Similarly, Girma (1997) estimated 40.0% tree death on young coffee replanted
as refill on sites of previously uprooted infected trees of the same coffee
line at Teppi. Among some socio-economic factors noted contributing to the
spread of CWD, particularly in Ethiopia, was frequent replacing/replanting
with several seedlings (three to eight) per uprooted wilted trees (CABI, 2003).
The infection of the young replants undoubtedly suggests that the fungus survives in stump and other debris and the soil for some 2 years after uprooting.
Perithecia of G. xylarioides, containing a great number of viable ascospores with a germination rate of 90% to 95%, are very abundant in the coffee
fields. These sexual structures are the most important sources of inoculum in
62
Girma et al.
CWD epidemics. High infection of susceptible arabica seedlings was noticed

after inoculating with field-collected ascospores, suggesting that the perithecial state is the primary source of inoculum in the field (Girma, 2004).
4.9. Country-Specific Management Practices in Ethiopia

4.9.1. Cultural control
Unlike with other coffee diseases, namely, CBD and CLR, coffee trees infected
by CWD cannot be saved. Successful control of the disease depends on the
principles of disease prevention (avoid wounding of any part of the plant) and
phytosanitation. The conventional phytosanitary approach of uprooting and
burning the whole infected coffee tree on the spot is strongly recommended
to coffee farmers to contain the disease as soon as symptoms are seen, but this
relies on early diagnosis. Use of CWD-infected trees for any purpose is prohibited, and replanting with susceptible coffee seedlings should be delayed at
least for 2 years (Girma et al., 2001; Girma, 2004). Cultural weed control activities like slashing and digging should be avoided in CWD-prone coffee fields,
and agronomic practices (pruning and stumping) that bring about wounding
in coffee trees should be done with efficiently disinfected tools. Disinfection
of farm implements such as machetes, bow saws and pruning shears with
potent disinfectants (>75% alcohol) followed by intense heating with fire is
strongly recommended to farmers whenever pruning, rejuvenating old coffee
trees and thinning newly suckers. Farmers field schools recommend growing
cover crops such as Desmodium sp. and haricot bean, which are very efficient
in suppressing weeds (so reducing the need for slashing) and as legumes, promote the growth of coffee trees. Applying ash, mulch and slashing between
plots with hand weeding around coffee trees were also promising treatments
in CWD control trials in Ethiopia (Fig. 4.6) (CABI, 2005, unpublished).
4.9.2. Chemical control

Preliminary results of on-station and on-farm participatory trials implicate
that coffee stem paint with copper oxychloride (Kocide) and weed control
with herbicide reduced CWD incidence (Fig. 4.5) (CABI, 2005, unpublished).
Fungicide paint of the cut surface of stumped coffee with Kocide is observed to
effectively protect the large wounds from infection by the CWD pathogen.
4.9.3. Biological control

The result of a recent in vitro study conducted by Muleta et al. (2007) on antagonistic effects of some rhizobacteria against the three Fusarium pathogens
including F. oxysporum, Fusarium stilboides and F. xylarioides were promising.
Of 23 bacterial isolates obtained from rhizospheres of arabica coffee trees
63
18
16
14
12
10
8
6
4
2
0
Ash
Mulch
Fungicide
paint
Fungicide
spray
Herbicide
Slash +
hand weed
Slashing
only
Fig. 4.6. Effect of different agronomic practices and control methods on CWD in Ethiopia (error bars indicate standard deviations) (CABI, unpublished).
in south-west Ethiopia, 21 significantly inhibited the mycelial spread of F.

xylarioides. Bacillus subtilis, designated as isolate AUBB20, was the most antagonistic to this pathogen and was indicated as a potential candidate for
biological control of CWD (Muleta et al., 2007).
4.9.4. Deployment of CWD-resistant cultivars

In fields affected by CWD, it is possible to observe considerable variation in
disease severity among coffee cultivars (Van der Graaff and Pieters, 1978;
Merdassa, 1986; Girma, 1997; Girma and Hindorf, 2001; Girma et al., 2001;
Girma, 2004). Cultivar SN-5, F-51/53 and 248/71 showed 100% tree loss,
whereas F-35 and F-51 had significantly (P < 0.05) lower mortality rates of
9.3% and 27.9%, respectively, at Gera (Girma, 1997; Girma et al., 2001). At
the same locality in another highly infested field where some CBD-resistant
selections (n = 30) were planted, disease incidence ranged from 12% for selection 8150 to 96% for 74304 (susceptible check) (Girma, 1997; Girma and
Hindorf, 2001).
Apparently, disease-free (tolerant) trees in the field can be selected and
evaluated for true resistance through repeated seedling inoculation tests under
controlled conditions (Chapter 10). Artificial inoculation tests have shown that
cultivars 1579, 200/71 and 8136 were resistant to CWD with low-percentage
deaths (12.7%, 15.2% and 25.2%, respectively) accompanied by long incubation
periods (Table 4.4) before symptoms appeared (Girma and Chala, 2008). Culti-
64
Girma et al.
vars 146/71, 206/71 and 8144 showed moderate CWD infection, whereas others
including Caturra and Geisha had the highest wilt severity (>90%) indicating susceptibility to the disease. There was a correlation between the lowest seedling death rates in the greenhouse and wilt severity observed in the
fields (Table 4.4). Thus, those cultivars demonstrating resistant reactions
under both field and greenhouse conditions can be recommended for use in
CWD-prone areas provided that they have other desirable traits like resistance to CBD, high yield and improved quality.
4.9.5. Training and information dissemination on CWD management

Because a large number of coffee farmers (>85%) were not aware of CWD
prior to the Regional CWD Program (CABI, 2003), the first step was training
Table 4.4. Reactions of some coffee cultivars to CWD.
Coffee cultivars
1185
1785
1579
2179
4/70
36/70
146/71
200/71
206/71
8112
8133
8136
8143
8144
F-27
F-35
Caturra
Geisha
7440c
SN-5d
Mean
LSDe (P < 0.05)
CV (%)
a
Actual value
(mean % death)
Transformed
valuea
Incubation period
(mean no. of days)
Incidence
in the fieldb
86.0
78.7
12.7
63.3
77.2
60.9
34.6
15.2
52.8
74.9
64.2
25.3
61.6
40.2
81.0
85.7
68.9
88.1
40.4
69.7
68.8
75.1 ab
67.9 ah
16.9 s
53.4 io
62.0 bl
56. fn
35.1 qr
20.3 s
46.1 mq
63.2 ak
54.1 gn
29.6 rs
52.7 jp
39.1 or
67.0 aj
70.9 ae
59.2 dm
73.9 ac
38.7 pr
56.7 en
58.3
14.4
21.8
90.0 op
80.0 p
157.5 a
140.8 ad
117.5 dm
92.5 np
122.5 dk
152.5 afb
125.0 dj
112.5 fo
122.5 dk
150.0 ac
125.0 dj
137.5 ae
90.0 op
97.5 lp
130.0 bh
97.5 lp
135.0 af
119.2 dl
115.5
23.5
17.9
75.0
75.0
10.2
20.5
56.2
15.3
68.4
28.2
48.9
63.1
19.2
29.4
42.4
37.0
10.9
26.2
74.5
29.1
20.3
99.9
The actual data (2 years) were transformed to arcsine square root values before analysis.
Means followed with the same letter(s) are not significantly different from each other.
b
CWD incidences (mean %) summarized from various fields and localities (Girma, 1997; Girma
and Hindorf, 2001; Girma, 2004).
c
Resistant/tolerant cultivar.
d
Susceptible checks.
e
LSD = least significant difference.
Brochure (Amharic)
20,000, 23%
65
Brochure (Oromiffa)
Leaflets
6500, 8%
Caps, T-shirts
Posters
5000, 6%
5000, 6%
50,000, 57%
Fig. 4.7.
Proportions of CWD publicity materials produced and disseminated in Ethiopia.
extension agents who in turn trained coffee farmers about the disease. The
training courses included diagnosis and identification of CWD-infected trees
based on symptoms and signs of the pathogen, transmission/spread mechanisms and control methods. Besides, a large number different types of publicity materials such as leaflets, brochures, manuals, posters, wall calendars,
T-shirts, caps and stickers with brief CWD information were produced and
disseminated to almost all coffee stakeholders in the country. Exercise books
describing CWD were provided to school students in the vicinity of most
affected areas (Fig. 4.7). There have been publications in newspapers and
bulletins and broadcasts on television and radio (national and local FM) notifying farmers of CWD prevention and control. Efforts were made to raise
awareness and sensitize policy/decision makers at various levels via conferences and national workshops. Participatory on-station and on-farm trials
and season-long farmers field schools on CWD management practices were
conducted, and farmers field days were organized to popularize promising
results of the trials to coffee farmers in highly affected districts (weredas) in
south and south-west Ethiopia ( see Chapter 11).
4.10. Conclusion
The primary foci of CWD in Ethiopia are not known because the disease was
reported to occur concurrently in many areas of the country. In general, it appears that CWD existed for many years as an endemic disease of C. arabica but
has gained importance over time in almost all coffee-growing regions. The
nationwide biological survey of CWD showed that on average, 27.9% of 1607
sample coffee farms were affected, with disease incidence ranging from 15%
to 34.0% and disease severities varying between 1.3% and 5.0% (CABI, 2003;
66
Girma et al.
Oduor et al., 2005). However, it has been realized that CWD incidence was
relatively higher and more severe in the garden-based and plantation farming systems than in the semi-forest and forest coffee production systems.
Among other factors, varietal composition and human activities profoundly
aggravate the distribution and significance of the disease in Ethiopia.
The soil-borne nature of the pathogen and perennial character of coffee
have made management of the disease difficult through the conventional
control approach of uproot and burn infected trees at the spot. Avoidance
of using infected trees for firewood, for construction of huts and fences or
for other agricultural uses and avoidance of immediate replanting/replacing
with susceptible coffee seedlings have been recommended in Ethiopia (Girma
et al., 2001; Girma, 2004). Any practice that create wounds in coffee trees like
slashing and digging of coffee fields should be avoided, whereas hand weeding around the trees, spraying herbicides, application of mulch and growing cover crops such as Desmodium sp. are encouraged. All farm implements
need to be efficiently disinfected for use in coffee fields. Knowledge transfer
and information dissemination especially descriptions of symptoms have
enhanced the early diagnosis of this disease and have greatly contributed to
containing the disease. In fact, in the short term, sensitization of major coffee
stakeholder of the threat of CWD remains one of the key practices in managing this disease because early diagnosis can aid management.
The longer-term prospects of successful management of CWD depend
principally upon employing resistant coffee cultivars. In this regard, the genetic variability of arabica coffee populations presents a great opportunity
to develop a number of CWD-resistant varieties. Some cultivars, such as
1579, 200/71 and 8136, have shown high-resistance levels in artificial seedling inoculation tests that well correlated with that of natural infection in
the field. To exploit the enormous genetic potential in the control of CWD,
independent selection and screening programme should be designed and
implemented as experienced in the ever successful CBD programs.
In addition, from a research perspective, biocontrol agents including
Rhizobacteria, Trichoderma spp. and other potential microorganisms should
be tested against CWD pathogen as for other Fusarium spp.
Acknowledgements
We appreciate CABI for the excellent coordination and execution of the Regional Coffee Wilt Programme and the European Union and the Common
Fund for Commodities for funding and fully supporting implementation of
the programme. The German Academic Exchange Service is also acknowledged for financial support of the research work conducted on CWD in
Ethiopia. This work would not have been possible without the commitment
and determination of the staff of Jimma Agricultural Research Centre in particular and the Ethiopian Institute of Agricultural Research in general. Dr
J. Flood (CABI, UK) is indeed highly appreciated for critically reading and
improving this chapter.
67
References
Admasu, S. (2006) Economic analysis of conserving wild coffee genetic resource in
southwestern Ethiopia. PhD thesis. University of Bonn, Bonn, Germany.
Arega, Z. ( 2006) Diversity of arabica coffee populations in afromontane rainforests of
Ethiopia in relation to Colletotrichum kahawae and Gibberella xylarioides. MSc
thesis. School of Graduate Studies, Department of Biology, Addis Ababa University, Addis Ababa, Ethiopia.
CABI (CAB International) (2003) Surveys to assess the extent and impact of coffee wilt
disease in East and Central Africa. Final technical report. CABI Regional Centre,
Nairobi, Kenya. pp. 149.
Eshetu, D., Teame, G. and Girma, A. (2000) Significance of minor diseases of Coffea
arabica L. in Ethiopia: a review. In: Proceedings of the workshop on control of coffee berry disease (CBD) in Ethiopia. Addis Ababa, Ethiopia, pp. 5865.
Flood, J. 1997. Tracheomycosis or vascular wilt disease of coffee in Uganda. Report
Getachew, Z. (1991) A study of coffee weed flora and possible control measures in
coffee state farms of Kaffa, southwestern Ethiopia. MSc thesis. Alemaya University
of Agriculture, Alemaya, Ethiopia.
Girma, A. (1997) Status and economic importance of Fusarium wilt disease of arabica
coffee in Ethiopia. In: Hakiza, G.J., Birkunzira, B. and Musoli, P. (eds.) Proceedings
of the First Regional Workshop on Coffee Wilt Disease (tracheomycosis). International Conference Centre, Kampala, Uganda, pp. 5361.
Girma, A. (2004) Diversity in pathogenicity and genetics of Gibberella xylarioides (Fusarium xylarioides) populations and resistance of Coffea spp. in Ethiopia. PhD dissertation. University of Bonn, Bonn, Germany.
Girma, A., Flood, J., Hindorf, H., Bieysse, D., Simons, S. and Mike, R. (2007) Tracheomycosis (Gibberella xylarioides) a menace to world coffee production: evidenced
by cross inoculation of historical and current strains of the pathogen. In: Proceedings of the 21st International Scientific Conference on Coffee Science (ASIC). Montpellier, France, pp. 12681276.
Girma, A. and Hindorf, H. (2001) Recent investigation on coffee tracheomycosis, Gibberella xylarioides (Fusarium xylarioides) in Ethiopia. In: Proceedings of the 19th
International Scientific Conference on Coffee Science (ASIC). Trieste, Italy, pp.
12461252.
Girma, A. and Chala, J. (2008) Resistance levels of arabica coffee cultivars to coffee
berry disease, coffee wilt and leaf rust diseases in Ethiopia. In: Proceedings of the
12th Conference of the Crop Science Society of Ethiopia (CSSE). Addis Ababa,
Ethiopia. Sebil 12, 92103.
Girma, A., Mengistu, H. and Hindorf, H. (2001) Incidence of tracheomycosis,
Gibberella xylarioides (Fusarium xylarioides) on arabica coffee in Ethiopia. Journal
of Plant Diseases and Protection 108, 136142.
Hein, L. and Gatzweiler, F. (2006) The economic value of coffee (Coffea arabica) genetic resources. Ecological Economics 60, 176185.
Kassahun, H. and Getnet, G. (2008) Performance, structure and prospects of coffee marketing in Ethiopia. In: Adugna, G., Bellachew, B., Shimber, T., Taye, E. and Kufa, T.
(eds.) Coffee Diversity and Knowledge: Proceedings of a national workshop four
decades of coffee research and development in Ethiopia. Addis Ababa, Ethiopia,
pp. 416423.
Kranz, J. and Mogk, M. (1973) Gibberella xylarioides Heim & Saccas on arabica coffee
in Ethiopia. Phytopathology Z. 78, 365366.
68
Girma et al.
Lejeune, P. 1958. Rapport au Gouvernement Imprial dEthiopie sur la production cafire. FAO, Rome. 49 pp.
Merdassa, E. (1986) A review of coffee diseases and their control in Ethiopia. In: Abate, T.
(ed.) Proceedings of the First Ethiopian Crop Protection Symposium. Institute of
Agricultural Research, Addis Ababa, Ethiopia, pp. 187195.
Meyer, F.G. (1965) Notes on wild Coffea arabica from southwestern Ethiopia, with some
historical considerations. Economic Botany 19, 136151.
Million, A. (2000) Significance of arthropod pests of coffee in Ethiopia. In: Proceedings
of the Workshop on the Control of Coffee Berry Disease (CBD) in Ethiopia, Addis
Ababa. Institute of Agricultural Research, Addis Ababa, Ethiopia, pp. 6671.
Muleta, D., Assefa, F. and Granhall, U. (2007) In vitro antagonism of Rhizobacteria isolated from Coffea arabica L. against emerging fungal coffee pathogens. Engineering
in Life Sciences 7, 111.
Oduor, G., Phiri, N., Hakiza, G., Million, A., Asiimwe, T., Kilambo, D.L., Kalonji-Mbuyi,
A., Pinard, F., Simons, S., Nyasse, S. and Kebe, I. (2005) Surveys to establish the
spread of coffee wilt disease, Fusarium (Gibberella) xylarioides, in Africa. In: Proceedings of the 20th International Scientific Conference on Coffee Science (ASIC).
Bangalore, India, pp. 12521255.
Paulos, D. and Demel, T. (2000) The need for forest germplasm conservation in Ethiopia and its significance in the control of coffee diseases. In: Proceedings of the
Workshop on Control of Coffee Berry Disease (CBD) in Ethiopia. Addis Ababa,
Ethiopia, pp. 125135.
Stewart, R.B. (1957) Some plant diseases occurring in Kaffa province, Ethiopia. Imperial
Ethiopian College of Agriculture and Mechanical Arts, Alemaya, Ethiopia.
Tadesse, E. (2001) Weed flora and weed control practices in Ethiopia: a review. In: 19th
International Scientific Colloquium on Coffee (ASIC). Trieste, Italy.
Van der Graaff, N.A. and Pieters, R. (1978) Resistance levels in Coffea arabica L. to
Gibberella xylarioides and distribution pattern of the disease. Netherlands Journal
of Plant Pathology. 84, 117120.
Van der Graaff, N.A. (1979) Breeding for stable resistance in tropical crops: strategies
to maintain balanced pathosystems in modern agriculture. FAO Plant Protection
Bulletin 27, 16.
Van der Graaff, N.A. (1983) Durable resistance in perennial crops. In: Lamberti, L.,
Waller, J.M. and van der Graaff, N.A. (eds.) Durable resistance in crops. Plenum
Press, New York, pp. 263276.
Workafes, W. and Kassu, K. (2000) Coffee production systems in Ethiopia. In: Proceedings of the Workshop on Control of Coffee Berry Disease (CBD) in Ethiopia. Addis
Ababa, Ethiopia, pp. 99107.
Wrigley, G. (1988) Coffee. Tropical agriculture series. Longman Scientific and Technical Publisher, New York.
D.L. Kilambo, N.M. Nghoma, J.M. Teri and L. Masumbuko

Tanzania Coffee Research Institute (TaCRI), Lyamungu PO Box 3004, Moshi, Tanzania
www.tacri.org
5.1. Introduction
Coffee species grown in Tanzania are Coffea arabica and Coffea canephora. Arabica coffee is grown in the northern and southern highlands, Tarime and the
west (Kigoma), whereas robusta coffee is grown in the western zone, mainly
in the Kagera region. The total area under coffee in Tanzania is estimated to
be 265,343 ha (Tanzania Coffee Board, 2006), which is composed of 77% arabica coffee and 23% robusta coffee. Most of the coffee farms produce around
250 kg of clean coffee per hectare.
Since its appearance in Tanzania in 1997, coffee wilt disease (CWD) has
clearly demonstrated both its ability to spread rapidly to new areas and cause
serious losses. From 1997 to date, 44 of 100 wards in the Bukoba, Muleba,
Misenyi and Karagwe districts are reported to be affected by the disease,
causing monetary losses of $197,551. In Tanzania, control of the disease has
involved uprooting and burning of the affected trees in situ, with particular
emphasis on early diagnosis of the disease to reduce the risk of spread to
other trees and farms.
This chapter describes the history of the disease in Tanzania, its geographical distribution, economic importance and control measures.
5.2. Importance of Coffee in Tanzania

Coffee is currently Tanzanias second most important export among traditional exports accounting for 23% of the countries total foreign exchange of
US$267.1 (Fig. 5.1). Coffee contributed about US$61.4 million to export earnings (Tanzania Coffee Board, 2006). The coffee industry in Tanzania provides
employment to 420,000 families, of which 90,000 are from the robusta-growing
69
70
Kilambo et al.
Fig. 5.1. Contribution proportional of Tanzania Traditional exports in 2006. (From the Tanzania
Coffee Board.)
areas of the Kagera region, and benefits about 2,000,000 other people indirectly (Baffes, 2003).

There are a number of constraints that face coffee farmers in Tanzania. Productivity tends to be low with most of the coffee farms producing around
250 kg of clean coffee per hectare. Some of the factors that contribute to low
coffee productivity include fluctuating world coffee prices and low farm gate
prices resulting in small returns to coffee producers. In contrast to these low
incomes, the costs of inputs such as fertilizers, fungicides and insecticides are
extremely high, making them unaffordable to most of the farmers. Yet, these
inputs are needed since the crop is threatened by a range of insect pests and
diseases. In addition, due to a lack of information, many farmers use inappropriate husbandry practices.
5.4. Most Important Pests and Disease Problems

The production of arabica coffee in Tanzania is to a great extent limited by diseases and pests. These include coffee berry disease (Colletotrichum kahawae Waller
and Bridge sp. nov.), coffee leaf rust (CLR; Hemileia vastatrix Berk and Br) and in
some areas Fusarium bark disease (Fusarium stilboides). Among the insect pests,
white stem borer (Monochamus leuconotus), antestia bug (Antestiopsis lineaticollis)
and coffee berry borer (Hypothenemus hampei Ferrari) are the most serious pests
of arabica coffee in Tanzania.
For robusta, coffee production is influenced by disease problems such
as red blister disease on the berries caused by Cercospora coffeicola Cooke and
Berk and CLR (H. vastatrix Berk and Br), which may cause serious leaf defoliation and dieback. The major insect-pest is coffee berry borer (H. hampei
Ferrari), which if left unmanaged can cause yield losses of up to 90%.
71
Robusta coffee production trend in Kagera
Metric tonnes
30000
25000
20000
15000
10000
5000
0
93
2/
9
19
94
3/
9
19
95
4/
9
19
96
5/
9
19
97
6/
9
19
98
7/
9
19
99
8/
9
19
00
9/
9
19
01
0/
0
20
02
1/
0
20
03
2/
0
20
04
3/
0
20
05
4/
0
20
06
5/
0
20
07
6/
0
20
Seasons
Fig. 5.2.
Board.)
Robusta coffee production trends in Kagera for 15 years. (From the Tanzania Coffee
More recently, robusta production in the Kagera region of Tanzania has

been badly affected by a vascular wilt disease coffee (CWD). Since its appearance in 1997, coffee production trends in Kagera has been highly variable
(Fig. 5.2).
5.5. Brief History of CWD in Tanzania

Incidences of a wilt-like disease in robusta coffee were first reported in Misenyi division in Minziro ward at Kigazi village near the Ugandan and Tanzanian border towards the end of 1996. A team of researchers and extension
workers visited the affected areas in January 1997. A fungus, Gibberrella
xylarioides Heim and Sacc, was isolated from robusta stems and branches collected from diseased plants (Kilambo and Kaiza, 1997; Kilambo et al., 1997)
and confirmed by CABI Europe (UK). Additional surveillance was made between 1998 and 2003 (Nghoma, 2003; Mohamed et al., 2002; Swai, 1998), and
CWD was again confirmed in the Karagwe and Muleba districts. CWD has
continued to be a serious threat to the robusta coffee industry in Kagera since
this first report in 1997. Its introduction in Tanzania could be associated with
human activities prevailing between residents of Kagera and neighbouring
countries where there is CWD.
5.6. Coffee Wilt Distribution in Tanzania

A detailed biological survey was carried out in Tanzania between September
2002 and February 2003 to determine the geographical distribution and host
range of CWD in Tanzania (Kilambo et al., 2004; Oduor et al., 2003). CWD was
found to be confined only to robusta coffee in the Kagera region (Fig. 5.3).
72
Kilambo et al.
Fig. 5.3. Map showing distribution of CWD in Tanzania.
In the north part of Kagera where CWD is present, groups of infected farms
are detected in close proximity to each other but about 100 km away, and in
Muleba, only isolated farms were affected. This distribution could reflect the
dispersal of the pathogen. Waller and Holderness (1997) indicated that spore
dispersal (conidia and ascospores) can occur through different mechanisms,
including rain, soil, plant-to-plant contact (relatively short distances), wind,
humans and possibility of insects (for longer distance spread).
In Kagera, average incidence and severity were 13.5% and 4.6%, respectively, whereas higher incidences and severity were recorded in Bukoba
(39.1% and 16.1%, respectively). Among divisions in Kagera, Bugabo was
73
Table 5.1. Estimated number of robusta trees infected and uprooted from 1997
to 2007. (From the Tanzania Coffee Research Institute Lyamungu, Moshi.)
District
Bukoba
Misenyi
Karagwe
Muleba
Total
Infected coffee
trees
Uprooted coffee
trees
Percentage of
uprooted due to CWD
28,155
17,277
6,279
2,422
54,133
14,079
4,566
6,279
2,348
27,272
50
26
100a
97
50
CWD is still prevailing.
the worst affected (incidences of 80.7% and severity of 46%). In 1997, CWD
was detected in three wards, but to date (2008), the disease has spread and
a total of 44 wards out of 100 in Kagera are now reported to be affected by
the disease.
5.7. Importance of CWD in Tanzania

Prior to the outbreak of CWD, robusta coffee was the major source of income
to more than 90% of households in Kagera. Since the outbreak of CWD, most
of the coffee growers are in despair about the devastation caused by CWD. The
disease has caused substantial losses of coffee trees in the region (Table 5.1).
At the same time, cost of hiring labour for uprooting infected stems has
increased from US$0.42 in 2002 to US$0.83 in 2007. Many farmers have
switched livelihood as a coping strategy to the disease, with farmers shifting to fishing, growing bananas and brewing beer. In addition, many coffee
farms have been abandoned. Losses of trees can be equated to yield loss,
Table 5.2. Estimated CWD infected trees, yield losses and cost of uprooting. (From
Tanzania Coffee Research Institute Lyamungu, Moshi.)
District
Bukoba
Misenyi
Karagwe
Muleba
Total
a
Infected
coffee trees
19972007
Estimated yield
losses from
CWD-infected
treesa (kg)
Yield losses
in monetary
terms+
(US$)
Cost of
uprooting
infected
trees# (US$)
Total losses
(US$)
28,155
17,277
6,279
2,422
54,133
84,465
51,831
18,837
7,266
162,399
141,057
86,558
31,458
12,134
271,207
23,369
14,339
5,212
2,010
44,930
164,426
100,897
36,670
14,144
316,137
Estimated that a mature robusta coffee tree 25 years old can produce 3 kg of clean coffee.
Estimated that the price of 1 kg of clean coffee is US$1.67 (as per OctoberDecember 2007).
#
Cost of uprooting a stool of infected robusta tree is US$0.83.
+
74
Kilambo et al.
which is approximately 162,399 kg of clean coffee lost due to the death of

54,133 trees from CWD. It is estimated that the disease has caused a financial
loss of approximately US$316,137 over the last 10 years (Table 5.2).
5.8. Factors Affecting Severity or Incidence of CWD in Tanzania

The CWD epidemic in Tanzania can be mainly attributed to the following
factors:
5.8.1. Improper gap filling procedures

New coffee seedlings are being planted immediately after uprooting of
CWD-diseased trees. It is suspected that in the presence of G. xylarioides
propagules, the roots of the new seedlings may start to be infected shortly
after replanting.
5.8.2. Source of planting materials

Some farmers use volunteer seedlings that can either be already contaminated by G. xylarioides or be susceptible to the pathogen because they are
from a narrow genetic base.
5.8.3. Use of infected material for fuel

CWD-infected robusta trees are being used as a source of firewood. Sometimes, this firewood is stored in the backyards of the farmers houses. Following conducive weather conditions, the conidia of the asexual stage of the
pathogen can be disseminated and induce infections of previously healthy
coffee trees.
5.8.4. Quarantine procedures

Farmers are not following Plant Protection Act no. 13 of 1997, which prohibits movement of coffee planting materials and importing from neighbouring
countries, from district to district and from farm to farm.
5.9. Country-Specific Management Practices in Tanzania

Since the outbreak of CWD, the government of Tanzania through the Ministry of Agriculture, in collaboration with other sectors of the coffee industry under the coordination of Tanzania Coffee Research Institute (TaCRI),
75
has implemented strategies to minimize the effect of CWD in collaboration

with regional partners (CABI Africa and Advanced Research Institutes in
the region) and with international partners such as scientists based in the UK
(CABI Europe, UK).
5.9.1. Exclusion
To restrict the spread of CWD, in 2002, the Ministry of Agriculture and Food
Security by then enforced Plant Protection Act no. 13 of 1997 to effect quarantine measures to prevent movement of plant materials from neighbouring
countries and also movement of plant material, soil and farm implements
from district to district and from farm to farm. Achievements have been very
low because enforcement was not very effective. Tanzania shares borders
with other countries like Rwanda and Burundi, which do not yet have CWD,
so there is a need to strengthen quarantine procedures to prevent spread of
the disease to these countries also.
5.9.2. Eradication
During the past 10 years, joint efforts between TaCRI, CABI and internal
stakeholders has greatly facilitated a reduction in the impact of CWD. In
May 2003, stakeholders including smallholders, districts and regional leaders, agricultural extension staff, representatives from the Ministry of Agriculture and Food Security and Cooperative Unions met at Maruku Agricultural
Research Institute and made a crucial decision that a campaign to eradicate
CWD should be a top priority. A simple methodology was used, this involves
the uprooting and burning of affected trees in situ, with particular emphasis
on early diagnosis of the disease to reduce the risk of spread to other trees
and farms. Results since the campaign have been impressive, with concerted
awareness and education campaign helping to get the message to farmers
effectively and rapidly. Eradication of CWD-affected trees has significantly
reduced the incidence of the disease (Fig. 5.4). This shows that if eradication
of CWD-diseased trees is done promptly in newly infected trees, it can assist
in the management of the disease on the farm and minimize its spread to others. Since the outbreak of CWD, a total of 27,272 of diseased trees have been
uprooted and destroyed. This is about 50% of infected coffee trees.
Coffee stakeholders supported efforts in locating areas affected by
CWD, training programmes and eradication of affected trees. The support
came from the Tanzania Coffee Board to initiate surveys in January 1997
by providing about US$8000; the Ministry of Agriculture and Food Security
donated US$10,000 in 2002 to continue with the surveillance and eradication
programme and districts councils in Karagwe and Muleba provided US$6000
for training of extension workers and for the eradication programme in
20022004. From 2002 to 2007, the Common Fund for Commodities donated
over US$70,000 to conduct training of farmers and extension workers on the
76
Kilambo et al.
Changes in CWD incidence between
2002 and 2007
80
60
% CWD
incidence 40
20
0
2002
2007
Bugabo Kaisho Muleba
Divisions
Fig. 5.4. Changes in incidence of CWD in three administrative divisions of Kagera between
2002 and 2007.
identification of early symptoms of CWD and dissemination of technologies

to combat the disease. In addition, the Coffee Research Network (CORNET)
supported by EU funds, contributed US$39,000 in 2002 to 2003 to carry out
biological and socio-economic surveys. TaCRI (using STABEX funds) spent
approximately US$10,000 in the eradication programmes.
5.9.3. Protection
Stem paint using copper oxychloride (300 g/l of water) applied 50 cm from
ground level and four times a year as a preventive measure was shown to
be effective in minimizing levels of CWD infection. This was tested in onfarm agronomic trials, and some farmers have started to adopt it. In addition, farmers local practice of heaping ash at the collar position of robusta
plants prevents infection of G. xylarioides in some of the areas, particularly
in Bukoba. However, in Uganda, use of ash was not effective in preventing
CWD in almost all agronomic trial sites. Efficiency of use in Tanzania may
be associated with the soil pH of the particular areas where use of ash was
effective. Observations in the Karatu district where lime is being applied in
coffee farms have indicated that the treatment is effective for lowering the incidence of collar rot disease (Fusarium lateritium), but this requires further investigation. Measures to improve the health of trees by mulching, manuring
and fertilizer application also assist the plants to tolerate CWD infection.
5.9.4. Host plant resistance

Commercial robusta clones MS 1, 2, 3, 4 and 5 released to coffee growers in
recent years are all susceptible to CWD in varying degrees. Efforts to evaluate robusta coffee germplasm for their resistance to CWD using different
isolates and the root dip technique developed by Hakiza et al. (2004) were
initiated in May 2004. Seedlings (56 months old) were inoculated using the
77
Table 5.3. Eleven best-performing resistant robusta

lines to CWD combining CLR resistance.
S/no
Coffee line/
cultivar
1
2
3
4
5
6
7
8
9
10
11
12
ML 26
NG 08
ML 35
NG 13
BK 27
KR 21
NG 12
KR 11
NG 17
ON BK 02
MR 10
MS 1a
CWD
resistance
CLR
resistance
R
R
R
R
R
R
R
R
R
R
R
S
R
R
R
R
MR
MR
MR
MR
MR
MR
MR
R
Commercial cultivar, R = resistant, MR = moderately

resistant, S = susceptible.
stem nicking method and the root dip method. In both methods, the inoculum was standardized at 1.3 106 conidia per millilitre of water. After 9
months of assessment, the survivors were considered as resistant and developed further for future selection (Table 5.3). To date, 273 clones have been
established in a clonal mother garden at Maruku with the intention to obtain
enough cuttings to establish in the National Performance Trials and on farms
(Kilambo et al., 2006).
5.9.5. Variability of the CWD pathogen

Twenty isolates were sent to CABI Europe (UK) for assessing genetic variability of the pathogen by studying the DNA banding patterns produced
by PCR amplification of the ribosomal DNA intergenic spacer region of
G. xylarioides isolates from coffee. The isolates were found to be genetically
similar to larger C. canephora group from the Democratic Republic of the
Congo and Uganda but different from that found to attack C. arabica in Ethiopia (Rutherford, 2005). Observation of CWD pathogenic variability in Kagera
reveals that only C. canephora is being attacked by the pathogen.
Pathogenicity studies conducted in Tanzania showed that all 14 isolates
collected from various ecosystems were very aggressive (Kilambo et al.,
2006) (Table 5.4). The isolates are currently being used to search for durable
resistance.
A comparison of the methodologies for assessing resistance of coffee
germplasm at Maruku was undertaken. Two methods for screening for
78
Kilambo et al.
Table 5.4. Pathogenicity test results of G. xylarioides on MS 1 and MS 2.

CWD isolate
code no.
TaCRI
CABI UK
Number of
dead seedlings
at end of test
Location collected
District
Coordinates
Altitude (m)
MS 1
MS 2
2004/10
T1
Muleba
S 0145.901; E 3135.491
1547
2004/13
T 2a
Muleba
S 0146.827; E 3134.541
1545
10
9
10
2004/07
T 3a
Muleba
S 0149.702; E 3141.137
1395
2004/08
T4
Muleba
S 0143.159: E 3138.078
1510
10
2004/02
T 5a
Muleba
S 0141.172; E 3137.731
1287
10
2004/06
T 8a
2004/01
Bukoba
S 0100.595; E 3146.582
1189
10
10
Bukobaa
S 0114.836; E 3150.682
1200
10
2004/12
T 9a
Bukoba
S 0101.612; E 3132.758
1256
10
2004/14
T 12a
Karagwe S 0118.600; E 3047.205
1424
2004/03
T 13a
Karagwe S 0126.166; E 3052.801
1317
10
2004/05
T 14a
Karagwe S 0117.308; E 3053.896
1659
2004/09
T 15a
Karagwe S 0115.309; E 3057.347
1354
2004/09
T 15b
Karagwe S 0115.309; E 3057.347
1354
2004/09
T 15c
Mean
SE
CV
LSDb (P 0.05)
Karagwe S 0115.309; E 3057.347
1354
9
10
9.57
0.14
5.30
0.30
9
10
9.21
0.11
4.50
0.23
Isolate 2004/1 was used for CWD resistance evaluation.

LSD = least significant difference.
resistance, namely, root dip and stem nicking, were evaluated for their ability to differentiate between susceptible and resistant material by producing
a high selection pressure. In the study, the root dip method produced higher
selection pressure by allowing fewer survivors than stem nicking (Kilambo
et al., 2006, Table 5.5).
5.9.6. International and regional collaboration

With support from DFID, two TaCRI scientists and a technician received
in-depth training in appropriate pathogenicity methods and screening techniques at the Coffee Research Institute, Kituza, Uganda (Plate 32). This support also financed the collection and shipment of CWD isolates in UK and
CWD screening for resistance. TaCRI is now applying these techniques for
screening CWD resistance in this country. From 2004 to date, TaCRI used
US$17,000 to support development of varieties with durable resistance to
CWD and maintenance of robusta coffee germplasm.
79
Table 5.5. Robusta lines that survived after artificial

inoculation with G. xylarioides using stem nicking and
root dipping procedures.
S/no
1
2
3
4
5
6
7
8
9
10
11
12
Mean seedling survival

at end of test
Coffee line/
cultivar
Stem nicking
Root dipping
ML 26
NG 08
ML 35
NG 13
BK 27
KR 21
NG 12
KR 11
NG 17
ON BK 02
MR 10
MS 1*
Mean
SE
CV
LSD (P 0.05)
4
5
2
4
4
3
4
5
3
3
5
0
3.50
0.40
40.00
1.08
3
3
1
3
4
3
3
3
1
3
4
0
2.58
0.35
35.40
0.95
*Commercial variety.
5.9.7. Dissemination and training

As part of the programme, an awareness campaign was initiated which involved CWD training for 5659 growers and 192 extension staff in Kagera,
and 171 in non-CWD areas. Training encompasses symptoms identification,
safe handling and destroying of diseased robusta trees to minimize further
spread. This was complemented by publication of articles in newspapers,
radio and local television in Kagera Region and the publication of posters
and leaflets alerting stakeholders (particularly farmers) of the threat from the
disease. Approximately 1622 posters and leaflets have been distributed.
5.9.8. Farmers field schools/participatory groups

Twenty-three groups have been formulated in CWD hot spot areas. The
groups meet once a month, and farmers have an opportunity to share experiences in managing the effects of CWD. It is expected that knowledge of disease management will be disseminated to some other farmers. TaCRI has an
innovative approach of participatory extension working with extension staff
and farmer groups. This has assisted highly in the dissemination of technologies in minimizing the effect of CWD and boosting coffee production
80
Kilambo et al.
in the region. There are 23 participatory groups in Bukoba and Misenyi, 16 in

Karagwe and 13 in Muleba. An excellent example of farmer-to-farmer extension is exemplified by Chabuhora, which had 30 participants and formulated
five groups for training other farmers. This shows how a participatory approach through farmer-to-farmer contact, extension workers and researchers is effective in the diffusion of coffee husbandry practices. Manyafubu
is another good example of a participatory farmer approach. The group
have 25 members (13 men and 12 women) and was formed in March 2004.
It conducted training to allow farmers to identify robusta trees infected with
CWD and attempted to minimize spread through an eradication campaign.
To date, the group has conducted training in 18 villages in the Bugabo division, and 1839 farmers including village leaders and ward representatives
have been trained. Seven farmer field schools have been initiated; ten farmer
groups formed in 20 training centres.
5.9.9. Open/field days

On 24 September 2004, TaCRI instituted an open day at Maruku. The major
emphasis was to stress the threat of CWD, its impact on livelihoods and initial steps to consider on how to contain its further spread to new areas. More
that 500 participants from all levels attended. Since then, 23 field days have
been conducted in CWD-diseased areas in Kagera.
5.9.10. Sensitization of policy makers

Policy maker fora were used to disseminate information on CWD. For example,
on 22 May 2003, TaCRI and representatives from the Ministry of Agriculture
and Food Security were invited to attend the Regional Coordinating Committee
of Kagera, chaired by the then Regional Commissioner Major Gen. Tumainiel
Kiwelu. Highlights and strategies on management of CWD were presented. The
meeting had significant impact on policy makers who supported the eradication
programme and training of extension workers and farmers using funds from
district councils.
5.9.11. Prioritizing CWD research programmes

In 2003, TaCRI formulated its 5-year Strategic Action Plan 20032008, whose
vision is to contribute to the transformation of the Tanzanian Coffee Industry
to sustainable prosperity with a major goal of creating a profitable and sustainable coffee industry in Tanzania. The vision of TaCRI is in line with the
Agriculture Vision and the Tanzania Development Vision 2025. In this strategic action plan, CWD has been given a high priority. Since 2003, TaCRI has
been addressing CWD by screening robusta germplasm present at Maruku
and facilitating eradication and training programme at village levels.
81
5.9.12. Extensive replanting programme

Tanzania has embarked on an extensive coffee replanting programme with
five carefully selected, indigenous robusta coffee clones being distributed to
farmers, and from 1999, more than 1,500,000 coffee plantlets from a relatively
narrow genetic base have been replanted by more than 1000 farmers. However, the coffee varieties used in this programme were later known to be
susceptible to CWD, so there are renewed efforts to get new CWD-resistant
clones to the farmers, as a replanting programme will have great impact in
Kagera.
5.10. Conclusion
G. xylarioides is the most damaging pathogen of robusta coffee in Tanzania.
Despite efforts already undertaken to minimize the effect of the disease, the
answers related to the rational control measures will depend on successful
execution of the following activities:
1.
Concentration on already long-term breeding strategies for robusta coffee, involving:

Identification of more CWD-resistant accessions from Maruku germplasm by artificial inoculation.
Initiation of hybridization schemes and selection and advancement
of elite materials.
Selection of local robusta survivors from CWD hot spot areas.
Expansion of clonal mother garden already established at Maruku,
hence providing more CWD-resistant materials for distribution.
2. Continue to disseminate information and train farmers about minimizing further spread and socio-economic impact of CWD.
3. Continue with national and international collaboration to have an indepth knowledge of the CWD pathogen and to search for durable resistance through exchange of germplasm.
Acknowledgements
The management of CWD in Tanzania since its first report 10 years ago has
been an excellent example of partnership and collaboration. We are grateful
to the Tanzania coffee stakeholders, European Commission, Dar es Salaam
and the Government of Tanzania for generous financial support to TaCRI
during the implementation of this work. We are also grateful to DFID, CFC,
CABI Africa and Europe-UK, the Coffee Research Centre, Uganda and Centre de Cooperation Internationale en Recherche Agronomique pour le Dveloppement (CIRAD) for support to carry out activities in Kagera to combat
the threat of CWD.
82
Kilambo et al.
References
Baffes, J. (2003) Tanzania coffee sector: constraints and challenges in global environment. Africa Region Working Paper Series No. 56, The World Bank, Washington,
DC.
Hakiza, G.J., Kyetere, D.T. and Olal, S. (2004) Mode of penetration and symptom expression in robusta coffee seedlings, inoculated with Gibberella xylarioides, the
cause of coffee wilt disease in Uganda. In: ASIC 2004 20th International Conference on Coffee Science. Bangalore, India.
Kilambo, D. and Kaiza, D. (1997) Investigation of Fusarium xylarioides on robusta coffee in Misenyi division Bukoba district. Trip report, Lyamungu misc. report.
Kilambo, D., Kaiza, D. and Swai, F.B. (1997) Observations of wilt disease in some coffee growing areas in Tanzania. In: Proceedings of the 1st Regional workshop on
coffee wilt disease (tracheomycosis). International Conference Centre, Kampala,
Uganda.
Kilambo, D.L., Nghoma, N.M., Mohamed, R., Teri, J.M., Poole, J., Flori, A. and Pinard,
F. (2004) Coffee disease surveys in Tanzania. In: Proceedings of the 20th International Conference on Coffee Science. Bangalore, India, pp. 12631266.
Kilambo, D.L., Nghoma, N.M., Mtenga, D.J., Teri, J.M., Nzallawahe, T., Rutherford,
M. and Masumbuko, L. (2006) Progress towards searching for durable resistance to
Fusarium wilt (Fusarium xylarioides) in Coffea canephora germplasm on Tanzania.
In: Proceedings of the 21st International Conference on Coffee Science. Montpellier, France, pp. 13861389.
Mohamed, R.A., Nghoma, N.M., Sayi, B. and Kabumbire, A. (2002) Report on coffee
wilt disease Surveys in Kagera. Ministry of Agriculture and Food Security, United
Republic of Tanzania.
Nghoma, N.M. (2003) Survey and mapping of areas affected by coffee wilt disease
(CWD) in Karagwe district. Technical report, Ministry of Agriculture and Food Security, supported by Karagwe District Council.
Oduor, G., Simons, S., Phiri, N., Njuki, J., Poole, J., Pinard, F., Kyetere, D., Hakiza, G.,
Musoli, P., Lukwago, G. Abebe, M., Tesfaye, A., Kilambo, D., Asiimwe, T. and
Munyankere, P. (2003) Surveys to assess the extent and impact of coffee wilt disease in East and Central Africa. Final Technical Report EU Contract No. ASA-RSP/
CV-006. CAB International, Egham, UK.
Rutherford, M.A. (2005) Epidemiology and variability of Gibberella xylarioides, the coffee wilt pathogen. DFID Crop Protection Programme, Final Technical Report
CABI Uk. 24 pp.
Swai, F.B. (1998) Surveillance of wilt disease in Kagera. Lyamungu misc reports.
Tanzania Coffee Board (TCB) (2006) Data on coffee production. TCB miscellaneous
reports.
Waller, J.M. and Holderness, M. (1997) Fusarium diseases on coffee. In: Proceedings of
the First Regional Workshop on the Coffee Wilt Disease (Tracheomycosis). International Conference Centre, Kampala, Uganda, pp. 3139.
Socio-Economic Impact of
Coffee Wilt Disease
R.O. Musebe,1 J. Njuki,1 S. Mdemu,2 G. Lukwago,3

A. Shibru4 and T. Saiba5
CABI Africa, UN Avenue, ICRAF Complex, PO Box 633-00621, Nairobi, Kenya

Tanzania Coffee Research Institute, Lyamungu, PO Box 3004, Moshi, Tanzania
3
National Agricultural Research Organization, PO Box 421, Kabale, Uganda
4
Jimma Agricultural Research Centre, Ethiopian Institute of Agricultural Research
5
Office National du Caf, BP 8931, Kinshasa Kingabwa, Congo
2
6.1. Introduction
Coffee is a major export crop in Africa and accounts for the bulk of the export
earnings of most economies in the Eastern and Central Africa (ECA) regions.
It is also important because of its contribution to farm income, employment
and food security. The crop is a major source of livelihood particularly for
small-scale producers. Coffee is a leading export sector in Ugandas economy,
accounting for over US$100 million in export sales in the 2000/2001 market
year (GoU, 2001). In Kenya, coffee ranks fourth after horticulture, tea and
tourism, accounting for 10% of the total export earnings (Karanja and Nyoro,
2002). Similarly, the commodity is Tanzanias largest export crop, contributing approximately US$115 million to export earnings and provides employment to over 400,000 families (Baffes, 2003). In Ethiopia, coffee contributes
around 50% to the countrys foreign exchange earnings and about 10% of the
gross domestic product. More than 25% of the population of Ethiopia, which
represents 15 million people, are dependent on coffee for their livelihoods,
including 8 million people directly involved in coffee cultivation and 7 million in the processing, trading, transport and financial sectors (Beintema and
Solomon, 2003; Oxfam, 2002; Charveriat, 2001). In the Democratic Republic of
Congo (DRC), coffee is the number three export after copper and cobalt, and it
83
84
Musebe et al.
represents approximately 9% of the total value of DRC exports and almost 60%
to 65% of the total value for the whole of exports of the agricultural produce
(Banque Centrale du Congo, 2005). Growth in coffee production and income
has been one of the main engines for the development of coffee-producing
areas. Thus, coffee plays a significant role in the national economies and the
livelihood of the rural poor in the subregion.
Coffee production, productivity, quality and earnings are under threat
from coffee wilt disease (CWD). Incidence and severity of CWD are highest
in Uganda and lowest in Tanzania (Table 6.1). In Tanzania, the disease was
only found in the Kagera region, which borders Uganda.
CWD is observed on robusta coffee only in DRC, Tanzania and Uganda;
whereas in Ethiopia, it was observed on arabica coffee only. This finding suggests that the diseases found in the two coffee types, which are also separated
geographically, are genetically distinct (Oduor et al., 2003). Following the reemergence of CWD, farmers have experienced changes in their livelihoods
and resource endowments and have adopted various coping strategies as a
result of the disease. This chapter examines some of the changes that have
occurred in respect to coffee since the re-emergence of CWD as farmers try
to deal with the problem in DRC, Ethiopia, Tanzania and Uganda and draws
conclusions and recommendations aimed at improving coffee production
through containment of the disease. The changes analysed include (i) the importance of coffee as a source of income, (ii) coffee production, (iii) input use
in coffee, (iv) liquidation of assets, (v) coping strategies and (vi) household
expenditure.
Different sampling approaches were used in different countries to identify the households to provide the requisite data. The sampling approach
used depended on the distribution of CWD in the country, with purposive
sampling being used where incidence of the disease was relatively low or focused in certain areas of the country and random sampling where CWD was
widespread and severe. In Tanzania, CWD was found in the Kagera region
only, and cluster sampling was used in villages known to have CWD, and
these were identified through key informants; purposive sampling was used
to target 99 affected farms. In Uganda, a multistage sampling procedure was
followed in which 21 robusta-coffee-growing districts affected by CWD were
targeted; subcounties, parishes and villages were subsequently selected randomly. Three farmers were then selected from each village. A total of 356
households were interviewed. In Ethiopia, the socio-economic survey was
Table 6.1. Incidence and severity of CWD in
countries in the ECA region.
DRC
Ethiopia
Tanzania
Uganda
Incidence (%)
Severity (%)
90.0
27.9
2.2
90.3
40.5
3.0
0.7
44.5
2a
2b
Plate 1. Coffee field attacked by CWD as shown by stumps of dead trees. (Photograph
courtesy of G. Hakiza.)
Plate 2. (a) Healthy coffee field in 1993. (b) The same field in 1998. (Photograph courtesy of
G. Hakiza.)
4a
4b
Plate 3. Robusta plant affected by CWD in Uganda. (Photograph courtesy of S. Olal.)

Plate 4. (a and b) Arabica coffee trees with complete wilt symptoms. (Photographs courtesy
of A. Girma.)
Plate 5. Veinal necrosis, caused by coffee wilt disease (Gibberella xylarioides), clearly visible
on the underside of a coffee leaf. (Photograph M. Rutherford.)
Plate 6. Wounds (arrows), originating from slashing during weed control, at the base of a
coffee stem. (Photograph courtesy of N. Phiri.)
Plate 7. Stromatic fruiting bodies (perithecia and ascospores of Gibberella xylarioides in the
bark of dead Arabica coffee tree. (Photograph courtesy of A. Girma.)
Plate 8. Characteristic blue-black colouration under the coffee bark when an infected stem
is scraped with a knife. (Photograph courtesy of N. Phiri.)
10
11
12
13
14
Plate 9. Perithecioide ascomata (perithecia) produced by the coffee wilt pathogen,

Gibberella xylarioides, in cracks of the tree bark. (Photograph J. Flood.)
Plate 10. Diagrammatic representation of asci (left) and the typically curved conidia (right)
of the coffee wilt pathogen, Gibberella xylarioides. Asci contain eight two-celled ascospores.
Illustration from Booth, C. and Waterston, W.M. 1964. CMI Descriptions of Plant pathogenic
Fungi and Bacteria No. 24, CABI, Wallingford.
Plate 11. Conidia of Gibberella xylarioides on Spezieller Nhrstoffarmer Agar (SNA)
medium. (Photograph P. Lepoint and H. Maraite.)
Plate 12. Asci of Gibberella xylarioides, the coffee wilt pathogen, containing two celled
ascospores (arrowed). (Photograph J. Flood.)
Plate 13. Failure to produce orange pigmentation, a characteristic used to rapidly
differentiate biological species BS 1 from other representatives of the G. xylarioides complex.
In this case, MAT-1 C. canephora-associated (BS 2, orange colonies) mating type tester
strains were confronted with a MAT-2 C. arabica-associated G. xylarioides sensu lato strain
(BS 1, white colony) on carrot agar. Incompatibility between the two BS is denoted by the
absence of perithecia or protoperithecia along the confrontation zones. (Photograph P.
Lepoint and H. Maraite.)
Plate 14. Production of perithecia along confrontation zones of G. xylarioides sensu lato
strains of opposite mating type and of the same biological species (BS). (Photograph
P. Lepoint and H. Maraite.)
15
16
Plate 15. Purple black G. xylarioides sensu lato perithecia producing an orange cirrhus of
ascospores on carrot agar medium when C. canephora associated strains of opposite mating
type are crossed. (Photograph P. Lepoint and H. Maraite.)
Plate 16. Graphical representation of the spatial and temporal development of coffee wilt
disease on coffee trees on a farm in Mayuge, Uganda. This depicts the point, during a 17
month study involving ten assessments, at which external disease symptoms were first
observed. Points where trees were missing at the outset of monitoring, dead trees and trees
that did not develop symptoms throughout the study are denoted by white, black and grey
spots respectively. Trees that developed symptoms in the earlier, intermediate and latter
stages of the study are denoted by red, orange and yellow spots respectively, the numbers
denoting the actual assessment point (1-10) at which symptoms were first observed.
(Illustration courtesy of M. Rutherford.)
17
18
19
Plate 17. Host specificity of Gibberella xylarioides isolates collected from Coffea arabica and
C. canephora on seedlings of the respective host species. (Photograph courtesy of A. Girma.)
Plate 18. Host specificity of Gibberella xylarioides isolates from Coffea arabica (G3P22),
C. canephora (CAB003, TZ009) and C. excelsa (DSMZ 62457) on seedlings of C. arabica
(cv. SL28) in growth room. (Photograph courtesy of A. Girma.)
Plate 19. Specificity of Gibberella xylarioides isolate DSMZ62457 (an historical Excelsa
strain), G3P22 (Arabica strain) and CAB003 (Canephora strain) on Catimor seedlings (1570)
in growth room. (Photograph courtesy of A.Girma.)
20
21
22
Plate 20. The characteristic unilateral wilting symptoms caused by the historical strain DSMZ
62457 on C. liberica seedlings. (Photograph courtesy of A. Girma.)
Plate 21. Gibberella xylarioides -Coffea canephora interaction at cellular level (cross section
of a 9-month old infected seedling (Photograph courtesy of D. Bieysse.)
Plate 22. Gibberella xylarioides-Coffea canephora interactions at cellular levels (arrow
hyphal colonization of vascular tissues of 18-month old seedlings 45 days after inoculation.
(Photograph courtesy of D. Bieysse.)
23
24
25
Plate 23. Screening Coffea arabica collections by seedling inoculation test in the green
house in Ethiopia (Jimma). (Photograph courtesy of A.Girma.)
Plate 24. Coffea arabica seedlings re-inoculated to verify the resistance in the screen house.
(Photograph courtesy of A.Girma.)
Plate 25. C. canephora seedlings that survived (resistant) CWD infection after the first bout
of inoculation. Left foreground, susceptible plants: right foreground, survivors. (Photograph
courtesy of P. Musoli.)
26
27
Plate 26. Diversity of Ugandan C. canephora and other diversity groups. (Illustration courtesy of P. Musoli.)
Plate 27. Eracta and nganda C. canephora trees. (Photograph courtesy of P. Musoli.)
28
29
30
31
Plate 28. Uprooting and burning is the most effective eradication strategy for CWD infected
trees. (Photograph courtesy of N. Phiri.)
Plate 29. FFS farmers practicing pruning technique. (Photograph courtesy of N. Phiri.)
Plate 30. A coffee farm mulched with dry grass which suppressed weeds. (Photograph
courtesy of N. Phiri.)
Plate 31. A coffee farm planted to Desmodium intotum which suppressed weeds.
(Photograph courtesy of N. Phiri.)
32a
32b
33a
33b
34
35b
35a
35c
Plate 32. (a and b) TaCRI team receiving training at Kituza. (Photograph courtesy of
H. Mutenyo.)
Plate 33. (a and b) Spreading CWD technologies through participatory approach.
Plate 34. Farmers carrying out agro-ecosystem observation and analysis in small groups.
Plate 35. Media interviews with (a) Ethiopian researchers, (b) farmers and (c) CABI
scientists.
85
conducted in four regions confirmed to have CWD, namely, the Southern

Nations Nationality Peoples, Oromiya, Gambella and Amhara. Households
with a CWD severity of at least 10% were randomly selected from each of the
regions. A total of 137 households were interviewed in Ethiopia. In DRC, the
survey was conducted in North Kivu and eastern provinces where a total of
436 households were randomly selected and interviewed.
Structured questionnaires were used to collect data from all the selected
households in all the countries between March 2002 and January 2003. Data
collected included household characteristics, farm characteristics, farmers
perceptions of CWD, awareness and ability to identify the disease, severity
of CWD and its progression on farms, coping strategies by farmers, income
changes, input use, asset portfolio and changes in household expenditure.
Data collected were analysed using descriptive statistics, regression analyses
and chi-square tests to establish relationships. Analysis was conducted using the Statistical Package for Social Scientists. Information from other minor
studies on the disease carried out in individual countries has also been incorporated for comparison purposes and to enrich the chapter.
6.2. Farmer Perceptions of CWD in ECA

Farmer perceptions of CWD varied across the region. Overall, awareness
was moderate (40.6%) except for Uganda (90%), where the disease had been
particularly severe and the majority of farms (90.3%) had CWD. Its early
detection in the country was also partly responsible for this high awareness
(Hakiza, 1995, unpublished). In DRC, farmers knowledge of the disease is
moderate, and about 40% of the interviewed farmers know CWD. Awareness of CWD was similar in Ethiopia (17%) and Tanzania (15%). Although
CWD is caused by a fungus, some farmers attributed the disease to insects,
soil-borne pathogens, nematodes, ants, soil nutrient depletion and the environment. There was lack of consistency on when farmers thought the disease
spreads most, whether in the rainy season or in the dry season. In Tanzania
and Uganda, more than half of the farmers reported that the disease spreads
more during the dry season, whereas more than half of the farmers in Ethiopia indicated the opposite, that the disease spreads more during the wet season. About 34.8% of the farmers in DRC indicated that the disease spreads
most during dry weather, whereas 22.2% thought that the disease spreads
during wet periods. The remainder of the farmers did not know when the
disease spreads most. Dry weather resulting in moisture stress makes coffee
plants more vulnerable to CWD, which may lead to more observable effects
during the dry season compared to the wet season. The fungus is however
known to germinate and spread more during the wet rather than the dry season. All farmers interviewed noted that CWD was causing major damage to
the coffee crop on their farms. On the trend of the disease since farmers first
noticed it, majority of farmers in all the countries (93.8% in Tanzania, 73.7%
in Uganda and 97% in Ethiopia) indicated that the disease was increasing.
This meant that control measures would be necessary to reduce crop losses.
86
Musebe et al.
6.3. Changes in the Importance of Coffee as a Source of Income

Prior to the re-emergence of CWD, there was already significant variation in
the importance of coffee as a source of income across the region. Overall, it was
the major source of income for 70% of coffee farmers, although the importance
was much lower in Uganda (15%) compared to DRC (95%), Ethiopia (91%)
and Tanzania (95%). The relative contribution of coffee to household income in
the region has, however, declined by an average of 12% following the onset of
CWD, and currently, only an average of 58% of coffee farmers consider coffee
to be their major source of income. It is expected that if CWD is not managed,
the importance of coffee as a source of income is likely to decline further. Similarly, the percentage of farmers with coffee as their only source of income has
fallen from 50% to 44% across the region. There is a general decline in earnings
from coffee in the ECA region, which is attributed to CWD. The fall in number
of farmers that had coffee as the main source of income and decline in coffee
incomes was aggravated by the fall in world coffee prices during the period.
Before the onset of CWD in Tanzania, coffee was the major source of
income for 95% of the farmers and the second most important source of income for the remaining 5%. After the onset of CWD, the percentage dropped
to 91%, with some of the farmers opting for fishing and fish trade and growing bananas and beer brewing. The alternative to coffee growing adopted
by farmers depended on level of accessibility to the enterprise. Whereas the
percentage of farmers with only one source of income was 58% before CWD,
this fell to 53% after CWD. The small decline in the relative importance of
coffee in Tanzania is because most of the coffee grown here is arabica coffee,
which was not affected by CWD strain in the country. Only the CWD strain
that attacks robusta coffee is present currently in Tanzania. Hence, the arabica coffee productivity masked the effect of the CWD.
In Ethiopia, the percentage of farmers who rated coffee as the most important source of income was 91% before CWD but fell to 84% after CWD.
Eighty percent of the farmers indicated that income from coffee was decreasing, whereas 10% indicated that it was increasing. The farmers that reported
increases in income may have escaped the effect of CWD and also had better
access to marketing services. Another 1.5% reported that there was no change
in the income from coffee. Considering the proportion of income from coffee
for households, 33% reported that the share when compared to other sources
of income was decreasing, 51% reported that it had not changed while 15%
reported that it was increasing.
The contribution of coffee to household income greatly diminished following the CWD epidemic in Uganda. Twelve percent of the farmers interviewed rated coffee as the most important source of income. Livestock
production and trading in non-agricultural goods, such as household consumables, became more important as income sources following the CWD
epidemic, implying that these are important strategies adopted as a means
of coping with CWD in Uganda. The reason for increased livestock keeping
may be because of its importance as an alternative income source and a store
of wealth (asset). Before CWD, coffee plantations served as an asset and
87
income source held by many farm household currently investing in livestock,

especially cattle, as an additional asset. This created an opportunity for increasing the integration of crops and livestock production to exploit synergistic interactions between the two enterprises to achieve sustainable agricultural growth.
In DRC, coffee was generating about 95% of income for farmers before reemergence of CWD. However, a combination of fluctuating and falling prices
of coffee in the world market and CWD has reduced the contribution of coffee
to farmers income. Coffee now contributes 20% of income of the farmers.
6.4. Change in Coffee Production and Income

Production of coffee at the farm level fell by an average of 35.0% after CWD
infestation, which was reflected in significant (P < 0.01) reduction of income
across the region. Decline in production for individual countries were 38.6%,
37.0% and 29.4% for Uganda, Ethiopia and Tanzania, respectively.
There are a number of factors that may explain yield loss at the farm
level. The fall in world market prices of coffee has significantly reduced the
amount of money from coffee that farmers take home (Charveriat, 2001). As
a result of this, management of coffee has declined due to a reduction in the
amount of money that farmers have for purchase of inputs and for carrying out management practices on their coffee. The yield losses given above
are therefore a result of a multiplicity of factors and may not be entirely explained by the incidence or severity of CWD. To determine how much of the
yield loss was attributable to CWD, a model was developed using data for
farmers perceptions of the changes in yield before and following the onset of
CWD and data for the percentage of trees lost to CWD. This showed a direct
correlation between the changes in yield and the percentage of trees lost due
to CWD as depicted for Tanzania in Fig. 6.1.
The percentage annual yield loss attributable to CWD was then calculated as follows:
Predicted % Change in Yield
Number of trees lost to CWD per annum

100
Initial number of trees per farm
Fig. 6.1.
100
90
80
70
60
50
40
30
20
10
0
80
20
40
60
% Trees Infected of CWD
100
Predicted yield loss based on the incidence of CWD in Tanzania.
88
Musebe et al.
Using this method, annual reductions in coffee yield of 7.4%, 1.6% and
2.6% were calculated to be directly attributed to CWD in Uganda, Ethiopia
and Tanzania (Kagera Region only), respectively. Assuming that the loss due
to CWD is constant and taking into account the average number of years the
farmers have had CWD, this equates to a total yield loss of 39%, 16.7% and
10.5% in Uganda, Ethiopia and Tanzania, respectively. It should be noted
that coffee trees that had been recently planted and were not yet bearing
were excluded from the calculations.
In Tanzania, the amount of coffee produced fell significantly (P < 0.01)
after CWD infestation from an average of 950 to 686 kg/ha, which is a change
of 28.0%. A comparison of the mean coffee production before and after CWD
showed significant difference (P = 0.001). Coffee income fell from an average
of US$531 to US$132, which is a drop of 75%. Comparing the mean income
before and after CWD, the difference was found to be significant (P = 0.001).
In Uganda, coffee income reduced as a result of CWD. Among the small
farmers, income from coffee significantly fell after CWD for all severity levels (low, medium and high) and so did the quantity of coffee produced and
sold, except in the low-disease-severity areas where the change in coffee production and sales was insignificant. The area under coffee fell significantly
after CWD in the high-severity areas. At medium- and low-disease-severity
levels, the area under coffee rose, but the rise was statistically insignificant.
Overall coffee production reduced significantly after CWD (Fig. 6.2).
Changes in area, production, sales and income for coffee were separately
analysed for small-scale farmers (<5 acres) and large-scale farmers (>5 acres).
These changes were analysed at different levels of disease severity (low [L],
medium [M] and high [H]), wherein below 5% severity is considered low, 5%
to 20% severity is considered medium and above 20% severity is high (Table
6.2). In DRC, coffee production experienced a sharp fall of about 71.5% in
500000
production (mt)
400000
300000
200000
100000
0
1990
1992
1994
1996
With CWD
Fig. 6.2.
Impact of CWD at national level in Uganda.
1998
2000
Without CWD
2002
89
Table 6.2. Coffee area, production, sales and income before and after CWD infestation for
small farmers (<5 acres of coffee).
Category
Area (acres)
Incidence
level
L
M
H
Quantity produced
(60-kg bags)
L
M
H
Quantity sold
(60-kg bags)
L
M
H
Income (Ushs)
L
M
H
Before CWD
1.42 (0.91)
(n = 16)
1.87 (1.34)
(n = 36)
2.06 (1.35)
(n = 160)
7.19 (3.16)
(n = 14)
18.13 (17.79)
(n = 32)
19.17 (16.98)
(n = 154)
7.59 (2.90)
(n = 13)
16.71 (15.51)
(n = 29)
19.12 (16.95)
(n = 143)
338,542 (278,137)
(n = 12)
475,136 (317,844)
(n = 25)
714,202 (710,740)
(n = 123)
After CWD
1.73 (1.35)
(n = 18)
1.88 (1.39)
(n = 39)
1.19 (1.13)
(n = 139)
6.58 (4.74)
(n = 16)
11.78 (11.25)
(n = 33)
5.32 (5.25)
(n = 118)
7.92 (7.15)
(n = 17)
10.27 (8.02)
(n = 32)
5.05 (5.06)
(n = 112)
101,800 (97,626)
(n = 15)
163,559 (162,458)
(n = 29)
30,199 (27,508)
(n = 90)
T value
0.78, NS
0.03, NS
6.0***
0.4, NS
1.72, NS
8.5***
0.16, NS
2.06**
8.49***
3.08***
4.63***
9.12***
Figures in parentheses are standard deviation.

*, ** and *** imply statistically significant differences between means at 10%, 5% and 1% levels
of significance, respectively; NS = implies not significant; n = number of observations.
North Kivu province and 78% in eastern provinces. This is attributed to an

average destruction of 487 and 1029 coffee trees per farm in North Kivu and
eastern provinces, respectively.
A Uganda Coffee Development Authority (UCDA) survey confirmed
that all robusta-producing districts have been affected by CWD. In some
cases, entire fields of coffee have been destroyed. A 2003 review study estimated that of the total of 240,000 ha in all 21 robusta-growing districts,
122,400 ha have been infected. UCDA estimated a loss equivalent to 61,200 t
of coffee, which is around 40% of the output in recent years (Baffes, 2006).
The number of living trees in Uganda is smaller than it was in the early 1990s
(EPL, 2005). However, in its effort to contain the disease, Uganda has undertaken a replanting programme where farmers who uproot infected trees
are given coffee-planting materials. Likewise, coffee has been introduced in
northern Uganda as a cash crop (UCDA, 2004).
90
Musebe et al.
In the case of Ethiopia, there was a 37% loss in yield from coffee, which
fell from a yield of 481 to 303 kg/ha after CWD. The income from coffee fell
by 67% from US$553 to US$181 after CWD. The mean yield before and after CWD were significantly different (P < 0.05). These yield differences also
translate into production (Fig. 6.3).
Geiser et al. (2005) also asserted that there is a substantial loss in coffee
yield in Ethiopia, which is to a large extent determined by the wide spread of
CWD. The income before and after CWD were also found to be significantly
different. The yield loss attributable to CWD was then calculated by multiplying the average trees lost due to CWD and the average yield per tree.
This gave a yield loss of 23.1%. Thus, of the 37% change in yield, 23.1% was
a result of CWD.
In DRC, CWD infection caused a decrease in funds available for health
and education resulting in high reduction of schooling population and increase in infant death rate. Similarly, malnutrition and the number of homes
living below poverty levels increased. To compensate for fall of revenues
formerly generated by coffee, the farmers in DRC increased land for food
crop production for sale at the expense of household food consumption.
CWD destroyed almost 28,853 ha formerly setaside for robusta in DRC. This
was accompanied by the abandoning of much of the infrastructure leading
to immobilization of huge amounts of capital meant to contribute to socioeconomic development through employment. In 1989, territories of Beni and
Lubero in North Kivu had 24 coffee-processing factories with a capacity of
452 t per day and 24 warehouses with total capacity to stock 50,900 t. Currently, more than 90% of employment that these infrastructures were creating is non-existent.
Using the annual percentage yield loss (computed in section 5.4) for
each country, the impact of CWD at the national level was calculated using
FAO coffee production figures for 2002 (FAO, 2004) and International Coffee Organization indicator prices for 2002 (ICO, 2004). Annual losses attributed to CWD were 14,573 t amounting to US$9,644,279 in Uganda and 3360 t
350,000
Production (mt)
300,000
250,000
200,000
150,000
100,000
50,000
0
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
With CWD
Fig. 6.3.
Impact of CWD at national level in Ethiopia.
Without CWD
91
amounting to US$3,750,976 in Ethiopia. In the case of Tanzania, the national

level impact of CWD was calculated using the production figures for the
Kagera region because this was the only region where CWD was found. The
annual yield loss for Tanzania was 299 t amounting to US$197,551. CWD
has caused a disaster to the coffee industry at national level in DRC leading
to a vicious circle of poverty. There has been a reduction in investment and
resource mobilization leading to stagnation in the level of production and
productivity and a fall in revenues. The annual loss in DRC is 68,000 t of coffee, which translates to US$46.6 million.
6.5. Change in Input Use and Management of Coffee

Across the region, there has been a significant decrease (P < 0.01) in the use
of inputs on coffee. Prior to the onset of CWD, an average of 24.2% of farmers were using inputs on coffee (19.4% in Uganda and 29% in Tanzania). Of
the farmers that were using inputs, 46.2% were using organic manure, 15.9%
were using pesticides, 43.5% were using herbicides (only in Uganda) and
13.7% were using fertilizers. After the onset of CWD, the percentage of farmers using inputs on coffee across the region decreased to only 7.1% (11.2% in
Uganda and 3% Tanzania). Of those farmers still using inputs, use of manure
fell from 46.2% to 14.2% and the use of fertilizers also decreased from 13.7%
to 5.8%. Table 6.3 provides changes for individual countries.
Labour inputs into coffee production fell by an average of 33.5%. This is
because as income from coffee declines, farmers are discouraged from investing in the crop. A few farmers uprooted diseased plants (except in Ethiopia)
to remove coffee and replace it with other crops such as pineapple, bananas
and vanilla. In Uganda, the number of households spending on non-labour
inputs for coffee production after CWD reduced by 52% in the high-diseaseseverity areas and by 39% in the medium-disease-severity areas. Most farmers in Ethiopia did not use any inputs on coffee either before or after CWD.
There was no use of fertilizer, pesticides or insecticides. The only input on
coffee was labour. Before CWD, family labour input on coffee was an average of 1625 h per household per year, but after CWD, this decreased to 1297 h
per household, which is a 20% decrease. There was however a drop of 69% in
Table 6.3. Changes in level of input usage (%) in different countries.
Ethiopia
Tanzania
Uganda
Type of input
Before
CWD
After
CWD
Before
CWD
After
CWD
Before
CWD
After
CWD
Organic manure
Pesticides
Herbicides
Fertilizers
Family labour
0.0
0.0
0.0
0.0
100.0
0.0
0.0
0.0
0.0
80.0
62.0
31.0
0.0
7.0
100.0
2.0
31.0
0.0
7.0
83.0
30.4
5.8
43.5
20.3
100.0
0.0
2.0
45.0
0.0
75.0
92
Musebe et al.
hired labour from 4031 to 1244 h per household per year. Overall, there was
a 51% drop in the labour allocated to coffee.
Management practices on coffee in Tanzania include weeding, mulching, application of manure, stumping, pruning and spraying. Before CWD,
93% of the farmers weeded their coffee. This percentage did not change even
after the onset of CWD. However, the percentage of farmers who practiced
mulching fell from 47% to 38%. The highest reduction was in the percentage
of farmers who were pruning, which fell from 93% to 33%.
Management practices on coffee fields in Uganda include weeding,
mulching, pruning/de-suckering and spraying. Stumping is usually done
once in 5 to 6 years. Although stumping is one of the recommended practices,
farmers are reluctant to use it for fear of loosing the crop for several seasons.
Only a small proportion of farmers mulch their coffee. Coffee management
practices in the affected countries are declining due to neglect of coffee.
6.6. Asset Liquidation as a Result of CWD

Twenty-seven per cent of the coffee farmers throughout the region liquidated some of their assets to meet their financial obligations as a result of
reduced income from coffee. Assets included land (16% of farmers), livestock
(12% of farmers), electronic equipment (10% of farmers) and bicycles (12%
of farmers). Despite the sale of assets, some farmers still failed to meet their
financial obligations, the most common of which was completion of houses
that they had started to build prior to the re-emergence of CWD and school
fees for their children. The money obtained from the liquidated assets was
used mainly on investment in non-crop-farming enterprises such as starting
poultry farms, transportation business, house construction and for solving
household needs such as purchasing food and paying for medical expenses
and financing burial arrangements.
After the onset of CWD, 39.4% of farmers in Tanzania disposed some of
their assets. Of these, 5% disposed more than one asset. Most of the farmers
sold livestock, electronics and bicycles/motorcycles. The value of assets sold
varied from US$4 to US$1250. The liquidation of the assets is mainly due to
reduced income from coffee and increased expenditure on CWD.
In Uganda, a few farmers, particularly in the high-disease-severity areas,
liquidated some of their assets following the outbreak of CWD so as to cope
with the disease. The most commonly liquidated assets were land, followed
by communication equipment then bicycles and livestock of all types. The
money obtained from the liquidated assets was mainly used on investments
in non-crop-farming enterprises such as starting poultry farms, purchasing
motor bike to start transportation business, constructing houses and solving household and domestic needs such as purchasing food and paying for
medical expenses, school fees and financing burial arrangements. The value
of assets sold varied from US$20 to US$857 in Uganda. In addition, 16.6% of
coffee farmers had started borrowing money after CWD to meet their shortterm financial requirements, whereas 7% rented out land.
93
Twenty-six per cent of households interviewed in Ethiopia liquidated

their assets as a result of CWD. Most of the assets liquidated were livestock
including cows, oxen, sheep, goats, donkeys and chickens.
6.7. Changes in Household Expenditure Using Income From Coffee

Information was sought on changes in proportion of expenditure for various
items that were coming from coffee before and after CWD to measure the impact of CWD on farmers family welfare. Patterns of household expenditure
with income from coffee changed following the general reduction in income
from coffee, although the trends were not consistent across the region. In
Uganda, there was a significant reduction in the proportion of expenditure
on all items (education, health, food consumption, etc.) except labour following the outbreak of CWD. In Ethiopia, the proportion of income from coffee
spent on household items decreased significantly (P < 0.01) from 85.6% to
81.2% following the onset of CWD. Income from coffee was used primarily
on food, clothing and health, in order of priority, with some farmers depending on it for almost all of their food supply. The proportional reduction in
expenditure on food is leading to an increase in food insecurity, especially
protein. This in turn negatively impacts on productivity. In Tanzania, there
was very little change in the amount spent on items such as food and health,
but the proportion spent on education increased from 33.1% to 41%, and
there was a decrease in the proportion spent on labour (from 13.9% to 8.5%),
investments (from 26.3% to 9.1%) and leisure (from 13% to 6%). Thus, following the onset of CWD and the subsequent reduction in income from coffee,
households in Tanzania opted to cut down on non-essential items and reallocate the available income to essential items.
6.8. Farmer Coping Strategies and Factors Influencing Them

Following the onset of CWD, farmers adopted a range of different strategies
for coping with the effects of the disease. These included the following:
Re-allocating labour to other on-farm activities

Diversifying into other crops not previously grown including bananas,
beans, cassava, fruit, groundnut, maize, millet, sorghum, onion, peas,
sweet potato, tomato, vanilla and yams
Starting other non-farm activities including brewing, fishing and trade
Opening up new coffee fields
Replacing infected coffee trees
Renting out coffee fields
Borrowing money from accessible sources.
There were notable variations in the relative importance of different coping strategies across the region. Farmers in Uganda coped with CWD by
re-allocating labour to other on-farm activities (57.6%), re-allocating labour
94
Musebe et al.
to off-farm activities (22.4%), abandoning coffee fields after CWD (18.2%)

and renting out coffee fields (1.8%). In Ethiopia, however, farmers gave their
coping strategies as replanting coffee (76.5%) and replacing infected coffee
trees (23.5%), which is distinguished from replanting in that replacing means
planting new coffee at the same place where the uprooted coffee was. This
is common in Ethiopia as farmers often attempt to replant seedlings from
forest or other coffee fields. Unlike other countries in the region, farmers in
Ethiopia did not start other off-farm activities nor diversify into alternative
crops. This is because of the socio-cultural significance of coffee in Ethiopia,
which is deeply rooted in traditional culture and sentiment and farmers do
not easily resort to production of other crops. Regardless of what happens
to productivity, farmers keep their coffee at least for their own consumption. Moreover, production of other crops is highly constrained by wildlife
because most of the coffee areas have forests harbouring wild animals. In
Tanzania, the main coping strategies included diversifying into other crops
not previously grown (33%); starting off-farm activities such as brewing,
fishing and trading (19%); opening up new coffee fields (4%) and replacing
infected coffee trees (48%). In addition to the yield loss and decline in income
from coffee, farmers have been obliged to spend money especially for hiring
labour to uproot and burn infected trees. Some farmers also incurred additional costs to purchase chemicals and other control methods they felt would
control the disease. In Uganda for example, 58% of farmers adopted uprooting and burning of infested trees as a control measure. Other technological
control measures included the use of ash, urine and tobacco.
Socio-economic factors influencing the choice of coping strategies adopted by coffee farmers were evaluated using a binary logistic regression
model for each country. In Uganda, the choice of re-allocating labour from
coffee to other on-farm activities as a coping strategy for CWD was influenced by the number of years of experience of coffee farming, the yield of
coffee prior to the onset of CWD and whether coffee was the only source of
income for the household. Switching to non-farm activities as a coping strategy for CWD was influenced positively by the income of coffee prior to the
onset of CWD and the proportion of household labour inputs into coffee. Relatively higher coffee income before CWD meant the farmer was more likely
to switch to non-farm activities after the onset of CWD. Higher household
labour input into coffee prior to CWD meant that farmers were more likely
to switch to non-farm activities after infestation of CWD. The likelihood of
adopting non-farm activities as a coping strategy was also increased by the
proportion of trees lost and quantity of coffee produced before CWD.
In Ethiopia, the choice of coping strategy was influenced by the level of
education of the farmer. Farmers with a lower level of education were more
likely to replant coffee (97%) than farmers with higher levels of education (P =
0.012). Although replacing coffee trees is not common in Ethiopia, farmers
with no education were more likely to replace (29%) than farmers who had
primary school or higher levels of education (P <0.05).
Coping strategies were also influenced by two biological factors, namely,
severity of CWD on-farm and the yield of coffee prior to the onset of CWD.
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 10 20 30 40 50 60 70 80 90 100
Severity of CWD
95
Predicted probability of farmer
replacing wilted tree

replacing wilted tree

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
0.5
1
1.5
2
Yield per Tree (kg)
2.5
Fig. 6.4. Effect of severity of CWD (left) and coffee yield (right) on the predicted probability of a
farmer replacing a wilted coffee tree in Ethiopia.
0 20 40 60 80 100
Age of Farmer (years)
100
90
80
70
60
50
40
30
20
10
0
0
20
40
60
80
Years of coffee farming

growing alternative crops
100
90
80
70
60
50
40
30
20
10
0

growing alternative crops

engaging in non-farm activities
The higher the severity of CWD, the higher the probability of the farmer
replacing the infected coffee trees (P = 0.024) (Fig. 6.4). Similarly, the higher
the yield before CWD, the higher the probability that farmers would replant
their coffee (P = 0.037; Fig. 6.4).
In Tanzania, the likelihood of a farmer choosing non-farm activities as a
coping strategy was significantly reduced as the age of the farmer increased
(Fig. 6.5). Of the non-farm activities, fishing was the most practiced, and
because it was done mainly at night, it was particularly suitable for young
and/or unmarried men as compared with the older ones. Farmers who had
been growing coffee for a longer period of time were also more likely to grow
alternative crops (Fig. 6.5). Rather unexpectedly, the length of time CWD had
been present on a farm was negatively correlated with the likelihood of a
farmer switching to alternative crops (Fig. 6.5).
90
80
70
60
50
40
30
20
10
0
10
15
20
0
5
Years with CWD on farm
Fig. 6.5. Effect of age of farmer (left), number of years of coffee farming (centre) and number of
years of CWD on farm (right) on predicted probability of farmer diversifying from coffee to alternative crops as a result of CWD in Tanzania.
96
Musebe et al.
6.9. Conclusions
Coffee is an important commodity in the region because of its contribution
to income, foreign currency, employment and livelihood. Coffee production
and productivity are affected by CWD, which is a serious coffee disease in
the region, although its effect is not uniform and some places are affected
worse than others. CWD appears to have a significant impact, but this is
confounded by impacts on the coffee business at the global level. Farmer
understanding of the disease and its control methods are still low in the region, although higher in Uganda where the disease has been for long and
widespread. Low understanding of the disease coupled with financial limitations and knowledge of CWD occurrence has resulted in limited farmer
capacity to control CWD in the region. Low technical know-how and the low
availability of a diverse range of affordable control methods have also contributed to a situation where there is little attempted management of CWD,
and this has led to reduced coffee yields and overall coffee production. The
importance of coffee as a major source of income in the region declined due
to the effects of CWD. Use of inputs for coffee fell significantly in all the
countries as did use of labour for coffee. Similarly, production of coffee at
farm level fell significantly in all the countries. Following the re-emergence
of CWD, income from coffee declined by over 50% in the affected countries.
This has led to losses at farm, household and national levels. The livelihood
of farmers has therefore declined further because they were forced to sell
their assets and restrict expenditure to basic requirements only. Reduction in
regular income as a result of reduced income from coffee led farmers to liquidate their assets after CWD to meet their financial obligations. Household
expenditure fell in all the affected countries after CWD. Many coping strategies were adopted for CWD. There were however notable variations in their
relative importance across the region.
The rating of coffee as an important source of income in the region and
individual countries means that efforts should be directed at managing CWD
and increasing coffee production. CWD control requires concerted efforts.
An integrated approach involving farmers and the institutions concerned
with crop protection in the respective countries is essential. Farmer awareness of the disease has to be increased coupled with an increase in alternative methods of control. Continuously increasing awareness about CWD and
how to manage it is likely to lessen the rate of wilt spread. Raising awareness
is likely to be important where the severity of the disease is high and for
farmers that are most likely to persevere with coffee rather than diversify
to other activities. Awareness campaigns through national extension programmes, farmer groups and farmers field schools are needed. Awareness
creation should be associated with networking on how to manage the disease. Similarly, farmer capacities need to be improved by facilitating access
to the requisite resources for control of the disease. An appropriate farmers
field school curriculum to empower coffee farmers using participatory approaches to raise awareness of CWD and its management would help reduce
CWD. Farmer training and implementation of good husbandry procedures
97
and practices can increase potential for stabilizing the social situation in the
long term. In this way, rural incomes can also be improved through value
addition.
Host resistance would offer the best prospect for control and demand
very little effort from the farmers. This means that a long-term solution to
CWD would be development of wilt-disease-resistant varieties. Intensification and acceleration of research for wilt-resistant varieties should receive attention (see Chapter 10). In the short term, breeding should be accompanied
with provision of free and clean planting materials. Revival of research to
generate a diverse range of control strategies should also be considered.
Acknowledgements
Collection of socio-economic data and information was facilitated by many persons in different organizations in target countries. The country level contacts
for the CWD project are highly acknowledged. The coordinator of the coffee
wilt project is highly acknowledged for providing assistance in data collection
and provision of most of the literature reviewed while preparing this chapter.
We wish to thank CABI Africa and other organizations to which the co-authors
are affiliated. This work was funded by the European Union and the Common
Fund for Commodities. The financial support is highly appreciated.
References
Baffes, J. (2006) Restructuring Ugandas coffee industry: why going back to basics matters. Development Policy Review 24, 413436.
Baffes, J. (2003) Tanzanias coffee sector: constraints and challenges in a global environment. Africa Region working Paper No. 56, June 2003. www.worldbank.org/
afr/wps/wp56/htm.
Banque Centrale du Congo (BCC) (2005) 19952007 Rapports Annuels dActivits.
Beintema, N.M. and Solomon, M. (2003) Agricultural science and technology indicators for Ethiopia. ASTI country Brief No. 9. October 2003.
Charveriat, C. (2001) Bitter coffee: How the poor are paying for the slump in coffee
prices. Oxfam, Oxford.
EPL (The Evaluation Partnership Limited) (2005) Impact evaluation of the four CFC funded projects in Uganda. The Evaluation Partnership Limited, Middlesex, UK.
FAO (2004) Agriculture and food trade statistics. www.faostat.fao.org.
Geiser, M.D., Lewis, I.L.M., Hakiza, G., Juba, H.J. and Miller, A.S. (2005) Gibberella
xylariodes (anamorph: Fusarium xylariodes), a causative agent of coffee wilt disease in Africa, is a previously unrecognized member of the G. fujikuroi species
complex. Mycologia 97, 191201.
GoU (Government of Uganda) (2001) The path forward in Ugandas coffee sector. A
report prepared for the presidential conference on export competitiveness. Prepared by the complete project and Alice Agowa (Uganda coffee) on behalf of the
European commission, February 2001.
ICO (2004) Trade statistics. Available at: www.ico.org/frameset/traset.htm.
98
Musebe et al.
Karanja, A.M. and Nyoro, J.K. (2002) Coffee prices and regulation and their impact on
livelihoods on rural community in Kenya. Tegemeo Institute of Agricultural Policy
and Development. Working paper, 64 pp.
Oduor, G., Simons, S., Phiri, N., Njuki, J., Poole, J., Pinard, F., Kyetere, D., Hakiza,
G., Musoli, P., Lukwago, G., Abebe, M., Tesfaye, A., Kilambo, D., Asiimwe, T.
and Munyankere, P. (2003) Surveys to assess the extent and impact of coffee wilt
disease in East and Central Africa. Final Technical Report EU contract No. ASARSP/CV-006. CAB International, Egham, UK.
Oxfam (2002) Crisis in the birth place of coffee. Oxfam International research paper,
September 2002. Available at: http://www.maketradefair.com/en/assets/english/
coffeecrisisKafaEthiopia.pdf.
UCDA (2001) UCDA Annual Report, Industrial Graphics Limited, Entebbe, Uganda.
UCDA (2004) UCDA Annual report, Industrial Graphics Limited, Entebbe, Uganda.
Biology, Taxonomy and Epidemiology

of the Coffee Wilt Pathogen Gibberella
xylarioides sensu lato
M.A. Rutherford,1 D. Bieysse,2 P. Lepoint3 and H.M.M. Maraite3

1
CAB International (CABI), Europe-UK, Bakeham Lane, Egham,

Surrey, TW20 9TY, UK
2
Centre de Coopration Internationale en Recherche Agronomique pour le
Dveloppement (CIRAD), TA41/K, Campus International Baillarguet,
34398 Montpellier, Cedex 5, France
3
Unit de Phytopathologie, Universit Catholique de Louvain (UCL),
Croix du Sud 2/3, B-1348 Louvain-la-Neuve, Belgium
7.1. Introduction
Coffee wilt disease (CWD) was first reported on excelsa coffee (Coffea liberica
var. dewevrei, formerly Coffea excelsa) in Central African Republic (CAR) in
1927 (Figueres, 1940). Although subsequent epidemics in CAR, Cte dIvoire
and Democratic Republic of Congo (DRC) between the 1920s and 1960s
caused extensive damage to C. liberica and Coffea canephora (robusta coffee),
the disease was successfully reduced mainly by sanitation and elimination of
the most susceptible genotypes to the level of a minor disease (Fraselle, 1950;
Jacques-Felix, 1950; Saccas, 1951; Jacques-Felix, 1954; Stewart, 1957; Meiffren,
1961; Kranz, 1962) (see also Chapter 9). However, renewed and devastating
outbreaks on C. canephora in DRC, Uganda and Tanzania, coupled with worsening problems on Coffea arabica in Ethiopia (Kranz and Mogk, 1973; Girma
et al., 2001; Girma, 2004; Girma et al., 2006), are major cause for concern for
future sustainability of coffee production in Eastern and Central Africa.
Until recently, little was known about the fungus responsible for CWD,
Gibberella xylarioides Heim & Saccas, in comparison with other fusarial pathogens possessing similar capacity for destruction and economic loss. In 2000,
the Regional Coffee Wilt Programme (RCWP) was initiated to manage CWD
through the development and implementation of an effective and sustain-
99
100
Rutherford et al.
able disease management programme based principally on host resistance.

As part of this programme, extensive multidisciplinary research was undertaken to improve our understanding of the occurrence of CWD and of the
biological nature of G. xylarioides.
This chapter provides a review of current knowledge of the pathogen,
incorporating comprehensive new information acquired through international research efforts supported by the INCO-Dev Programme of the
European Community and the UK Department for International Development, Crop Protection Programme.
7.2. The Cause of CWD

CWD has been attributed to the fungus G. xylarioides Heim & Saccas (anamorph Fusarium xylarioides Steyaert; syn. Fusarium oxysporum forma xylarioides [Steyaert] Delassus). Steyaert (1948) described the anamorph,
F. xylarioides, as a new species causing wilt of coffee (C. excelsa).
7.3. Symptoms
The external symptoms exhibited by coffee plants affected by CWD are generally similar to those of other vascular wilt pathogens. Symptoms may appear at any stage of crop growth, and the rate at which they develop may
vary. Once affected, death of the plant is inevitable, and in mature trees, it
usually occurs between 3 and 15 months after first appearance of symptoms.
Young plants, however, may be killed within a matter of a few weeks of
infection.
In young plants, leaves initially exhibit chlorosis (yellowing), begin to
droop or wilt and develop brown necrotic lesions, initially along the margins
and veins. Lower leaves tend to be affected first, lesions gradually enlarging
until the leaves dry, shrivel and abscise, often within a matter of days. In
more mature plants, a general chlorosis, wilt and folding or rolling of leaves
are among the earliest symptoms observed. Symptoms are often initially restricted to, or more pronounced on, one part of the tree, a likely consequence
of initial infection occurring on one of several main stems. Dark brown or
black streaking of shoots, clearly visible on younger green shoots, and discolouration of leaf veins (Plate 5) may also occur. Symptoms again culminate
in drying and loss of leaves, constituting a progressive and often unilateral
dieback and defoliation that gradually extend to the entire plant. Chlorosis,
leaf droop and leaf abscission in particular may be more pronounced under
natural conditions of water stress (e.g. dry season, drought), less so with the
onset of the rains. Coffee berries that would normally be green may redden
as if ripening prematurely but often remain intact on shoots following shedding of leaves.
Other external symptoms include swelling of the trunk and the appearance of vertical or spiralling cracks in the bark of mature trees. Small blackish-
Biology, Taxonomy and Epidemiology of the Coffee Wilt Pathogen
101
brown perithecia of the fungus, similar in appearance to dark soil particles,

may be produced in the cracks of the bark (Plate 9), particularly during the
rainy season or otherwise moist conditions. As described below, perithecia,
produced primarily as sexual reproductive structures, may also constitute
a means of survival for the pathogen and are commonly observed on dead
and decaying material of plants affected by CWD.
It is important to note that many of the external symptoms associated
with CWD are not unique to the disease and may also be caused by other
pathogens, including other Fusarium species (see 7.10), root rot, physiological disorders, such as overbearing dieback, and adverse environmental conditions, including water stress. However, confirmation of CWD can usually
be confirmed by the development of bluish-black staining of the wood directly beneath the bark. Staining is usually most pronounced in the collar
region but may extend below soil level and/or towards the apex of the tree.
Such staining may be confirmed by careful removal of the bark, by hand or
with an implement.
7.4. Disease Cycle and Epidemiology

Although recent research has provided further information on the life
cycle of G. xylarioides and on disease epidemiology, our understanding of
the pathogen remains limited in comparison with many other fusaria and
vascular wilt pathogens. Some studies of hostpathogen interactions have
been undertaken (Chapter 8), but further in-depth research is required to
clarify the primary infection points, investigate mechanisms of tissue colonization following infection, determine what physiological and biochemical
responses are elicited by the plant and investigate the existence, role and
interactions between genes conferring resistance in the host and virulence in
the pathogen.
7.4.1. Host range

G. xylarioides would appear to have a narrow host range, coffee being its
primary host and perhaps the only plant on which it is pathogenic. It is
clear from observations of the occurrence of CWD in the field (including
those of comprehensive and systematic surveys completed across East and
Central Africa), the results of pathogenicity testing and isolations made
from host substrates that the fungus is prevalent on the two most important commercially cultivated coffee species, C. arabica and C. canephora
(Kilambo et al., 1997; Girma, 2004; Kilambo et al., 2004; Oduor et al., 2004;
Tshilenge-Djim et al., 2004; Kilambo et al., 2006) (see also Chapters 3, 4 and
5). The disease was also recently observed on excelsa coffee in Uganda,
but only on plants held at a coffee germplasm conservation site (to the
authors knowledge, commercial production of excelsa coffee in Africa is
very limited). G. xylarioides was reported on wild C. canephora exhibiting
102
Rutherford et al.
CWD symptoms in forest areas of Kibale and Itwara in west Uganda and
confirmed as pathogenic to C. canephora in subsequent pot-based tests
(Bieysse, 2005; Bieysse, 2006).
A range of other crops and weed species commonly found or cultivated
in coffee farms affected by the disease in Uganda have been investigated as
possible alternate hosts for G. xylarioides (Serani, 2000; Kangire et al., 2002;
Serani et al., 2007). However, the fungus was only recovered from the roots
of the banana cultivar kayinja (syn. pisang awak) (Serani, 2000; Serani et
al., 2007), often intercropped by farmers alongside coffee for production of
banana juice and preparation of beer. This strain was recovered from within
the root tissues and found to be pathogenic to C. canephora in subsequent host
inoculation studies (Serani, 2000; Serani et al., 2007; G. Hakiza, personal communication, 2004). Pathogenicity to banana was not assessed. The extent of
similarity, including genetic, between this strain and those recovered from coffee affected by CWD is otherwise unclear.
In DRC, a fungus identified as G. xylarioides on the basis of morphological
and genetic characteristics (i.e. comparison of tef 1a sequences against those
held for G. xylarioides on the FUSARIUM-SEQFUSARIUM-ID database;
Geiser et al., 2004), has been found on cinchona affected by Phytophthora canker
(A. Buddie, P. Cannon and P. Kelly, personal communication, 2007). As yet, it
is unknown whether this fungus is pathogenic to cinchona or indeed to coffee. G. xylarioides has also been found on cotton seed (Pizzinatto and Menten,
1991) and on rotting tomatoes obtained from fruit markets in Nigeria (Onesirosan and Fatunla, 1976). However, information supporting these reports
is limited, and it is unknown whether these fungi differ morphologically,
genetically or otherwise to those currently observed on coffee. Studies to investigate and identify other potential hosts for G. xylarioides and to assess the
pathogenic nature of the fungus on such hosts are otherwise limited. Insects
have been suggested as possible vectors for CWD, but attempts to recover
G. xylarioides from common pests of coffee (including coffee berry borer,
Hypothenemus hampei), as well as from bees, termites and other insects found
on affected coffee farms, have been unsuccessful (Rutherford and Flood, 2005;
G. Hakiza, personal communication, 2004).
7.4.2. Infection and colonization of coffee by G. xylarioides

G. xylarioides is considered to be an endemic, soil-inhabiting fungus. Although the life cycle and epidemiology of the fungus are not fully understood, the mechanisms of host infection, colonization and symptom
development are thought to be similar to those of other vascular wilt
pathogens. Initial entry is considered primarily via the roots and lower
stem, facilitated by the presence of wound sites as entry points. These
may occur naturally as a consequence of farm management practices (see
6.4.3) or by livestock foraging at the base of the coffee tree. Once within
the host tissues, the fungus moves within the vascular system and surrounding tissues, leading to restriction of water transfer through direct
103
physical occlusion with fungal material or by the host plant as a response

to invasion. As with other vascular wilts, staining of the wood and roots
is probably due to changes in phenol metabolism and may constitute
a resistance response by the plant (Pegg, 1981; Beckman, 1987; Hillocks,
1992).
7.4.3. Fungal transmission and dynamics of disease spread

Vegetative and reproductive structures of the fungus, namely, microconidia,
macroconidia, ascospores and mycelial fragments, may be readily dispersed
by air, water (including rain splash and run-off ) and through human activity, perhaps attached to soil and plant debris (Van der Graaff and Pieters, 1978). A profuse production of conidia is noticed on faint mycelium
patches developing mostly at the stem base under high-humidity conditions
(Fraselle, 1950). Perithecia, often observed on coffee wood under natural
conditions and often abundant during the rainy season, may constitute as
a means of survival given the sparsity or absence of the thick-walled chlamydospores produced by other fusaria (Flood and Brayford, 1997). Perithecial production has been reported to commence on CWD-affected trees
within 2 weeks of the onset of the rainy season and to be more pronounced
towards the stem base (Musoli and Hakiza, 2007). Day-to-day agricultural
practice, exchange of planting material between growers, dissemination of
germplasm by nurseries and distribution centres and transportation of coffee by traders are likely to disseminate fungal propagules and contaminated
soil, providing an efficient means of disease spread on a local, national or
regional scale (Chapters 3, 4 and 5). Wood cuttings, obtained as a means
of vegetative propagation, are also exchanged and sold by growers and
nurseries to establish new plantings. Coffee berries are obviously widely
distributed and marketed and, if contaminated by G. xylarioides, may again
provide an effective means of dispersal of the fungus. Isolation of G. xylarioides from within coffee berries has been reported in Uganda (Serani et al.,
2007).
The machete is an essential implement for coffee farmers in Africa and is
used on a daily basis to perform a range of tasks including pruning and weeding. On-station trials confirm that wounding of coffee trees with a machete
previously used on infested coffee wood is sufficient to transmit the pathogen
to, and result in CWD development on, healthy and mature trees under field
conditions. Coffee wood, an important source of fuel for farmers, also constitutes a source of infection for young coffee plants, while infested field soil may
remain infective to young plants for several months following uprooting and
removal of affected trees (Rutherford and Flood, 2005; Musoli and Hakiza,
2007; Chapter 3, this volume).
Studies to monitor the spatial and temporal development of CWD symptoms at on-station and on-farm locations affected by CWD provide an insight
on disease spread under natural conditions (Rutherford and Flood, 2005; Musoli, 2007; Musoli et al., 2008). In one instance, incidence of CWD in mature
104
Rutherford et al.
plants of a range of susceptible clones increased from below 3% (9 plants) to

45% (166 plants) over a 2-year period with, on average, six plants developing
external symptoms per month (Rutherford and Flood, 2005). Thirty months
later, more than 90% of plants had either been killed or showed symptoms of
the disease. The rate of increase in disease incidence also varied from clone
to clone, suggesting different levels of susceptibility. The studies indicate
that although initial foci of disease (in some cases individual trees) may be
randomly distributed, these enlarge over time but in no particular direction
(Plate 16). They also suggest that infection of adjacent trees, perhaps through
root-to-root contact or short-distance dispersal of fungal material, is the primary means of on-farm disease development (Rutherford and Flood, 2005;
Musoli and Hakiza, 2007). Although the point at which individual plants
became infected in these studies is unknown and the precise mode of infection is unclear, they provide the most accurate information to date on localized spread of CWD and do confirm the rapid and destructive effects of the
disease once introduced. They also have important implications in terms of
cultural management and the potential for eradicating CWD through early
removal and destruction of plants as soon as symptoms become apparent
(see Chapter 9). An illustration of the putative life cycle of G. xylarioides is
provided in Fig. 7.1.
Fig. 7.1. Putative life cycle of the coffee wilt pathogen G. xylarioides (anamorph F. xylarioides),
as composed by H. Maraite based on drawings from Van den Abeele and Vandenput (1956),
Meiffren (1957), Booth (1971), von Blittersdorff and Kranz (1976) and Pochet (1988). (Courtesy of
H. Maraite, Universit Catholique de Louvain, Belgium.)
105
7.5. Isolation of G. xylarioides from Plant Tissues and Soil

G. xylarioides may be readily isolated from the stems, twigs and roots of young
coffee plants and mature trees exhibiting disease symptoms. The fungus may
also be grown on a range of suitable agar media. As with other fusaria, this is
usually achieved by placing small pieces of surface sterilized wood excised
from symptomatic plant tissues on to the surface of tap water agar medium
(Booth, 1971). After a few days, a small amount of fungal material from an
emerging colony may be transferred on to fresh Spezieller Nhrstoffarmer
agar medium (Nirenberg, 1976) to induce more prolific growth and sporulation. Isolations are most successful when made from wood pieces excised
from immediately below the bark and exhibiting the blueblack discolouration typical of CWD (see 7.3). Fungal cultures should be incubated at 25C
and, if possible, under daylight fluorescent light tubes. Some exposure of
cultures to ultraviolet light may improve sporulation.
In contrast, and while it can be successful, isolation of G. xylarioides from
soil has often proven problematic. This is largely due to more rapid colonization of the agar medium by other fungi, including other fusaria (see 7.10),
and is a constraint to in vivo research. These fungi may compete with and
inhibit the normal growth of G. xylarioides or may simply obscure the presence of the pathogen in culture. Colonization of, and development of CWD
symptoms on, susceptible coffee grown in either naturally contaminated or
artificially inoculated soil is an indication of the presence of the fungus. It
also highlights the potential of this approach, albeit time consuming, as a
means of baiting for G. xylarioides. More rapid and straightforward procedures for isolating the fungus or simply confirming its presence in soil (and
indeed in plant material) are required, including molecular and biochemical
approaches similar to those developed for fungal pathogens of other crops
(Doohan et al., 1998; Alves-Santos et al., 2002; Knoll et al., 2002).
Following isolation of the fungus from a substrate, and prior to attempting identification or more in-depth characterization, purified colonies should
be prepared by transfer and culture of a single spore (microconidium or macroconidium) on fresh agar medium.
7.6. Storage and Preservation of G. xylarioides

Strains of G. xylarioides may be maintained over the short to medium term
on Spezieller Nhrstoffarmer agar plates or slants stored at 20oC to 25oC
(for more routine use) or at 5oC (i.e. normal refrigerator temperature). For
longer-term storage and preservation, strains may be preserved by depositing at very low temperatures under liquid nitrogen (cryopreservation) and/
or in a freeze-dried state (lyophilization) (Ryan and Smith, 2007). Representative strains or those exhibiting characteristics of interest or value should be
deposited in this manner, as should strains intended for future use. This is, of
course, dependent on access to suitable facilities, something that is currently
beyond the scope of many African countries. As yet, the viability and condi-
106
Rutherford et al.
tion of strains stored for prolonged periods by cryopreservation and lyophilization has not been thoroughly evaluated. G. xylarioides strains, representing
specific biological traits, are currently held in Genetic Resource Collections
at CAB International (CABI) E-UK (Egham, UK), Centre de Coopration Internationale en Recherche Agronomique pour le Dveloppement (CIRAD;
Montpellier, France) and MUCL/BCCM (Louvain-la-Neuve, Belgium).
7.7. Cultural and Morphological Characteristics of G. xylarioides

As with many fusaria, the cultural and morphological characteristics exhibited in vitro by G. xylarioides depend on a number of factors, including strain,
growth medium and environmental conditions, and can vary considerably.
A description of the main cultural and morphological characters, sufficient
to enable identification of the fungus to species level, is provided in Booth
and Waterston (1964) (see below): this is based on growth of the fungus on
potato sucrose agar (Booth 1971), a medium routinely used for identification
of Fusarium species. For a full explanation of terminology, refer to Ainsworth
and Bisbys Dictionary of the Fungi (Kirk et al., 2008).
Fungal colonies are initially pale beige with sparse white mycelium.
Purple discolouration may develop as the colony matures, accompanied by
the production of dark bluish-black, discrete stromata, some representing
ascomatal initials. Microconidia and macroconidia are produced in slimy
masses on short conidiogenous cells, pionnotes or sporodochia on the vegetative mycelium. Microconidia, generally abundant, are unicellular, allantoid, curved and 5 to 10 2.5 to 3 mm (Plates 10 and 11). Macroconidia, which
tend to be less abundant, are fusoid, falcate, 23 septate and 20 to 25 4 to
5 mm. Chlamydospores may be produced but generally are not abundant.
They are oval to globose, smooth or roughened and 10 to 15 8 to 10 mm.
Globose perithecia, 200 to 400 180 to 300 mm, violaceous in colour, are embedded singly or in groups in dark purple stromata. Asci within the perithecia are cylindrical, thin-walled, shortly pedicellate and 90 to 110 7 to 9.5 mm
(Plate 12). Each ascus contains eight monostichous, hyaline to straw-coloured
ascospores that are fusoid, 13 septate, finely roughened and 12 to 14.5 4.5
to 6 mm.
7.8. Variability Within G. xylarioides

7.8.1. Cultural and morphological variability
Cultural and morphological variability is apparent among strains of
G. xylarioides grown on the same culture medium and maintained under
uniform environmental conditions. This can be beneficial in delineating
strains into intraspecific subgroups, and although these can relate to genetic
traits, field behaviour and other attributes, a clear link is often not apparent.
Differences, however, have been observed in pigment production (Plate 13),
107
growth rate, growth response to temperature (optimum and maximum

for growth in culture) and teleomorph form between strains associated
with C. arabica and those associated with C. canephora and excelsa coffee
(Girma, 2004; Tshilenge-Djim et al., 2004; Lepoint, 2006), some of which reflected newly defined biological species (BS) designations (Lepoint, 2006) and
formae speciales (Girma, 2004) (see 7.9). Although examination of such traits
provides opportunities for identification and differentiation of strains, there is
a need to exercise caution, given that environmental conditions alone, which
vary considerably across laboratories even within East Africa, can influence the
cultural characteristics of a strain significantly.
For further information on methods of isolation, culture, purification,
identification, storage and preservation of G. xylarioides and other Fusarium
species occurring on coffee, refer to Rutherford (2003).
7.8.2. Genetic variability and the role of the sexual cycle

The presence of a perfect state in G. xylarioides implies at least some degree
of variability within the coffee wilt pathogen, but until recently, little was
known about the underlying genetics of the fungus, the extent of genetic
variability (if any) that exists or how it relates to other attributes, including
field behaviour. This was partly a consequence of limited expertise and methodology to investigate genetic diversity. However, the recent re-emergence
and devastating effects of the disease in Africa, coupled with development
and improved access to modern molecular approaches, has renewed interest in the fungus. As a consequence, a number of research centres across
the world have now completed comprehensive studies to investigate genetic variability within G. xylarioides and to explore the relationship between
G. xylarioides and other fusaria (namely from coffee) from a diagnostic and
taxonomic perspective.
As facilities and expertise required to undertake molecular research
in Africa remain limited, the majority of such work has been undertaken at
CABI, CIRAD and the Universit catholique de Louvain (UCL) in the UK,
France and Belgium, respectively. A broad range of primarily PCR-based approaches e.g. analyses of microsatellite loci, presumptive mitochondrial
DNA restriction fragment length polymorphisms, amplified fragment length
polymorphisms, inter-simple sequence repeat [ISSR] anchored sequences,
nuclear housekeeping genes [translation elongation factor, tef, calmodulin, CL, histone H3] and intergenic spacer and internal transcribed spacer
regions were applied to explore intraspecific genetic variability within designated regions of the G. xylarioides genome and within specific cell structures
(e.g. nuclei, mitochondria, ribosomes). Analysis of internal transcribed spacer,
for example, has been successfully used to delimit species, whereas analysis
of ISSR, intergenic spacer and microsatellites can reveal interspecific and intraspecific variability. ISSR analysis in particular can reveal genetic variability
within fungi in a similar manner to random amplified polymorphic DNA but is
deemed to provide more reliable discrimination of individual fungal strains.
108
Rutherford et al.
As may be expected, the various molecular approaches differed in the

manner and extent to which they exposed genetic variability. Nevertheless,
the findings of these studies, which examined several hundred strains of
G. xylarioides, were generally consistent and provided an intriguing insight into genetic diversity and how it relates to other recognized traits.
Although overall variability within the pathogen would appear to be limited, the research suggests that two major genetically distinct populations
are responsible for current outbreaks across DRC, Uganda, Tanzania and
Ethiopia (Girma, 2004; Tshilenge-Djim et al., 2004; Adugna et al., 2005; Lepoint, 2006; Rutherford, 2006; Bieysse, 2007). The first of these consists of
G. xylarioides obtained from C. canephora affected by CWD in Uganda, DRC
and Tanzania since re-emergence of the disease. Interestingly, this group
also includes a single strain obtained from excelsa in Uganda. The second
population consists of G. xylarioides obtained from C. arabica affected by
CWD in Ethiopia. Despite the level of resolution possible with some of the
molecular approaches applied, genetic diversity is not apparent within either of the two populations. As yet, therefore, it has not been possible to
more precisely define the origin of strains in each population either geographically or in terms of the coffee species or cultivar of origin. This is
perhaps surprising, given that production of perithecia by the fungus is
frequently observed on coffee trees under suitable conditions in the field
(Van der Graaff and Pieters, 1978; Girma et al., 2001) and assuming, of
course, that the perithecia so observed are a reliable indicator of sexual
reproduction.
Genetic variability also exists among G. xylarioides strains recovered from
C. canephora and C. excelsa during earlier CWD epidemics in CAR, Guinea
and Ivory Coast, all of which differ genetically to the two populations responsible for current outbreaks. Unfortunately, few strains are available for
study from these earlier outbreaks, and these may have become altered due
to routine sub-culturing on nutrient-rich culture media since their isolation.
Uncertainty therefore remains as to how and why this variability may have
arisen and how it relates to field behaviour. Indeed, whether any of these
strains were responsible for the earlier outbreaks, and to what extent, remains unknown. Furthermore, we do not know precisely how these strains
relate to those currently affecting C. canephora, C. excelsa and C. arabica in
DRC, Uganda, Tanzania and Ethiopia. It is feasible that the limited genetic
diversity observed in current strains may have arisen as a consequence of
(i) selection pressure imposed by widespread cultivation of resistant coffee
genotypes (species or varieties) to counteract the earlier epidemics and/or
(ii) the relative fitness and rapid spread of a single or small number of survivors or variants, perhaps in response to cultivation of resistant germplasm.
Apparent sexual incompatibility observed in the fungus (see below) may
have also played a role in emergence of the genetic variants encountered to
date.
The existence of a functioning sexual cycle in G. xylarioides, as in any
organism, can play a significant role in the exchange of genetic information and hence the ability of the pathogen to adapt to changing conditions,
109
including those brought about by modification of crop management practices. Booth (1971) considered G. xylarioides to be a heterothallic fungus composed of sex-linked male and female forms that could be distinguished by
their morphological features. As described below, Booths inclusion of
the male form within G. xylarioides now appears inappropriate, as genetic
studies show it to be more closely related to Fusarium lateritium than to
G. xylarioides (Geiser et al., 2005; Lepoint et al., 2005). Booths original description may in some ways be understandable, given that F. lateritium is
also a recognized pathogen of coffee, and simultaneous colonization of CWD
affected plants by several fusaria is observed (Tshilenge-Djim et al., 2004,
Lepoint 2006).
Recent research nevertheless confirms that G. xylarioides is indeed heterothallic. At UCL, the teleomorph of the fungus was successfully generated when two tester strains of opposing mating type, each derived from a
single conidium or ascospore, were paired in vitro (Lepoint et al., 2005). Pairing strains obtained from different geographic locations and coffee species
enabled discrimination of three BS, BS1, BS2 and BS3, within which strains
were sexually compatible (Plate 14). A sterility group (SG4), composed of
a number of reproductively sterile strains, was also identified. The latter
are sexually incompatible with each other and with strains in each of the
BS, whereas strains belonging to any one BS are sexually incompatible with
those of another. It should be noted that in these mating tests, teleomorph
production and sexual compatibility are confirmed by production of fertile
ascospores and not merely fusion of fungal mycelium or production of empty
perithecia (protothecia) (Plate 15). To date, morphological discrimination of
mating types has not been possible, all strains exhibiting the characteristics
previously described for Booths female form.
In these studies, a clear relationship between the genetic and biological attributes of the pathogen was observed, in that BS1 and BS2 comprised strains associated, respectively, with C. arabica in Ethiopia and with
C. canephora in DRC, Uganda and Tanzania (Lepoint, 2006). BS3 comprised
the small number of strains associated with CWD during the earlier and very
damaging outbreaks in Central and West Africa. Of importance, and given
the time required to perform in vitro mating assays, molecular characterization (PCR amplification and sequencing) of the mating type gene (MAT),
based on primer pairs previously developed for F. oxysporum and the Gibberella fujikuroi species complex (GFC), enabled the identification of mating type
idiomorphs, MAT1-1 and MAT1-2, thereby supporting the hypothesis that
G. xylarioides sensu lato is heterothallic (Lepoint et al., 2005; Lepoint, 2006).
This now enables mating types to be identified before confirmatory crosses
are performed or where these are not successful. Taxonomically, and although
there are still differing opinions as to their suitability and relevance, sequencing and phylogenetic analysis of MAT loci also place the fungus within the
African clade of the GFC, an important taxon comprising pathogenic fusaria,
and resolves four distinct phylogenetic species corresponding to the BS and
SG revealed in the mating tests. These lineages were also resolved by random
amplified polymorphic DNA analysis and amplification and sequencing of a
110
Rutherford et al.
combination of non-MAT nuclear genes (Fig. 7.2). Of note, infertility of mating crosses between G. xylarioides and recognized GFC mating populations
suggests the coffee wilt pathogen to be a new BS within the GFC. Mating
type tester strains identified for the four lineages have been deposited for
secure, long-term storage at MUCL/BCCM.
The findings concerning the reproductive cycle may, at least in part, provide an explanation as to why perithecia production readily occurs in nature
but (to the authors knowledge) has never been observed on coffee plants
inoculated with fungal cultures derived from a single conidiospore. The ability to cross representative strains of G. xylarioides in such mating tests also
offers new possibilities to further explore the nature, origins and transfer
Fig. 7.2. Maximum-parsimony phylograms based on the combined autosomal (tef + CL + H3)
data set of representative strains of the G. xylarioides complex (GxC) and Gibberella indica complex from diverse geographical and host origins. BS and SG defined in carrot agar crosses are
indicated by a coloured box for the GxC and by a dotted line for the G. indica complex next
to terminal clades resolved. Trees were generated with PAUP v.4.0 b10 (Swofford, 2001) using
F. oxysporum as outgroup and available National Center for Biotechnology Information sequences for closely related species belonging to the GFC African clade (ODonnell et al., 1998). Bootstrap values based on 1000 replications are indicated in percentages at internodes when replication frequencies exceed 50%. (Courtesy of P. Lepoint and H. Maraite, UCL, Belgium.)
111
of important characteristics of the CWD pathogen and to more accurately

assess the likelihood of new variants emerging and implications for future
disease management.
7.8.3. Pathogenic variability

Extensive host range testing has been undertaken both under natural (e.g.
screen house) environments in Africa and controlled (glasshouse) conditions
in Europe to clarify the nature of the plantpathogen relationship in CWD (Pieters and Van der Graaff, 1980; Tshilenge-Djim et al., 2004; Adugna et al., 2005;
Girma et al., 2006; Kilambo et al., 2006; Musoli et al., 2006; Bieysse, 2007). The
primary aim of these studies was to evaluate specificity and aggressiveness
in the pathogen and resistance within coffee, partly as a means of identifying
possible sources of resistance for use in ongoing breeding programmes. They
involved inoculation, by various means, of a range of coffee species, cultivars
and clones obtained from Africa and other coffee-producing areas, including C. canephora, C. arabica and C. excelsa, with purified strains of G. xylarioides
obtained from different coffee species and locations in Africa where CWD is,
or was, prevalent. Strains associated with both the current outbreaks in DRC,
Uganda, Tanzania and Ethiopia and with the earlier outbreaks in West and
Central Africa were included in these studies.
Host specificity within the pathogen was clearly shown, with the majority of strains inducing CWD symptoms only on the coffee species from
which they had been derived (Chapter 8). The findings reflect those of other
approaches to pathogen characterization, including the studies of genetic
diversity described above, as they again delineated G. xylarioides into two
geographically defined populations one comprising strains originating
from Ethiopia and pathogenic towards C. arabica, and the other strains originating from DRC, Uganda and Tanzania and pathogenic to C. canephora. The
findings also confirm earlier suspicions of host specificity based solely on
the observed occurrence of CWD on C. arabica and C. canephora in the respective countries, including the findings of comprehensive and systematic
surveys undertaken across East and Central Africa in 2002 and 2003 (Oduor
et al., 2004; see also Chapters 3, 4 and 5). It is also of significance that where
C. arabica and C. canephora are cultivated in close proximity and even under
conditions of severe disease pressure, CWD will only affect one of the two
species (Kilambo et al., 2006).
Of significance, the few strains associated with the earlier CWD outbreaks
in Central and West Africa were also assessed for pathogenicity but do
not show the same degree of host specificity. One strain in particular, isolated in the 1960s from excelsa coffee affected by CWD in the CAR, is crosspathogenic to excelsa, C. arabica and also C. canephora.
The new knowledge acquired on genetic diversity within G. xylarioides,
on the pathogenic nature of the fungus and on the sexual cycle has led to
new species and formae speciales concepts being proposed, as described
below.
112
Rutherford et al.
7.9. Taxonomic Status of G. xylarioides

G. xylarioides was previously recognized as a member of the Fusarium section
Lateritium. However, recent phylogenetic studies (including those described
above) on strains that match both the original morphological description of
the species (Steyaert, 1948) and of the female form as described by Booth
(1971) have shown that these belong to the African clade of the GFC, a clade
replacing Fusarium section Liseola (ODonnell et al., 1998; ODonnell et al.,
2000; Geiser et al., 2005; Lepoint et al., 2005). Strains exhibiting the morphology of the male strain, as described by Booth, have been found to be genetically more closely related to F. lateritium.
As described here and in other chapters of this book, recent research on
the CWD pathogen has combined conventional approaches with some of the
most up-to-date analytical methods currently available and has provided invaluable new information on diversity within the fungus and how it relates
to behaviour. It is clear that a number of geographically isolated forms exist
that are genetically diverse, sexually incompatible, pathogenic to particular
coffee species and show at least some cultural and morphological variability.
Each could represent a distinct phylogenetic (evolutionary) lineage in what
was originally considered to be a single species, suggesting the existence of
a species complex (GxC) composed of several cryptic species. Based on the
findings, the taxonomic status of the fungus has been reviewed and to date,
the following taxa proposed (Lepoint et al., 2005; Lepoint, 2006):
G. abyssiniae (anamorph, F. abyssiniae) sp. nov. responsible for current

CWD outbreaks on C. arabica in Ethiopia;
G. congoensis (anamorph, F. congoensis) sp. nov. responsible for CWD
outbreaks on C. canephora in the DRC;
F. guineensis sp. nov. described using C. canephora-associated strains
obtained during the earlier outbreaks in Guinea, Cte dIvoire and possibly CAR;
G. xylarioides (F. xylarioides) neotypified using Central African C. excelsa
strain BBA 62457 (=DSMZ 62457) and strain ATCC 15664 collected during the first CWD epidemic.
The first three species are previously undescribed.

Similarly, and principally based on hostpathogen interactions and molecular characterization, two formae speciales have been proposed for strains
affecting C. arabica, C. canephora and C. excelsa (Girma, 2004; Girma et al., 2006)
(Chapter 8). These correspond to the first two taxa described above and are,
respectively:
G. xylarioides f. sp. abyssiniae (anamorph: F. xylarioides f. sp. abyssiniae)

G. xylarioides f. sp. canephorae (anamorph: F. xylarioides f. sp. canephorae).
Caution must be exercised when interpreting the findings of the research

studies completed to date. Given the number and geographically diverse
origins of strains examined and the breadth of characterization methods
employed, there is good evidence for two biological populations or species
113
being responsible for current CWD problems in East and Central Africa.
Although observed pathogenic specificity to coffee suggests that these may
represent two formae speciales, a designation that implies a strict relationship
with a particular host, evaluation of strains against other potential hosts has
been limited. History has shown that a number of fungi assigned to formae
speciales have subsequently been found to have a broader host range (albeit
encompassing primary and secondary hosts) than originally described. Fusarium wilt of cotton is but one example (Hillocks, 1992).
The limited number of strains available for study from the earlier epidemics also limits our ability to draw firm conclusions as to their role in CWD
development and their relationship with current strains. Have, for example,
these earlier strains been eradicated? Are they still present and perhaps still
responsible for CWD outbreaks, albeit localized and perhaps unrecognized or
not reported (on either cultivated or wild coffee)? Do other variants capable of
attacking C. arabica or C. canephora already exist and simply await dispersal to,
or the arrival of, a suitable, susceptible host? Further assessment of the current
situation in CAR, Guinea, Ivory Coast and neighbouring countries in Central
and West Africa is required to address these questions, coupled with access to a
greater number of G. xylarioides originating from these regions.
7.10. Other Fusaria on Coffee

A number of other Fusarium species occur on coffee, including Fusarium
solani (Mart.) Sacc. (Nectria haematococca), Fusarium stilboides, F. lateritium,
F. oxysporum (a true vascular wilt) and Fusarium decemcellulare (Waller and
Holderness, 1997; Rutherford, 2003; Tshilenge-Djim et al., 2004; Rutherford
and Phiri, 2006; Serani et al., 2007; Waller et al., 2007). All are recognized as
pathogens of coffee in their own right, and all have been recovered from
trees affected by CWD, in some cases from the same excised wood pieces
from which G. xylarioides was isolated (Rutherford and Flood, 2005; Lepoint,
2006). The nature of the association of these fungi with G. xylarioides is not
clear. They may, under some circumstances, coexist with G. xylarioides as
components of a CWD complex (Rutherford, 2006; Lepoint, 2006) and may
therefore have some influence not only on development of what are considered to be CWD symptoms but also on crop yield and quality. Importantly,
the external and internal symptoms that these fusaria induce can be very
similar to those of CWD and may lead to an inaccurate diagnosis (Waller
and Holderness, 1997; Tshilenge-Djim et al., 2004; Serani et al., 2007). F. solani, for example, also causes wilt symptoms accompanied by a dry root rot.
Coffee bark disease, caused by F. stilboides Wollwenw. (Gibberella stilboides),
is denoted by a progressive decline in host vigour. Discolouration of the
wood beneath the bark may also develop in both diseases, F. solani causing
a reddish-purple or brown discolouration in the roots and collar region
for example, but this is usually unlike the blueblack staining typical of
G. xylarioides. Tshilenge-Djim et al. (2004) noted that a number of fusaria originally isolated from wilted coffee could induce wilt-like symptoms when
114
Rutherford et al.
re-inoculated on to coffee plantlets. However, death of plantlets was only

observed following inoculation with G. xylarioides.
Invasion of coffee by several fusaria may obscure symptoms, external
and internal, that would otherwise be readily visible and allow a clear diagnosis. This highlights the need for sufficient awareness of the possible occurrence of these various fusaria on coffee and of symptoms characteristic for
each. It also stresses the need to confirm the presence, identity and pathogenic nature of an apparent causal organism using appropriate methods, including those outlined here and in other chapters of this book. G. xylarioides,
F. solani, F. stilboides and the other fusaria highlighted here are distinguishable
by the cultural (e.g. colony pigmentation) and morphological characteristics
they exhibit when grown on an appropriate agar medium such as potato
sucrose agar (Waller and Holderness, 1997; Rutherford, 2003). Generally,
G. xylarioides is readily distinguishable from these other species by its curled
or crescent-shaped microconidia. However, and although not normally associated with coffee, care should be taken not to confuse microconidia of
G. xylarioides with G. indica (anamorph Fusarium udum), a known pathogen
of pigeon pea (Cajanus cajan) (Hillocks and Songa, 1993). Although examination of morphological characters remains a fundamental requirement for
species identification, molecular tools and other approaches to identification
and characterization of G. xylarioides and other fusaria associated with CWD
are now available. Some have the added benefit of discriminating within, as
well as between, the various fungal species.
7.11. Conclusion
Recent studies on CWD, largely completed as part of the RCWP, have greatly
enhanced our understanding of the occurrence, form and behaviour of the
CWD pathogen, G. xylarioides, including from a historical perspective. It has
provided important information on coffee production areas already affected
by the disease and those under threat. A number of genetically and biologically distinct forms, proposed as distinct species, have been shown to be responsible for current epidemics across East and Central Africa. These are
geographically delineated and exhibit pathogenic specificity to one or other
of the two coffee species of economic importance in the region, C. arabica and
C. canephora. Although access to forms of the pathogen associated with earlier epidemics in West and Central Africa has been limited, characterization
and comparison of these with those prevalent today have provided some
clarification as to their importance to current epidemics and how these may
have arisen. This new knowledge, coupled with a global search to obtain coffee germplasm with resistance to CWD and the outcome of research to clarify
the underlying genetic basis of resistance, provides a basis for redefining our
approach to managing CWD through both prevention and eradication. Swift
action is required. CWD has already devastated coffee production across
large parts of Africa, destroying the livelihoods of millions of families already
facing hardship, and continues to sweep across the continent. In parallel,
115
awareness within the coffee community of the disease and how it can be
tackled must be increased, particularly in areas not yet suffering but under
imminent threat. Research efforts must also continue to allow us to build
upon what we have already learned and to seek an effective and sustainable
means to eradicate CWD or, in the short term, maintain disease levels within
an acceptable economic threshold.
Acknowledgements
The authors would like to thank those who provided information required
for the preparation of this chapter, in particular colleagues and collaborators
from: CABI, UK and Kenya; UCL, Belgium; CIRAD, France; Coffee Research
Institute, Uganda; Ethiopian Agricultural Research Organisation; Tanzania
Coffee Research Institute and Universit de Kinshasa, DRC. Coordinated research on CWD in Africa and Europe was performed in the framework of
European Community-funded project International Cooperation-COWIDI
ICA4-CT-2001-10006, coordinated by CIRAD, and UK Department for International Development Crop Protection Programme-funded project ZA0505,
coordinated by CABI E-UK. The Common Fund for Commodities is also
acknowledged for providing financial support to the overall RCWP. The
authors would also like to acknowledge the considerable efforts made by
research scientists, extension services, farmers and other coffee stakeholders in Africa and elsewhere who completed the many challenging studies of
coffee wilt to help improve our understanding of the disease. Without their
contribution, the new knowledge presented here and in other chapters of
this book would not have become available. Thanks also go to Dr J.M. Waller
(CABI Emeritus Fellow) and Dr J. Flood (CABI E-UK) for kindly reviewing
and providing advice on this chapter.
References
Adugna, G., Hindorf, H., Steiner, U., Nirenberg, H.I., Dehne, H.W. and Schellander,
K. (2005) Genetic diversity in the coffee wilt pathogen (Gibberella xylarioides) populations: differentiation by host specialization and RAPD analysis. Zeitschrift fur
Pflanzenkrankheiten und Pflanzenschutz 112, 134145.
Alves-Santos, F.M., Ramos, B., Garca-Snchez, M.A., Eslava, A.P. and Daz-Mnguez,
J.M. (2002) A DNA-based procedure for in planta detection of Fusarium oxysporum f. sp. phaseoli. Phytopathology 92, 237244.
Beckman, C.H. (1987) The Nature of Wilt Diseases of Plants. The American Phytopathology Society, St. Paul, MN.
Bieysse, D. (2005) (ed.) Development of a Long-Term Strategy Based on Genetic Resistance and Agroecological Approaches Against Coffee Wilt Disease in Africa:
Fourth Annual Report. INCO-DEV Contract ICA4-CT-2001-10006. CIRAD-AMIS,
Montpellier, France.
Bieysse, D. (2006) (ed.) Development of a Long-Term Strategy Based on Genetic Resistance and Agroecological Approaches Against Coffee Wilt Disease in Africa: Fifth
116
Rutherford et al.
Annual Report. INCO-DEV Contract ICA4-CT-2001-10006. CIRAD-AMIS, Montpellier, France.
Bieysse, D. (2007) (ed.) Development of a Long-Term Strategy Based on Genetic Resistance and Agroecological Approaches Against Coffee Wilt Disease in Africa: Final
Report. INCO-DEV Contract ICA4-CT-2001-10006. CIRAD-AMIS, Montpellier,
France.
Booth, C. (1971) The Genus Fusarium. Commonwealth Mycological Institute, Kew,
UK.
Booth, C. and Waterston, W.M. (1964) CMI Descriptions of Plant Pathogenic Fungi and
Bacteria No. 24. CAB International, Wallingford, UK.
Doohan, F.M., Parry, D.W., Jenkinson, P. and Nicholson, P. (1998) The use of speciesspecific PCR-based assays to analyse Fusarium ear blight of wheat. Plant Pathology
47, 197205.
Bulletin Agricole du Congo Belge XLI, 361372.
Figueres, R. (1940) Sur une maladie trs grave du cafier en Oubangui. Rapport. Ministre des Colonies, Paris, France.
Flood, J. and Brayford, D. (1997) Current knowledge of pathogen variability, epidemiology and control of Fusarium xylarioides. In: 1st Regional Coffee Wilt Workshop,
Kampala, Uganda.
Geiser, D.M., del Mar Jimnez-Gasco, M., Kang, S., Makalowska, I., Veeraraghavan,
N., Ward, T.J., Zhang, N., Kuldau, G.A. and ODonnell, K. (2004) FUSARIUM-ID
v. 1.0: a DNA sequence database for identifying Fusarium. European Journal of
Plant Pathology 110, 473479.
Geiser, D.M., Lewis Ivey, M.L., Hakiza, G., Juba, J.H., and Miller, S.A. (2005) Gibberella xylarioides (anamorph: Fusarium xylarioides), a causative agent of coffee wilt
disease in Africa, is a previously unrecognized member of the G. fujikuroi species
Girma, A.S. (2004) Diversity in Pathogenicity and Genetics of Gibberella xylarioides
(Fusarium xylarioides) Populations and Resistance of Coffea spp. in Ethiopia. PhD
dissertation. University of Bonn, Bonn, Germany.
Girma, A.S., Flood, J., Hindorf, H., Bieysse, D., Simons, S. and Rutherford, M. (2006)
Tracheomycosis (Gibberella xylarioides)a menace to world coffee production:
evidenced by cross inoculation of historic and current strains of the pathogen. In:
Proceedings of the 21st International Scientific Conference on Coffee Science (ASIC).
Montpellier, France, pp. 12681276.
Girma, A., Mengistu, H. and Hindorf, H. (2001) Incidence of tracheomycosis, Gibberella xylarioides (Fusarium xylarioides), on arabica coffee in Ethiopia. Zeitschrift
fur Pflanzenkrankheiten und Pflanzenschutz 108, 136142.
Hillocks, R.J. (1992) Fusarium wilt. In: Hillocks R.J. (ed.) Cotton Diseases. CAB International, Wallingford, UK, pp. 127160.
Hillocks, R.J. and Songa, W. (1993) Root-knot and other nematodes associated with
pigeonpea plants infected with Fusarium udum in Kenya. Afro-Asian Journal of
Nematology 3, 143147.
Jacques-Felix, H. (1950) Premire action contre la trachomycose du cafier en Cte
dIvoire. Agronomia Tropical Nogent, 12 pp.
Jacques-Felix, H. (1954). La carbunculariose. Bulletin ScientifiqueCentre de Recherches Agronomiques (Francia) 5, 296344.
Kangire, A., Kytere, D., Hakiza, G., Warren, H., Erbaugh, M. and Kabole, C. (2002)
Determination of alternate hosts for F. xylarioides, the causal agent of Coffee Wilt
Disease (CWD) in Uganda. Internal Report. Coffee Research Institute, Uganda.
117
Kilambo, D.L., Kaiza, D.A. and Swai, F.B. (1997) Observation of wilting disease in some
coffee growing areas in Tanzania. In: Proceedings of the 1st Regional Workshop on
Coffee Wilt |Disease (Tracheomycosis). International Conference Centre, Kampala,
Uganda, pp. 6165.
Kilambo, D.L., Nghoma, N.M., Mohamed, R.A., Teri, J.M., Poole, J., Flori, A. and
Pinard, F. (2004) Coffee disease survey in Tanzania. In: Proceedings of the 20th
International Conference on Coffee Science (ASIC). Bangalore, India, pp. 1263
1266.
Kilambo, D.L., Nghoma, N.M., Mtenga, D.J., Teri, J.M., Nzallawahe, T., Rutherford, M.
and Masumbuko, L. (2006) Progress towards searching for durable resistance to Fusarium wilt (Fusarium xylarioides) in Coffea canephora germplasm in Tanzania. In:
Proceedings of the 21st International Scientific Conference on Coffee Science (ASIC).
Kirk, P.M., Cannon, P.F., Minter, D.W. and Stalpers, J.A. (2008). Dictionary of the Fungi, 10th edn. CABI Publishing, Wallingford, UK.
Knoll, S., Mulfinger, S., Niessen, L. and Vogel, R.F. (2002) Rapid preparation of Fusarium DNA from cereals for diagnostic PCR using sonification and an extraction kit.
Kranz, J. (1962) Coffee disease in Guinea. FAO Plant Protection Bulletin 10, 107109.
Kranz, J. and Mogk, M. (1973) Gibberella xylarioides Heim et Saccas on arabica coffee
in Ethiopia. Phytopathologische Zeitschrift 78, 365366.
Lepoint, P.C.E. (2006) Speciation within the African coffee wilt Pathogen. PhD thesis.
Universit catholique de Louvain, Louvain, Belgium.
Lepoint, P.C.E., Munaut, F.T.J. and Maraite, H.M.M. (2005) Gibberella xylarioides sensu
lato from Coffea canephora: a new mating population in the Gibberella fujikuroi
species complex. Applied and Environmental Microbiology 71, 84668471.
Meiffren, M. (1957) La trachomycose. Les maladies du cafier en Cte dIvoire. Haut
Commissariat de lA.O.F., Centre de Recherches Agronomiques de Bingerville,
Cte dIvoire. Bulletin de la recherche agronomique du Bingerville 13, 954.
Meiffren, M. (1961) Contribution aux recherches du cafier en Cte dIvoire. Caf Cacao Th 5, 2837.
Musoli, P.C. (2007). Recherche de sources de rsistance la trachomycose du cafier
canephora Pierre, due Fusarium xylarioides en Ouganda. Thesis. Universit de
Montpellier II, Montpellier, France.
Musoli, P.C., Aluka, P., Cubry, P., Dufour, M., de Bellis, F., Nakendo, S., Nabaggala,
A., Ogwang, J., Kyetere, D., Leroy, T., Bieysse D. and Charrier, A. (2006) Fighting
against coffee wilt disease: Uganda wild C. canephora genetic diversity and its
usefulness. In: Proceedings of the 21st International Scientific Conference on Coffee
Science (ASIC). Montpellier, France, pp. 12681276.
Musoli, P.C. and Hakiza, G. (2007) Individual final report, National Agricultural Research Organisation (NARO). In: Bieysse, D. (ed.) Development of a Long-Term
Strategy Based on Genetic Resistance and Agroecological Approaches against
Coffee Wilt Disease in Africa: Final Report. INCO-DEV Contract ICA4-CT-200110006. CIRAD-AMIS, Montpellier, France, pp. 5670.
Musoli, P.C., Pinard, F., Charrier, A., Kangire, A., ten Hoopen, G.M., Kabole, C., Ogwang,
J., Bieysse, D. and Cilas, C. (2008) Spatial and temporal analysis of coffee wilt disease caused by Fusarium xylarioides in Coffea canephora. European Journal of
Plant Pathology (in press).
Nirenberg, H.I. (1976) Untersuchungen ber die morphologische und biologische Differenzierung in der Fusarium Sektion Liseola. Mitteilungen aus der Biologischen
Bundesanstalt fur Land- und Forstwirtschaft (Berlin-Dahlem) 169, 1117.
118
Rutherford et al.
ODonnell, K., Cigelnik, E. and Nirenberg, H.I. (1998) Molecular systematics and phylogeography of the Gibberella fujikuroi species complex. Mycologia 90, 465493.
ODonnell, K., Nirenberg, H.I., Aoki, T. and Cigelnik, E. (2000) A multigene phylogeny
of the Gibberella fujikori species complex: detection of additional phylogenetically
distinct species. Mycoscience 41, 6178.
Oduor, G., Phiri, N., Hakiza, G.J., Abebe, M., Asiimwe, T., Kilambo, D.L., Kalonji-Mbuyi,
A., Pinard, F., Simons, S., Nyasse, S. and Kebe, I. (2004) Surveys to establish the
spread of coffee wilt disease, Fusarium (Gibberella) xylarioides, in Africa. In: Proceedings of the 20th International Scientific Conference on Coffee Science (ASIC).
Bangalore, India, pp. 12521255.
Onesirosan, P.T. and Fatunla, T. (1976) Fungal fruit rots of tomatoes in Southern Nigeria. Journal of Horticultural Science 51, 473479.
Pegg, G.F. (1981) Biochemistry and physiology of pathogenesis. In: Mace, M.E., Bell,
A.A. and Beckman, C.H. (eds.) Fungal Wilt Diseases of Plants. Academic Press,
New York and London, pp. 193252.
Pieters, R. and Van der Graaff, N.A. (1980) Resistance to Gibberella xylarioides in Coffea arabica: evaluation of screening methods and evidence for the horizontal nature of the resistance. Netherlands Journal of Plant Pathology 86, 3743.
Pizzinatto, M.A. and Menten, J.O.M. (1991) Pathogenicity of eight Fusarium species
isolated from seeds to cotton seedlings. Summa Phytopathologica 17, 124134.
Pochet, P. (1988) La trachomycose du cafier robusta. Publications Agricoles 19, 28 pp.
Rutherford, M. (2003) Isolation and Identification of Fusaria from Coffee. A Training
Manual. CABI International, Egham, UK.
Rutherford, M. (2006) Current knowledge of coffee wilt disease, a major constraint to
coffee production in Africa. Phytopathology 96, 663666.
Rutherford, M.A. and Flood, J. (2005) Epidemiology and variability of Gibberella xylarioides, the coffee wilt pathogen. UK Department for International Development
(DFID) Crop Protection Programme Project R8188: Final Technical Report. CABI
International, Egham, UK.
http://www.research4development.info/SearchResearchDatabase.asp?ProjectID
=3636
Rutherford, M.A. and Phiri, N. (2006) Pests and Diseases of Coffee in Eastern Africa: A
Technical and Advisory Manual. CAB International, Egham, UK.
http://www.research4development.info/PDF/Outputs/CropProtection/U3071
Coffee Manual.pdf
Ryan, M.J. and Smith, D. (2007) Cryopreservation and freeze-drying of fungi employing centrifugal and shelf freeze-drying. In: Day, J.G. and Stacey G.N (eds.)
Cryopreservation and Freeze Drying Protocols, 2nd edn. Hna Press, New Jersey,
pp. 127140.
Saccas, A.M. (1951) La trachomycose (carbunculariose) des Coffea excelsa, neo-arnoldiana et robusta en Oubangui-Chari. Agronomie Tropicale 6, 453506.
Serani, S. (2000) An investigation of Fusarium spp. associated with coffee and other
plants as potential pathogens of robusta coffee. MSc dissertation. Makerere University, Kampala, Uganda.
Serani, S., Taligoola, H.K. and Hakiza, G.J. (2007) An investigation into Fusarium spp.
associated with coffee and banana plants as potential pathogens of robusta coffee.
African Journal of Ecology, 45, 9195.
Swofford, D.L. (2001) PAUP* Phylogenetic Analysis Using Parsimony (*and other
methods), Version 4. Sinauer Associates, Sunderland, Massachusetts.
Stewart, R.B. (1957) Some diseases occurring in Kaffa Province, Ethiopia. Imperial Ethiopian College of Agriculture and Mechanical Arts, Alemaya, Ethiopia, pp. 1516.
119
Steyaert, R.L. (1948) Contribution ltude des parasites des vgtaux du Congo Belge.
Bulletin de la Societe Royale de Botanique de Belgique 80, 1158.
Tshilenge-Djim, P., Munaut, F., Kalonji-Mbuyi, A. and Maraite, H. (2004) Caractrisation des Fusarium spp. associes au dprissement du cafier robusta en Rpublique Dmocratique du Congo. Parasitica 60, 1934.
Van Den Abeele, M. and Vandenput, R. (1956) Les principales cultures du Congo Belge,
3rd edn. Royaume de Belgique, Ministre des colonies, Publication de la Direction
de lagriculture, des forts et de llevage, Bruxelles, Belgium. 932 pp.
Van der Graaff, N.A. and Pieters, R. (1978) Resistance levels in Coffea arabica to Gibberella xylarioides and distribution patterns of the disease. Netherlands Journal of
von Blittersdorff, R. and Kranz, J. (1976) Vergleichenden.Untersuchundgen an Fusarium xylarioides (Gibberella xylarioides Heim & Saccas), dem Erreger der Tracheomykose des Kaffes. Zeitschrift fur Pflanzenkrankheiten und Pflanzenschutz 83,
529544.
Waller, J.M., Bigger, M. and Hillocks, R.J. (2007) Coffee Pests, Diseases and their Management. CAB International, Wallingford, UK.
Waller, J.M. and Holderness, M. (1997) Fusarium diseases of coffee. In: Proceedings of
the First Regional Workshop on the Coffee Wilt Disease (Tracheomycosis). Kampala, Uganda, pp. 2736.
HostPathogen Interactions in
CoffeaGibberella xylarioides Pathosystem
A. Girma,1 D. Bieysse2 and P. Musoli3

2
34398 Montpellier, Cedex 5, France
3
8.1. Introduction
Gibberella xylarioides Heim & Saccas (anamorph: Fusarium xylarioides Steyaert)
causes a vascular wilt disease of coffee referred to as tracheomycosis. The
disease caused large-scale damage to plantations of various Coffea spp. in Africa in the early 1950s (Wellman, 1961; Kranz, 1962; Muller, 1997). Coffee wilt
disease (CWD) has been observed only on Coffea canephora (robusta coffee) in
the Democratic Republic of Congo (DR Congo), Uganda and Tanzania and
on Coffea arabica (arabica coffee) in Ethiopia (Girma et al., 2001; Oduor et al.,
2003), suggesting some form of host specialization in the pathogen populations. Tracheomycosis re-emerged as a serious disease of coffee in Africa, and
certain factors (shift in the host and/or in the pathogen populations, acting
either independently or in concert) have been postulated about its reappearance and becoming a major constraint to coffee production in this continent
(Flood and Brayford, 1997; Girma and Hindorf, 2001; Girma, 2004).
Management of soil-borne disease is a continual challenge to growers,
and one of the major reasons for the limited success is our limited knowledge of the genetic structure of the pathogen populations (Watson, 1970;
McDonald, 1997). According to Leung et al. (1993), population structure refers to the amount of genetic variation among individuals in a population, the
ways in which this variation is partitioned in time and space and the phylogenetic relationships among individuals within and between subpopulations.
120
CAB International 2009. Coffee Wilt
HostPathogen Interactions in CoffeaGibberella xylarioides Pathosystem
121
The most widely used method for characterizing pathogen populations is the
determination of the virulence spectrum on a set of differential varieties carrying different resistance factors, and this measure of genetic variability provides
information on the pathotype structure of the populations (Leung et al., 1993).
Understanding pathogen population structure can contribute to improved
disease management by allowing resistance genes and plant genotypes to be
identified and characterized relative to the spectrum of the pathogen population and by providing a knowledge base important to the design of strategies
for variety development (Leung et al., 1993). Resistance to wilt diseases depends in part on genetic potential for virulence within the pathogen populations and the inoculum concentration. The occurrence of distinctly different
levels of virulence in the wilt fungus indicates that certain mechanisms of
resistance are not effective against the highly virulent strains, even though
they are effective against most strains of the fungal species (Beckman, 1987).
Resistance of a plant (or tissue) changes sequentially during growth and development; thus, certain growth stages are more favourable than others for
comparison of resistant and susceptible cultivars (Beckman, 1987). According to Beckman and Talboys (1981), two sets of events have to take place
before a vascular wilt disease can develop: the pathogen must gain access to
the xylem of the host, and it must continue to colonize the xylem more or less
extensively. However, the fact that a particular mechanism contributes to
disease resistance does not necessarily mean that it is implicated in the determination of specificity in hostparasite interaction (Beckman, 1987; Beckman
and Talboys, 1981).
Interaction between coffee and G. xylarioides (F. xylarioides) has been thoroughly studied through several artificial seedling inoculation tests. These
studies were exclusively conducted in well-designed experiments in the
greenhouse or growth room where disease parameters such as infected and
dead seedlings and number of days to first symptom appearance after inoculation (incubation periods) were documented and then subjected to statistical analysis and biometrical interpretations of the results. The hostpathogen
combination genetic analyses were further backed by molecular studies,
fungal morphology and in vitro mating test (Chapter 7). Thus, this chapter
elucidates the current knowledge on hostpathogen interactions in the Coffeavs.-G. xylarioides pathosystem gained through repeated cross-inoculation
studies (pathogenicity tests), supported by some preliminary histopathological observations. This chapter also includes appraisal on seedling inoculation protocols, which are the basis for testing and verifying host resistance
that can be used as a standard procedure for mass screening and breeding
programs.
8.2. Appraisal of Seedling Inoculation Protocols for Screening for

Resistance and Studies of the HostPathogen Interaction
The study of hostpathogen associations involves a three-dimensional interaction between host varieties, pathogen strains and environmental variables that
122
Girma et al.
can affect disease expression emanating from compatible/incompatible combinations. To be able to limit the effect of these factors on hostpathogen interactions, artificial inoculation tests have been developed. Artificial inoculation
allows more genotypes/cultivars to be tested. Besides, reliance on screening
for resistance in the field (based on natural inoculum) is unsatisfactory due to
irregular occurrence of the pathogen in time and space, lack of understanding
of which strains or races are present and unfavourable conditions that may interfere with proper appraisal of host resistance. Yet, breeding programs with a
perennial crop are long term, and at least 7 to 10 years is needed to select material on the basis of natural infection in field trials; thus, alternative, more rapid
tests have to be investigated to help speed up screening of germplasm (Flood,
2006). Therefore, developing standard artificial screening protocols that discriminate between resistant and susceptible genotypes is an imperative and is
very relevant for such soil-borne vascular diseases as CWD.
In this particular hostpathogen interaction, many different inoculations
techniques are and have been employed in various laboratories in different
countries over many years. Stem-nicking method of young coffee seedlings
with inoculum suspension (22.5 106 concentration) of G. xylarioides isolate
at cotyledon stage (2 to 2.5 months old) using a scalpel has been adopted as
the preferred standard practice on C. arabica in Ethiopia (Pieters and Van der
Graaff, 1980; Girma and Mengistu, 2000; Girma et al., 2005, 2007), whereas syringe injection of inoculum (1 106 concentration) into the stem of growing
seedlings (at 9 to 10 months old) has been routinely used at Centre de Coopration Internationale en Recherche Agronomique pour le Dveloppement
(CIRAD; D. Bieysse, personal communication, 2003) and the root dip inoculation procedure is employed in screening C. canephora seedlings (5 to 6 months
old) in Uganda, DR Congo and Tanzania (Hakiza et al., 2004; Kilambo et al.,
2007; Chapter 10).
The stem injection method is labourious and time consuming, necessitating the piercing of the stem with a syringe and deposition of the required inoculum concentration. This may have the disadvantage of disease escape and
is inconvenient to screen a large number of coffee genotypes. The stem-nicking
method, although allowing screening of many genotypes in a short period,
does have the disadvantage of sometimes causing seedling damage. Further
problems associated with artificial screening techniques include differences in
resistance response (symptom expression) among coffee types due to growth
stage of seedlings, which in turn relate to appearance and development of
the primary and secondary xylem tissues of the plant. In addition, inoculum
concentration and pre- and post-inoculation conditions (mainly temperature,
relative humidity and optimum moisture of the media) are also known to influence infection and disease development (Girma and Mengistu, 2000). Thus,
standardizing the inoculation protocols (methodologies) so that they can be
uniformly applied with reproducible results across different laboratories/
greenhouses, identifying proper growth stages of the host that show differential reactions, selection of aggressive strain/isolate and conditions that favour
infection and wilt disease development is of paramount importance in designing an effective screening and breeding programme for CWD management.
123
To optimize the protocol, a comprehensive investigation was carried out

on three C. arabica cultivars (designated resistant, intermediate and susceptible) from known field observations, which were inoculated with a standard
arabica isolate (Gx2) (IMI 71975) at four different growth stages of seedlings
(between soldier and two pairs of true leaves) employing five inoculation methods, namely, stem nicking, stem injection, root cut and dip, root
wound and transplant into artificially and naturally infested soil. The fungal
inoculum was uniformly adjusted to 2 106 spore per millilitre for the treatment combinations. There were highly significant differences between coffee
cultivars, inoculation methods and growth stages of the seedlings among
the three variables (Tables 8.1 and 8.2). As indicated in Table 8.1, cotyledon
and one pair of true leaf stages showed significantly higher seedling death,
whereas seedlings at soldier stage failed to open cotyledon leaves, which
may be due to physiological shock and did not reveal typical wilting although the seedlings eventually died.
When means of inoculation methods were compared, significantly
higher percentage deaths were recorded in stem nicking followed by the
root dip and stem injection methods and revealed differences among coffee
cultivars in their response to the disease (Table 8.2). Seedlings inoculated by
the root dip method developed symptoms very slowly and sometimes failed
to recover after transplanting and did not show typical CWD symptoms. The
stem injection method was technically inefficient in screening large number of accessions, as it is difficult to pierce the seedling stems and place the
Table 8.1. Percentage seedling deatha in three C. arabica cultivars inoculated with
G. xylarioides isolates at different growth stages in the greenhouse++.
Growth stagesc
Coffee
cultivarb
Soldier
Cotyledon
One pair of
true leaves
Two pairs of
true leaves
Mean
200/71
21/79
20/85
Mean
24.5 fg
32.5 ef
48.1 c
35.1 X
46.3 cd
45.2 cd
68.3 a
53.3 Y
48.6 c
48.1 c
58.8 b
51.8 Y
21.6 g
35.8 e
38.9 de
32.1 X
35.3 A
40.4 A
53.5 B
43.1
Percentage death is calculated from cumulative number of dead plants over total number of
seedlings (20 per treatment) 6 months after inoculation, and the actual wilt values were arcsine-square root transformed to normalize the data.
b
Coffee cultivars 21/79, 201/71 and 20/85 had resistance, moderate resistance and susceptible
reactions under field conditions, respectively.
c
Growth stages were soldier or hypocotyls, when the seeds emerge from the soil before opening of the cotyledon; cotyledon, fully opened butterfly; one pair of true leaves, when 1st pair
true leaves are fully opened; two pairs of true leaves, when two pairs of true leaves were fully
opened.
++
Means followed with the same letter(s) are not significantly different from each other at
P = 0.05. Least significant difference values for mean comparison for cultivars, stages and
interactions are 6.93, 4.61 and 7.99, respectively.
124
Girma et al.
Table 8.2. Percentage seedling deatha in three C. arabica cultivars inoculated with
G. xylarioides isolates with different inoculation methods greenhousec.
Inoculation methods1
Coffee
cultivarsb
200/71
21/79
20/85
Mean
Stem nicking
Stem injection
Root dip
Artificially
infested soil
Naturally
infested soil
Mean
53.1 df
70.5 ab
61.1 cd
61.6 w
45.0 f
55.5 ce
62.1 bc
54.2 x
46.9 ef
57.8 cd
71.5 a
58.7 wx
24.0 g
17.0 g
47.2 ef
29.4 y
7.3 h
1.2 h
25.7 g
11.4 z
35.3 A
40.4 A
53.5 B
43.1
As Table 8.1.
Coffee cultivars (as Table 8.1).
c
Means followed with the same letter(s) are not significantly different from each other at
P = 0.05. Least significant difference values for mean comparison for cultivars, methods and
interactions are 6.93, 5.61 and 8.93, respectively.
1
Inoculation methods were as described in the text.
b
required amount of inoculum at the later growth stages. Overall, the mean
percentage seedling death in cultivar 200/71 (resistant) and cultivar 21/79
(moderately resistant) were significantly different from the susceptible cultivar (20/85) with seedling deaths of 35.3%, 40.4% and 53.5% (Tables 8.1
and 8.2).
The stem-nicking method at the cotyledon stage using spore concentration of 2 106 was adopted as a standard protocol for screening coffee
germplasm and studying the hostpathogen interaction. Inoculated seedlings were maintained at high humidity (>95%) and temperature of about
23oC for 1 week as optimum conditions for infection. After such a postinoculation period, the seedlings were transferred into a glasshouse for
further disease development. The seedlings were not uprooted and transplanted, as this practice disturbs the plant system and may predispose them
to the disease.
8.3. C. arabicaG. xylarioides Isolates Interactions (Arabica Cultivars

vs. Isolates)
These studies involved inoculating seedlings of nine C. arabica cultivars that
had been observed to possess various levels of resistance to the disease under field conditions with four G. xylarioides isolates obtained from infected
C. arabica trees (Gx12 [IMI 375906], Gx26 [IMI 375907], Gx31 [IMI 375908]
and Gx43 [IMI 375909]). The result showed highly significant (P < 0.01) differences among coffee cultivars and G. xylarioides isolates and a significant
(P < 0.05) cultivarisolate interaction both in terms of percentage seedling
death and incubation period (from inoculation to first symptoms) (Girma
and Mengistu, 2000; Girma and Hindorf, 2001). According to Van der Plank
125
(1984), highly significant differences among the main effects indicate the
existence of horizontal resistance in the host and variation in aggressiveness
in the fungus population, but a significant interaction between the cultivars
and the isolates (i.e. a differential effect) in both disease parameters implies
vertical resistance in the host and virulence in the pathogen (Girma and
Mengistu, 2000; Girma and Hindorf, 2001).
When comparing all cultivars, 61/85, 24/85 and F-17 showed significantly (P < 0.05) higher disease levels with 62.6%, 60.5% and 51.4% seedling death, respectively (Table 8.3), with a shorter incubation period of about
3060 days (Table 8.4). In contrast, a significantly (P < 0.05) lower percentage of seedling death (28.8%, 24.3%, and 12.0%) was observed on cultivars
35/85, 74165 and 7440, respectively (Table 8.3), with longer incubation periods ranging between 84 and 112 days (Table 8.4). Isolates Gx26, Gx43 and
Gx31 caused more seedling death (58.2%, 53.4% and 52.2%, respectively) than
isolate Gx12, which induced no symptoms throughout the trial (Table 8.3).
The Teppi isolate (Gx26) induced wilting symptoms in a significantly shorter
incubation period (82 days) as compared to Jimma isolate (Gx31), which induced symptoms in around 100 days (Table 8.4). Comparing the combined
effect of cultivar vs. isolate interactions (Table 8.3), Gx26 (Teppi isolate) induced a higher rate of death on cultivars SN-5 (70.8%), 74304 (64.6%) and
Table 8.3. Percentage seedling deatha in C. arabica cultivars inoculated with G. xylarioides
isolates collected from representative fields in south-west Ethiopia.
G. xylarioides isolatesc
Coffee
cultivarsb
Gx12
Gx26
Gx31
Gx43
Mean
74165 (R)
7440 (MR)
74304 (S)
F-17 (R)
F-61 (MR)
SN-5 (S)
35/85 (R)
61/85 (MR)
24/85 (S)
Mean
0.0 j
0.0 jd
0.0 j
0.0 j
0.0 j
0.0 j
0.0 j
0.0 j
0.0 j
0.0 N
40.5 ei
17.1 hj
64.6 af
77.8 ac
54.8 ag
70.8 ae
43.8 dh
80.4 ab
73.8 ad
58.2 M
33.9 fi
11.6 ij
48.8 bh
52.6 ag
57.1 af
62.4 af
35.3 fi
85.1 a
83.0 a
52.2 M
22.6 gj
19.3 hj
38.0 fi
75.1 ad
70.8 ae
46.8 ch
36.0 fi
85.1 a
85.2 a
53.4 M
24.3 E
12.0 F
37.8 CD
51.4 AB
45.7 BC
45.0 BC
28.8 DE
62.6 A
60.5 A
Percentage seedling deaths were transformed to angular values to normalize the data before
analysis. Means (column, row) followed with the same letter(s) are not significantly (P < 0.05)
different from each other according to Duncans multiple range test. Least significant difference
values for the cultivars, the isolates and the interactions comparisons were 10.8, 9.2, and 27.6,
respectively.
b
Cultivars 74165, F-17 and 35/85 were resistant (R); 7440, F-61 and 61/85 were moderately
resistant (MR); 74304, SN-5 and 24/85 were susceptible (S) to CWD under field conditions.
c
Gx12, Gx26, Gx31 and Gx43 were G. xylarioides isolates obtained from Bebeka, Teppi, Jimma
and Gera, respectively.
d
0.0 indicates no external symptoms were observed during the trial.
126
Girma et al.
Table 8.4. Incubation periods (in days) for C. arabica seedlings inoculated with
G. xylarioides isolates collected from representative fields in south-west Ethiopia.
G. xylarioides isolatesb
Coffee
cultivarsa
Gx12
74165 (R)
7440 (MR)
74304 (S)
F-17 (R)
F-61 (MR)
SN-5 (S)
35/85 (R)
61/85 (MR)
24/85 (S)
Mean
0.0 h
0.0 h
0.0 h
0.0 h
0.0 h
0.0 h
0.0.h
0.0 h
0.0 h
0.0 R
Gx26
Gx31
Gx43
Mean
126 ac
133 ac
35 gh
84 cg
56 eh
98 bf
105 bf
56 eh
49 fh
82.4 Q
154 ab
168 a
112 ae
98 bf
77 cg
77 cg
133 ac
35 gh
49 fh
100.3 P
154 ab
147 ab
119 ad
63 dg
84 cg
105 bf
98 bf
28 gh
28 gh
91.8 PQ
108.5 A
112.0 A
66.5 CD
61.2 CD
54.2 D
70.0 C
84.0 B
29.7 E
31.5 E
Gx12, Gx26, Gx31 and Gx43 were G. xylarioides isolates obtained from Bebeka, Teppi, Jimma
and Gera, respectively.
b
Cultivars 74165, F-17 and 35/85 were resistant (R); 7440, F-61 and 61/85 were moderately
resistant (MR); 74304, SN-5 and 24/85 were susceptible (S) to CWD under field conditions.
c
0.0 indicates no external symptoms were observed over the test period. Means followed with
the same letter(s) are not significantly (P < 0.05) different from each other according to Duncans multiple range test. Least significant difference values for the cultivars, the isolates and
the interactions comparisons were 13.7, 17.5 and 55.6, respectively.
74165 (40.5%) than Gx43 (Gera isolate), suggesting that the Teppi isolate was
more aggressive than the Gera isolate (Gx43) on these cultivars, although
Gx26 was less aggressive than the Gera isolate (Gx43) on cultivar F-61. The
Jimma isolate (Gx31) was moderately aggressive on cultivars 74165, 74304,
F-17 and SN-5, whereas the Bebeka isolate (Gx12) was least aggressive in
all the cultivars (Table 8.3) (Girma and Mengistu, 2000; Girma and Hindorf,
2001).
The results concur with the conclusions of Zadoks and Van Leur (1983)
but contradicted the findings of Pieters and Van der Graaff (1980). These latter
authors determined percentage germination of conidia from four G. xylarioides
isolates on the wood of branch internodes of six C. arabica varieties and concluded the presence of horizontal resistance (non-significant interaction).
However, Zadoks and Van Leur (1983) found significant cultivarisolate interaction indicating vertical resistance after re-analysing the same data reported
by Pieters and Van der Graaff (1980). Thus, our results corroborated the existence of variation in both levels of resistance of coffee genotypes and in the
aggressiveness of pathogen strains. A significant cultivarisolate interaction
(differential effect) further evidenced the existence of some vertical resistance
in the C. arabica and G. xylarioides pathosystem (Girma and Mengistu, 2000;
Girma et al., 2001). Further investigations were then undertaken to compare
fungal isolates from C. arabica and C. canephora to examine any host specificity in the hostpathogen interactions.
127
8.4. Interactions Among C. arabica and C. canephora vs.

G. xylarioides Isolates From Both Host Species
In this study, ten isolates from C. arabica trees in Ethiopia were compared with
an isolate from C. canephora in Uganda. The C. arabica isolates represented ten
major arabica coffee-growing districts in Ethiopia, at varying altitudes (1000
to 2000 m above sea levels) and production systems, vis. semiforest, garden
and plantation coffee. The isolates were inoculated on coffee seedlings in two
sets of experiments following the standard inoculation protocols. The data
analyses showed highly significant (P < 0.001) differences among cultivars,
isolates and cultivarisolate interactions both in percentage seedling wilt
and incubation period (even excluding C. canephora line and isolate) (Girma,
2004; Girma et al., 2005).
C. canephora appeared to be very susceptible and was severely attacked
(84.1%) by isolate GxU12 in a short incubation period of 44 days (Table 8.5).
However, no deaths were observed when C. canephora seedlings were inoculated with any arabica isolates (Gx1Gx9 and Gx11) even up to 12 months
Table 8.5. Percentage seedling deatha among C. arabica cultivars and one C. canephora
line inoculated with 11 G. xylarioides isolates collected from various geographical origins.
C. arabica cultivarsc
Isolatesb
Gx1
Gx2
Gx3
Gx4
Gx5
Gx6
Gx7
Gx8
Gx9
Gx11
GxU12
Mean
a
Catimor-J19
7440
F-59
Caturra Rojo
24/85
C. canephora
Mean
30.6 pr
19.9 ru
17.4 ru
27.9 qs
8.8 uv
8.3 uv
24.3 rt
14.4 tu
15.0 su
77.2 af
0.0 v
22.2 T
66.2 fj
52.5 jn
30.8 pr
65.8 fj
30.5 pr
42.3 np
62.7 gk
27.1 qt
57.5 im
86.0 ab
0.0 v
47.4 S
90.0 a
78.2 af
64.6 fj
83.9 ab
67.3 ei
77.4 af
81.7 ad
75.0 bg
85.7 ab
81.5 ad
0.0 v
71.4 P
83.5 ac
81.7 ad
64.9 fj
85.7 ab
47.6 lo
65.2 fj
81.8 ad
58.7 hl
81.6 ad
90.0 a
0.0 v
67.4 Q
78.2 af
69.6 ci
50.3 ko
80.3 ae
44.6 mo
68.4 di
68.9 di
38.0 oq
62.5 gk
72.2 bh
0.0 v
57.6 R
0.0 v
0.0 v
0.0 v
0.0 v
0.0 v
0.0 v
0.0 v
0.0 v
0.0 v
0.0 v
84.1 ab
7.6 U
58.1 B
50.3 C
38.0 E
57.3 B
33.2 E
43.6 D
53.2 BC
35.5 E
50.4 C
67.8 A
14.0 F
Percentage death was calculated from cumulative number of dead over total number of seedlings (20 per treatment) 6 months after inoculation, and the actual wilt values were arcsinesquare root transformed to normalize the data. Means followed with the same letter(s) are not
significantly different from each other. Least significant difference values (P = 0.05) for the
cultivars, the isolates and the interactions comparisons are 3.5, 4.7, and 11.6, respectively.
Coefficient of variation (CV) = 15.8%.
b
Gx1, Gx2, Gx3, Gx4, Gx5, Gx6, Gx7, Gx8, Gx9, Gx11 and GxU12 designate G. xylarioides
isolates collected from Jimma, Gera, Chira, Gechi, Yayu, Mettu, Tepi, Bebeka, Ayraguliso,
Yirgacheffe and Uganda (C. canephora strain), respectively.
c
Cultivars Catimor-J19 and 7440 were resistant; F-59 was moderately resistant; and Caturra
Rojo and 24/85 were susceptible to CWD under field and greenhouse conditions. C. canephora
was not affected by the disease in the field in Ethiopia.
128
Girma et al.
after the trial had been completed. In contrast, C. arabica isolates were compatible with seedlings of all the arabica cultivars and caused varying percentages of death (Table 8.5) but induced no symptoms on C. canephora seedlings
(Plate 17). The compatibleincompatible interactions of the isolates with the
respective Coffea spp. were confirmed by re-isolating the fungus from the inoculated coffee seedlings. Results of both sets of the experiment proved host
specificity or specialization of G. xylarioides populations to each Coffea spp.
(Girma, 2004; Girma et al., 2005). Among the C. arabica cultivars, Catimor-J19
showed a significantly (P < 0.05) lower mean percentage of dead seedlings
(22.2%) followed by cultivar 7440 (47.4%) (Table 8.5), with incubation periods of 70 and 83 days, respectively, indicating high and moderate levels of
resistance. Cultivars F-59 and Caturra Rojo were susceptible to the disease
with the highest seedling deaths of 71.4% and 67.4%, respectively (Table 8.5)
(Girma, 2004; Girma et al., 2005).
There were significant (P < 0.05) ranges of variation in aggressiveness
among the isolates of C. arabica. Isolates Gx3, Gx5 and Gx8 caused low seedling infections of 38.0%, 33.2% and 35.6%, respectively, with long incubation periods (Table 8.5). On the contrary, significantly high mean death rates
of 58.1%, 57.3% and 67.8%% were induced by isolates Gx1, Gx4 and Gx11
(Table 8.5), respectively, suggesting that the former group of isolates were
more aggressive than the latter group (Gx3, Gx5 and Gx8). The differential
effects (interactions) indicated relatively low levels of infections on cultivar 7440 with Gx3, Gx5 and Gx8, with respective wilt incidences of 30.8%,
30.5% and 27.1%; whereas death rates of 66.2%, 65.8%, 62.7% and 86.0% were
caused on the same cultivar with isolates Gx1, Gx4, Gx7 and Gx11, respectively (Table 8.5). Catimor-J19 showed some horizontal resistance, although
it was highly infected by one isolate (Gx11). Cultivar 7440 was moderately
resistant, whereas Caturra Rojo and 2485 were susceptible to most arabica
isolates, but were moderately tolerant to Gx3, Gx5 and Gx8 isolates (Table
8.5) (Girma, 2004; Girma et al., 2005).
In conclusion, in both sets of experiments, all G. xylarioides isolates derived
from C. arabica induced disease in seedlings of cultivars of C. arabica (with
varying levels of aggressiveness), but these isolates did not induce disease
in C. canephora. Conversely, the strain from C. canephora was highly pathogenic to seedlings of its host but not to the C. arabica cultivars. Re-isolation
from samples of inoculated seedlings of C. arabica and C. canephora with
the respective strains showed a similar trend. These experiments represent
the first time that cross-inoculation experiments between C. arabica and
C. canephora plant material with their respective fungal isolates were undertaken and clearly evidenced host specialization of G. xylarioides (F. xylarioides)
populations to the two commercial coffee species. Such host specificity has
been previously speculated from field observations in DR Congo and Uganda
(Flood, 1996, 1997) and in Ethiopia (Girma, 1997; Girma et al., 2001). In addition to the indication of host specificity, the results further support the previous findings of Girma and Mengistu (2000) confirming horizontal resistance
in C. arabica, aggressiveness in arabica isolates and vertical resistance/virulence in the hostpathogen combinations. However, because only one isolate
129
of G. xylarioides from C. canephora had been studied, further investigations

were conducted with more strains.
8.5. Interactions Among Coffea Species (C. arabica/C. canephora/

Coffea excelsa) With G. xylarioides Isolates From Either Host
Species
The objectives of this study were to determine regional diversity of the current and historical strains of the coffee wilt pathogen collected from Coffea
spp. in almost all of the regions where tracheomycosis has been a serious
problem (Tables 8.68.8) and to further verify the reactions of the respective
host species and host specialization of the fungus populations. Three independent sets of cross-inoculation experiments were conducted using recently
isolated strains from C. arabica in Ethiopia (G3P22) and from C. canephora in
Uganda (CAB003), DR Congo (RDC002) and Tanzania (TZ008, TZ009). A
historical strain (DSMZ 62457) collected from C. excelsa in the Central African
Republic (CAR) in the 1960s and a recent isolate (OUG152) from the same
host species in Uganda were included. Seedlings of six C. arabica genotypes
obtained from Ethiopia, Kenya and Costa Rica; eight C. canephora lines from
DR Congo and Ivory Coast and Coffea liberica in Costa Rica were inoculated
with suspensions of each isolate following the standard stem nicking method
and then kept in a growth room with 12-h light/dark cycle at 25C temperature throughout the study period (Girma et al., 2007).
Table 8.6. Percentage seedling deatha of C. arabica and C. canephora inoculated with
G. xylarioides strains.
G. xylarioides strainc
Coffea spp.
C. arabica
C. canephora
Cultivars/
Linesb
Arabica strain
G3P22 (E)
CAB003 (U)
Canephora strain
TZ009 (T)
DSMZ62457 (C)
Excelsa strain
K7
SL28
LR/R1P2 (7)
LR/R1P3 (17)
LR/R1P4 (25)
TR CI17/37
90.0
93.8
0.0
0.0
5.4d
0.0
0.0
0.0
61.3
52.9
73.3
20.6
0.0
0.0
34.5
80.6
61.9
14.3
58.3
62.5
13.3
6.9
5.6
0.0
Percentage wilt was calculated from cumulative number of dead over total number of seedlings (ten per treatment) 6 months after inoculation.
b
C. arabica cultivars (K7 and SL28) widely grown in Kenya and C. canephora lines obtained
from DR Congo and Ivory Coast.
c
G. xylarioides strains collected from C. arabica (arabica strain) in Ethiopia (E), C. canephora
(canephora strains) in Uganda (U) and Tanzania (T) and a historical strain from C. excelsa
(excelsa strain) in CAR (C).
d
G. xylarioides was not re-isolated from seedlings (seedling death by other factors).
130
Girma et al.
Table 8.7. Percentage seedling deatha of C. arabica and C. canephora inoculated with
different G. xylarioides strains.
G. xylarioides strainsc
Arabica strain
Coffea spp.
C. arabica
C. canephora
Canephora strain
Excelsa strain
Cultivars/
Linesb
G3P22 (E)
CAB003
(U)
TZ008
(T)
RDC002
(C)
OUG152 (U)
7454
74165
TR CII7/1
TR CI17/16
53.3
30.0
4.0d
8.3d
0.0
0.0
63.0
21.7
0.0
0.0
48.1
38.5
0.0
0.0
59.3
46.2
0.0
0.0
65.4
40.9
Percentage wilt was calculated from cumulative number of dead over total number of seedlings
(ten per treatment) 6 months after inoculation.
b
C. arabica cultivars (7454 and 74165) grown in Ethiopia and C. canephora lines obtained from
DR Congo.
c
G. xylarioides strains collected from C. arabica (arabica strain) in Ethiopia (E); C. canephora
(canephora strains) in Uganda (U), Tanzania (T), and DR Congo (C) and a recent strain from
C. excelsa (excelsa strain) in Uganda (U).
d
In the first trial set, C. arabica isolate G3P22 (IMI 392680) from Ethiopia
was shown to be aggressively pathogenic to seedlings of the two C. arabica
cultivars K7 and SL28 from Kenya, with 90% and 94% seedling deaths, respectively, at the end of the trial after 6 months (Table 8.6). Historical strain
DSMZ 62457 (IMI 127629) of C. excelsa in CAR also infected seedlings of these
cultivars (58% and 63% deaths, respectively). C. canephora isolates CAB003
(IMI 392263) and TZ009 (IMI 392679), obtained from recent CWD outbreak
areas in Uganda and Tanzania, respectively, attacked the four C. canephora
(robusta lines) from DR Congo and Ivory Coast with seedling mortality
ranging from 14.3% to 80.6% (Table 8.6). Neither of these isolates induced
symptoms in C. arabica (Plate 18). A number of the C. canephora seedlings
inoculated with the historical strain also developed symptoms (13.3% death)
(Girma et al., 2007). Similar hostpathogen combinations were observed on
grown seedlings (9 months old) of the catimor line (1570) inoculated with the
three strains DSMZ, G3P22 and CAB003 (Plate 19).
In the second trial set, C. arabica isolate G3P22 infected seedlings of the
two arabica coffee cultivars 7454 and 74165, which were obtained from Ethiopia, but did not induce symptoms on C. canephora lines (Table 8.7). Conversely, these C. arabica cultivars were not infected by any pathogen strains
from C. canephora, namely, CAB003, TZ008 (IMI 392679) and RDC002 (IMI
392268). Strain OUG152 (IMI 392681), isolated from symptomatic excelsa
coffee at Kituza, Uganda, caused up to 65% seedling death in C. canephora
(robusta lines) but did not cause wilting in C. arabica (Girma et al., 2007). In
the third set of experiments, the arabica isolate G3P22 killed seedlings of
the two C. arabica accessions known as Yemen/Java and E-238, which were
131
Table 8.8. Percentage seedling deatha of C. arabica, C. canephora and C. liberica

inoculated with G. xylarioides strains.
G. xylarioides strainsc
Coffea spp.
C. arabica
C. canephora
(robusta type)
C. liberica
Cultivars/
Linesb
Arabica strain
G3P22 (E)
Canephora stain
CAB003 (U)
Excelsa strain
DSMZ62457 (C)
Yemen/Java
E-238
KR 10/7
KR 15/4
T-1984
T-1872
90.9
77.8
0.0
0.0
5.0d
21.1
0.0
0.0
51.9
42.3
100
95.8
4.2d
37.0
15.4
0.0
80.0
94.4
Percentage wilt was calculated from cumulative number of dead over total number of seedlings
(ten per treatment) 6 months after inoculation.
b
C. arabica cultivars (Yemen/Java and E-238) grown in Costa Rica, C. canephora (robusta
line) obtained from DR Congo and C. liberica collected in Costa Rica.
c
G. xylarioides strains collected from C. arabica (arabica strain) in Ethiopia (E), C. canephora
(canephora strains) in Uganda (U) and a historical strain from C. excelsa (excelsa strain) in
CAR (C).
d
received from Costa Rica, with 91% and 78% death, respectively (Table 8.8).
Seedlings of C. liberica lines, T-1984 and T-1872, from the same country were
susceptible to the historical strain DSMZ 62457 and the current robusta isolate CAB003. The characteristic wilting symptoms appeared on these strain
coffee line combinations within less than 30 days of inoculation, and most of
the seedlings collapsed at cotyledon stage (Plate 20). Accession T-1872 also
exhibited moderate susceptibility to the arabica isolate.
In summary, there were no wilting symptoms, and the fungus was also
not re-isolated from control (uninoculated) seedlings in the three trial sets;
successful re-isolation of G. xylarioides from plant tissues of most of the inoculated and symptomatic seedlings (inoculated with their respective isolates)
confirmed the hostpathogen compatibility. The results of all these crossinoculation trials further evidenced that, with the exception of the historical strain, host specialization occurs in the G. xylarioides populations. Those
strains collected in the field from C. arabica trees showing CWD symptoms
were pathogenic to C. arabica cultivars obtained in Ethiopia, Costa Rica and
Kenya and suggest that this disease could be a serious problem on arabica
coffee if introduced into countries that are tracheomycosis-free (both in Africa and beyond). In addition, the strains isolated from C. canephora trees
in countries with recent CWD outbreak, namely, DR Congo, Uganda and
Tanzania, were aggressively pathogenic to all the seedlings of C. canephora
(robusta lines) collected in Ivory Coast and DR Congo (Girma et al., 2007).
The recent strain OUG152 (IMI 392681) isolated from C. excelsa in a
clonal trial site of the Coffee Research Institute at Kituza, Uganda, also induced disease in C. canephora lines but did not affect C. arabica. Excelsa strains
132
Girma et al.
OUG151 (IMI 392265), OUG154 (IMI 392682) and OUG155 (IMI 392266) isolated from the same host species showed varying levels of seedling death
on C. canephora (robusta line, 1331) (D. Bieysse, personal communication).
In contrast, the historical excelsa strain DSMZ 62457 (IMI 127629), originally
collected from C. excelsa trees in CAR in the 1960s, induced typical CWD
symptoms and exhibited a wide range of pathogenicity (seedling death) on
C. arabica, C. canephora and C. liberica. Pathogenicity tests of other historical
strains ATTC 15664 (IMI 392676) and CBS 749.79 (IMI 392675) collected from
C. canephora in Ivory Coast and CBS 258.52 (IMI 392674) isolated from an
unknown Coffea sp. in Guinea caused death of 10%, 30% and 35% of plants
of C. canephora (robusta line 1406), respectively (D. Bieysse, personal communication). These results suggest that historical strains collected during the
1960s remained aggressively pathogenic to Coffea spp. after 40 to 50 years of
preservation. Girma et al. (2005) hypothesized that the compatibility and infection of C. canephora by excelsa strains imply that G. xylarioides population
presently causing CWD outbreaks in DR Congo, Uganda and Tanzania may
have arisen from older populations in CAR or, alternatively, that a separate
divergent population is evolving on C. excelsa in nature (Girma et al., 2007).
8.6. Histological Studies of C. canephora/G. xylarioides Interactions

Detailed understanding of infection processes from penetration of the pathogen to colonization of the host tissues is very important to define hostpathogen interaction at cellular levels. Thus, preliminary observations were carried
out on C. canephora (robusta line) seedlings 9 and 18 months old that were
inoculated with G. xylarioides strain CAB003 (IMI 392263) by injection of a
droplet of spore suspension of conidia (106 conidia/ml) into the trunk under
the cotyledon leaves using a syringe. Light microscopic studies were carried
out on 2-mm sections excised from the inoculated seedlings showing early
symptoms (leaves were partially wilting) 8 weeks after inoculation. The five
internodes were sampled successively upwards starting from the cotyledon
leaves per sample seedlings.
8.6.1. Histological study on 9-months-old C. canephora seedlings

Microscopic observations of inoculated plants at the first internode level
above the cotyledons revealed that the epidermis and external parenchyma
had disappeared and the cork and phellogen layers had partially disintegrated. The wood tissues were found disturbed in some places but had not
become totally disorganized; no mycelium could be seen in these sections.
The epidermis, external parenchyma and phellogen were all absent at the
2nd internode position, and slight hypertrophy of the parenchyma was observed, but the xylem tissue was intact and no mycelium was detected. At
the 3rd internode level, the epidermis was absent, and swelling of the external and upper parenchyma cells had produced a clear tissue deformation
133
and some disintegration. Xylem and pith cells had also disintegrated and
contained a large quantity of starch grains, but no mycelium was observed. At the 4th internode, a swollen discoloured structure was observed,
which spread over 1 cm along the stem. The swelling of the epidermis and external parenchyma cells was due to an increase of the volume of the vacuole.
Although phellogen cells were damaged, the cork cell structure remained
intact except for a slight swelling of cell nuclei. The fungus was observed in
the wood tissues, more especially in fibres, and appeared as a blue, intracellular mycelium colonizing fibres closer to the pith (linearly following the
rays) (Plate 21). Cells that contained mycelium had little or no lignin on their
walls. Mycelial colonization appeared absent from vessels. At the 5th internode, the youngest level, the epidermis was absent; parenchyma cells were
strongly hypertrophied leading to tissue distortion; the phellogen was absent
and no mycelium observed. A significant accumulation of starch granules
could be observed in pith and in the wood cells. The presence of mycelium
at the 4th node from the infection point indicates that it had colonized the
plant probably through translocation of conidia that are transported in the
sap towards the apical part of the plant.
8.6.2. Histological interaction study on 18-month-old C. canephora seedlings

In 18-month-old plants, where the symptoms were advanced, the mycelium
was observed to be invading the xylem, and dense colonization was seen in
the vessels and in the intercellular spaces (Plate 22). However, in sections
taken from those plants exhibiting early symptoms, the fungus was not
observed.
8.7. Summary and Conclusions

The prospects of successful control of CWD rely principally upon deployments of resistant coffee cultivars/lines. In this regard, full-fledged, independent, large-scale coffee collection and screening strategy should be planned
hand in hand with efficient breeding programme. Because of the difficulty in
screening for resistance under field conditions for soil-borne diseases, standard seedling inoculation protocols should be employed. For a number of
technical reasons, stem nicking (at 2 cm above the soil level) of late-cotyledonstage (2 to 2.5 months old) coffee seedlings with inoculum concentration of
2 106 conidia/ml is recommended. The inoculated seedlings need to be
placed in an air-conditioned room with higher humidity (>95%) and temperature of about 23C for 1 week that ensure infection and then be transferred
to the greenhouse (Plate 23). Disease parameters such as number of dead
seedlings and number of days between inoculation and external symptom
appearance (chlorosis, retarded plant growth and reduced leaf size, wilting
and finally death) should be recorded for at least 6 months (preferably at
fortnightly intervals), from which percentage death and incubation periods
134
Girma et al.
can be computed and used for statistical analyses. Pieters and Van der Graaff
(1980) reported the highest correlation between the field scores and death
rates and incubation time in the seedling tests 6 months after inoculation.
The resistance of surviving seedlings of a particular cultivar/line needs to
be further verified by re-inoculating grown seedlings (12 months old) in the
screen house (following the same technique) and evaluated at least for one
more year (Plate 24) because of the difficulties of proving host resistance under field conditions.
Following on from the detailed hostpathogen interactions supported by
random amplified polymorphic DNA analysis, Girma et al. (2005) introduced
the epithet formae speciales (special forms), G. xylarioides f. sp. abyssiniae (anamorph: F. xylarioides f. sp. abyssiniae) for strains attacking C. arabica confined
to Ethiopia and G. xylarioides f. sp. canephorae (anamorph: F. xylarioides f. sp.
canephorae) for strains specifically pathogenic to C. canephora and C. excelsa.
Rutherford (2006), based on the various molecular studies, reported that two
clonal populations are responsible for the current CWD outbreaks in Africa,
one composed of isolates obtained from affected C. arabica in Ethiopia (variant A), and the other are isolates from affected C. canephora in DR Congo,
Uganda, and Tanzania (variant C) (Chapter 7). Phylogenetic analysis of F.
xylarioides including the number of isolates obtained from CWD-affected
coffee trees in Uganda has shown these to belong to the African clade of
the Gibberella fujikuroi species complex, a clade replacing Fusarium section
Liseola (Geiser et al., 2005). The appearance of new populations or strains of
the pathogen also needs to be monitored in the field because G. xylarioides
produces very abundant perithecia and ascospores in the stem bark of dead
trees and stumps of coffee (Van der Graaff and Pieters, 1978; Flood and Brayford, 1997; Girma et al., 2001; Girma, 2004).
The preliminary histopathological observation of the interaction of
C. canephora vs. G. xylarioides highlighted the hyphal colonization of xylem
tissues and the intercellular spaces, blocking water movement that leads finally to wilting of infected coffee trees. However, the underlying resistance
mechanisms that may be morphological, physiological and/or biochemical
involved in host defense systems of different cultivars/lines of Coffea species
should also thoroughly be further studied. The resistance in C. canephora and
C. liberica was supposed to be associated with rapid suberization in wounds
and occurrence of caffeine and chlorogenic acid in high concentrations (Booth,
1971; Holliday, 1980). Understanding the gene(s) that govern resistance to
CWD and mode of inheritance of the resistance mechanisms is of paramount
importance to design and implement successful breeding programme.
Acknowledgements
The authors acknowledge the Department for International Development,
UK, for funding part of the work done at CIRAD, Montpellier, and we are
very thankful to all the staff at Jimma Agricultural Research Center (EIAR,
Ethiopia) and CIRAD (France) for their unreserved technical assistance.
135
References
Beckman, C.H. (1987) The nature of wilt diseases of plants. APS Press, St. Paul, MN,
USA.
Beckman, C.H. and Talboys, P. W. (1981) Anatomy of resistance. In: Mace, M.E., Bell,
A.A. and Beckman C.H. (eds.) Fungal Wilt Diseases of Plants. Academic Press,
New York, pp. 487521.
Booth, C. (1971) The Genus Fusarium. Commonwealth Mycological Institute, Kew, Surrey, England.
Report presented to Zairean Coffee Organization (OZACAF) August 1996, 13 pp.
Flood, J. (1997) Tracheomycosis or vascular wilt disease of coffee in Uganda. Report
Flood, J. (2006) A review of Fusarium wilt of oil palm caused by Fusarium oxysporum f.
sp. elaeidis. Phytopathology 96, 660662.
Flood, J. and Brayford, D. (1997) Reemergence of Fusarium wilt of coffee in Africa.
In: Proceedings of the 17th International Scientific Conference on Coffee Science
(ASIC). Nairobi, Kenya, pp. 621627.
Geiser, D.M., Lewis Ivey, M.L., Hakiza, G., Juba, J.H. and Miller, S.A. (2005) Gibberella
xylarioides (anamorph: Fusarium xylarioides), a causative agent of coffee wilt
disease in Africa, is a previously unrecognized member of the G. fujikuroi species
Girma, A. (1997) Status and economic importance of Fusarium with disease of Arabica
coffee in Ethiopia. In: Hakiza, G. J., Birkunzira, B. and Musoli, P. (eds.) Proceedings
of the First Regional Workshop on Coffee Wilt Disease (Tracheomycosis). International Conference Centre, Kampala, Uganda, pp. 5361.
Girma, A. (2004) Diversity in pathogenicity and genetics of Gibberella xylarioides
(Fusarium xylarioides) populations and resistance of Coffea spp. in Ethiopia. PhD
dissertation. University of Bonn, Bonn, Germany.
Girma, A., Flood, J., Hindorf, H., Bieysse, D., Simons, S. and Mike, R. (2007) Tracheomycosis (Gibberella xylarioides) a menace to world coffee production: evidenced
by cross inoculation of historical and current strains of the pathogen. In: Proceedings of the 21st International Scientific Conference on Coffee Science (ASIC).
Girma, A. and Hindorf, H. (2001) Recent investigation on coffee tracheomycosis,
Gibberella xylarioides (Fusarium xylarioides) in Ethiopia. In: Proceedings of the
19th International Scientific Conference on Coffee Science (ASIC). Trieste, Italy,
pp. 12461252.
Girma, A., Hindorf, H., Steiner, U., Nirenberg, H., Dehne, H.-W. and Schellander, K.
(2005) Genetic diversity in the coffee wilt pathogen (Gibberella xylarioides) populations: differentiation by host specialization and RAPD analysis. Journal of Plant
Diseases and Protection 112, 134145.
Girma, A. and Mengistu, H. (2000) Cultural characteristics and pathogenicity of
Gibberella xylarioides isolates on coffee. Pest Management Journal of Ethiopia 4,
1118.
Girma, A., Mengistu, H. and Hindorf, H. (2001) Incidence of tracheomycosis, Gibberella
xylarioides (Fusarium xylarioides) on arabica coffee in Ethiopia. Journal of Plant
Diseases and Protection 108, 136142.
Hakiza, G.J., Kyetere, D.T. and Olal, S. (2004) Mode of penetration and symptom expression in robusta coffee seedlings, inoculated with Gibberella xylarioides, the
136
Girma et al.
cause of coffee wilt disease in Uganda. In: Proceedings of the 20th International
Scientific Conference on Coffee Science (ASIC). Bangalore, India, pp. 12321233.
Holliday, P. (1980) Fungal Diseases of Tropical Crops. Cambridge University Press,
London.
Kilambo, D.L., Nghoma, N.M., Mtenga, D.J., Teri, J.M., Nzallawahe, T., Mike, R.
and Masumbuko1, L. (2007) Progress towards searching for durable resistance to
Fusarium wilt (Fusarium xylarioides) in Coffea canephora germplasm in Tanzania.
In: Proceedings of the 20th International Scientific Conference on Coffee Science
(ASIC). Bangalore, India, pp. 13861389.
Kranz, J. (1962) Coffee diseases in Guinea. FAO Plant Protection Bulletin 10, 107
109.
Leung, H., Nelson, R.J. and Leach, J.E. (1993) Population structure of plant pathogenic
fungi and bacteria. Advances in Plant Pathology 10, 157205.
McDonald, B.A. (1997) The population genetics of fungi: tools and techniques. Phytopathology 87, 448453.
Muller, R.A. (1997) Some aspects of past studies conducted in western and central
Francophone Africa on coffee tracheomycosis. In: Hakiza, G. J., Birkunzira, B. and
Musoli, P. (eds.) Proceedings of the First Regional Workshop on Coffee Wilt Disease (Tracheomycosis). International Conference Centre, Kampala, Uganda, pp.
1830.
Musoli, P., Lukwago, G. Abebe, M., Tesfaye, A., Kilambo, D., Asiimwe, T. and
Munyankere, P. (2003) Surveys to assess the extent and impact of coffee wilt
disease in East and Central Africa. Final Technical Report EU contract no. ASA-RSP/
Pieters, R. and Van der Graaff, N.A. (1980) Gibberella xylarioides on arabica coffee:
evaluation of testing methods and evidence for the horizontal nature of resistance.
Netherlands Journal of Plant Pathology 86, 3743.
Rutherford, M.A. (2006) Current knowledge of coffee wilt disease, a major constraint to
coffee production in Africa. Phytopathology 96, 663666.
Van der Graaff, N.A. and Pieters, R. (1978) Resistance levels in Coffea arabica L. to
Gibberella xylarioides and distribution pattern of the disease. Netherlands Journal
of Plant Pathology 84, 117120.
Van der Plank, J.E. (1984) Disease Resistance in Plants, 2nd edn. Academic Press, London.
Zadoks, J.C. and van Leur, J.A. (1983) Durable resistance and hostpathogenenvironment interaction. In: Lamberti, L., Waller, J.M. and Van der Graaff, N.A. (eds.)
Durable Resistance in Crops. Plenum Press, New York, pp. 125140.
Watson, I.A. (1970) Changes in virulence and populations shifts in plant pathogens.
Annual Review of Plant Pathology 8, 209230.
Wellman, F.L. (1961) Coffee Botany, Cultivation and Utilization. Leonard Hill (Books)
Ltd., London.
N. Phiri, M. Kimani, E. Negussie, S. Simons and G. Oduor

9.1. Introduction
Until the early 1990s, coffee (Coffea spp., Rubiaceae) was the worlds most important internationally traded commodity in terms of monetary value, after oil (Rice and Ward, 1996). Currently, however, coffee is ranked as only
the 5th most important traded commodity after oil, aluminium, wheat and
coal. In 2003, global green coffee production was 7,221,255 t, of which approximately 60% was arabica (Coffea arabica) and 39% robusta coffee (Coffea
canephora) (Rice and Ward, 1996). Of the coffee produced in 2003, 5,233,064 t
were exported, with a value of approximately $5.5 billion (FAO, 2004). Despite its decline in rank as a traded commodity, coffee continues to be an important source of foreign exchange earnings, and it is the primary export of
many developing countries. Cultivation, processing, trading, transportation
and marketing of coffee provide employment for millions of people worldwide (International Coffee Organisation [ICO], 2005).
In sub-Saharan Africa, coffee is the economic backbone of more than
50 countries and central to the livelihoods of more than 20 million rural families
(Oduor and Simons, 2003). It is also a major source of income for millions of
smallholder coffee growers and their households who are responsible for an
estimated 80% of coffee production in Africa (Oduor and Simons, 2003). Coffee is the most important cash crop for Africa as a whole, contributing some
10% of the total foreign exchange earnings in the continent (FAO, 2004).
A number of coffee-producing countries in sub-Saharan Africa, including
Uganda, Ethiopia, Rwanda and Burundi, depend on the export of this commodity for more than half of their foreign exchange earnings.
However, African coffee suffers from a range of co-evolved pests and
diseases including coffee berry disease, coffee leaf rust (CLR), coffee berry
borer etc. Global coffee production has nearly doubled since 1970, but Africas share of total production declined from 30% to less than 15% (FAO,
137
138
Phiri et al.
2004) despite accounting for one third of the worlds coffee hectarage. The
average yields in Africa are generally low and declining, ranging from 0.3 to
0.38 t/ha, half that achieved in Latin America and almost one third of Asias
productivity. This is partly due to history: some 40% of African plantations
date from the pre-independence era and have not been renewed since. The
continued reliance on outdated and often unproductive varieties in the face
of widespread prevalence of pests and diseases has contributed to this decline, as has the recent re-emergence of coffee wilt disease (CWD).
CWD or tracheomycosis is a highly destructive disease of robusta coffee (C. canephora) and arabica coffee (C. arabica) throughout East and Central Africa. Coffee production in Uganda and the Democratic Republic of
Congo (DRC) has been decimated following the re-emergence of CWD in
the 1970s and 1980s in DRC and 1990s in Uganda (Flood and Brayford, 1997;
G. Hakiza, 1995, unpublished data; Oduor et al. 2003; Office Zairois du Cafe
[OZACAF], 1995), and the disease has since spread to neighbouring countries in the region including Tanzania. CWD is caused by a fungus called
Gibberella xylarioides R. Heim & Saccas, which has Fusarium xylarioides as its
anamorphic phase. Following the re-emergence of CWD, income from coffee
declined by over 50% in three CWD-affected countries. A number of control
methods are available for the management of CWD, which include use of
resistant varieties, chemical control and cultural methods.
This chapter discusses available management methods and those which
could be available in the near future.
9.2. Cultural Practices Used in Managing CWD

A range of cultural practices can be used in managing CWD. Perhaps the
commonest of all and a well-known method is the uprooting of infected
coffee trees and burning in situ. Other cultural methods include quarantine
(intra- and interfarm, inter- and intracountry quarantine), use of diseasefree planting materials, prevention of coffee tree wounding, control of insect
pests, disinfecting farm implements (e.g. secateurs, pruning saws) and general plant health. These methods are discussed below.
9.2.1. Uprooting and burning

Uprooting and burning is probably one of the oldest CWD management
methods. It dates back to the first CWD outbreak in the 1950s (Saccas, 1956)
and contributed to the control of CWD when coupled with use of resistant
varieties. The method works by removing the source of pathogen inoculum
and involves frequently inspecting the coffee farm to identify infected coffee
bushes. Once identified, the infected coffee bushes are uprooted by digging
out as much of the root system as possible then burning them on the spot at
the earliest opportunity in the hole where the coffee bush was uprooted
(Plate 28). Although early identification of symptoms, uprooting and burning
139
the infected coffee trees at this early stage is crucial for CWD control, it
poses problems because the coffee materials are not dry enough at this stage
and do not burn easily. The best way for carrying out uprooting and burning newly infected coffee trees is, therefore, to cut the coffee materials into
smaller pieces, which promotes drying, and these pieces can then be heaped
in the hole from where the infected coffee tree was uprooted. Dry firewood
could also be used to facilitate the burning. To reduce further risk of inoculum in adjacent coffee trees, those coffee trees surrounding infected coffee
trees also need to be uprooted and burnt. Uprooting the coffee tree that has
already died from CWD is too late; the pathogen will have already spread to
surrounding and distant coffee trees in the farm and even to adjacent farms.
Conidia and ascospores of G. xylarioides spread through wind, rain and human activities (e.g. harvesting, pruning) (Jacques-Felix, 1954). When CWD
infection levels are high, such as 70% of coffee trees being infected in a coffee garden, it is advisable to uproot and burn all coffee trees in the farm and
replant with resistant or tolerant coffee germplasm. This may pose problems
due to loss of the little remaining source of income for the farmer. Educating the farmers about the advantages and disadvantages of uprooting and
burning all coffee trees in heavily infected farms should help in this case. The
practice of burning infected coffee plant material on the spot helps in limiting the spread of G. xylarioides inoculum from the plant material and the contaminated soil. G. xylarioides is considered to be an endemic soil-inhabiting
fungus (Chapter 7). Although most of the root system may be dug out, much
still remains in the soil and is potential source of inoculum. Such inoculum
can be eliminated by observing a fallow period of at least 6 months (Hakiza,
personal communication) to 2 years (Wrigley, 1988) before replanting with
a susceptible coffee seedling. By this time, the viability of the soil inoculum
will have declined (Wrigley, 1988).
Training of farmers in the identification of CWD symptoms, especially
the early stages of symptom expression, is crucial for early diagnosis and
therefore for better chances of eradication. CWD symptoms include leaf curling, chlorosis and wilting (Plate 3), with eventual leaf defoliation leading
to dieback (Plates 3 and 4). The symptoms, in most cases and especially in
multiple-stem coffee bushes, start on one side of the coffee stem where the
infested vascular bundles are, which have been blocked by a combination of
fungal colonization and host responses. These symptoms are confirmed by
scraping the diseased stem with a knife. A blueblack stain is characteristic
of an infected coffee stem (Plate 8). Training of farmers and extensionists is
so important in the management of CWD that it formed a major component
of the Regional Coffee Wilt Programme (RCWP), Chapter 11.
Yet despite clear guidance about burning in situ, there is a common practice by people within the coffee-growing areas to transport infected coffee
stems to their homesteads for use as firewood, as staking material for climbing beans and even as fencing material around coffee gardens (Chapters 2, 3,
4 and 5). These practices help spread CWD further in the area. This poses a
huge threat to the spread of CWD, and limiting these practices will improve
management of CWD. Proper disposal of infected coffee material is crucial
140
Phiri et al.
in the management of CWD because the bark, roots and stems all contain the
pathogen in the form of spores, hyphae and fruiting bodies (CAB International [CABI], 2006).
There are a number of examples of use of uprooting and burning. When
CWD first appeared in West Africa from 1927 to the 1950s, it was controlled
in Cameroon by mobilizing soldiers to uproot and burn all infected coffee in
the country (Deassus, 1954). In recent outbreaks, CWD is being controlled
by practicing uprooting and burning in addition to other control methods.
In Tanzania, for example, farmers are carrying out uprooting and burning
campaigns of CWD-infected coffee trees and are being supported by their
extensionists, members of Tanzania Coffee Research Institute (TACRI) and
the local, district and regional leaders in the Bukoba, Misenyi, Karagwe and
Muleba districts of the Kagera region (Chapter 5). Uprooting and burning is
used in combination with use of copper stem paint and mulching. In addition, uprooting and burning of infected coffee trees is also being practiced in
DRC, Ethiopia and Uganda (Chapters 3 and 4).
However, where governments need to bring in soldiers or others to uproot and burn farmers CWD-infected coffee as part of a concerted action,
appropriate legislation to back up such an action may be needed, although
many countries already have such legislation in place.
9.2.2. Quarantine as a tool for managing the spread of CWD

Quarantine is an important cultural control method for CWD and is aimed
at preventing the movement of infected plant material, infected soil and infected implements to clean areas where it can be a source of new disease
outbreaks. Quarantine can be used at farm, national and international levels. It is rare for most people to think of quarantine at farm level. However,
it is crucial that a farmer observes quarantine to limit spread of CWD in
or among his/her farm(s). The practice includes carrying out interfarm activities from the non-affected farm and finishing with the affected farm. In
addition, the farmer should have separate implements for the affected and
non-affected farms to reduce the risk of spreading CWD to the clean farm.
When working within the affected farm, the farmer should start working in
the clean section and finish with the affected section of his/her coffee farm,
thus minimizing spread of CWD within the farm. This is because the pathogen can spread through contaminated farm implements, even on clothes and
shoes/feet when contaminated soil sticks to the shoes/feet or farm implements. Spores of G. xylarioides can also stick to clothes if a farmer is in contact
with infected coffee bushes, especially if they have been left in the garden for
a long time. Quarantine can also be used between farmers; exchange of coffee plant material and farm implements should be discouraged to reduce the
risk of spreading CWD. The practice of using diseased material as firewood
(see 9.2.1) has had a huge impact on the spread of CWD, and steps to limit
these practices are being implemented mainly through raising awareness
among farmers.
141
Quarantine also involves monitoring and restricting movement of plant

material between affected and non-affected countries. For example, coffeegrowing countries like Burundi, Rwanda and Kenya, which have neighbours
with CWD outbreak, should be diligent in preventing cross-border movement
of coffee material (germplasm, coffee in parchment or green coffee) to prevent
spread of CWD from the infected neighbouring countries. This has already occurred between DRC and Uganda. CWD was observed on abandoned coffee
farms in eastern DRC in 1986 and is believed to have moved or been moved to
Uganda, as it was recorded in the Ugandan district of Bundibujyo in the early
1990s (Flood and Brayford, 1997). However, implementation of quarantine
measures at intercountry levels may be difficult to enforce in Africa because
of the many porous borders and unmonitored informal cross-border trade,
which could have accounted for the spread of the current outbreak of CWD
from DRC to Uganda and subsequently from Uganda to Tanzania.
Enforcement of intercountry quarantine is probably the only chance for
Kenyas C. arabica, which faces threat from the arabica strain of G. xylarioides
that is affecting C. arabica in Ethiopia. Studies by Girma et al. (2001) (Chapter 8) showed that most of the existing Kenyan coffee varieties, except for
Ruiru 11, which was not included in the study, were susceptible to the arabica strain, which is referred to as variant A. However, the desert between
the two countries provides a natural barrier and helps in reducing the risk of
the arabica strain from spreading to Kenya and the rest of the arabica-coffee
growing countries south of Ethiopia.
9.2.3. Use of disease-free planting materials

CWD is spread through planting materials, among others. Although it is a
subject of debate, transmission through seed is a potential way of spreading CWD. Wrigley (1988) quotes Clowes and Hill (1981), who claimed that
a berry and branch blight in Zimbabwe was due to G. xylarioides and that
this was spread in seed from berries that may contain the CWD pathogen.
However, CWD has not been confirmed in Zimbabwe, and further observations showed that the disease in Zimbabwe was similar to that caused by
Fusarium bark disease (Fusarium stilboides) and was probably confused with
CWD. Studies by Girma and Hindorf (2001) showed that seeds did not transmit CWD in arabica coffee in Ethiopia. It is important to note, however, that
G. xylarioides is a systemic pathogen affecting the vascular elements, and
therefore, seed transmission may be possible even though it has not been confirmed so far. Seedlings may also transmit G. xylarioides through soil adhering to their roots and spores carried on their stems and branches. Seedlings
from reliable sources, such as registered clonal propagators, may limit this
risk. In addition, farmers and seedling producers should be trained in using
potting soil from an area without the potential of harbouring G. xylarioides,
such as a virgin land or an area without a history of coffee. Soil sterilization
is another option for eliminating soil-borne inoculum, but this may be expensive and time consuming.
142
Phiri et al.
The availability of coffee planting materials is a challenge at the moment

because of lack of available varieties or clones resistant to CWD. The current
available resistant clones in Uganda are still being propagated, and it will be
sometime before these clones will be widely available to farmers. Resistant
clones in Tanzania and DRC are yet to be evaluated for other traits such as
red blister (Cercospora coffeicola) and CLR and agronomic traits.
9.2.4. Prevention of coffee tree wounding

G. xylarioides penetrates through wounds, so any agency causing wounds
will aid the spread of the fungus. Kranz and Mogk (1973) noted that most
dying and dead trees had been wounded during weeding. The main author observed wounds caused by weeding practices, in particular slashing with a bushmans knife (slasher) on most of the infected coffee trees in
Ethiopia (Plate 6). This is worsened by the fact that most of the coffee are
grown using very close spacing in Ethiopia, and some of the coffee are actually semiforest type, i.e. it was not planted by farmers, hence may grow
at even closer spacing than the recommended one. Coffee tree wounding
is also prevalent in most countries in Africa, and it is mainly due to weed
management.
In addition, there are other coffee farm operations that have to be carried out which cause wounds on the coffee bush. Particular operations that
pose a huge challenge in the prevention of CWD spread is the rejuvenation
of coffee or changing of a coffee cycle that involves stumping. Changing a
coffee cycle is a routine practice that has to be carried out when coffee bushes
have aged or have become unproductive. The operation requires cutting of
coffee stems at about 45 cm above the soil level, at an angle, with a pruning
saw so that new coffee shoots that sprout below the cut point are selected
and one or two are allowed to form the next coffee cropping cycle. Changing
the coffee cycle is carried out several times for over a 50-year period, and the
technique allows the bushes to be maintained over many years. Although
this is a very important way of turning unproductive coffee bushes into productive ones, there is a high risk of transmitting G. xylarioides through the
pruning saw. In addition, the freshly cut surface of the coffee stump provides
a very large surface area that can be used as an entry point for G. xylarioides.
The main author here has observed farms where coffee trees were completely destroyed by CWD after changing the cycle in Ethiopia. The new
suckers that were sprouting from the cut stumps were all infected with CWD
and hence died before they even grew up to 1 m from the growing point.
To prevent this catastrophe, the implements should be flamed over fire before cutting each stem or after cutting a few stems so that they are sterilized, particularly in farms with the history of CWD. Coffee trees showing
CWD symptoms need to be uprooted and burnt. If a farmer can afford, it
is advisable for the stumps to be painted with a fungicide paste. An ordinary copper-based fungicide can be used for this purpose. The copper-based
fungicide should be mixed with water at the rate of 300 g copper-based
143
fungicide to 1 l of water. The paste can be applied with a paint brush on the
cut surface of the stump.
Pruning to cut off dead or interlocking branches or branches that are
pointing inwards is also a potential way of spreading CWD and should be
managed properly. Secatuers that are used in the pruning process should be
sterilized over fire as described above.
However, wounding also occurs naturally at the collar area of the coffee
stem (the area of the coffee stem at the soil level). These microscopic wounds
develop mostly when the trees sway due to heavy winds. The wounds can
therefore be reduced through provision of wind breaks, for example planting a band of bananas or closely planted trees, in particular fast-growing
trees around the edge of the coffee farm, could help reduce the level of this
type of wounds.
Another cause of wounding is wood-boring insect pests such as the
white coffee stem borer (Monochamus leuconotus [Pascoe] or Bixadus sierricola
[White]) and the yellow-headed borer (Dirphya nigricornis [Olivier]), which
create entry points, ring barking and entry and exit holes that could provide
entry points for G. xylarioides. In addition, insects are suspected to transmit
CWD from tree to tree. Insect pests ought to be controlled to prevent the
spread of CWD in addition to the primary objective of preventing direct crop
losses caused by the insect pest damage. Fipronil, at a dilution rate of 1.25 ml
in 1 l of water, was found effective in controlling the white stem borer in
Malawi and Zimbabwe when applied as stem paint to the first 50 cm of the
stem from the soil level. In addition, chlorpyrifos diluted at 35 ml in 1 l of
water and applied to the first 50 cm of the coffee stem was also found effective to white stem borer in Zimbabwe (Integrated Stem Borer Management
in Smallholder Coffee Farms in India, Malawi and Zimbabwe Project [CFC/
ICO/18] Final Technical Report, 2008).
Livestock, in particular goats and cattle, also wound coffee stems. Goats
were observed, on a number of occasions, eating the coffee bark in Uganda
when left to graze in the coffee gardens. In addition, cattle were found tethered to coffee stems resulting in the rope, used for tethering the animals,
wounding the coffee stem as the tethered animal moved around the coffee
tree when grazing. It is therefore imperative that goats ought to be prevented
from grazing in coffee gardens, and cattle should not be tethered to coffee
stems to prevent wounding.
Wounding can be avoided by a number of careful management techniques that help in managing CWD.
9.2.5. Weed management practices for preventing coffee stem wounding

Mulch (Plate 30) suppresses weeds but limits the use of herbicides, and combined with hand-picking weeds around the coffee stems instead of slashing
significantly reduced the incidence of CWD in recent CWD management studies, which were part of the Improvement of Coffee Production in Africa by
the Control of Coffee Wilt Disease (Tracheomycosis) Project (CFC/ICO/13).
144
Phiri et al.
This resulted in farmers adopting these methods for controlling weeds. In addition, live mulch, Desmodium intotum, which suppresses weeds during its
growing condition (Plate 31), was also adopted as an alternative to weed
control after substantial studies by the weed scientist at Jimma Agriculture
Research Centre in Ethiopia.
Trials on evaluation of cultural and chemical control methods for the
management of CWD in Ethiopia, Uganda, Tanzania and DRC
Following biological and socio-economic surveys that were carried out
in Uganda, Ethiopia, Tanzania and DRC, some factors from the surveys
showed some positive correlation to CWD. It was therefore decided that the
country-specific factors be evaluated on farm and on station in these countries. However, some management options also came from literature, scientists personal experience and experience of farmers and extensionists. Management options that were agreed upon in participatory workshops attended
by scientists, farmers and extensionists are presented in Table 9.1.
Details for applying the treatments:
Herbicides (Roundup) applied at 150 ml Roundup in 15 l of water with

a sprayer. Applied when weeds appear (roughly applied twice a year).
Copper oxychloride spray mixed at the rate of 40 g per 15 l of water
and sprayed to the coffee stem only.
Copper oxychloride stem paint mixed at 300 g of copper oxychloride
in 1 l of water in a small bucket. A paint brush was used to apply the
fungicide paint from soil level up to 50 cm on the coffee stem.
Frequency:
Copper spraying was applied once a month during rainy
season and once every 3 months during dry season.
Copper stem paint was applied to the stem at the frequency
of once in 4 months.
Table 9.1. CWD management options agreed upon and evaluated in each country (a tick
[] under each country represents options that were tried in the country).
Agreed and tried options in each country
Management options
Mulch
Herbicide
Ash
Copper stem paint
Copper stem spray
Hoe weeding
Slash and hand weeding
Slashing only
Pruning
DRC
Ethiopia
Tanzania
Uganda
Fig. 9.2.
Mulch
Herbicide
Fungicide
spray
Fungicide
paint
25
15
10
0
Herbicide
Fungicide
spray
Fungicide
paint
Mulch
Ash
Coffee wilt disease incidence (%)
Slashing
only
20
Mulch
CWD incidence on C. arabica in Ethiopia.

Slash +
hand
weed.
Management option
Herbicide
Fungicide
spray
Fungicide
paint
Contol hoe
25
Contol hoe
Ash
Coffee wilt disease incidence

(%)
Fig. 9.1.
Ash
Coffee wilt disease incidence

(%)

145
25
20
15
10
Management option
20
15
10
Management option
CWD incidence on C. canephora in Tanzania (a) and Uganda (b).
146
Phiri et al.
NPK (nitrogenphosphoruspotassium) fertilizer was applied at the rate

of 250 g per tree twice per year during the rainy season.
Some promising results: It was apparent that all management options

were effective in managing CWD in Ethiopia (Fig. 9.1). In particular, the result
was particularly encouraging in that the addition of hand weeding around
the coffee stem as opposed to the normal practice of slashing in coffee fields
significantly reduced the incidence of CWD on C. arabica in Ethiopia (Fig. 9.1).
It was concluded that hand weeding prevents wounding, which is associated
with use of a bushmans knife in Ethiopia. Farmers have since adopted hand
slashing plus hand weeding, mulching and use of herbicide for the management of CWD in Ethiopia.
Studies on Coffea robusta in Tanzania and Uganda (Fig. 9.2a and b) did
not give as marked differences as those on C. arabica in Ethiopia. However,
copper oxychloride applied as stem paint and herbicide lowered the incidence of CWD. These studies were carried out over a 3-year period. Because
they were carried out in the field under natural inoculum, there is need to
carry them out for a much longer period, between 5 and 7 years.
9.3. Use of Resistant Varieties (Clones)

Use of resistant varieties is the most appropriate, efficient and economical
method for the management of CWD. It is as environmentally friendly as
most cultural control methods. When CWD first occurred in 1927, use of resistant varieties in combination with cultural methods reduced the impact
of the disease in the 1950s, and CWD became regarded as a minor disease
until its re-emergence in 1986 in DRC. Several authors have reported varietal
differences in resistance to CWD and suggested the use of resistant varieties as a means of controlling CWD (Fraselle, 1950; Deassus, 1954; Bouriquet,
1959; Porteres, 1959). Cultivars of C. canephora (notably robusta), which were
resistant, formed the basis of breeding programmes in many of the West African countries. Van der Graaff and Pieters (1978) reported that coffee lines of
C. arabica in Ethiopia showed differences in resistance to the CWD pathogen,
thus providing potential for controlling CWD using resistant varieties in
arabica coffee. They suggested that resistance in C. arabica was quantitative
in nature and horizontal, and there was no evidence of single-gene (vertical)
resistance that could be readily overcome by pathogen adaptation. During
the current studies under the RCWP, a series of screening trials were carried
out in participating countries, which aimed at finding resistant coffee varieties and clones. In many cases, seedlings and rooted cuttings were raised
from coffee bushes, which were apparently healthy in coffee fields where
nearly 100% of the coffee bushes had been destroyed by CWD. In addition,
seedlings and rooted cuttings were also obtained from existing coffee germplasm collections in these countries. Collections from the two sources were
screened by artificial inoculation in a spore suspension of the two strains of
G. xylarioides (from arabica and robusta coffees). Screening was also carried
147
out in fields by assessing for over 5 years. In Uganda, 8 resistant clones were
identified through artificial inoculation and field screening, which have been
released for growing by farmers but are yet to be commercialized. Further
resistant clones are also under multilocational field evaluation (for further
details, see Chapter 9).
In Tanzania, 851 lines were screened, and 273 robusta lines have been
identified to have resistance to CWD. The materials are currently being raised
in a clonal mother garden for further assessment. The materials from these
mother gardens will be assessed further for other agronomic traits through
on-farm and national trials.
Screening studies in DRC reported that 35 coffee lines were already collected and established in two areas with CWD.
9.4. Chemical Control

CWD is a systemic disease that infects vascular bundles; therefore, controlling it with chemicals is a very difficult option. The best way of controlling
CWD is by preventing infection from occurring or by using resistant varieties. However, resistant varieties are not yet available to every farmer in
the affected countries. In addition, susceptible varieties or clones will still be
available even after resistant varieties are available to farmers, thus warranting alternative strategies such as fungicides. The pathogen requires wounds
for infection, and because of its systemic nature, one method for managing
CWD is therefore to treat pruning wounds (either when changing a cycle or
for rejuvenation, or during normal pruning operations). Copper-based fungicides can be used for sealing the wounds after forming a paste from the
fungicide (Phiri, personal observation).
In addition to using fungicides for sealing wounds, studies on spraying
coffee stems or painting the first 50 cm of a coffee stem from the ground level
with a copper-based fungicide showed some promise in reducing the incidence
of CWD in on-station and on-farm trials in a number of participating countries
(see 9.2.5, Trials on Evaluation of Cultural and Chemical Control Methods for
the Management of CWD in Ethiopia, Uganda, Tanzania and DRC).
Earlier studies by Gaudy (1956) showed that spraying coffee bushes
with copper oxychloride was effective in controlling CWD. There is therefore potential in using copper-based fungicides, but the economics of using
a fungicide will be costly.
Control of CWD with systemic fungicides can be an alternative option,
but again, its economic viability is likely to be low, and the risk of contaminating the coffee beans with the fungicide should be investigated. There is
also the possibility of the pathogen adapting to systemic chemicals. Systemic
fungicides are already being used in the control of CLR, mostly by large-scale
coffee farmers. Even if systemic fungicides are found to be uneconomical to
use for controlling CWD or are found in the coffee bean, they may be the
only solution for saving valuable coffee germplasm materials that have useful agronomic traits but are susceptible to CWD. Many CWD-affected coun-
148
Phiri et al.
tries are losing their valuable germplasm. Uganda has lost coffee germplasm
due to CWD (Hakiza, personal communication). In addition, Uganda has
lost Coffea kapagota at the Entebbe Botanical Gardens due to this disease. An
effective systemic fungicide could therefore be used to protect these valuable
materials. Preliminary studies with benomyl under greenhouse conditions
in Uganda demonstrated that the fungicide may control CWD, but frequent
drenching every month was required (Hakiza, personal communication).
9.5. Biological Control

Biological control is defined as the use of a living organism to control or manage another living organism. Natural enemies include parasites, predators,
fungi, nematodes and viruses. Biological control agents (natural enemies)
are currently unavailable for G. xylarioides. However, earlier studies by
Rabechault (1954) reported four actinomycetes, one bacterium and Corticium,
Marasmius and Trichoderma spp., which had inhibitory effects to G. xylarioides.
Biological control could have great potential for the control of CWD as a
sealant for the wounds brought about by rejuvenation (stumping), and this
should be investigated further.
9.6. Integrated Crop and Pest Management

Integrated crop and pest management (ICPM) is an integrated approach to
crop health management. ICPM is ideal for management of CWD and involves use of as many management methods as possible to minimize problems caused by insects, plant diseases and crop management. Management
of CWD includes cultural management (prevention of injury and ensuring proper crop nutrition), management of insect pests such as stem borer
and management of CWD itself. The methods include cultural, mechanical,
physical, environmental, chemical and biological control. The components of
ICPM must be used in a systematic way; it must include a proper monitoring
programme for CWD and for insect pests and a proper identification of these
so that ideal components are chosen and combined to manage the disease
effectively. Use of information systems, in particular training of extensionists and farmers, and dissemination of information are crucial for the ICPM
approach to manage CWD.
The main objectives of an ICPM programme are to eliminate or reduce
the initial inoculum for the disease, to reduce the effectiveness of initial inoculum, to increase the resistance of the crop, to delay the onset of the disease
and to slow down the secondary cycles. ICPM involves the selection and
application of a harmonious range of disease-control strategies that minimize losses and maximize returns. ICPM of CWD can therefore include a
combination of some of the following methods: (i) use of resistant or tolerant
varieties; (ii) provision of balanced crop nutrition; (iii) use of healthy (clean)
planting material; (iv) effective quarantine; (v) management of infected coffee
149
plant materials; (vi) control of insect vectors or those that predispose the coffee tree; (vii) use of cultural practices, for example preventing tree wounding
during weed management; (viii) choosing a clean field to establish coffee
planting; (ix) suppression of the pathogen in infected fields; (x) prevention of
the spread of disease in the field and (xi) use of pesticides. It is this approach
which was advocated during the RCWP, and methods for disseminating
these messages are given in more detail in Chapter 11.
9.7. Governance Issues Related to CWD and Other Disease Epidemics

A detailed study was carried out by Quinlan et al. (2006) that included governance issues related to CWD and other epidemic diseases. Although coffee is the most important cash crop, a major source of household income
and is fundamental to livelihoods in more than 20 African countries, little
attention was paid by national governments to managing the problem of
CWD. It took too long from the time of initial detection, recognition of the
severity of the problem and conducting of surveys. Initially, this was due
to the challenge of conducting any surveillance in areas of conflict in DRC.
Yet, there was also a lack of preparedness and absence of an early warning
system to trigger responses in other neighbouring areas. Existing institutions
dedicated to coffee research supported CWD detection, identification and
management after the initial discovery of the disease. However, even after
the disease was identified, the lag time in responding suggests a lack of technical and institutional capacity and a lack of an adequate strategy to deal
with the problem. Inadequate budgetary support was also a big problem.
9.7.1. Preparedness before the outbreak of CWD

It is difficult to judge the preparedness of a country or countries for a disease
outbreak once it has occurred. Perhaps the best measure is whether relevant
institutions were in place before the outbreak occurred. Coffee research institutes were already established in all affected countries before the re-emergence of CWD. However, for a disease that spreads rapidly, the existence
of coffee research institutes alone was simply not sufficient. The ongoing
conflict in DRC at the time when CWD was identified in that country exacerbated the problem and is undoubtedly a major factor in the disease not
receiving the attention it deserved, especially in the initial foci of infection
where it could easily have been eradicated.
Initial investigations were also carried out in Uganda by a team of scientists from the then Coffee Research Centre (COREC) (now CORI) after they
received reports of a disease outbreak on coffee from the Bundibujyo and Rukungiri districts, which border DRC. CWD was not confirmed during these
initial investigations, and the outbreak was thought to be of relatively minor importance (Birikunzira and Lukwago, 1997) (see Chapter 3 for further
details).
150
Phiri et al.
In 1996, DRC through ICO requested funds to confirm the identity of

the disease outbreak before their coffee was completely destroyed. CABI,
together with scientists from DRC, plus officials from OZACAF, undertook a
preliminary survey mainly in Isiro and Beni. The survey confirmed the presence of CWD, and a formal report was produced (Flood, 1996). Although
there was a coffee research capability under the National Institute of Agricultural Studies and Research (INERA) prior to the outbreak of CWD, activities
concerning the management and control of CWD were not initiated until the
launch of the RCWP. This was largely due to the civil conflict that erupted
in DRC in 1996.
There were no early warning systems in place in DRC, and although
the first reports of CWD emerged in the 1970s, it was nearly 25 years before
OZACAF requested the assistance from ICO in confirming a possible outbreak of CWD (Chapter 2). In contrast, the response to the outbreak of CWD
in Uganda, once it had been officially confirmed, was rapid. The Ugandan
government through COREC funded preliminary surveys to establish the
occurrence of CWD, and other surveys were undertaken under the RCWP.
9.7.2. Initial detection of CWD resurgence

In Uganda, the CWD outbreak was first noticed by local agriculture personnel in the Bundibujyo and Rukugiri districts who reported it to COREC
through their district offices. Initial visits failed to confirm the identity of
the disease probably due to the high incidence of other pests and diseases.
This changed dramatically when the same symptoms were reported in the
Mukono district, 300 km away. The samples collected by COREC yielded
F. xylarioides (G. Hakiza, 1995, unpublished data). Other government institutes that played an active role when the outbreak occurred were the Uganda
Coffee Development Authority and the Faculty of Agriculture of Makerere
University (Chapter 3). Following the launch of RCWP, the aforementioned
institutes continued to collaborate in the management of and research aimed
at controlling CWD in Uganda. COREC led the Ugandan team.
In DRC, the confirmation of the re-emergence of CWD was carried out by
CABI (Flood, 1996) at the request of OZACAF (now Office National du Caf
[ONC]) through ICO. OZACAF prepared an initial report that outlined the
scale of the problem and submitted it to ICO with a view to getting external
funding to identify and manage the disease. ONC has been responsible for
coffee extension services in DRC since 1991, but these services have not been
particularly effective in rural areas in recent years due to problems associated with the deterioration of the infrastructure in the rural areas, political
instability and armed conflict (ICO, 2005). This to a great degree also affected
the monitoring and surveillance of CWD in the early stages of the outbreak.
Partners involved in CWD at a national level include the University of Kinshasa, the National Programme of Research on Coffee, under the INERA and
ONC, who are leading the programme in DRC. INERA and ONC are both
under the Ministry of Agriculture.
151
Based on the aforementioned experiences in Uganda and DRC, there are

a number of strategies that should now be put in place to avoid a recurrence
of CWD or any other invasive pest/disease:
1.
2.
3.
4.
5.
6.
7.
Legislation should provide for the destruction of infected coffee crops,

together with coffee crops that may not necessarily be infected but are in
the proximity of infected crops and could eventually become a secondary source of inoculum.
It is important that scientists and extensionists are trained in the identification of CWD even in the absence of an outbreak to reduce the possibility of incorrect diagnoses.
Raising awareness of the existence of CWD is of paramount importance
in containing an outbreak. Given the high percentage of rural communities, a farmer-participatory approach to raising awareness is essential.
Communities along the borders between affected and non-affected countries should be sensitised so that they do not cross with the concerned
coffee materials/products using unofficial routes.
Once an outbreak of CWD has been reported in a particular region, it
is essential that trained scientists undertake routine surveillance, especially in border areas, so that any outbreaks can be contained rapidly
and more cost-effectively using early warning systems.
Once an outbreak has occurred, the area must be cordoned off/quarantined immediately to prevent further spread. This could have prevented
the spread of CWD in Uganda where the initial disease outbreak was
on the boarder of DRC. Instead it was carried rapidly to the Mukono
district, 300 km from the first disease foci, and subsequently to all coffeegrowing districts.
All borders with affected countries should be closed for the movement
of any coffee materials/products to reduce the risk of contaminated coffee material entering the country. This may involve employing extra
quarantine personnel to man the borders.
9.8. Conclusion
CWD remains a huge threat to coffee production in Africa, and it is a potential threat to coffee production in the other coffee-growing continents, such
as South America and Asia. However, a number of methods that can help
in managing CWD are available. Notable ones are uprooting and burning
CWD-infected coffee trees; prevention of tree wounding; use of protective
fungicides in sealing wounds, particularly those from changing crop cycles
and pruning; quarantine (very important especially for preventing intercontinent spread of CWD); biological control and use of resistant varieties. Use
of systemic or curative fungicides has potential particularly in saving in situ
coffee germplasm from decimation by CWD. The use of systemic fungicides
on a commercial scale needs to be investigated further. Biological control
has a very high potential, especially as a sealant for the coffee stumps after
152
Phiri et al.
changing the coffee cycle. However, use of resistant varieties is probably the
most cost-effective method for the management of CWD and is relevant to
smallholder production. Resistant varieties and clones have been identified,
although a lot more work needs to be done before they become available
to all coffee farmers in the affected countries. It is therefore imperative that
government institutions in all coffee-growing countries, particularly in Africa, take breeding for resistance as a priority for the control of CWD, and
governments should provide the necessary financial and political support.
Quarantine is crucial at this stage when a number of countries are sharing
borders with CWD-affected countries. Monitoring and surveillance activities
ought to be maintained to curb a possible spread of CWD from the affected
countries. Dissemination of information and training of farmers are crucial
in the process of controlling CWD and any other pest epidemics of CWDs
magnitude and should continue as long as coffee is being grown. What has
happened in Central Africa should be taken as a lesson-learned exercise for
other coffee-growing states. Governments of CWD-affected countries were
not prepared for the resurgence of CWD and did not act fast enough to curb
further spread of CWD in their countries. It is a matter of time before CWD
spreads to other countries. The control of CWD therefore requires joint efforts from all coffee-growing countries, not only in Africa but also in the other
continents as well. They need to set aside funds for surveillance activities.
For example, Rwanda, although not affected directly, was supported to carry
out surveillance activities, and this would greatly help in limiting the spread
and impact if there should be a CWD outbreak in the country.
In conclusion, it is important that governments are willing to share the resistant coffee germplasm with those that do not have. This is the only way of
making CWD history as happened in the earlier decades of the 20th century.
Acknowledgements
This chapter is an output of Project 4 activities of the RCWP, which was
funded by the Common Fund for Commodities. The ICO supervised the
project, whereas CABI was the project execution agency. The authors sincerely thank all the institutions, scientists, extensionists and farmers who
participated in implementing the project activities from which most of the
information in this chapter have come.
References
Birikunzira, J.B. and Lukwago, G. (1997) The status of coffee wilt disease (tracheomycosis) in Uganda: a country report. In: Proceedings of the first regional workshop on
the coffee wilt disease (Tracheomycosis). Kampala, Uganda, pp. 98.
Bouriquet, G. (1959) Plant diseases and pests in some African territories. FAO Plant
Protection Bulletin 7, 6163.
CABI (2006) Crop Protection Compendium, 2006 edn.
153
Clowes, M.St.J. and Hill, R.H.K. (1981) Coffee Handbook, 2nd edn. Zimbabwe Coffee
Growers Association, Harare, Zimbabwe.
Deassus, E. (1954) La tracheomycosis de Cafeier. Bull. Sci. Minist. Colon. Sect. Agron.
Trop. 5, 345348.
FAO (2004) FAOSTAT data, 2004.
Flood, J. (1996) Observations on a coffee disease in North Eastern Zaire (JulyAugust
1996) Report to OZACAF August 1996.
Flood, J. and Brayford, D. (1997) Re-emergence of Fusarium wilt of coffee in East and
Central Africa. In: Proceedings of the 17th International Conference on Coffee
Science. Association Scientifique Internationale du Cafe (ASIC), Paris, France,
pp. 621628.
Fraselle, J. (1950) Observations preliminaires sur une tracheomycosis de Coffea robusta.
Bulletin Agricole du Congo Belge 41, 361372.
Gaudy, M.R. (1956) Contribution du techniques, scientifique ou dveloppement de
lagriculture en Afrique Occidentale Francaise. Journal of the West African Science
Association 2, 172197.
Girma, A. and Hindorf, H. (2001) Recent investigation on coffee tracheomycosis,
Gibberella xylarioides (Fusarium xylarioides) in Ethiopia. In: Proceedings of the
19th International Scientific Conference on Coffee Science (ASIC). Trieste, Italy,
pp. 12461252.
Girma, A., Hulluka, M. and Hindorf, H. (2001) Incidence of tracheomycosis, Gibberella xylarioides (Fusarium xylarioides), on arabica coffee in Ethiopia. Zeitschrift fr
Pflanzenkrankheiten und Pflanzenschutz 108, 136142.
ICO (2005) Exports by exporting countries to all destinations. Available at: http://www.
ico.org/trade/m1.htm.
Jacques-Felix, H. (1954) La carbunculariose. Bull. Sci. Minist. Colon. Sect. Agron. Trop.
5, 296344.
Krantz, J. (1962) Coffee diseases in Guinea. FAO Plant Protection Bulletin 10, 107
110.
in Ethiopia. Phytopathologische Zeitschrift 78, 365366.
Oduor, G. and Simons, S.A. (2003) Biological control in IPM for coffee. In: Neuenschwander, P., Borgemeister, C. and Langewald, J. (eds.) Biological Control in IPM
Systems in Africa. CAB International, Wallingford, UK.
Musoli, P., Lukwago, G., Abebe, M., Tesfaye, A., Kilambo, D., Asiimwe, T. and
Munyankere, P. (2003) Surveys to assess the extent and impact of Coffee Wilt Disease in East and Central Africa. Final Technical Report EU contract No. ASA-RSP/
OZACAF (1995) Project de lutte contre la tracheomycose du Cafeier au Zaire. Report for
Office Zairois du Cafe (OZACAF) on coffee wilt in Zaire.
Porteres, R. (1959) Valeur agronomique des Cafeiers des types Kouilou et Robusta cultivars en Cote dIvoire. Caf Cacao 3, 313.
Quinlan, M.M., Phiri, N., Zhang, F. and Wang, X. (2006) Foresight, Infectious Diseases:
Preparing for the future. The Influence of Culture and Governance on Detection, identification and Monitoring of Plant Disease, A comparative assessment
of the United Kingdom, China and Sub-Saharan Africa. Office of Science and
Innovation, London, United Kingdom. Available at: http://www.foresight.gov.uk/
Infectious%20Diseases/d4_1.pdf.
Rabechault, H. (1954) Sur quelques facteurs de resistance du Cafier a la Tracheomycose. Bull. Sci. Minist. Colon. Sect. Agron. Trop. 5, 292295.
154
Phiri et al.
Rice, R.A. and Ward, J.R. (1996) Coffee, Conservation and Commerce in the Western
Hemisphere. In: Smithsonian Migrator Bird Centre and Natural Resources Defence
Council, Washington, DC. p 4.
Saccas, A.M. (1956) Recherches experimentales sur la tracheomycose des cafeiers en
Oubangui-Chari. Agronomia Tropical Nogent, 11, 738.
Van der Graaff, N. and Pieters, R. (1978) Resistance levels in Coffea arabica to Gibberella xylarioides and distribution pattern of the disease. Netherlands Journal of
Wrigley, G. (1988) Coffee. Longman Press, London, UK.
10
Breeding for Resistance Against

Coffee Wilt Disease
P.C. Musoli,1 A. Girma,2 G.J. Hakiza,1 A. Kangire,1 F. Pinard,3

C. Agwanda4 and D. Bieysse3
1

e-mail: cori@africaonline.co.ug
2
3
34398, Montpellier, Cedex 5, France
4
e-mail: c.agwanda@cabi.org
10.1. Introduction
Coffee wilt disease (CWD) is very destructive to coffee trees, which leads to
enormous loss of investment in coffee farming. The disease was first reported
in 1927 on Coffea liberica var. dewevrei, formerly C. liberica type excelsa (Davis
et al., 2006), in the Central African Republic (CAR) (Figueres, 1940). It progressively destroyed this crop during the 1930s to the 1950s in other Central
African countries, particularly in Cameroon (Guillemat, 1946; Fraselle, 1950;
Saccas, 1951; Muller, 1997). During the same period, it destroyed Coffea canephora in the Ivory Coast, causing serious losses and disappearance of some
local varieties (Delassus, 1954). In contrast, several varieties of C. canephora
imported from the Democratic Republic of Congo (DRC) between 1914 and
1933 exhibited some level of field resistance, which was later confirmed
through artificial inoculation (Meiffren, 1961). Meiffren (1961) also reported
apparent differences for the same materials planted in different areas of the
region, i.e. certain C. liberica and C. canephora varieties showing resistance in
Ivory Coast were completely susceptible in CAR, suggesting the resistance
155
156
Musoli et al.
was either being influenced by environmental conditions or there were different physiological races of the pathogen in different localities of this region.
Fraselle (1950) reported CWD attack on C. canephora at Yangambi in DRC in
1948, and subsequently, the disease became a potential problem for the country. In 1957, similar symptoms were reported on Coffea arabica in Ethiopia
(Lejeune, 1958), and later, Kranz and Mogk (1973) confirmed that the disease
on C. arabica was also caused by Gibberella xylarioides. Pieters and Van der
Graaff (1980) reported that CWD was endemic in all coffee-growing areas
of Ethiopia. Van der Graaff and Pieters (1978) reported resistance among
C. arabica lines in fields, which was later confirmed after artificial inoculation
(Pieters and Van der Graaff, 1980).
As CWD threatened the coffee industry throughout Africa, affected
countries decided in 1956 to implement systematic elimination of all affected
plants over large areas and to search for resistance both in wild and cultivated
varieties. Following this initiative, C. canephora-resistant varieties identified
in DRC were used for replanting within DRC and Ivory Coast (Saccas, 1956;
Meiffren, 1961). In Cameroon, the disease was eliminated by rigorous systematic uprooting (Muller, 1997) of the Coffea excelsa var. dewevri plantations.
These strategies proved to be successful as the disease had declined drastically
by the end of the 1950s, and it eventually disappeared from Cameroon and
Ivory Coast, and probably DRC and CAR. However, the disease continued affecting C. arabica in Ethiopia, and it is doubtful if any of these strategies were
applied there. In 1986, new large-scale outbreaks of CWD were reported on
C. canephora in the north-east of DRC (Flood and Brayford, 1997), from where
it spread rapidly into Uganda (1993) and north-west Tanzania (1996). Because
the disease appeared in these countries for the first time, there were no resistant varieties available for replanting in infected areas and all available commercial varieties were susceptible to CWD, so replanting with these materials
in contaminated soils was not an option. In addition, there were no effective
phytosanitory management practices. Thus, following the successful use of
resistance in Ivory Coast and the CAR, in Uganda, a breeding programme was
initiated at the Coffee Research Centre (COREC) (now CORI) which aimed at
developing resistant germplasm for managing the disease. Similar breeding
programmes were initiated by TaCRI in Tanzania and the University of Kinshasa in DRC. The national breeding programmes in the respective countries
were implemented independently, although the programmes in DRC and
Uganda had a strong linkage through their collaboration with Centre de Coopration Internationale en Recherche Agronomique pour le Dveloppement
(CIRAD) in France.
10.2. Potential Sources of Resistance

Intra- and interspecific differences among and between coffee species respectively provide potential genetic variability, which is exploited for resistance
against CWD. Intraspecific variability is the best and easiest to exploit since
resistant individuals are easily released as new varieties without undergoing
Breeding for Resistance Against Coffee Wilt Disease
157
hybridization, provided they posses other agronomic traits such as being high
yielding; having resistance to other major diseases, mainly leaf rust and red
blister disease and coffee berry disease (arabica only) and having good market
qualities (big beans and good cup qualities). Where necessary, intraspecific hybridization is easier to carry out, and it is usually more successful.
C. canephora is particularly genetically variable (Plate 26), and the variability is very diverse even among genotypes from the same locality (Musoli,
2007) or members of the same progenies, mainly because it is out breeding
and its wide geographical distribution (Leakey, 1970). Musoli (2007) found
that most of the genetic diversity among C. canephora populations at molecular level was attributed to variations (heterozygosity) within individuals. Similar studies carried out in DRC revealed diverse variability among C.
canephora populations in that country. C. arabica is relatively less genetically
diverse, but nevertheless, the available diversity is high enough to be exploited for resistance against CWD.
Open pollinated seedlings of different C. canephora populations, which
included two cultivated distinct morphological types (erecta and nganda;
Plate 27) plus wild populations from Kibale and Itwara forests and a feral
population from Kalangala, an isolated island in Lake Victoria, were assessed
for CWD resistance through artificial inoculation in a screen house at COREC.
These studies showed significant (P = 0.001) genetic differences between the
populations for CWD resistance. The disease (measured by plant mortality)
progressed at different rates and to different final levels for the different populations (Fig. 10.1). This illustrates the usefulness of a diverse germplasm population when sourcing for resistance against CWD. Artificial inoculation of
different genotypes of C. canephora in DRC and Tanzania also showed varying
levels of resistance among the genotypes. Similar studies carried out on C. arabica in Ethiopia revealed differential reactions to CWD by genotypes from different localities (Chapter 8). In Uganda, exploitation of intraspecific variability
in C. canephora led to selection of eight high-performing, CWD-resistant clones,
70.0
60.0
% mortality
50.0
40.0
Kibale
Nganda
Erecta
Itwara
Kalangala
30.0
20.0
10.0
0.0
wk4
wk5
wk6
wk7
wk8
wk9
wk10
Time in weeks
Fig. 10.1. Progression of CWD among open pollinated progenies of different C. canephora
populations.
158
Musoli et al.
which also have good bean and cup qualities and resistance to red blister disease and leaf rust. These have been recommended for commercial cultivation.
In some parts of Uganda where C. canephora and C. arabica are cultivated
side by side, CWD affected only the former species. Similar responses have
been reported in Ethiopia, Tanzania and DRC. The difference observed between the two species in Uganda was confirmed through artificial inoculation
of seedlings of the two species in the screen house (Musoli et al., 2001). Interspecific hybrids (F1, backcross 1 and backcross 2; arabica as a recurrent parent) of C. canephora and C. arabica (arabusta) grown close to C. canephora fields
affected by CWD have remained unaffected since 2001, when the disease was
first observed on C. canephora. The resistance of the arabusta hybrids was also
confirmed through artificial inoculation of clones of their F1 progenies and F2
progenies at COREC. This implies that the resistance to CWD in C. arabica is
dominant, and it can be introgressed into C. canephora through interspecific
hybridization. However, deriving C. canephora-resistant varieties through
arabusta interspecific hybrids needs several backcrosses to the C. canephora recurrent parents This backcrossing is complex because the two species have different ploidy levels and the crosses normally involves C. canephora artificially
generated tetraploids.
Some level of host specificity to contemporary strains of G. xylarioides
was noted through hostpathogen interaction studies. The strain currently
affecting C. canephora in Uganda, DRC and Tanzania is specific to this species. Artificial cross inoculations carried out on young C. canephora and
C. liberica half sib progenies in growth rooms showed that G. xylarioides isolate (CAB003), which was obtained from C. canephora, causes severe mortality only on C. canephora (Fig. 10.2a) (Musoli, 2007). Similar studies showed
that the strain affecting C. arabica in Ethiopia is specific to this species
(Chapter 8). In mixed coffee cultures of C. canephora and C. arabica within
Ethiopia, the pathogen strains infect only C. arabica. Artificial inoculation of
C. canephora under controlled growth room conditions using isolate CAB007
collected from C. arabica in Ethiopia showed that this strain induces early
symptoms of CWD on C. canephora, but it is not fatal (Fig. 10.2b). The interaction of a historical strain, DSMZ62457, isolated in the previous epidemic
during the 1950s on C. canephora in CAR and evaluated through artificial
inoculations in the growth room gave mixed reactions. Although Musoli
(2007) showed that this strain was an aggressive pathogen of C. liberica and
moderate on C. canephora (Fig. 10.2b), other studies showed that C. liberica
isolate DSMZ62457 was pathogenic to C. canephora, C. liberica and C. arabica
and that other isolates from C. arabica and C. canephora induced some symptoms on C. liberica. Therefore, further studies are required to validate these
findings. Field reports from DRC and Uganda suggest that certain C. liberica
spp. are susceptible to the C. canephora isolate within these countries. These
observations thus exclude C. liberica from being a source of resistance to
CWD for introgression into arabica and robusta. Given such hostpathogen
specificity, varieties with durable resistance, which can be used across the
entire African region, irrespective of the prevailing pathogen strain, can be
derived through Arabusta interspecific hybridization.
% mortality
159
100
90
80
70
60
50
40
30
20
10
0
Cab003
DSMZ
C.
7
72
a-1
ric
libe
4
C.
72
a-1
ric
libe
0
C.
45
a-1
ric
libe
9
C.
60
a-1
ric
libe
8
C.
60
a-1
ric
libe
4
73
C.
a-1
ric
libe
ya
ma
Ne
ol-
ntr
Co
/6
P/3
/1
J/1
Cab003
Progenies
70.0
60.0
% mortality
50.0
40.0
30.0
CAB7
20.0
DSMZ
10.0
CAB3
0.0
C/6/1 J/1/1
B/1/1 B/2/1
Q/6/1 Q/1/1
B/6/2 H/4/1
G/3/7 P/3/6
Progenies
Fig. 10.2. (a) Mortality among C. canephora and C. liberica var. dewevrei seedlings inoculated
with G. xylarioides strains CAB003 (C. canephora) and DSM62457 (C. liberica). Nemaya is C. canephora progeny given for this purpose by CIRAD. J/1/1 and P/3/6 are C. canephora progenies obtained from Uganda. (b) Mortality of C. canephora seedlings inoculated with G. xylarioides strains
CAB003 (C. canephora), CAB007 (C. arabica) and DSMZ62457 (C. liberica).
10.3. Assessment of Resistance

Use of a quick and effective procedure is crucial to the successful identification of reliable resistance. Where breeding for resistance against CWD is
being initiated for the first time, it is necessary to have the basic skills and
knowledge, which can lead to successful screening of the germplasm and
160
Musoli et al.
identifying genuinely resistant genotypes. This will include having easy but
effective protocols for testing and quantifying the resistance. The screening
protocols involve assessing genotypes for resistance under natural and artificial infections. There are almost no special manipulations involved when assessing coffee plants for resistance in naturally infected gardens, and limited
variation in methodologies is anticipated, except perhaps in the quantification of resistance. There are more methodological variations when assessing
genotypes for CWD resistance in artificial conditions, perhaps due to variation of costs required by different methods and the facilities available for the
studies. Thus, a greater part of this section deals with methods involved in
artificial inoculation.
10.3.1. Inoculation methods

Different inoculation procedures have been used by different scientists for
artificially inoculating coffee plants when testing them for CWD resistance
(Meiffren, 1961; Van der Graaff and Pieters, 1978; Bieysse, 2005; Musoli
et al., 2001; Musoli, 2005). In Uganda, a number of artificial inoculation methods were evaluated to determine their efficacy when testing for CWD resistance in C. canephora (Hakiza et al., 2002). The inoculation methods evaluated
include
Root dip method: involved inoculation by dipping the entire root system
of the plants into a standard conidia inoculum of G. xylarioides;
Stem wounding: involved introducing a standard inoculum through artificial wounds on the stems of the study plants;
Soil infection: involved planting test plants in soil contaminated with the
CWD pathogen;
Soil drenching without root wounding: involved drenching the soils
where test plants are growing with a standard inoculum;
Soil drenching with root wounding: involved drenching the soils where
the test plants are growing with a standard inoculum but after wounding the roots while the plants remain in situ; and
Leaf wounding: involved applying the inoculum on to the leaves through
artificially made wounds.
All the inoculations were carried out using inoculum derived from a single conidium obtained from the same host plant and standardized at 1.3 106
conidia per millilitre of water. The inoculum was applied on 6- to 8-monthold seedlings or cuttings of known C. canephora genotypes (Plate 25). Plants
inoculated by root dipping developed the disease symptoms earlier than
plants inoculated by other methods and had a higher incidence of diseased
plants, and there was clear contrast between susceptible and resistant genotypes. This method was therefore adopted for large-scale germplasm screening by scientists in Uganda and Tanzania. However, root dipping required a
lot more inputs (polythene pots, soils and manpower) than the other methods
and therefore is more costly. It was also suspected that some of the plants
161
infected by this method could have developed the disease because of extra
stress resulting from root damage incurred when stripping off soils from roots
in preparation for dipping. At CIRAD, where labour is costly, stem nicking
was adopted. This method can also clearly enable differentiation of resistant
and susceptible genotypes. Stem nicking was also adopted for germplasm
screening in Ethiopia, as it is considered to be less expensive although the
disease levels among plants inoculated by root dip were always higher.
10.3.2. Inoculum concentration
At COREC, it was thought that a high inoculum concentration such as the
1.3 106 used by scientists during the past epidemics in Central and West
African countries for routine germplasm screening could kill plants with
moderate resistance, leading to discarding of useful material. Conversely,
lower inoculum concentrations may not be effective enough to select plants
that are resistant enough to the disease pressure in fields over long periods, as in the case where the variety is planted in heavily infested gardens.
To test the optimum concentration, 6- to 8-month-old C. canephora seedlings
were artificially inoculated by dipping their entire root system into inoculum (derived from a single conidium) at concentrations of 1.3 101, 1.3
102, 1.3 103, 1.3 104, 1.3 105, 1.3 106, 1.95 106 and 2.6 106 conidia per
millilitre of water. Results of these tests revealed that infection occurs even
at the lowest concentration (1.3 101), but this inoculum concentration was
associated with long incubation periods (for symptom development) and the
lowest disease incidence among the inoculated plants (Fig. 10.3). Increase in
100
90
% Wilt Incidence
80
70
60
50
40
30
20
10
0
40
50
1.3 x 101
1.3 x 105
60
70
Days after inoculation
1.3 x 102
1.3 x 106
1.3 x 103
1.95 x 106
80
90
1.3 x 104
2.6 x 106
Fig. 10.3. Effect of inoculum concentration on incidence of CWD on artificially inoculated open
pollinated seedlings of C. canephora.
162
Musoli et al.
inoculum concentration resulted in reduced incubation times and increased

wilt incidence. It was decided that concentrations of 1.3 103, 1.3 104 and
1.3 105 conidia per millilitre should be adopted for routinely screening for
resistance against CWD.
10.3.3. Duration of exposure to the inoculum for root dipping

To maximize the number of plants inoculated by root dipping, plants were
assessed after various times of exposure to the inoculum standardized at
1.3 106 conidia per millilitre of water. The exposure times tested included
immediate root dip (which involved withdrawal of the plants immediately
after their roots had been dipped into the inoculum), 20 min of exposure
(which involved withdrawal of plants after 20 min in the inoculum), 40 min
of exposure, 60 min of exposure, 4 h of exposure, 8 h of exposure, 12 h of
exposure and 24 h of exposure. C. canephora seedlings of 68 months were
used in these tests, and the tests were all conducted twice. All inoculated
seedlings developed CWD symptoms irrespective of duration of the exposure time. The differences in incidence of diseased plants, especially during
the first 40 days after inoculation, were insignificant. These results showed
no advantage in leaving plants in the inoculum for more than 20 min since
shorter exposure periods allow more plants to be inoculated within a given
period.
10.3.4. Quantifying CWD resistance in artificial inoculation

An effective and reliable method of quantifying resistance was necessary
for comparison of results among experiments and among scientists, and it
was necessary for the selection of genuinely resistant genotypes, irrespective
of whether the evaluation was carried out on mature plants in the field or
young plants in the screen house. The method adopted will depend on the
purpose of the study. Where the study aims at determining relative resistance between progenies or clones, resistance can be assessed as numbers of
infected plants compared to those uninfected. The numbers can be expressed
as percentage infection. Relative resistance between progenies and/or clones
can also be expressed on a disease symptom severity scale. In Uganda, the
plants studied in artificial inoculations were commonly assessed on a scale
of 1 to 5, where 1 = no disease, 2 = curling leaves and stunted growth, 3 = leaf
wilting and yellowing, 4 = leaf necrosis, leaf wilting, and abscission and 5 =
plants are dead. Mature plants studied in fields were also assessed on a scale
of 1 to 5, but the quantification of the disease levels in the field was slightly
different. In field assessment (mature coffee trees) 1 = no disease, 2 = 1%25%
defoliation, 3 = 26%50% defoliation, 4 = 51%75% defoliation and 5 = 76%
100% defoliation. Plants scored as level 5 are normally considered dead. The
studies carried out at COREC (Uganda) and CIRAD showed that coffee trees,
even of the same genotype and in heavily infected gardens, get affected at
163
Table 10.1. Correlation of field resistance of C. canephora clones to

CWD measured by different traits.
Disease severity
(15 scale)
% plant mortality
Disease severity
(15 scale)
Disease period
0.997 (<0.0001)
Disease period
Sqr. AUDPC
-0.446 (0.064)
-0.454 (0.059)
0.935 (<0.0001)
0.931 (<0.0001)
-0.281 (0.274)
Sqr. AUDPC is coefficient of correlations performed on the square root of

AUDPC. Figures in parentheses are probability values. AUDPC = area under disease progress curve.
different rates (Section 10.3.5), and that the time lapse between the first observed symptoms and the death of trees varies between genotypes. Thus,
there was an attempt to quantify relative resistance among C. canephora clones
using duration of the disease period (average time taken by plants from
appearance of first symptoms to death). However, this quantification did not
give consistent results because some moderately resistant clones had a shorter
disease period than some highly susceptible clones and vice versa, and its correlations with other traits were not significant (Table 10.1).
Where the aim was to identify genotypes with total resistance, plants
are classified as either diseased or not diseased. Since all plants that develop
CWD symptoms eventually die, only plants without the symptoms after a
long period of infection (6 months for plants in artificial inoculations and
not less than 5 years for plants evaluated in heavily infested fields) were
considered resistant. Where necessary, these plants can be re-inoculated and
assessed again for another 6 months to ascertain their resistance. The plants
that remained healthy after the re-inoculation were considered to have complete resistance, and such plants were planted in mother gardens for cloning
and further assessments.
10.3.5. Evaluation for CWD resistance in the field

Studies carried out on mature trees of 20 C. canephora clones in a field at
COREC revealed variable responses of the clones to CWD infection (Fig. 10.4).
The tree mortality for different clones was at varying levels when assessment started, and it progressed at varying rates to varying final levels. This
showed that the clones had variable levels of resistance and the disease resistance is most likely imparted by many genes, which are variably available in the clones. Clone J/1/1, which did not succumb to CWD throughout
the assessment period, and clones Q/3/4, R/1/4 and 1s/3, whose disease
levels were comparatively very low, were considered resistant. Clone 1s/2
appeared to be resistant until April 2002, when its trees started dying in high
numbers. This shows that some genotypes require higher inoculum concen-
Musoli et al.
% mortality
164
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
Highly
susceptible
Moderately
susceptible
Moderately
resistant
Resistant
% mortality
Time in months
H/4/1
/1/7
C
Q/1/1
257/53
/6/1
C
P/5/1
1S/3
Q/3/4
R/1/4
B/2/1
Highly
susceptible
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
Moderately
susceptible
Moderately
resistant
Resistant
Time in months
E/3/2
223/32
1S/2
B/1/1
J/1/1
G/3/7
Q/6/1
P/3/6
L/2/7
B/6/2
Fig. 10.4. Tree mortality of 20 C. canephora clones evaluated for CWD resistance in a field at
Kituza.
trations for infection, and thus, there is need to assess the plants for many
years, to allow the inoculum build up, when screening them for resistance
in the field. Analysis of variance revealed significant genetic differences between C. canephora clones for resistance to CWD in the field.
165
10.3.6. Comparison of resistance in the field and in artificial inoculations

Rooted cuttings and open pollinated seedlings of some of the clones studied
for field resistance were studied for resistance in artificial inoculations in a
screen house at COREC, and their final per cent tree mortality were correlated
with the final per cent tree mortality in the field. It was observed that even
clone J/1/1, which had no diseased trees in the field, had 15% of its cuttings
killed in artificial inoculation, but it was among the least affected (Table 10.2).
Clone B/2/1 had 54.2% mortality in the field but did not have any dead
cuttings after artificial inoculation. This clone could be having differential
reactions to infection under the different conditions. However, there was an
overall significant correlation (P = 0.002) between mortality in the field and
mortality of cuttings in the screen house (Table 10.3). This indicates that resistance detected in artificial inoculation is depicted in the field, and therefore, the artificial inoculation is a good protocol for screening germplasm
for CWD resistance. Where correlations between mortality among open pollinated progenies in the screen house and mortality of their parents in the
field and in the artificial inoculations (rooted cuttings) were not significant,
Table 10.2. Mortality of C. canephora clones in the field and

their rooted cuttings and progenies in artificial inoculation.
Clone
Field mortality
J/1/1
Q/3/4
1S/3
R/1/4
C/6/1
Q/6/1
B/2/1
223/32
L/2/7
Q/1/1
B/1/1
257/53
G/3/7
P/5/1
E/3/2
1S/2
P/3/6
B/6/2
H/4/1
C/1/7
0.0a
4.2b
33.3c
33.3c
50.0cd
50.0cd
54.2cd
58.3cde
62.5def
66.7defg
75.0defgh
83.3efgh
83.33efgh
87.5fgh
87.5fgh
87.5fgh
91.7gh
91.7gh
94.4gh
95.8h
Rooted cuttings
Open pollinated
progenies
15.0b
20.0b
44.4bc
0.0a
90.0d
50.0bcd
25.0b
80.0cd
100e
70.0cd
63.6cd
35.0abcd
35.0abcd
65.0def
80.0fg
10.0a
65.0def
25.0abc
53.0cde
85.0fg
40.0bcde
50.0cde
25.0abc
20.0ab
69.0efg
68.0efg
60.0def
95.0g
Means separated by Student NewmanKeuls mean separation

test. Figures represented by different letters are significantly
variable.
166
Musoli et al.
Table 10.3. Correlations of C. canephora resistance to CWD in
field and artificial inoculations.
Clone in field
Cuttings
Progeny
Progeny (E41)
0.965 (0.002)
6
0.132 (0.698)
11
0.791 (0.209)
4
0.708 (0.010)
12
0.879 (0.121)
4
0.099 (0.798)
9
Rooted cuttings
Open pollinated
progeny
Figures in parentheses are probability values; E41 and E42 are

tests 1 and 2 within experiment 4, respectively. The number below
each correlation coefficient shows data pairs used to generate the
coefficient.
then the response of open pollinated seedlings does not effectively represent
CWD resistance of their parents.
10.4. Inheritance of CWD Resistance

Understanding the type of resistance and inheritance of the resistance was
considered necessary for breeding CWD-resistant varieties, which posses
other commercial attributes. Analysis of variance performed on the disease
symptom severity data of clones in the field, rooted cuttings and open pollinated progenies in artificial inoculation found highly significant (P < 0.0001)
genetic differences between the clones, rooted cuttings and open pollinated
progenies (Table 10.2). These results and the responses of the different C. canephora populations (Fig. 10.1) indicated that CWD resistance in C. canephora
is quantitative; thus, it is controlled by many genes that are variably distributed among the genotypes and populations.
Inheritance of CWD resistance in C. canephora was calculated at 50%
65% plant mortality using disease data recorded on mature trees of full sib
progenies. The data were analysed using diogene quantitative genetics software (Baradat and Labb, 1995) using Garretsen and Keuls random model
adapted for incomplete/half diallel (Keuls and Garretsen, 1977). This model
was able to calculate variance components and narrow (h2n) and broad (h2b)
sense heritabilities. The broad sense heritability calculated from this analysis
was moderate (0.329), and the narrow sense heritability was low (0.112). This
shows that CWD resistance is heritable, but its transmission from parents
to progenies is only about 33%; therefore, the genetic gains of choosing a
progeny of resistant parents as source of planting materials for a production
garden are low. Commercial CWD-resistant robusta coffee varieties should
therefore be propagated vegetatively to retain the resistance. It should however be noted that resistant parents of the progenies used in this study were
partially resistant, and therefore, analyses of data on progenies derived from
167
crosses of completely resistant and highly susceptibility parents are needed

to validate the current observations.
10.5. A Perspective for Developing CWD-Resistant Varieties

Basing on the information above, a practical perspective for successfully developing CWD-resistant varieties is given below. The perspective outlined
is for C. canephora, and it might not be suitable for C. arabica because the two
species have different reproductive systems and the interaction of C. arabica
with the G. xylarioides strain is different. We have not attempted to give a
possible breeding strategy for developing CWD-resistant C. arabica commercial varieties because we lack basic information on type of resistance and
heritability of the resistance in this species. A schematic representation of the
breeding outline is given in Fig. 10.5.
10.5.1. En masse germplasm screening

As already indicated, there is variability among C. canephora genotypes and
populations for CWD resistance. Thus, en masse screening of seedlings
and clones derived from different populations through artificial inoculation leads to identification of CWD-resistant genotypes quickly. In Uganda,
1519 CWD-resistant genotypes have so far been identified through en masse
screening of coffee seedlings and cuttings derived from germplasm available
in conservation plots at the research institutes and from coffee trees surviving in wilt-devastated gardens. The conservation plots at Kawanda Agricultural Research Institute were the immediate and most available source
and therefore provided the biggest amount of seedlings and cuttings so far
screened. Open pollinated seedlings and rooted cuttings were raised from
various genotypes in these collections at different dates following routine
nursery procedures of raising coffee seedlings and rooted cuttings (MAAIF,
1995). The seedlings and cuttings, when at 68 months old, were inoculated
with field isolates of the pathogen prepared from a specimen obtained from
a commercial C. canephora clone 257/53. The plants were inoculated using
the root dip method and a standardized inoculum (1.3 106 conidia per
millilitre). All inoculated plants were incubated at room conditions in the
screen house as they were monitored for CWD symptoms (Plate 25). At the
end of each trial, any healthy-looking plants were re-inoculated 6 months
after the first inoculation. These re-inoculated plants were again incubated
in the screen house and monitored for the disease symptoms. Again after
another 6 months, healthy-looking plants were selected for another inoculation. Survivors of these re-inoculations were considered resistant and
potential clones for future varieties. All the resistant genotypes were then
planted in mother gardens and each of them cloned through cuttings to raise
clones for establishing field evaluation trials (preliminary variety evaluation
trials).
168
Musoli et al.
Screening local
germplasm (2)
Screening exotic
germplasm (4)
Vegetative
propagation (3)
On-station field
assessments and
selection (6)
Vegetative propagation (2)
On-farm adaptation & adoption

evaluation & selection (6)
Hybridization (1)
1. Local R x CC
2. Local R x R
3. Local R x exotic R
4. CC x exotic
On-station field assessment

and selection (6)
Clonal varieties
Vegetative propagation (3)
Vegetative propagation of
mother stock for farmers (2)
On-station polycross seed

gardens (3)
Seedling varieties
Tissue culture +
nodal cuttings
On-farm progeny evaluation &
selection (6)
Mass vegetative propagation
Mass seed propagation
Fig. 10.5. Proposed scheme for developing C. canephora varieties resistant to CWD. Figures in
parentheses are the minimum time in years a particular breeding stage could take. R = resistant,
CC = current commercial varieties.
169
Open pollinated seeds and semihard wood cuttings were obtained from
coffee trees surviving in wilt-devastated gardens. Seedlings and rooted cuttings were, respectively, raised from these materials for inoculation and selection of resistant genotypes following the procedure described above.
It was observed that a very small proportion (3%15%) of the plants
raised from the on-station germplasm plots survived the first inoculation. The
rate of survival varied between progenies (Fig. 10.2), perhaps due to relative
resistance of their parents. However, the proportion of survivors after the first
inoculation from among seedlings and cuttings raised from genotypes surviving in wilt-devastated gardens was relatively higher (15%20%). This implies
that mother plants surviving in wilt-devastated gardens have some degree of
resistance, although this resistance might not be strong enough to withstand
the inoculum pressure given in the screen house. It has also been observed
that in the re-inoculation, about 50%70% of the re-inoculated plants die.
In Tanzania, more than 270 resistant genotypes have been identified
through mass screening of germplasm. The resistant genotypes have been
planted in mother gardens for vegetative multiplication.
10.5.2. Field germplasm screening

Germplasm screening can also be carried out in field observation plots.
However, field evaluations tend to handle relatively fewer plants, and the
assessment is prolonged over many years, yet one cannot ascertain that resistant genotypes selected through field observations can withstand high
disease pressures under different agroecological conditions unless the trials
are validated under a range of agroecological conditions. At COREC, clones
in a field trial were assessed for CWD resistance starting 1999 to 2006. It was
noted from this field trial that clones respond to the disease differently and
one clone (J/1/1) was still resistant by November 2007 (Table 10.2). However, rooted cuttings raised from the resistant clone, when tested for resistance in the screen house, succumbed to the disease although the level of
disease among its cuttings was lower than that among cuttings of the other
clones within this trial.
10.5.3. Multiplication of the CWD-resistant varieties

Because C. canephora is predominantly an out-crossing crop, it is anticipated
that all the individuals are heterozygous, and therefore, the CWD-resistant
genotypes are expected to segregate for most traits including the resistance
to CWD. It is therefore recommended that these genotypes are propagated
entirely using vegetative methods, particularly rooted nodal cuttings or tissue culture. At the COREC, the former is being used for propagating materials earmarked for both further evaluation and farmers use whereas the latter
exclusively dedicated to raising planting materials for farmers use from somatic embryogenesis.
170
Musoli et al.
10.5.4. Field evaluation of the resistant varieties

It is obvious that not all the CWD-resistant genotypes have all desired
qualities of commercial varieties. Therefore, the CWD-resistant genotypes
must undergo preliminary field evaluation, where they are evaluated for
yield, resistance to other diseases (coffee leaf rust [CLR] and red blister
disease in Uganda) and liquor and physical bean qualities. Because this
initial evaluation involve many genotypes, it is ideal to first have a simple
trial where all the available genotypes are assessed and genotypes with
higher potential are selected for further evaluation in different agroecological localities. As of November 2007, 1519 CWD-resistant clones had been
identified in Uganda, and these have been planted in single rows of six
trees in an on-station trial at Kituza. To date, 25 superior clones have been
selected from among 167 clones that have undergone the initial on-station
evaluation. The 25 clones are being multiplied through nodal cuttings for
further evaluation in different agroecological areas so as to validate their
performance and adaptation to different localities. Good yielding (at least
2 t of green beans) clones that are resistant to CLR and red blister disease
and have good liquor and physical bean qualities will be recommended
for farmers use, either across the entire country or for particular agroecological localities, depending on the performance of individual clones in the
multilocation trials.
10.5.5. Hybridization for CWD resistance
Although CWD takes precedence at the screening stage, yield, quality and
resistance to CLR, red blister disease and other major coffee diseases are important traits for commercial C. canephora varieties, particularly in Uganda.
CWD-resistant clones that might not be high yielding and have low potential
for cup and bean qualities and resistance to other diseases can be crossed
with resistant genotypes and other lines that posses these complementary
traits. The hybrid progenies can then be evaluated for all the traits starting
either at the screen house through field trial testing or in field trials only. The
former approach ensures that genotypes that proceed to the field are all resistant to CWD. The latter approach is less efficient for CWD resistance since
it relies on field responses only, but it retains more genes especially those in
moderately resistant genotypes to be tested that could have been eliminated
in the more exacting screen house tests.
The current commercial clones (1s/2, 1s/3, 1s/6, 223/32, 257/53 and
258/24) used in Uganda have the complementary traits, but they are susceptible to CWD. The commercial clones yield 2.53 t of green coffee beans per
hectare per year, have good bean qualities (1822 g hundred beans weight,
over 90% retained by screen 18/64) and have good cup qualities. Hybrids
have been generated through artificial pollinations to combine the CWD resistance and the complementary traits found in these commercial clones. The
progenies generated have been planted in field trials at Kituza for evaluation
171
as individual trees. The individual genotypes that will be within acceptable

limits of the traits will be selected, multiplied as clones and planted in multilocation trials in different agroecological localities for re-assessment. Clones
that shall perform well in these trials will be selected for release to farmers
as clonal varieties either at agroecological zone level or for the entire country
depending on the performance of the variety.
Hybridization can also be between the CWD-resistant clones possessing
complementary traits. The hybrid progenies generated in such crosses are
also evaluated as individual trees for resistance against CWD and for field
performance in the other traits. Good-performing individuals can be selected,
cloned and planted in multilocation trials for adaptation and adoption tests.
Good-performing clones will then be selected and released to farmers.
If there is an entire progeny of a cross between resistant and susceptible
parents that perform well (resistant to CWD, CLR and red blister disease;
have good qualities and are high yielding), then parents of such progenies
can be planted in polycross seed gardens for production of seeds to be given
out to farmers. It is anticipated that progenies involving parents from different populations, particularly from distant geographical locations, shall benefit from hybrid vigour derived from double heterozygosity of the parents. As
already indicated above, given the specificity of the pathogen populations
affecting the different commercial coffee species, resistant varieties can be
derived through interspecific hybridizations, bearing in mind the complications associated with such hybridizations and the difficulty to derive a variety of desired quality.
As known from many breeding programmes, incorporation of CWD resistance genes into commercial clones that should also have good quality
traits is likely to take a very long time, and at times, it is a gamble. Therefore,
molecular techniques can be adopted to facilitate the breeding and selection
process. Studies had been initiated at CORI to characterize CWD resistance
using molecular markers, and it is our hope that a follow-up of these studies
shall continue, which should lead to mapping of the resistance genes. Mapping studies could be initiated using the double haploid CWD-susceptible/
resistant parents and their progenies to identify molecular markers and
or quantitative trait loci (QTL) associated with CWD resistance to assist in
breeding resistant varieties and isolation of resistance genes for creating bacterial artificial chromosome libraries. Use of double haploids shall minimize
the effect of heterozygosity.
10.5.6. Grafting
Because currently the mechanism of resistance against the CWD pathogen
is not known, we are not certain whether a resistant rootstock can prevent
the pathogen from reaching the scion. However, if it is established that
this phenomenon is possible, then grafting scions of varieties with good
agronomic characteristics but which are susceptible to CWD on to a CWDresistant rootstock should be a good and probably quicker means of deriving
172
Musoli et al.
appropriate planting materials for farmers of both C. canephora and C. arabica.

In Uganda, grafting of current commercial clones on to the CWD-resistant
clones is being evaluated for (i) compatibility between the scion and rootstock and (ii) the ability of the rootstock to prevent the CWD from getting
through the vascular system of the rootstock to the scion. If these results are
satisfactory, then grafting should be adopted for continuing with large-scale
multiplication of the current commercial varieties.
Rootstock of other Coffea species can also be explored for this purpose. Successful interspecific grafting involving C. canephora and C. liberica spp. has been
reported (Couturon, 1993). However, Bertrand et al. (2001) reported depressing effects of C. liberica ssp. rootstock on yield and quality of C. arabica scion
varieties. Therefore, other agronomic properties of the grafted varieties such as
yield and quality should be studied and well understood before recommending grafting for producing planting materials for farmers on a large scale.
10.6. Conservation of CWD-Threatened Coffee Genetic Resources

Currently, the C. canephora and C. arabica gene pools in some Eastern and
Central African countries are being depleted at varying rates by CWD. It is
therefore necessary to protect the genetic resources of these species, possibly
in their current diversity state. Because genetic differentiation among C. canephora populations is high, samples of a few genotypes with diverse genetic
variability can be obtained from many populations so as to capture as much
of the natural genetic variability as possible for conservation and utilization
in future breeding programs. In Uganda, the C. canephora samples can be collected from different sites in their natural habitats such as Itwara, Kibale forests and other relic forests where wild Coffea exists. Samples should also be
collected from isolated coffee localities such as Kalangala Islands in Lake Victoria. Because of the threat from CWD and other unforeseen natural disasters
on coffee genetic resources, new locations should be identified for local ex situ
conservation. In addition, representative samples of the genetic diversity of
these species should be conserved in multiple international germplasm collections located in different countries through international collaboration.
Coffee genetic resources are threatened further by the likely adoption,
on a large scale, of the incoming CWD-resistant varieties, which are likely
to be derived from a few genotypes that represent a very narrow genetic
base. This implies that even the few coffee plants that have withstood CWD
in farmers fields shall be abandoned and most likely destroyed as the fields
are planted with the new CWD-resistant varieties. Initiation and implementation of in situ conservation and local germplasm utilization programmes,
at regional levels within affected countries, for the remnant diversity found
in devastated gardens would minimize depletion of gene pools. Such programmes are expected to sensitize the local populations and ensure that
farmers multiply, conserve and beneficially utilize the plants surviving in
their fields. Through such programmes, diversity at regional levels within
countries can be conserved.
173
10.7. Conclusions
CWD has been and remains an important constraint of coffee production in
Africa, and if not controlled, it could spread to affect coffee growing in other
continents. Through concerted efforts, the previous epidemics in some Central and West African countries were controlled by developing and planting
resistant varieties. Variety resistance is so far considered the most appropriate
and effective method for controlling CWD. There has been enormous effort
to develop more resistant varieties to control CWD in Uganda and Tanzania.
Through these works, possible sources of intra- and interspecific resistance
have been identified. In both Uganda and Tanzania, a number of resistant C. canephora genotypes have been identified, and Uganda in particular has already
CWD clones fronted for release to farmers. Availability of these varieties will
certainly help to revitalize the coffee production, not only in Uganda but also
in neighbouring countries through regional cooperation, especially DRC and
Tanzania. The authors hope that the highlights of challenges encountered by
the breeding programmes since the re-emergence of CWD and the potential
and actual procedures taken while trying to develop resistant varieties, given
here, shall stand to be quick reference for future work when the same or similar outbreaks occur again, whether in Africa or elsewhere. Some challenges
remain: super CWD-resistant varieties have to be checked for complementary
traits and adaptation for cultivation in a wide range of geographical locations,
and another challenge is the continuing depletion of the gene pools caused by
CWD, whether directly or indirectly. Uganda is at the centre of diversity for C.
canephora, so any threat to its diversity is a threat not only to Ugandan coffee
production but also to the coffee sector as a whole.
Acknowledgements
We are grateful to the governments of Uganda, DRC, Tanzania and Ethiopia and our national institutes, where the studies that provided information
given in this chapter were conducted, for all the support provided while
these studies were being carried out. We also thank the donor communities, particularly the European Union, Common Fund for Commodities and
United States Agency for International Development, for the financial support given for the studies that led to the generation of the information given
in here. Last but not least, we acknowledge CAB International Africa for facilitating the writing up of this work.
References
Baradat, P. and Labb, T. (1995) OPEP: un logiciel intgr pour lamlioration des plantes prennes. In: Traitements statistiques des essais de slection. CIRAD, Montpellier, France, pp. 303330.
Bertrand, B., Etienne, H. and Eskes A. (2001) Growth, production and bean quality
of Coffea arabica as affected by interspecific grafting: consequences for rootstock
174
Musoli et al.
breeding. In: Proceedings of American Society for Horticulture Annual Conference
No. 96. Minneapolis, MN, pp. 206238.
Bieysse, D. (2005) Development of long term strategy based on genetic resistance and
agro ecological approaches against coffee wilt disease in Africa. EU-INCO project.
Third annual report.
Couturon, E. (1993) Mise en vidence de diffrents niveaux daffinit de greffes interspcifiques chez les cafiers. In: 15me colloque de l Association Scientifique
Internationale du Caf. Montpellier, France, pp. 209217.
Davis, A.P., Govaerts, R., Bridson, D.M. and Stoffelen P. (2006) An annotated taxonomic conspectus of the genus Coffea (Rubiaceae). Botanical Journal of the Linnaean
Society 152, 465512.
Delassus, E. (1954) La trachomycose du cafier. Bulletin Scientifique du Ministre des
Colonies, Section Agronomie Tropicale 5, 345348.
Figueres, R. (1940) Sur une maladie trs grave du cafier en Oubangui. Rapport. Ministre des colonies, Paris, France.
Flood, J. and Brayford, D. (1997) The re-emergence of Fusarium wilt of coffee in Africa.
In: ASIC 1997 Proceedings of the International 17th Scientific Colloquium on Coffee Conference. Nairobi, Kenya, pp. 621627.
Fraselle, J. (1950) Observations prliminaire sur une trachomycose de Coffea robusta.
Bulletin Agricicole, Congo Belge XLI, 361372.
Guillemat, J. (1946) Quelques observations sur la trachomycose du Coffea excelsa.
Revue de Botanique Applique and dAgriculture Tropicale 26, 542550.
Hakiza, G. et al. (2002) CORI Annual Report 2001/2.
Keuls, M. and Garretsen, F. (1977) A general method for the analysis of genetic variation
in complete and incomplete diallels and North Carolina II designs. Part I: procedures
and formulas from the random model. Euphytica 26, 537551.
in Ethiopia. Phytopathology 78, 365366.
Leakey, C.L.A. (1970) The improvement of robusta coffee in East Africa. In: Leakey, C.L. (ed.)
Crop Improvement in East Africa. Government Press, Kampala, Uganda, pp. 250277.
Lejeune, J.B.H. (1958) Rapport au Gouvernement Imperial dEthiopie sur la production
cafire. Rapport du la FAO, Rome FAO158/3/1881.
MAAIF (Ministry of Agriculture, Animal Industry and Fisheries) (1995) Clonal Robusta
Coffee Handbook. Part 1: Nursery Management Practices. Communication Centre,
Entebbe, Uganda.
Meiffren, M. (1961) Contribution aux recherches sur la trachomycose du Cafier en
Cote dIvoire. Caf, Cacao, Th 5, 2837.
Muller, R.A. (1997) Some aspects of past studies conducted in Western and Central
Francophone Africa on tracheomycosis (Cote dIvoire, Cameroon and Central African Republic). In: Hakiza, J.G., Birikunzira, B. and Musoli, P. (eds.) Proceedings of
the first regional workshop on coffee wilt disease (Tracheomycosis). International
Conference Centre, Kampala, Uganda, pp. 1526.
Musoli, P.C. (2005) Breeding for resistance against coffee wilt disease. In: Bieysse, D. (ed.)
Development of long-term strategy based on genetic resistance and agro ecological
approaches against coffee wilt disease in Africa. Fourth Annual report, pp. 4872.
Musoli, P.C. (2007) Recherche de sources de rsistance la trachomycose du cafier
Coffea canephora Pierre, due Fusarium xylarioides Steyaert en Ouganda. PhD
thesis. Montpellier II University, Montpellier, France.
Musoli, P., Olal, S., Nabaggala, A. and Kabole, C. (2001) Screening robusta coffee
germplasm for resistance against coffee wilt disease. CORI progress report on coffee wilt disease research and development 19972000.
175
Pieters, R. and Van der Graaff, N.A. (1980) Resistance to Gibberella xylarioides in
Coffea arabica: evaluation of screening methods and evidence for the horizontal
nature of the resistance. Netherlands Journal of Plant Pathology 86, 3743.
Saccas, A.M. (1951) La trachomycose (Carbuncularise) des Coffea excelsa, neoarnoldiana et robust en Oubangui-Chari. Agronomie Tropicale 6, 453506.
Saccas, A.M. (1956) Recherches exprimentales sur la trachomycose des cafires en
Oubangui-Chari. Agronomie Tropicale 11, 758.
Van der Graaff, N.A. and Pieters, R. (1978) Resistance levels in Coffea arabica to Gibberella xylarioides and distribution patters of the disease. Netherlands Journal of
11
Extension Approaches and Information

Dissemination for Coffee Wilt Disease
Management in Africa: Experiences
From Ethiopia
E. Negussie,1 M. Kimani,1 A. Girma, 2

N. Phiri1 and D. Teshome2

2
11.1. Introduction
Coffee plays a vital role in the economies of many African countries. However, the performance of the sector has been depressed by a number of constraining factors. In particular, coffee wilt disease (CWD) has become an
issue of pressing concern in Central and Eastern African countries, including Ethiopia. Although the disease has expanded rapidly in recent years, the
farming communities hold different perception and distorted views of the
disease. The socio-economic baseline survey results revealed that the majority of the farmers in Central and Eastern African countries lack awareness
of CWD (Chapter 6). It was reported that in Ethiopia, only 17% of the coffee
farmers knew CWD as a disease but they did not know its specific name.
Farmers were also not sure of the causes and ways of transmission of CWD.
It was therefore realized by the Regional Coffee Wilt Programme (RCWP)
that both awareness and technical knowledge on how to manage CWD were
needed for farmers and other actors in the coffee sector. Consequently, awareness raising, training and dissemination of information were accorded top
priority in RCWP.
There are vast numbers of Ethiopian smallholder coffee producers that
operate under diverse socio-cultural, economic and natural conditions. At the
same time, there is a huge disparity in terms of literacy and access to infor-
176
Extension Approaches and Information Dissemination for Coffee Wilt Disease
177
mation and communication infrastructure. In addition to producers, there are

several actors that are directly or indirectly engaged in and influence the coffee sector. Thus, it was well recognized that relying on a few communication
channels and use of the traditional top-down and one-way approach would not
effectively reach all the target groups and bring positive changes. Therefore,
emphasis was placed on combining a range of complementary communication approaches and methods. These were believed to facilitate wider information dissemination and impact because most channels reinforce each other
and have synergetic influences. In addition, various channels have differential
impacts on different stakeholders within the coffee sector. Particularly, in view
of the absence of a comprehensive package of solutions for the CWD problem,
the project emphasized joint learning processes and sharing of information and
experiences among relevant stakeholders. To attain this, active participation of
all relevant parties was imperative. Therefore, participatory and interactive
approaches were employed with farmers, extensionists and researchers who
became partners and actors in the communication and learning processes.
This chapter describes the information communication strategies and
approaches used by RCWP for coffee wilt management in Ethiopia.
11.2. Communication Strategies and Channels Employed

Extension education involves the conscious use of communication of information to help people learn, form sound opinions, make appropriate
decisions and finally take necessary actions. Agricultural extension tries to
change farmers behaviour and practice through education and communication. More than natural resources, more than cheap labour, more than financial capital, knowledge is becoming the key factor of production (World
Bank, 1992). Therefore, choosing appropriate communication strategies
and methods is crucial to facilitate information and knowledge sharing for
change. The effectiveness of different channels varies according to the kind
of information to be transferred, the target user to be addressed, the level
of understanding of the receivers and other socio-economic factors. Studies
show that widespread responses are seen when people are exposed to information by several methods and approaches. The communication strategy of
RCWP employed both interpersonal and mass communication approaches.
11.2.1. Interpersonal communication approaches

Interpersonal communication channels involve a face-to-face exchange of information and interaction between two or more individuals. These channels
promote a two-way exchange of information, facilitate immediate feedback,
help to secure clarification and persuade individuals to adopt a new idea
or technology. In particular, the formation and change of strongly held attitudes are best accomplished by interpersonal channels (Rogers, 1983). These
channels have greater effectiveness in dealing with resistance or apathy on
178
Negussie et al.
the part of the target groups. The principal interpersonal communication

methods used by the project are described below.
Participatory hands-on training
Training sessions were the principal instruments used to create awareness
and build knowledge and skills with regard to CWD management. Several
training programmes were conducted for extension workers, farmers and
other stakeholders. In view of the drawbacks of the traditional training approaches and the lack of complete solutions to manage CWD at present, the
project employed a participatory training approach, which facilitated twoway communication and knowledge and experience sharing (Plate 33a). In
contrast to the conventional training approach, which considers trainees as
an empty bottle to be filled up by the trainer, the participatory approach
used by the project encouraged participants to see themselves as a source of
valuable information and knowledge and gave them a sense of ownership of
the acquired knowledge. It allowed trainees to build up on what they know
and played important role in empowering them. The training techniques and
methods used are discussed under this section.
Training of trainers for extension workers
Efforts were made to train a large number of extension workers drawn from
the agricultural development offices of various coffee-growing areas of Ethiopia, who in turn, train farmers and their colleagues. The trainings covered
technical aspects of CWD including identification of CWD, causes, mechanisms of spread and control methods. The extensionists were also equipped
with proper methods for training farmers such as adult learning principles,
communication, facilitation skills and participatory training methodologies.
The training sessions were characterized by active participation by the trainees and hands-on exercises. The training methodologies included interactive and brief lectures, brainstorming, buzz and large group discussions and
presentations, videos of practical evidence, question-and-answer sessions,
field practical, case studies and other exercises. Participants were given an
opportunity to identify diseased trees, demonstrate control measures and
how to practically train farmers on the field (Plate 33b). In addition, during
later programmes, some of the previously trained extension workers were
invited to the training sessions to share their experiences, efforts made and
challenges encountered in the course of information transfer and in implementing the acquired knowledge. In total, more than 6000 extension staff
received the training conducted by senior staff of the project-implementing
agencies (Jimma Agricultural Research Centre and CAB International [CABI]
Africa) in collaboration with the agricultural and rural development offices.
The training covered five regional states, 30 zones and 93 coffee-producing
districts. Extension workers trained by the project staff also played an active
role in training their colleagues in the offices and those based in the field.
During each training session, comprehensive handouts, posters, booklet and
leaflets were prepared and given to participants. These were intended to
179
help them remember the training messages and to effectively pass on the
information to their fellow workers and farmers.
Training of farmers
At the end of each training session, the trained extension workers were
encouraged to develop action plans to carry out training programmes for
farmers in their respective operational areas. Through successive training
programmes and various awareness creation fora, the trained extension
workers made efforts to reach all farmers under their supervision. Farmers
training focused on creating awareness about the disease: its damage, identification, ways of spread, measures to control it and on the importance of
mobilizing neighbours for collective action and integration of local by-laws
to combat CWD. Some of the farmers training sessions were provided by the
Jimma Agricultural Research Center and CABI Africa staff directly, although
most were covered by the local extension staff. In general, the trained extension workers showed remarkable commitment to reach as large number of
farmers as possible through all available means. Reports communicated to
Jimma Agricultural Research Centre indicated that close to half a million
farmers received direct training in CWD management. Although it was not
possible to obtain the exact figures, informal assessment and communication
show that a considerable number of farmers and extension workers were
made aware about the disease through various fora. It was also expected that
there would be a trickle-down and multiplier effect of information from the
trained farmers.
Training of master trainers (resource persons)
Towards the end of the project, the critical need to strongly build the capacity
of selected government staff to serve as resource persons was realized. To this
effect, a comprehensive training of trainers for selected coffee and crop protection experts drawn from regional, zonal and district agricultural and rural
development offices was conducted. This training was different from the other
training of trainers courses conducted for extension workers in several aspects, including educational level of the trainees, content, intensity and length
of the training course. The master trainees are expected to serve as resource
persons for their respective regions and play crucial roles in the continuation
of the training and information dissemination activities. They also offer advice
and support on issues related to CWD. Therefore, a 10-day-long intensive and
comprehensive training programme was organized for 20 participants drawn
from the major coffee-growing regions. The trainees were carefully selected
by considering their performance, experience and educational background.
During the training workshop, a number of areas related to coffee production and protection in general and on CWD in particular were presented and
discussed in detail. Moreover, relevant areas such as extension, communication, adult education, facilitation and participatory training methodologies
were addressed. In addition to classroom presentations, extensive field and
laboratory demonstration activities and practical exercises were carried out.
180
Negussie et al.
Finally, participants were encouraged to make policy analysis and recommendations in relation to CWD management. Then, they shared responsibilities to communicate these policy recommendations to their respective
regions and to influence higher officials in favor of CWD management.
Participatory technology development to identify and validate CWD
management options
Participatory on-farm trials were implemented for 3 years in different districts of Jimma (south-west Ethiopia), Gedeo and Sidama (Southern Ethiopia). The field trials enabled farmers to test and validate different CWD
management options jointly with extension workers and researchers, as
well as enhanced farmers experimentation capacity. The steps followed
in the Participatory Technology Development Process can be summarized as
follows:
U
U
U
U
U
U
Holding comprehensive stakeholders (farmers, extension workers and

researchers) training on causes, symptoms, transmission mechanisms
and control measures of CWD, as well as on concepts of on-farm experiment, field trial design and implementation;
Identification of local and scientific knowledge in relation to CWD management;
Determining trial design and implementation (treatment application);
Regular monitoring and data recording;
Setting criteria and joint evaluation by farmers, extension workers and
researchers;
Sharing results through workshops, field days and exchange visits;
Scaling out the promising ones through farmer field schools (FFS) and
other mechanisms.
Based on the knowledge and experiences of local farmers, extension workers and researchers, as well as by considering experiences from other countries, seven management options (treatments) were identified during the
workshop. Each treatment was applied on a plot with 15 coffee trees, and
the treatments were replicated across farmers. Below are the treatments and
their application.
U
U
Use of ash: applying 2 l per tree once per annum.

Mulch: applying once per annum at the end of the rainy reason preferably in October.
Fungicide (copper) spray: applying once per month during the rainy
season and just once every 3 months during the dry season. It is mixed
at the rate of 40 g/l of water.
Fungicide (copper) stem paint: painting the stem of a coffee tree up to
50 cm above the ground level every 4 months. 300 g of copper is added
to 1 l of water.
Herbicide (Roundup): adding 150 ml of Roundup in 15 l of water and
spraying as needed based on weed condition.
181
Slashing plus hand weeding: weeding by hand around the coffee tree
and slashing the other parts as needed based on weed condition.
Slashing (control): slashing the whole plot as needed.
After 3 years of running the trials, workshops were organized with the objective of reviewing the results of the field trials and to select those that can be
taken to the FFS study fields. In attendance were FFS and field trial hosting
farmers and facilitators, other extension staff at different levels and researchers. Researchers presented the results of the field trials during the workshop.
On-farm trial hosting farmers and facilitating extension workers were also
encouraged to share their observations and experiences. The participants
were given an opportunity to visit the field trials and evaluate the performance and effectiveness of the different treatments. Then they were guided
to set criteria against which to evaluate and choose the best CWD management options (Table 11.1).
As indicated in Table 11.1, the participants tended to choose mulch,
slashing + hand weeding and ash, in that order, in southern Ethiopia. Similarly, mulch was ranked first, followed by ash and slashing + hand weeding
in the Jimma area (Table 11.2). In view of the promotion of the organic coffee
concept and the costs of chemical inputs, participants tended to disfavour
fungicides and herbicides. In particular, mulch was ranked most highly with
highest scores for all criteria except affordability, with participants talking
of its additional advantages in terms of weed control, moisture and soil conservation, soil fertility improvement and avoiding slashing and other operations that would wound the coffee trees. However, in terms of affordability,
use of ash was more favoured than mulch at both areas. The participants
Table 11.1. Mean scores and ranks given by 12 farmers and 14 extension staff in choosing
different CWD management options in southern Ethiopia (Gedeo).
Treatments
Fungicide Fungicide
paint
spray
Slashing +
hand
Slashing
Herbicide
weeding
(control) Mulching
use
Criteria
Ash
Availability
Effectiveness/
efficacy
Applicability/
Simplicity
Acceptability
(locally)
Affordability
Overall (average
score)
Treatment rank
4.17
3.74
1.57
3.04
1.91
3.52
4.39
4.22
4.17
2.39
4.60
4.60
2.23
3.86
4.39
2.35
2.35
4.43
3.56
4.74
3.05
4.18
2.35
2.26
4.35
2.95
4.60
2.36
4.25
4.15
2.60
2.38
2.26
2.46
4.22
4.32
3.65
3.34
3.69
4.45
2.55
2.81
Scores: 1 = low (least preferred practice); 5 = very high (most preferred practice).
182
Negussie et al.
Table 11.2. Mean scores and ranks given by 17 farmers and 20 extension staff in choosing
different CWD management options in south-west Ethiopia (Jimma).
Treatments
Fungicide
spray
Slashing +
hand
weeding
Slashing
(control)
Herbicide
use
Criteria
Ash
Fungicide
paint
Availability
Effectiveness/
efficacy
Applicability/
simplicity
Acceptability
(locally)
Affordability
Overall (Mean
score)
Treatment
rank
4.60
3.70
1.60
3.30
1.60
3.10
4.10
3.70
3.60
1.70
3.90
4.90
1.80
3.90
4.20
2.70
2.40
4.00
3.50
4.20
2.90
4.30
2.90
2.70
3.90
2.70
4.20
3.20
4.40
4.24
1.90
2.48
1.70
2.30
3.50
3.84
3.10
2.92
4.10
4.26
2.10
2.78
Mulching
Score: 1 represents low (least preferred practice), and 5 represents very high (most preferred
practice).
decided to take the selected treatments (the first three) to FFS study sites and
to validate them.
After running for 3 years, some of the on-farm trials, especially those
which were far away from existing FFS, were converted into FFS study sites.
These helped to sustain the field trial efforts and to disseminate the preliminary findings from the field trials. It also enabled the utilization and/or sharing of the knowledge and experiences of field trial hosting farmers.
Farmer field schools
FFS have been used as an important participatory training and information
dissemination tool for CWD management. FFS is a participatory training approach that can be considered both as an extension tool and a form of adult
education (David et al., 2006). The approach focuses on building farmers
capacity to make well-informed crop management decisions through increased knowledge and understanding of the agroecosystem. On the basis
of this knowledge, farmers become independent, confident decision makers
and experts in their own fields (Fliert and Van de-Fliert, 1993). The training
is hands-on and is carried out almost entirely in the field. The approach
provides opportunity for learning by doing, through observation, discussion
and interaction among participating farmers. Extension workers or trained
farmers facilitate the learning process, encouraging farmers to discover key
agroecological concepts and develop integrated pest management (IPM)
skills through self-discovery activities practiced in the field (Ooi, 1996). More
importantly, it gives an opportunity for farmers, extension workers and re-
183
searchers to interact as partners in the development of IPM options. The four

major principles of IPM emphasized by FFS are the following: (i) grow a
healthy crop, (ii) conduct regular field observations, (iii) conserve natural
enemies of crop pests and (iv) farmers become IPM experts.
Implementation of FFS under RCWP in Ethiopia
The FFS implementation concentrated on areas with high incidence of CWD
such as the Gedeo, Jimma and Sidama zones. Apart from serving as a participatory learning platform, the FFS was also used as dissemination pathways
for the results of the field trials. Initially, three pilot FFS were established in
2004, with a further 21 FFS groups being formed in 2005 and 2006 in southern
and south-western Ethiopia. Some of the lately formed groups emerged from
the on-farm field trials.
During the application of FFS approach to coffee, which is a perennial
tree crop, some of the processes required changes and adaptations to the
original FFS concepts to fit the local conditions and the crop under consideration. The ground work, formation and implementation of the FFS groups
went through a number of processes and modifications.
Training of facilitators and curriculum development. One of the areas where
modification was made is training of FFS facilitators. Initially, a 4-day intensive training workshop was held for FFS facilitators (mainly extension
workers) to introduce concepts and practices of the FFS approach, adult education, group processes and management, communication and facilitation
techniques. The technical training on CWD management was addressed in
a separate session prior to the training of the facilitators. Moreover, technical skills were further developed in the course of actual implementation of
the FFS activities. Effort was also made to enhance facilitators knowledge
and skills through provision of regular back-stopping, refresher courses and
experience-sharing workshops. At the end of facilitators training workshop,
participants moved out to the field and developed a tentative curriculum
for the FFS activities together with farmers, which basically follows the crop
cycle or calendar. Although it primarily focused on IPM in relation to CWD,
the curriculum tried to address a broad range of coffee management practices. The curriculum was flexible and regularly updated by FFS members to
fit to local situations.
Community mobilization, FFS group formation and selection of study field. In general, FFS consists of groups of people with common interest who get together
on a regular basis to study the how and why of a particular topic (Gallagher,
2003). To establish FFS groups, the trained facilitators went back to their operational areas and held a village assembly and briefed the community about
the project and aspects of CWD and explained the objectives and concepts of
FFS. Then volunteers were asked to be members of the FFS group, and about
2530 coffee farmers were registered. The selection of farmers considered
interest, acceptance among the community and proximity to each other and
184
Negussie et al.
to the study plot. Local officials, extension workers and farmers representatives also played a role in the selection process.
One important component of FFS is the study field, which serves as a
training site and an experimental laboratory where participating farmers
make observations, pursue discovery-based learning and experiment with
various farming techniques. One of the modifications made is in terms of
plot size and crop stage. The study field, which in this case was about 0.50
ha of already established coffee farm, was provided by a group member. The
criteria used in selecting the study field were accessibility to most members,
proximity to the field trials (as they are expected to be a dissemination pathway for the results), trees that were relatively young and poorly managed to
allow quick and clear responses to the different management practices, uniformity among trees and occurrence of CWD in the area. The selected study
field was divided into two parts: coffee in one half received the farmers conventional practices, whereas the other plot received improved management
practices. This allowed farmers to compare the performance of the different
management practices. The improved crop and pest management practices
were determined jointly by researchers and extensionists and tested by farmers. Because of clear differences in the effects of the improved practices on the
incidence of CWD, as well as on the performance of the coffee, farmers were
convinced to try the improved practices/technologies on their own farms.
Regular meeting, facilitation and length of the group activity. Other areas that required adaptation was frequency of meeting and length of the group activity.
In its original form developed for rice farming in Asia, FFS is a season-long
activity with a weekly regular meeting. However, in view of the perennial nature of the coffee crop and its slow response to treatments, the CWD FFS were
designed to operate for 2 to 3 years. Most of the groups preferred to hold their
regular meeting monthly, whereas some, especially those established towards
the end of the project, decided to meet fortnightly. The meetings took place in
the morning between 8:30 a.m. and noon.
All RCWP FFS groups in Ethiopia have been facilitated by trained extension workers, with support from experienced farmers and researchers. It
was realized that the facilitators role and attitude are key factors in determining the success of an FFS. His or her duties include serving as catalyst,
encouraging analysis, setting standards, posing questions and concerns,
paying attention to group dynamics, serving as mediator and encouraging
participants to ask questions and come to their own conclusion (Braun et
al., 2000). In short, the role of the extension workers is to facilitate the learning, experimentation and reflection processes. However, it was noticed that
facilitators often tend to teach the group members in a traditional top-down
way rather than encouraging them to interact, explore, discover and learn
by their own. Efforts to address this problem were made through close follow-up and support by experienced project staff. Moreover, as they gained
experience in running the groups and through refresher courses and interaction with other facilitators, extension workers started to improve their
facilitation skills.
185
Why and how to conduct agroecosystem analysis?

The cornerstone of the FFS methodology is agroecosystem analysis (AESA),
which involves regular and systematic observation of the crop and fields
(Plate 34). FFS knowledge generation and dissemination are basically
through agroecological system analysis, which is a discovery learning process. The AESA process sharpens farmers skills in the areas of observation
and decision making and helps develop their powers of critical thinking
(Gallagher, 2003). During the meeting, the group was divided into small
groups of five and undertook AESA. This involves field observations, analysis, discussion and presentations. Aspects of agroecosystem that were observed and analysed include growth stage of the coffee tree, occurrence
and incidence of CWD and other diseases, insect pests, natural enemies,
weeds, weather, soil, moisture and shade conditions. Modifications made
to AESA in the CWD FFS were both in terms of frequency of observation (which was in this case fortnightly and monthly) and observation of
the plant itself. Observation of some of the parts, such as the root system,
was not possible so effort was made to make observation of the complete
canopy. After conducting observation of the entire parts of the two plots,
the groups had to randomly choose three coffee trees from each plot and
make close observation, record, draw and present. The members discuss the
recommendations of the small groups and take appropriate management
decisions.
The importance of special topics
The special topic is the topic that participants want to learn more about. The
facilitators encouraged members of FFS groups to continuously identify
special topics of interest to them. The topics were then addressed either by
facilitators providing information on these topics or engaging farmers in appropriate discovery learning exercises or, in most cases, inviting researchers
to attend meetings to discuss the topics. Researchers provided detailed information on the identified areas and sometimes introduced various improved
coffee production technologies. This particularly helped to enhance farmers
knowledge and boost adoption of improved coffee technologies. Improved
coffee management practices and associated technologies applied on the improved study plot and as special topics included
U
U
U
U
U
U
U
U
U
U
Pruning, stumping and sucker control (Plate 29);

Shade management;
Proper intercropping practices coffee with haricot bean or enset;
Proper weeding and hoeing;
Use of mulch (Plate 30);
Planting leguminous crops, such as Desmodium sp. (Plate 31);
Compost preparation and application;
Proper harvesting (selective picking of fully matured beans);
Soil and moisture conservation techniques;
Proper use of chemicals such as fungicides and herbicides.
186
Negussie et al.
What were the achievements and successes of CWD FFS?

The participatory, practical and flexible nature of the FFS approach was appreciated by participating farmers and created motivation and enthusiasm to
seek further information and knowledge. In general, the group learning exercises enhanced farmers awareness and knowledge about CWD. They have
become experts in CWD diagnosis and are able to easily identify CWD from
other diseases, such as root-rot, and tree death due to exhaustion and overbearing dieback. The process developed their critical thinking and experimental capacity. Members of the group have fully realized how improved
management practices improve tree vigour and yield, and thus, they started
practicing on their own farm what they have learned during the group learning activity. The process thus enhanced adoption of various improved coffee
production practices. Moreover, group work helped farmers to cooperate in
uprooting and burning infected coffee trees. The experience-sharing process
also created interest among other neighbouring farmers to obtain new information and technologies. Thus, it has been proven that FFS groups can
be promising dissemination pathways for information related to CWD in
particular and for improved coffee technologies in general. Involvement in
the implementation of FFS activities provided an opportunity for extension
workers to develop their technical knowledge and facilitation skills. Moreover, the process created better interactions and improved linkage among
farmers, extension workers and researchers.
Another significant contribution of the approach was that one of the
groups in Gera district of the Jimma zone has already converted into a more
permanent group that can cater for other issues. The group submitted an
application to the local cooperative development office with the assistance
of the facilitator. They got registered as a marketing cooperative, secured a
loan and started collecting and selling members coffee directly at the central
market in Addis Ababa. During the second year, the group started purchasing other farm products in addition to coffee. Above all, the FFS process produced motivated and committed farmers who have already started making
remarkable efforts to inform, teach and change other farmers. The FFS activities extend beyond the members of the FFS groups. After some suspicion in
the initial stages, the FFS approach has been well received by farmers and
the extension agency, and now, there is interest and demand for creation of
more FFS groups.
What were the challenges and limitations?
The application of the FFS approach to tree crops such as coffee and the FFS
approach in general are relatively new in Ethiopia. This somehow initially
caused suspicion about the effectiveness of the approach and lack of experience
in facilitating the learning process. There was a tendency among facilitators to
teach farmers in a traditional top-down way, especially in the early stages, but
this was gradually improved through mentoring, continuous backstopping
and experience sharing. Apart from efforts to improve facilitation and technical skills of the extension workers, there was a clear need for change of attitude
187
and mindset. In addition, because of lack of experience among facilitators,

the use of group dynamics exercises and icebreakers during FFS meeting was
minimal. High turnover of the trained and experienced facilitators negatively
affected the group activities at some of the locations, with processes being
slowed because new facilitators had to be trained and take over. There was
also a tendency to prepare and present AESA predominantly in writing than
in drawing, which limits involvement of illiterate farmers.
Sometimes, the process posed challenges to facilitators and researchers because farmers started coming up with certain puzzling questions and
comments for which they did not have ready answers. This in fact trigged
further learning, information seeking and research on the part of facilitators
and researchers. Lack of appropriate coffee technologies to satisfy demands
of farmers in some agroecological areas, such as the Gedeo and Sidama
zones, was a major challenge. Another major challenge was meeting demand for inputs such as herbicides (to control noxious coffee weeds), seeds
of cover crops such as Desmodium sp. (for farmers own coffee fields) etc.
Moreover, lack of effective CWD control methods such as resistant varieties, chemicals and labouriousness and ineffectiveness of the recommended
uprooting and burning practices were also among the challenges encountered in the process. The perennial nature of the crop also made change and
impact slow. In other words, it takes a long time to see impacts of the improved management practices. Another limitation was the low attendance
of regular meetings by FFS members during harvesting seasons in some of
the districts.
Awareness creation, sensitization and experience-sharing workshops
and seminars
In view of the low level of awareness about CWD among different stakeholders, mass awareness creation and sensitization activities were carried
out both on formally organized events and in all available informal fora.
Especially, the need to create awareness and sensitize officials at different
levels was frequently raised, especially by the trained extension workers.
The success of the training activities and the continuity of the efforts largely
depend on the attitude held and support rendered by these officials; hence,
their understanding was critical. As a response to this concern, workshops
were initially organized for district and zonal administrators and heads of
agricultural development offices at four locations of the major coffee-growing areas. The aim of the workshop was to create awareness regarding CWD,
the magnitude of the problem and consequences and to brief on overall project activities. The sessions involved introductory presentations and discussions followed by field visits to severely affected areas. In addition, national
workshops were organized for higher officials, researchers, planners, policy
makers and other actors. The sensitization and awareness creation activities
indeed played an important role in convincing and influencing officials and
policy makers. The awareness raising helped them to understand the seriousness of the problem, to put the issue of CWD at the top of their agenda and
188
Negussie et al.
to render necessary support and commitment. For instance, coffee stumping1

used to be implemented in a quota form, and the major concern of the extension workers was fulfiling the assigned quota by indiscriminately stumping as many coffee trees as possible. However, it is obvious that stumping
activities can play a significant role in spreading CWD unless proper care
and precautions are taken. As a result of the discussions on the sensitization
workshops, officials and policy makers decided to stop the quota system and
to undertake stumping activities with all necessary care.
In addition to the locally and nationally held workshops, researchers
and project implementers attended various workshops and conferences and
shared their experiences and findings with other development actors and
scientists. Moreover, farmers, extension staff, researchers and other partners
drawn from different countries took part in the annual regional workshops
organized by RCWP. These helped to exchange views, information and experience among participants of different countries.
Awareness-raising activities were also targeted at school children because
they play an important role in the coffee sector in Ethiopia. Their contribution
is especially prominent in coffee weeding, harvesting and marketing. Children
play a crucial role in informing and educating their parents about new developments such as incidence of diseases, new technologies, etc. Cognescent of
their pivotal role, the project produced exercise books with illustrations, descriptions of the disease and control measures and were distributed to school
children. To further harness this potential, agricultural teachers of different
primary and secondary schools were invited to the training sessions held for
extension workers. The trained teachers went back and created awareness
among their students. They were also given a large number of leaflets and
posters to distribute to their students. School children were also invited to take
part in field days. In addition, lectures and seminars on CWD were given at
higher agricultural learning institutions by researchers and senior students.
Field/open days and exchange visits
Field days were organized at some of the FFS study sites, in research stations
and on on-farm field trials, and a large number of stakeholders attended
the events. In total, more than 5000 stakeholders (farmers, extension staff,
researchers, local officials and school children) took part in the field days
organized by the project in different parts of Ethiopia. During the events,
efforts were made to create an atmosphere in which visiting farmers can inspect, inquire, question, interact and get to know what has been done and
the outcome. The hosting farmers played an active part in running the field
days by explaining their experiences. The events offered great opportunity
to exchange ideas, views and useful experiences, as well as helped to create
awareness and stimulate the interests of other farmers. They also played an
important role in sensitizing officials and policy makers.
1
Stumping is a rejuvenation technique whereby the stem of an old coffee tree is cut at 30- to
45-cm height above the ground at 45 angle.
189
In addition, exchange visits were organized for innovative farmers, extension workers and researchers to visit other districts and regions to share
experiences. The exchange visits were made to areas where there was high
incidence of CWD, and this clearly demonstrated the need to pay serious
attention to CWD control before huge damage has been made. Participating farmers and extension staff were able to see the severity, incidence and
damage caused by the disease; how other farmers deal with the disease and
coffee production systems and practices in other agroecologies. It was also
noted that based on the experiences they gained from the project, different
regional states and districts started organizing field days and exchange visits
for their farmers, extension staff and government officials.
11.2.2. Information dissemination using printed materials and mass media

Both an old Chinese saying and modern research agree that, We forget what
we hear, we remember what we see, but we understand what we hear, see
and do. Materials can support the talking, the seeing and the doing involved
in training (Zeitlyn, 1992). The use of different print and electronic media
was given due attention by RCWP in view of the difficulties of reaching all
coffee-farming community only through face-to-face communication due
to vastness of the coffee-growing areas, poor infrastructure and resource
Table 11.3. Differences between mass media and interpersonal communication. (Adapted
from Rogers and Shoemaker, 1971, and Van den Ban and Hawkins, 1996.)
Characteristic
Interpersonal channels
Mass media channels
Message flow
Communication context
Amount of feedback readily available
Ability to overcome selective
processes (selective exposure)
Speed to cover large audiences
Possibility to adjust message to
audience
Cost per person reached
Possibility for audience to ignore
Same message to all receivers
Possibility to persuade an individual
to adopt an innovation
Possibility to persuade an individual
to form or change a strongly held
attitude
Who gives the information?
Tends to be two-way
Face-to-face
High
High
Tends to be one-way
Interposed
Low
Low
Slow
Large
Rapid
Small
High
Low
No
High
Low
High
Yes
Low
High
Low
Everybody
Experts or power
holders
Knowledge change
Possible effects
Attitude formation and

change
190
Negussie et al.
limitation. Mass media can reach a large audience rapidly, spread information and create knowledge and lead to change in weakly held attitudes
(Rogers, 1983). The cost per individual reached is also very low (Table 11.3).
Realizing these benefits, they were used extensively to create awareness,
sensitize and stimulate interest among relevant stakeholders regarding
CWD.
Information dissemination through printed materials
The importance of illustrated visual aids in training and communication activities cannot be overemphasized. It was believed that the use of printed
extension materials can considerably complement the other dissemination
mechanisms and reinforce the communicated information. Their great advantage is that they provide a permanent and detailed record of information
that can be continually referred to. People can use them in their own homes
or offices and read at their own pace and whenever they want. This makes
them useful and lasting reminders of the messages communicated through
them. In this regard, large quantities of different types of printed materials
on CWD were developed and disseminated to farmers, extension staff officials, policy makers and other stakeholders. During their production, effort
was made to present the message in a comprehensible form by combining
words and pictures to convey simple and clear information.
Booklets, leaflets and training handouts
Well-designed and detailed booklets were prepared in two local languages
(Amharic and Oromiffa) and distributed to frontline extension workers, experts and relevant institutions. In total, more than 10,000 copies of booklets
were produced and distributed by RCWP. The booklets provide detailed
information on diagnosis, incidence, transmission mechanisms and management of CWD. The booklets were also given out to the trainees during
the training of trainers workshops. Moreover, two types of detailed training
handouts were prepared and given out to the participants of the training
sessions. One of the handouts focused on various aspects of CWD, such as
extent of damage, its causes, symptoms and control measures, whereas the
second one dealt with adult learning, extension/communication methods
and participatory training methodology. The booklets and handouts served
as reference materials for extension workers during training of farmers and
other field activities related to CWD.
Similarly, in view of their importance in communicating information,
different types of leaflets were developed in three languages (Amharic,
Oromiffa and English) and distributed to users. Leaflets are simple, informative and handy print extension materials. Leaflets and posters can be useful
reminders of the spoken words (FAO, 1985). Although the leaflets were brief
and prepared with simple languages, they contained complete information
regarding the disease, its causes, symptoms and control measures. More than
35,000 copies of these leaflets were distributed to users in different regions,
zones and districts. In addition, to those dealing with issues directly related
191
to CWD, leaflets were also produced on coffee stumping as related to CWD.

These emphasized how stumping activities spread CWD and the cares that
need to be taken prior to, during and post stumping.
Posters and calendars
Posters, when they are well designed with striking images and messages
that are short, catchy and to the point, can work as powerful public awareness tools (CTA, 2002). They can be used for creating awareness and to
persuade and promote action. Posters can be more memorable than other
print media because they are supported by pictures and diagrams. Both
illiterate and educated people can be informed through simple and welldesigned posters. Another advantage of posters is that once posted, they
can stay around for some time and can be seen by many people again
and again. Four types of posters were developed in Amharic and English,
and about 11,000 copies were distributed to farmers, extension workers and
different institutions. The posters were user-friendly, self-explanatory and
informative with simple words and good illustrations that catch peoples
attention. In addition, 10,000 copies of two types of wall calendars illustrated with pictures related to the symptoms and management of CWD
were produced and distributed to different stakeholders including policy
makers.
Newspapers, bulletins and newsletters
Media reporters were frequently invited, and news on different events and several articles related to CWD appeared in the two national newspapers (Addis
Zemen and The Ethiopian Herald), which have wide readership. These particularly played an important role in informing and sensitizing policy makers. In
addition, bulletins and newsletters were also targeted to disseminate information related to CWD. Among others, the bulletin of the Ministry of Agriculture
and Rural Development, Graduation bulletin of university students and quarterly newsletter of the Ethiopian Institute of Agriculture are the major ones.
These channels reached a large number of people in the agricultural or coffee
sector.
Other awareness creation materials
Coffee is a key commodity, and its production and use involve diverse actor groups in different ways. These stakeholders have important roles in the
spread and management of the disease. Thus, with the aim of reaching these
diverse groups, other awareness-creation materials such as caps, T-shirts,
exercise books, stickers, etc. were produced in large quantities and were
disseminated. In total, about 17,000 were produced and distributed. These
materials particularly targeted producers and those involved in processing,
transport, marketing, training, research and policy-makers. In particular, exercise books with illustrations and description of the disease were produced
and distributed to school children.
192
Negussie et al.
Table 11.4. Type of mass media used and audience coverage. (From producers of respective stations/medias.)
Medium/channel
Radio
Ethiopian national radio
Radio Fana
Metu FM radio
SNNP FM radio
Ghimbi FM radio
Addis FM radio
Television
Ethiopian Television
Newspaper and bulletin
Addis Zemen
Ethiopian Herald
MoARD Bulletin
EIAR Newsletter
Language
Audience coverage
Amharic and Oromiffa

Amharic
Oromiffa
Amharic
Oromiffa
Amharic
Over 25,000,000
Over 25,000,000
Over 5,000,000
Over 9,500,000
About 1,500,000
Information not available
Amharic, Oromiffa, English
Over 12,000,000
Amharic
English
Amharic
English
Over 1,000,000
10,000 copies disseminated
3,000 copies disseminated
500 copies produced; also
posted on web site
Information dissemination through electronic media: radio and television

Electronic mass media are communication channels that can expose large
numbers of people to the same information very quickly. Their high speed
and low cost in communicating information to audiences over a wide area
makes them attractive, although they are not suitable for teaching practical
skills or for obtaining feedback and answering questions immediately. Information can be communicated more quickly than through printed materials or extension agents. The issue of CWD got broad media coverage (Table
11.4). News about the various events such as training, workshops, field days,
cross-visits and detailed special programmes were prepared on the disease
and aired to the public.
In particular, radio programmes were extensively used to effect widespread awareness creation and information dissemination on CWD, as it was
felt that radio is the most widely available, accessible and affordable mass
medium for remote, isolated and less literate rural communities. Radios are
now common features in rural areas, even where there is no electric power.
In this regard, different national and local radio stations with different languages were used to reach farmers and other stakeholders.
On the other hand, the fact that television combines vision with sound
makes it attractive and influential in communicating information. Like radio, television is an instant medium that transmits information directly to a
mass audience. Although in many developing countries, such as Ethiopia,
television transmission and sets used to be restricted to urban areas, in recent years, they are found among some rural communities as well. Television
played an important role in informing extension workers, researchers, officials and policy makers about CWD (Plate 35). It was especially instrumental
in displaying the symptoms of the disease and in showing fields devastated
193
by CWD to sensitize and motivate them to pay adequate attention to CWD

management. It also helped in sharing the experiences and efforts made by
different stakeholders in controlling the disease.
11.3. Summary and Conclusion

Although coffee plays a prominent role in the economies of many African
countries, its productivity has been constrained by several factors. CWD has
become one of the major threats facing the coffee industry in many Central
and Eastern African countries. There has been general lack of awareness and
knowledge about the disease among coffee farmers. This calls for efforts to
enhance awareness and knowledge about CWD among different actors in
the sector. Realizing this need, massive communication and information dissemination efforts were made by RCWP to reach farmers, extension workers,
commercial agents, researchers and policy makers with information related
to CWD. The programme used different methods, combining both interpersonal and mass communication and dissemination approaches. The former
included training of trainers for extension workers, training of farmers and
other stakeholders; application of FFS approach and participatory technology development; holding of various workshops and field days and exchange visits. Apart from creating awareness, these methods helped to equip
farmers and extension workers with detailed knowledge and skills regarding CWD management. As they provide opportunity for face-to-face interaction and two-way exchange of information, interpersonal channels were
found to be effective in persuading farmers to take necessary measures to
control the disease. In particular, the FFS approach proved to be an effective
tool in facilitating farmers learning and in enhancing their critical thinking
and problem-solving skills. It played an important role in creating interest
among farmers to seek further information and knowledge. However, extensive effort needs to be made to improve facilitation skills of the extension
workers. Continuous training, mentoring and experience sharing are needed
to help them change their mindset and skills from the conventional approach
to a real FFS facilitator. In addition, to further upscale and sustain the initiatives, there is a need to develop a more viable institutional framework that
will provide support and ensure continuity of the FFS activities. This could
involve mainstreaming FFS activities into the regular government extension
system and/or forming second-order FFS associations that provide support
to the groups. It is also important to encourage development of FFS groups
into more permanent groups such as cooperatives or study clubs.
It was not possible to reach all stakeholders related to the coffee sector
through face-to-face contact. In this regard, the use of print and electronic
media played crucial role in spreading information to a wide range of stakeholders. They particularly played a vital role in sensitizing officials and policy makers about the seriousness of the problem. Among the print materials,
booklets, leaflets, posters, handouts, calendars, exercise books, newsletters
and newspapers were widely used. Because these media combined words,
194
Negussie et al.
pictures and diagrams, they helped to convey clear information about the
disease. They were also used as reference materials especially among extension workers. As regards electronic media, radio and television played a key
role in spreading information rapidly and in creating awareness among large
and diverse stakeholders.
In general, impressive results were achieved in terms of exposure and
awareness creation. The use of combination of communication channels played
a remarkable role in this regard. Although there have been some commendable initiatives, it is still imperative to further strengthen and secure continued
support, appreciation and commitment from higher-level officials in addressing the problems of CWD. Vigorous effort should be made to accelerate information dissemination and training activities to reach all relevant stakeholders,
as well as to reinforce the use and implementation of the communicated information. Farmer-to-farmer extension and information exchange through
field days, exchange visits, workshops and other fora need to continue. In
particular, it is important to mobilize the community to take necessary collaborative efforts because CWD management requires joint actions especially
to remove infected coffee trees from the area. Organizing campaigns on CWD
control in collaboration with local administrative body and other pertinent
institutions is imperative. Traditional communication systems can be important channels for facilitating learning, peoples participation and wider information dissemination. Therefore, their use should be seriously considered.
Lack of complete and comprehensive packages of technical solutions to control CWD is among the major challenges encountered. For instance, some of
the recommended control measures were found to be demanding in terms of
labour or were not effective; this limited their acceptance. Therefore, there is
a dire need to strengthen research efforts to generate appropriate and viable
CWD management options that can be communicated to end-users.
Acknowledgements
This chapter is an output of the Project 4 activities of RCWP, which was financed by the Common Fund for Commodities. The International Coffee Organization supervised the project, whereas CABI was the project execution
agency. In Ethiopia, the project was implemented by the Jimma Agricultural
Research Centre of the Ethiopian Institute of Agricultural Research. The authors sincerely thank all the institutions that participated in implementing
the project activities. The authors also wish to express their gratitude to Dr
Dannie Romney for her valuable comments on the draft article.
References
Braun, A.R., Thiele, G. and Fernandez, M. (2000). FFS and Local Agricultural Research
Committees: Complementary Platforms for Integrated Decision-making in Sustainable Agriculture. AgREN Network paper No. 105. ODI.
195
CTA (2002) Public awareness: a manual for agricultural NGOs and research institutions
in Africa.
David, S., Agordorku, S., Bassanaga, S., Couloud, J., Kumi, M., Okuku, I. and Wandji,
D. (2006) A Guide for Conducting FFS on Cocoa Integrated Crop and Pest Management. International Institute of Tropical Agriculture, Accra, Ghana.
FAO (1985). Guide to Extension Training. FAO Training Series. Rome.
Fliert, E.V.D. and Van de-Fliert, E (1993) Integrated pest management: FFS generate sustainable practices. Wageningen Agricultural University Papers 34, 177184.
Gallagher, K. (2003) Fundamental elements of a farmer field school. Magazine on Low
External Input and Sustainable Agriculture 19, pp. 56.
Ooi, P. A. C. (1996) Experiences in educating rice farmers to understand biological
control. Entomophaga 41, pp. 375385.
Rogers, E. M, (1983). Diffusion of Innovations, 3rd edn. Free Press, New York.
Rogers, E.M. and Shoemaker, F.F. (1971) Communication of innovations. Free Press,
New York.
Van den Ban, A.W. and Hawkins, H.S. (1996) Agricultural Extension, 2nd edn. Blackwell Science Ltd, Oxford, UK.
World Bank (1992) Policy Research Bulletin. Washington, DC. World Bank.
Zeitlyn, J., (1992) Appropriate Media for Training and Development. University Press
Ltd, Dhaka, Bangladesh.
12
Concluding Remarks
Julie Flood
CABI, Bakeham Lane, Egham, Surrey, TW209TY
From the proceeding chapters, it is obvious that much has been achieved
during the Regional Coffee Wilt Programme (RCWP). We know more about
the biology of the pathogen, the extent of its spread and its impact on the
livelihoods of smallholder producers across participating countries than we
did before the RCWP started. Producers and other stakeholders in the coffee
sector in participating countries are now better equipped to recognize the
disease so that sanitation measures can be conducted quickly. We also know
that breeding for resistance remains a very good option for management of
this disease (an approach used successfully in earlier epidemics in the 20th
century). All this information will certainly help the coffee sector in countries
currently affected by the disease. It will also help to alert other countries to
the threat of coffee wilt disease (CWD). We have learned much, but we also
should consider what lessons can be learned for the future.
It has been estimated that around 1 billion dollars has been lost from
this disease due to both direct losses from reduced coffee production and
in costs incurred to try to manage the disease. Affected countries have been
changed forever, with many farmers moving away from coffee production
completely either diversifying into other crops or moving out of agriculture altogether. What can we learn from the outbreak of CWD as an example
of a major epidemic of such a commercially important crop as coffee and
a crop so extensively grown in Africa? Essentially, the enormous impact it
has had across the continent was largely due to a lack of preparedness for
a major epidemic caused by a pest/disease in coffee-producing countries in
Africa. This was despite the fact that just a few decades previously, the disease had reached epidemic proportions in several African coffee-producing
countries and had caused major effects on production and on coffee diversity. Inadequate preparedness was compounded by failures to understand
the seriousness of the threat and failures to respond quickly with counteractive measures. National governments were initially slow to recognize the
threat to their economies and to producer livelihoods, and this was exacerbated by the failure of the international community to recognize the seriousness of the situation at an early stage. CWD was considered to be a minor
196
Concluding Remarks
197
problem and one that was already under control. Undoubtedly, if there had
been fewer delays, the disease could have been managed more effectively
and its spread limited. Because of the remote location of many of the initial
disease foci, many years elapsed from the start of the new epidemic in the
Democratic Republic of Congo (DRC; probably commencing in the 1970s)
to the international community being aware of the seriousness of the situation in that country and the likely spread to other countries. Uganda and
adjacent countries were not alerted to the seriousness of the situation until it
was too late.
Even when the international community was alerted, delays continued
with many years elapsing from the initial surveys conducted in the 1990s to
project activities being implemented in the field in the early 2000s. This illustrates the point that international funding agencies can be slow to respond to
issues relating to plant health, with far-reaching consequences. For diseases
such as swine flu or avian flu, which are seen to directly affect human health,
resources can be mobilized within days or weeks, but it can take years to mobilize funds to manage an epidemic plant disease despite the direct effect on
the livelihoods of hundreds of thousands of small producers. In the intervening years, the disease had become widespread in Uganda and had emerged
in Tanzania. In retrospect, although a programme such as the RCWP was
clearly urgently needed for the long-term understanding of the pathogen, to
raise awareness across the coffee sector and to initiate breeding programmes
for the sustainable management of CWD, there should also have been a
mechanism for developing countries to obtain international funds quickly
in order to respond much more rapidly and restrict disease spread. This
would have involved alerting coffee stakeholders in affected countries and
adjacent countries and conducting emergency sanitation measures. Some
national governments did not wait for intervention from the international
community. Given the gravity of their situation, a strategy was developed
in Uganda using internal funds. Further surveys were conducted to ascertain the full extent of the problem, and a programme of awareness raising
for farmers and a research strategy, including a breeding programme, were
initiated. Uganda was fortunate enough to have a good research base and
agricultural infrastructure to be able to implement these activities. Similarly,
Tanzania initiated surveys and alerted producers in robusta-growing areas
adjacent to Uganda rapid action that helped to reduce the impact of CWD
in the country to some extent. These activities undoubtedly received a much
deserved boost from international funds channelled through the RCWP, and
progress has been much more rapid. Participating countries also benefited
from institutional capacity building both through training programmes and
improvements in infrastructure. International and regional expertise were
mobilized through the RCWP. In the case of Uganda, lessons were learned
from CWD as can be seen from the speed of the national response to banana bacterial wilt that emerged in Mukono in 2002. Farmers were mobilized quickly to conduct sanitation, and a research agenda was identified and
funded internally by the national government, all within 1 year of the initial
report of the disease.
198
Julie Flood
Nevertheless, despite the successes, CWD remains endemic to many

parts of central and eastern Africa and is still spreading. To try to prevent the
same problems arising again in adjacent countries, regular surveys should
be undertaken and the coffee sectors in these countries should be alerted of
the threat to their industry. We need to improve national capacity to recognize and deal with threats like CWD and encourage adjacent countries to
put a strategy in place to facilitate this. Because early recognition of CWD
symptoms is a crucial part of successful management, making authorities
and producers in neighbouring countries aware of the presence of the disease
and its imminent threat is essential. The disease is spreading in the Equateur
Province in DRC, putting robusta coffee production there under threat, as
well as robusta production in adjacent countries such as Congo, Central African Republic, Gabon and Cameroun. In many remote areas, traffic of people and goods including coffee, across porous borders is likely to add to the
risk of further disease spread. There is a greater threat when mass movements of people take place in areas of conflict or due to natural disasters.
Spread of plant disease is often ignored in those circumstances and yet can
have far-reaching consequences, as has been highlighted here. The conflict
that erupted in eastern DRC during the 1990s undoubtedly contributed to
the spread of CWD on robusta coffee. In addition, isolates of the pathogen
from arabica coffee in Ethiopia are pathogenic to arabica coffees grown in
Kenya, and the Kenyan coffee sector needs to be aware of this. Indeed, this
disease is not just a threat to African coffee production. Coffee from other
producing countries has now been shown to be susceptible, so the disease
is a threat to coffee production globally and coffee producers worldwide
should be made more aware of this disease. It is the aim of the many contributors to this book, including myself, that it should be one of the avenues
of communication by which the global coffee industry is alerted to CWD and
the threat it presents.
The CWD pathogen also infects wild coffees. As such, it constitutes a
threat to coffee diversity, and this has significant implications for coffee improvement programmes worldwide. The disease is widely distributed in
Ethiopia, an important source of diversity for arabica coffee and where wild
forest arabica coffee is still found in the south-west of the country. Equally,
in Uganda, where relic forests exist, wild Coffea canephora can be found, e.g.
in Kibale forest. Wild C. canephora also exists in DRC. The existence of this
pathogen over large areas of the continent, which correspond to the centres
of diversity for both robusta and arabica coffees, could have a huge impact
on future coffee breeding globally because genetic diversity is reduced. Such
a reduction in diversity will not only restrict our ability to breed for pest
and disease resistance but will also impact on the improvement of many agronomic traits. The destruction of coffee germplasm collections during this
current epidemic has also affected national and regional strategies to manage pests and diseases, as well as improvements for other agronomic traits.
In short, the disease has reduced the diversity of material available for future
breeding programmes. Paradoxically, with the selection of CWD-resistant
material in-country, this effect may be further compounded, as farmers may
Concluding Remarks
199
opt to use resistant material extensively, further reducing diversity of material available for breeding programmes. Conserving the remaining coffee
diversity both in-country and in international collections should be an area
for future work.
Yet, one of the notable successes of the RCWP has been the identification and selection of resistant material in several countries. Multiplication of
this material is now needed for dissemination to growers. Further funding is
needed to support this, as it will necessitate mass multiplication of genotypes
currently considered as resistant to CWD and distribution of this material to
farmers. Finance will also be required to support future coffee-breeding programmes and coffee breeders. The situation highlights limitations relating
to the cycle of project funding and how progress made during one project
can be sustained into the future when investment is removed. We do not yet
know how long this germplasm will remain resistant to the disease in the
field. Improvement programmes should be ongoing. The disease is likely
to spread further in Africa, and resistant germplasm will be needed in more
countries. Strengthening coffee institutions and the coffee sector is needed
to enhance competitiveness, and consequently, long-term investment for revitalizing the African coffee sector is required. Currently, the coffee sector
in Africa is characterized by low productivity and an aging workforce who
need to know more about improved production methods, good agricultural
practices and improving their market access. Investment is required both
from national governments, so that their policies reflect support for the sector, and from other sources, including the private industry and international
funding agencies. In many respects, the decline in investment to the coffee
sector reflects a general lack of investment in agriculture, which has taken
place over many decades. Such a decline needs to be reversed. The net result
of decades of under-investment is a weakened agricultural infrastructure
and weakened institutions in many developing countries. The re-emergence
of CWD as a major production constraint for coffee production in several
African countries is an excellent illustration of what can go wrong due to the
many complex problems associated with this decline in agricultural investment that has taken place since the 1980s.
October 2009
Index
age of plantations 19
agroecosystem analysis (AESA) 184185
anthracnose (Colletotrichum sp.) 10
arabica coffee see Coffea arabica
Armillaria mellea 3, 10, 31, 33, 36
ascospores 33
ash heaping 76
Association for Strengthening Agricultural
Research in Eastern and Central Africa 4
Bafwasende 1517
Banalia 1517
bananas, hosting G. xylarioides 102
Bandundu 1315, 19
Bas Congo 1315, 19
Belgian Congo see DRC
biological control 6263, 66, 148
Bixadus sierricola (white stem borer) 31, 143
Bundibugyo 34, 37
CABI (CAB International) 3, 4
Cameroon, CWD history 2
Central African Republic (CAR)CWD
history 2
CFC (Common Fund for Commodities) 4
chemical treatments 2122, 62, 147148 see
also copper oxychloride; Roundup
cinchona, hosting G. xylarioides 102
CIRAD (Centre de Coopration Internationale
en Recherche Agronomique pour le
Dveloppement) 156, 161
clean weeding 21
clones see CWD resistant varieties, use of
CLR (coffee leaf rust, Hemileia vastatrix)
Uganda 29
Coffea abeokuta 7
Coffea abyssinica 50
Coffea arabica
absence of CWD, in Uganda 3536
CWD history 2, 4, 5, 53
in DRC 78
recommended spacing 28
in Tanzania 69
in Uganda 2829, 3536
Coffea arabica cv Catimor 59
Coffea arabica cv Catuai 59
Coffea arabica cv Caturra Amerello 59
Coffea arabica cv Caturra Rojo 59, 127128
Coffea canephora
CWD history 2, 5
in DRC 7
genetic variability 157
in Tanzania 69
Coffea canephora Pierre 7
Coffea canephora Pierre var. kouillou 8
Coffea canephora Pierre var. robusta
CWD history 1, 2
in DRC 78
recommended spacing 28
in Uganda 2829, 3536
Coffea kapagota 36, 148
Coffea liberica 7, 13, 36
Coffea liberica var. dewevrei (formerly
C. excelsa) 7
CWD history 1, 2
Coffea myrtifolia 7
Coffea neo-arnoldiana 2
Coffea stenophylla 7
coffee
Fusaria other than G. xylarioides
113114
importance of 8384, 8687, 137
DRC 89, 83, 86
Ethiopia 5152, 83, 86
Kenya 83
Tanzania 69, 83, 86
Uganda 29, 83, 8687
coffee leaf rust (CLR, Hemileia vastatrix)
Uganda 29
Coffee Research Centre (COREC) 35, 156
201
202
Coffee Research Network 4
coffee wilt disease (CWD)
asset liquidation resulting from 9293
cause 100
distribution 84
farmer coping strategies 9395
farmer perceptions 85
history 15, 99, 155156, 196197
impact on coffee as source of income
8687
impact on coffee production 8791
impact on household expenditure 93
impact on input use and management of
coffee 9192
management 137138, 151152
biological control 6263, 66, 148
chemical control 2122, 62, 147148
see also copper oxychloride; Roundup
cultural practices 138
DRC 2025
Ethiopia 6265
governance issues 149
initial detection of CWD
resurgence 150151
integrated crop and pest
management 148149
preparedness before outbreak 149150
prevention of tree wounding 142143
quarantine 140141
Tanzania 7481
Uganda 4347
uprooting and burning 138140 see
also sanitation
use of disease-free planting
materials 141142
use of resistant varieties 146147
overview 196199
pests and diseases mistaken for 31, 3536,
141
socio-economic impact study 9697
sampling approaches 8485
symptoms 2, 3, 1011, 13, 33, 53,
100101
treatment recommendations 2, 2021,
3234, 62
variability of pathogen 7778
see also Gibberella xylarioides
Control of Coffee Wilt Disease (Tracheomycosis) Project (CFC/ICO/13) 143
copper oxychloride (Cupravit, Kocide) 21,
62, 76, 144, 147
Index
Cte dIvoire, CWD history 2
cotton seeds, hosting G. xylarioides 102
cultivar susceptibility 59
Cupravit (copper oxychloride, Kocide) 21,
62, 76, 144, 147
CWD see coffee wilt disease
CWD resistant varieties
assessment 159160
comparison field and artificial
inoculation results 165166
duration of exposure for root
dipping 162
field evaluation 163164, 170
inoculation methods 160161
inoculum concentration 161162
quantifying resistance in artificial
inoculation 162163
breeding
conservation of CWD-threatened
genetic resources 172
overview 155156, 173
potential sources of resistance 156159
en masse germplasm screening 167169
field germplasm screening 169
grafting 171172
hybridization for 170171
inheritance of resistance 166167
multiplication of 169
University of Kinshasa research 2325
use of 146147
Department of International Development
Crop Protection Programme 4
Desmodium intotum 144
Desmodium sp. 62, 66, 185, 187
diagnostic symptoms, of CWD 3, 33, 101
disinfection of farm implements 20, 34, 62,
66, 138
DRC (Democratic Republic of Congo)
constraints to coffee production 910
country-specific management
practices 2025
cultural practices and chemical control
2022
CWD distribution 1315
CWD history 2, 5, 1013, 198
CWD impact on coffee production 90
CWD importance 1519
diseases and pests 10
factors affecting severity of CWD 1920
farmer perceptions of CWD 85
Index
importance of coffee 89, 83, 86
overview 78, 2526
replanting improved material 2223
stakeholders sensitization 25
use of variety resistance 2325
Entebbe Botanical Gardens 148
environmental factors affecting CWD 1920,
41
Equateur 9, 1315, 19, 198
CWD history 4
Ethiopia
asset liquidation resulting from CWD 93
practices 6265
CWD distribution 5458, 198
CWD history 2, 5, 53
CWD impact on household
expenditure 93
CWD impact on input use 91
CWD importance 5859
Farmer Field Schools 182189
information dissemination 6465, 176177
overview 5051, 6566
excelsa coffee see Coffea liberica var. dewevrei
Farmer Field School (FFS) approach 5, 46,
7980, 182189
fertilizer use 42, 91, 146
firewood, use of infected trees as 32, 42, 60,
66, 74, 139, 140
forest coffee 5455
Fusarium decemcellulare 113
Fusarium guineensis sp. nov. 112
Fusarium lateritium 113
Fusarium oxysporum 35, 53, 113
Fusarium solani 35, 113114
Fusarium stilboides 113114
Fusarium xylarioides 1, 3, 53 see also
Gibberella xylarioides
garden coffee 54, 5556
Ghimbi coffee 50
Gibberella abyssiniae 112
203
Gibberella congoensis 112
Gibberella indica 114
Gibberella xylarioides 40, 53, 99100, 112,
114115
cultural and morphological
characteristics 106
cultural and morphological
variability 106107
disease epidemiology 101
fungal transmission and dynamics of
disease spread 103104
genetic variability and the role of the
sexual cycle 107111
host range 101102
hostpathogen interactions 120121,
133134
arabica cultivars vs. isolates 124126
arabica/canephora vs. isolates from both
species 127129
Coffea spp. vs. isolates from either host
species 129132
histological studies 132
18-months-old C. canephora
seedlings 133
9-months-old C. canephora
seedlings 132133
seedling inoculation protocols 121124
infection and colonization of coffee
102103
isolation from plant tissues and soil 105
life cycle 104
pathogenic variability 111
storage and preservation 105106
symptoms 100101
taxonomic status 112113
see also Fusarium xylarioides
glyphosate/Roundup 21, 144
Guinea, CWD history 2
Harrar coffee 50, 52
Haut-Ul 11, 13
Hemileia vastatrix (CLR, coffee leaf rust)
Uganda 29
ICO (International Coffee Organisation) 3
ICPM (integrated crop and pest
management) 148149
INERA (Institut National dEtudes et de
Recherches Agronomiques) 8, 11
information dissemination
communication strategies 177
Ethiopia 6465, 176177
204
Farmer Field School 5, 46, 7980, 182189
field days 188189
importance for uprooting and burning 139
interpersonal communication
approaches 177178
mass media 191192
overview 192194
participatory hands-on training 178
participatory technology 180182
printed materials 189191
Tanzania 7980
training of farmers 179
training of master trainers 179
training of trainers for extension workers 178
Uganda 45, 4647
insects, wounding trees and transmitting CWD
143
integrated crop and pest management (ICPM)
148149
inter-governmental cooperation 152
intercropping 20, 102
International Coffee Organisation (ICO) 3
Isiro 3, 11, 12, 1517
Index
Muhangi 13, 1516, 18
Mukono 3234
mulching 43, 62, 76, 92, 140, 143, 146,
181182
Mutwanga 13, 1516, 18, 23
Nectria haematococca 113
North Kivu 9, 13, 1315, 1618, 19
CWD impact on coffee production 89, 90
Office National du Caf see OZACAF
Office Zarois du Caf (OZACAF) 3, 13
Oicha 1516, 18, 23
Opala 1517
Orientale 9, 11, 1316, 19
Oubangui-Chari see Central African Republic
OZACAF (Office Zarois du Caf) 3, 13
Participatory Technology Development
Process 180
plantation coffee 54, 5758
pruning 20, 42, 92, 103, 143
quarantine 140141
Jimma coffee 50
Kampala 1997 conference 4, 35
Kasai Occidental, CWD not present 1315
Kasai Oriental, CWD not present 1315
Kenya 198
importance of coffee 83
Kocide (copper oxychloride, Cupravit) 21,
62, 76, 144, 147
large-scale plantation systems 19
leaf wounding method of inoculation 160
Lekempt coffee 50
Limu coffee 50
livestock, increasing CWD incidence 42,
102, 143
MAAIF (Ministry of Agriculture Animal
Industry and Fisheries) 35
Makerere University 35
Mambasa 1617
Mangina 1516, 18
master trainers, training of 179
Ministry of Agriculture Animal Industry and
Fisheries (MAAIF) 35
Monochamus leuconotus (white stem
borer) 70, 143
RCWP (Regional Coffee Wilt Programme) 1,

45, 35, 176, 196, 198
replanting into infested fields 28
robusta coffee see Coffea canephora Pierre
var. robusta
root dip method of inoculation 160
root mealybugs 31, 33
root rot, Armillaria mellea 3, 10, 31, 33, 36
Roundup 21, 144
sanitation 2, 3, 11, 32, 66, 75 see also
uprooting and burning
seed, transmission through 141
seedling inoculation 121124
semi-domesticated coffee see semi-forest coffee
semi-forest coffee 5455, 142
semiplantation systems 19
shade 31, 42
Sidamo coffee 50
slashing 2122
soil infection method of inoculation
soil infection methods of inoculation 160
South Kivu, CWD not present 1315
spacing
recommended 28
and severity of CWD 6061
Index
spraying 21
stem injection method 122, 123124
stem-nicking method 122, 124
stem wounding method of inoculation 160
stumping 62, 92
after infection 33
not practiced in Uganda 30
recommendations for 142143
spreading CWD in Ethiopia 60
symptoms, coffee wilt disease (CWD) 2, 3,
1011, 13, 33, 53, 100101
TaCRI (Tanzania Coffee Research
Institute) 74, 80
Tanzania
action taken against CWD 197
practices 7481
CWD distribution 7173, 84
CWD history 5, 71
CWD impact on household expenditure
93
CWD importance 7374
overview 69, 81
tracheomycosis 1 see also coffee wilt disease
UCDA (Uganda Coffee Development
Authority) 3435, 4344
Uganda
action taken against CWD 197
coffee production move to non-traditional
areas 4445
country-specific management practices
4347
205
CWD distribution 3638
CWD history 5, 3236
CWD impact on household expenditure
93
CWD importance 3840
information dissemination 45, 4647
overview 2829, 4748
replanting programme 44
stakeholder participation 4344, 4546
Uganda Coffee Development Authority
(UCDA) 3435, 4344
University of Kinshasa 23
uprooting and burning 138140 see also
sanitation
weed management 20, 42, 60, 92
avoiding wounding of coffee trees 143146
clean weeding 21
white coffee stem borer 143
white stem borer (Bixadus sierricola) 31, 143
white stem borer (Monochamus leuconotus)
70, 143
wild coffees 198 see also forest coffee
wind breaks 143
wounding of coffee trees 5, 2122, 34, 60, 103
G. xylarioides as a wound pathogen 42
painting with fungicide 62
prevention 142146
and tree spacing 61
weed management practices 143146
xylem system 12
Yangambi plantations 10, 19
Yangambi Research Stations 8, 11
Yirgacheffe coffee 50
Zaire see DRC

Coffee Wilt Disease - J. Flood (CABI, 2009) WW PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Coffee Wilt Disease - J. Flood (CABI, 2009) WW PDF

Uploaded by

Copyright:

Available Formats

This page intentionally left blank